key: cord-0005321-lnqldz3a
authors: nan
title: Symposia: Symposium 1
date: 2005
journal: Qual Life Res
DOI: 10.1007/s11136-005-1711-8
sha: 897d4e88bbe93c4140c0533947a574b7e30b6e24
doc_id: 5321
cord_uid: lnqldz3a

nan

The PROMIS (Patient-Reported Outcomes Measurement Information System) Network is a group of 6 primary research sites and 1 statistical coordinating center in a US National Institutes of Health (NIH) Roadmap Initiative. The goal is to develop item banks to measure patient-reported health outcomes across many chronic diseases and to use Item Response Theory (IRT) to develop computerized adaptive testing (CAT) tools to facilitate assessments with high content validity and precision. While most of the PROMIS Network activity deals with adults, the University of North Carolina (UNC) research site will assess patient-reported health outcomes in children. UNC also contributes to PROMIS Network activities in helping to develop network-wide standards that items must meet to be included in the item banks. This symposium will focus on several aspects of research on PROMIS with emphasis on measuring patient-reported outcomes in children and adolescents. Presentation 1: The starting point for the PROMIS Pediatric Item Bank is PedsQL Ò 4.0, a widely used pediatric quality-of-life instrument. We will review the content of this instrument and the extensive experience with its use. Presentation 2: We will present new factor analyses examining unidimensionality and local independence in 8000 + children and 6000 + parents in English and Spanish (3000 + parents and children). Presentation 3: IRT analyses will review the item trace lines computed using the graded response model and will and examine the extent to which items functioned differently for reporters (child vs. parent) and languages (English vs. Spanish).

Presentation 4: We will next review ongoing work to set standards for the PROMIS Network to review items and determine what conditions items (for adults & children) must meet to be included in the Item Bank. Presentation 5 (Discussion): We will discuss all presentations in relation to new insights into child health assessment and IRT-based instrument development. Inclusion of children in clinical research is strongly supported by the American Academy of Pediatrics and the National Institutes of Health. Not only does research participation provide access to therapies available only in clinical trials, it is fundamental to the development of new interventions and diagnostics that are beneficial for children. As is the case for adults, clinical measures of disease do not fully capture ways in which chronic diseases and conditions affect children's lives. The importance of children's self-reports in evaluating interventions and treatment strategies is underscored by research finding parents' reports only mildly to moderately correlated with their children's self-reports. But outcome assessment in children with chronic diseases and disabilities presents unique challenges. For example, it must be established that the child has the cognitive ability and language skills necessary to respond to the instrument, and, for children and youth with disabilities, adapted procedures often are necessary. For instance, when interviewing youth with motor and visual limitations, scales are enlarged to facilitate pointing and readability. Computer adaptive scaling may be an attractive alternative to traditional paper and pencil measures in scaling children's selfreported outcomes. The computer interface is readily adaptable to the needs of children with disabilities. Further, the selection of items in a computer adaptive scale is tailored to the trait level of respondents, and precise estimates can be achieved with substantially fewer items. But data collection in children is complex, time-consuming, and resource intensive, making it difficult to collect sufficient data for item response theory analyses. In this symposium we will discuss ways to address challenges related to self-report assessment in children with disabilities/chronic diseases and explore opportunities for collaborations with the symposium participants.

Aims: The PROMIS network aims to develop of a comprehensive item bank for Patient Reported Outcomes (PRO) of chronic diseases. In this paper we describe the item review process for building the item bank. Methods: An essential step for the development of an item pool is the identification and evaluation of extant questionnaires and items and the development of new items for a core item pool. Four phases of item review will be completed: item acquisition; surface characteristics documentation; expert item review; and, cognitive testing. Item acquisition refers to the systematic search for existing items in extant scales. Surface characteristics refer to the basic properties of items including the instruction set, item text, response options, the period to be included in the response, the sample developed on, and whether or not disease-specific information is part of the item. Additionally, each item is analyzed for readability using the Lexile framework. Experts will rate the similarity of items to one another, the reference the item has to real-time experience, and whether or not the item should move to the next stage of evaluation. A pared down number of items will be subjected to cognitive testing to confirm their appropriateness and clarity prior to administration to large groups of patients. Results: The results of qualitative item review and analysis of existing data will be combined with focus group data to inform development of new items that will then undergo the same process of qualitative item review. Items successfully screened through this process will be tested in field trials and subjected to innovative scale construction procedures. Conclusions: By performing rigorous and transparent qualitative item review, the PROMIS Network will arrive at a core set of high quality items useful for computerized adaptive testing in a diverse patient population. One of the assumptions underlying most applications of item response theory (IRT) is unidimensionality. A closely related assumption, local independence, is also critical if the models are to be appropriately used. Several methods exist in the literature to assess the extent to which each of these assumptions is satisfied. The focus of this talk will be the use of item factor analysis to check IRT assumptions. In addition, general lessons regarding the factor analysis of polytomous items will be discussed. This illustration uses data from the PedsQLTM 4.0 Generic Core Scales. For each subscale, confirmatory factor analysis (CFA) was used to assess the extent to which a 1-factor model was appropriate. When conducting a CFA with ordinal data, standard maximum likelihood (ML) estimation methods are not appropriate. We explored the use of a number of alternative estimation methods for ordinal data. Additionally, modification indices were used to identify items that potentially violate the local independence assumption. On a general level, we found that there can be discrepancies among the estimation methods, as well as with the same estimation method implemented in different software packages. With respect to the PedsQLTM, we found generally strong support for 1-factor models for each subscale, but located several areas of concern regarding local dependence. We discuss the implications of local dependence for subsequent IRT analyses and ways to avoid violating the local independence assumption. This talk provides researchers interested in using IRT with a description methods for checking model assumptions, suggestions for using these methods in practice, and ways in which the results can inform item calibration, scoring, and future scale development. Item response theory (IRT) is a collection of latent variable models that explain the process by which people respond to items in terms of item and person parameters. Specifically, it can be used to identify items most related to the construct of interest, identify items most informative at specific levels of the construct, determine a scoring algorithm, and identify items that function differently across groups. The focus of this talk will be the use of item response theory in item evaluation and scale development, with an emphasis on practical application. This illustration uses data from the PedsQLTM 4.0 Generic Core Scales. Item parameters were calibrated separately for each of the four subscales using the graded response model. This analysis provides a slope and a set of threshold parameters for each item, which are used to assess reliability and calculate item and scale information functions. The results show that the items generally exhibit favorable characteristics and are related to the same construct within a given subscale. We describe the results in terms of trace lines, item information functions, and scale information functions. These tools allow one to assess how well the items measure the constructs of interest and the appropriateness of the scales for various ranges on the latent construct. We discuss the lessons that can be learned by comparing alternate forms of the same scale. Finally, using the results of item factor analysis, we assess the potential impact of local dependence on the item parameters. This study provides valuable insight into the information that IRT can provide about the items and scales being analyzed. In addition, we explore how this information can be used to evaluate an existing scale or aid in the construction of a new scale. The PedsQL (Pediatric Quality of Life Inventory) Measurement Model was designed as a modular approach to measuring pediatric health-related quality of life (HRQOL), developed to integrate the relative merits of generic and disease-specific approaches. The PedsQL Measurement Model emphasizes the child's perceptions. The items chosen for inclusion were derived from the measurement properties of the child self-report scales, while the parent proxy-report scales were constructed to directly parallel the child self-report items. The 23-item PedsQL 4.0 Generic Core Scales have resulted from an extensive iterative process, and include child self-report for ages 5-18 and parent proxy-report for ages 2-18. The large and continuously expanding PedsQL DatabaseSM provides the opportunity for comparisons across chronic disease groups and benchmarking with healthy populations. We present data on 16,417 healthy children and 1732 children with a chronic health condition (including children with asthma, cancer, cardiac conditions, cerebral palsy, diabetes, end-stage renal disease, obesity, and rheumatic conditions). The data are heterogeneous with regard to age, gender, race/ethnicity, and language (8038 forms administered in English, 5638 Spanish, 333 Chinese, 171 Korean, 145 Vietnamese). The PedsQL 4.0 Generic Core Scales distinguish between healthy children and children with pediatric chronic health conditions and have demonstrated sensitivity to disease severity, responsiveness through patient change over time, as well as an impact on clinical decision-making, and associations with quality of healthcare and barriers to healthcare. Given that the PedsQL Measurement Model integrates generic core and disease-specific items into one measurement system, opportunities exist for determining item parameters across healthy and patient populations. Recent books by well-respected scholars present the theme that business and medical research make a bad mix (e.g., Angell, 2004; Deyo and Patrick, 2005) and the credibility of privately sponsored research is intensively scrutinized. The aim of this symposium is to identify obstacles to ethical conduct of health outcomes research when funded by private interests and to offer solutions to enhance research credibility and quality from different perspectives: academic research, pharmaceutical outcomes, and private research firms. From the academic per-spective, recommendations include maintaining awareness of how funding sources may profit from the results, working for reform at governmental levels, and revamping the registry process to a system of conditional regulatory approval followed by careful data monitoring. Peer review is critical and health outcomes research can make a unique contribution to enhancing consumer decision-making. From the industry perspective, pharmaceutical companies can play an active role in maintaining research integrity by applying accepted scientific principles and meeting peer-review standards, conducting training and audits, and publishing results of all studies on corporate websites. Key to ensuring ethical conduct is inclusion of external collaborators to help eliminate potential industry bias. From the perspective of privately funded researchers, the conduct and dissemination of credible research is often hampered by the assumption of bias. The capitalist imperative must be borne in mind, along with the ways in which funding source priorities shape research agendas, dissemination, and intellectual property rights (e.g., public domain instrument access). A common theme across perspectives is the value of transparency and realism regarding bias.

S-2B/1751/CAPITALISM AND THE PRODUCTION OF HEALTH OUTCOMES RESEARCH Lori B. Frank, Center for Health Outcomes Research, The MEDTAP Institute at UBC The field of health outcomes research has developed over the past two decades and matured in the past few years. The development of the field to this point is the result of multiple factors including an increased consumer focus for healthcare, demand for health products data to better address the information asymmetry inherent in physician/patient interactions, public policy interest in the cost expansion of healthcare, technological developments beyond improving survival, and the consequent need for measuring a wider range of health outcomes. Outcomes research is funded publicly and privately. On the public side, there are well-established programs for knowledge discovery. On the private side, the majority of health outcomes research programs are funded by the pharmaceutical industry. High profile publications emphasize bias as a threat to medical research funded by pharmaceutical firms (e.g., Goozner, 2003; Angell, 2004; Deyo and Patrick, 2005; Kassirer, 2005) . Privately funded health outcomes researchers can be vigilant about bias by considering the following questions: Who's asking the research questions and why? What proprietary value might the research have? How are funding sources influencing research agendas and the evolution of health outcomes research? How do they influence selection of forum for dissemination (e.g., submission to peer-reviewed literature, direct to consumer, company proprietary) and decisions regarding intellectual property rights (e.g., public domain access to instruments vs. private control)? A high level of awareness of these issues can mitigate the reverse bias often encountered from government and academic researchers. Vigilance and participation in the peer-review process are key to ensuring that research credibility and quality remain high and can help minimize the limitations and maximize the potential that capitalist interests as funding sources present. Background: In both the professional and lay press, there is often an assumption that studies funded by the pharmaceutical industry are inherently biased. This symposium is based on this assumption, questioning whether privately-funded outcomes research can be ethical. However, there have been few objective evaluations of whether bias is present in pharma-funded studies, particularly in outcomes research evaluations. Aim: To review and discuss the evidence for bias in published research funded by the pharmaceutical industry and consider the potential impacts of an assumption of bias. Discussion: Little information is available on bias or quality of outcomes research studies funded by the pharmaceutical industry; most evaluations have focused on clinical trials. While some reviews have indicated bias among published studies funded by the pharmaceutical industry (e.g., greater likelihood of presenting positive results), others have not identified any significant bias. A number of reviews have reported higher quality among studies with industry funding; interestingly, this finding is often not included in published abstracts and/or mentioned only briefly in results sections. Despite the lack of evidence for systemic bias in industry-funded studies, this assumption can be seen in journal publication rules, letters to the editor, and double standards for conflict-of-interest guidelines. This can result in decreased opportunities for industry researchers to participate in professional activities (co-author manuscripts, serve as members of advisory committees) and unnecessary barriers in performing and disseminating important studies in outcomes research. Conclusions: For outcomes research to advance as an objective discipline, it is important to eliminate pervasive biases not supported by impartial evaluations. Current guidelines and standards regarding funding disclosure should continue to be used, but research should be evaluated on its merits rather than on the source of funding or affiliation of the authors.

Background: Pharma continues to be the greatest source of innovative drug products in the world. The development of new medicines demands not just the delivery of new pills to patients, but knowledge of their safety, effectiveness, impact to the health care system including patients, prescribers, payers and caregivers. This knowledge must be delivered with the product at launch, requiring either active study by the sponsoring company or funded research. Aim: What are the safeguards that industry can employ to maintain objectivity and guard against conscious and subconscious bias? Discussion: There are a number of activities that pharmaceutical companies can perform in the effort to maintain integrity in health outcomes research. (1) The first is to apply good scientific principles that are referenced in texts and the scientific literature. (2) These prin-ciples should be employed with sufficient rigor so as to meet the standards of scientific peers which [if sufficiently employed] should allow publication in the peer-reviewed literature. (3) The sponsoring company should reinforce these scientific principles with their own endorsed Principles of Medical Research (PMR). (4) These PMR should be reinforced within the company with training and audits to ensure knowledge as a prerequisite to compliance. (5) Internal accountability at the sponsoring company can be maintained through publication of results of all studies on their corporate website. Lastly, incentives to appropriately partner with external collaborators should not be discouraged so that the key leaders in the research area of interest can help eliminate potential subconscious bias of industry in the design, analysis and publication of health outcomes research. To meet financial targets, pharmaceutical companies must make and sell profitable products. Sometimes shareholder obligations supersede those to patients or the public. Investments may be made more with marketing in mind than with research. Recent evidence suggests some research may be stifled or manipulated. A drug's effectiveness may be exaggerated. Common aspects of life may be 'medicalized.' Privately funded health outcomes research exists and has grown with the needs of pharma, which are spawned also by public demand and the public's hope for long and healthy lives. Given limited public funding for health outcomes research, how can health outcomes researchers benefit from industry ties and maintain scientific integrity? Or is it, as some contend, a delusion? Hope or Hype: The obsession with medical advances and the high cost of false promises (Deyo and Patrick, 2005 ) makes specific recommendations for health outcomes researchers in improving public health while managing the profit imperative of pharma. First, maintain awareness of who is paying and how the funding source may profit from the results. Second, lead the call for reform to help push the balance of research/marketing back toward research. Third, revamp the registry process to a system of conditional regulatory approval followed by careful data monitoring in a phased release of the drug to the market. Health outcomes researchers can be particularly valuable in the design and analysis of the naturalistic trials that would be required. Fourth, health outcomes research plays and could play an even more critical role in informing consumers, through patient oriented evidence summaries and patient decision aids that clearly and fairly present benefits and risks. Health outcomes information is critical to shared decision making, in which physicians, insurers, governments, and consumers have access to fair information and a road map for using that information. Finally peer review and transparency are necessary for making industry-sponsored research a viable source of information to all consumers, purchasers, regulators, and the public.

As quantitative and technical advances move patient-reported outcome (PRO) instruments to new levels, critical questions have been raised about intellectual property (IP) and business models for sustaining instrument development. These questions are relevant not only to instrument developers who may find their instrument items in 'item banks' used by computerized adaptive tests developed by federally-funded initiatives or private businesses. They are also relevant to anyone who seeks to sustain instrument development and application after the grant funding has ended. This symposium will provide a primer on IP concepts, business models, and distribution processes for developing, servicing, and maintaining a growing constituency of end-users. Dr. Arnold, an attorney with IP expertise, will introduce key terms and protections relevant to IP. Dr. Gershon will then describe and discuss the types of business models applicable to PRO instruments, and will discuss how IP and public/private partnerships are relevant to these business models. Dr. Varni will focus on copyrighting and distributing an instrument, and the decision steps a developer faces once one's instrument gains recognition and use. An expert panel will then convene to engage with the audience in an extended moderated discussion. Co-moderated by Drs. Schwartz and Cella, this panel will be comprised of the above speakers and two scientists who have been involved in numerous tool development studies (Drs. Kaplan and Sloan). The panel will comment on the advantages and disadvantages of non-proprietary and proprietary PRO measures, and will discuss their personal experiences and philosophical viewpoints regarding IP in the context of static and CAT tools for PROs.

S-3B/1252/BE CAREFUL WHAT YOU WISH FOR: RESPONSIBILITIES FOR DISTRIBUTING A PATIENT-REPORTED OUTCOME INSTRUMENT James W. Varni, Pediatrics, Texas A&M University Although much of the academic training of scientists on instrument development focuses on the numerous steps for developing a useful and psychometrically sound measure, there are often unanticipated demands after a new instrument has been developed that go well beyond most prior experiences. This presentation will focus on the decision-making course of action that the instrument developer faces if his or her measurement instrument gains widespread recognition and use. Beyond the relatively simple but necessary steps for copyrighting the measure, there are increasingly complex issues that become the responsibilities for the developer. Once an instru-ment is copyrighted, there is a clear professional responsibility and expectation to be responsive to the needs of other investigators who seek to utilize the instrument in their research or clinical applications, often on a just-in-time basis (i.e., the grant is due on Monday, the IRB meets tomorrow, the clinical trial is ready to launch this week). Options will be discussed related to distribution, interpretation guidelines, translation, and infrastructure necessary to sustain continued access and development of a patient-reported outcome measure. Experiences with the PedsQL (''Pediatric Quality of Life Inventory'') will illustrate the decision-making process in selecting an optimal infrastructure solution.

S-3C/1251/WHAT HAPPENS WHEN THE GRANT RUNS OUT? SUSTAINABILITY AND BUSINESS MODELS FOR PATIENT-REPORTED OUTCOME TOOLS Richard Gershon, Center for Outcomes, Research, and Education, Northwestern University

Only a few quality of life instrument authors have successfully turned their development efforts into economically viable ventures. This presentation will focus on models for creating sustainable revenue streams for what is otherwise perceived to be non-profit material. Instrument developers are able to realize revenue in multiple venues including translating instruments, licensing English and/or translated copies of their instruments, licensing scoring algorithms, creating online service bureaus, providing scoring services, selling manuals, facilitating validity studies for a particular population and/or running industry studies which utilize a particular instrument. The concept of copyrighting items, complete instruments, scales, norms and software will be presented, along with strategies for patenting unique assessment processes. All of these revenue strategies can be conducted as part of a for-profit company, a non-profit company, or in partnership with a third-party who actually conducts the revenue-based activities and pays the instrument developer a license fee for the use of their assessments. An early overview of the public-private partnership opportunities being considered for the NIH Roadmap Patient Reported Outcomes Measurement Information System (PROMIS) will also be presented.

S-3D/1250/INTELLECTUAL PROPERTY: LEGAL PRO-TECTIONS, COMMON ISSUES AND PRACTICAL ADVICE David Arnold, Wonderlic, Inc. This presentation will discuss the four basic protections available for intellectual property, copyrights, patents, trade secrets and trademarks. While each form of protection can be important to test developers, the major focus of the presentation will be on the copyright issues surrounding the development, distribution, sale and use of assessments. Special attention will be paid to issues regarding individual test items and what does and does not constitute copyright infringement. In addition to discussing how test developers go about creating copyright protections for assessments, the doctrine of fair use (the lawful but unauthorized reproduction of a copyrighted work) will be discussed. Attendees will be provided with real-life examples of copyright infringement and pragmatic recommendations regarding how to identify and deal with such issues. There is no doubt that IRT has seen increasing application in QOL assessment research in recent years. The application of IRT to knowledge-based testing has long since been established and is a commonly-seen method applied to standardized testing. With any new technique (or more to the point in this case, a new application of an existing technique) it is important to put forward the best case scenario and also explore the devil's advocate position. Similar exercises were important to the use of factor analysis in the 1970s, ANOVA in the 1940s, structural equation models in the 1990s and led to the varying degrees of acceptance.

The key is that this new approach needs to be placed into the appropriate context alongside other methods so it can be used to its fullest extent possible and be applied intelligently. It is essential for the future of IRT that it be disseminated in a clear and consistent fashion if it is to be widely adopted. In this symposium, we will attempt to look at the application of IRT for PROs objectively by alternately posing and responding to a series of fundamental questions, including: What is the philosophical justification for IRT methodology? Can it hence readily apply to assessing patient's condition? What is the value added for IRT? Does it have to be traditional tool development VERSUS item response theory? What is the best way of communicating IRT results? We demonstrate the relative strengths and weaknesses of 'basic' approaches, IRT modeling, and alternative modeling approaches. We utilize two datasets for this purpose: an experimental setting using the well known SF36 instrument, and a dataset involving a number of different global assessment measures for QOL. We will also invite the audience to discuss their experiences with application of IRT to PRO assessment.

This presentation will demonstrate the advantages and challenges of applying 'classic' test theory to PRO assessment. Dr. Mandrekar will answer the fundamental questions: What does 'classic' test theory entail? What information does 'classic' test theory provide and what are its limitations? The three speakers will attempt to answer interactively, using data drawn from real clinical trails, the fundamental question: Does it have to be traditional tool development VERSUS item response theory? Both hopefully are complementary or supplementary segments of a comprehensive scientific investigation into how one could, would, and should measure patient QOL with optimal efficiency and effectiveness. The three speakers will present comparative analysis done in concert with each other highlighting the potential for integration and complimentary information that can be obtained through both analytical approaches. The three speakers will then engage the audience by asking for examples where they have applied IRT to QOL assessment and to discuss their successes and challenges.

This presentation will summarize the basic advantages and disadvantages of attempting to apply IRT to PRO assessment. Dr. Wilson will provide initial answers to the fundamental questions: What is the philosophical justification for IRT methodology? Can it hence readily apply to assessing patient's condition? What is the value added for IRT? What is the best way of communicating IRT results?

S-4D/1861/APPLICATION OF IRT TO PROS: AN HIS-TORICAL PERSPECTIVE Jeff Sloan, Health Sciences Research, Mayo Clinic This presentation will discuss the social history of item response theory as it has been applied to the assessment of patient-reported outcomes. Using examples from history such as the emergence of factor analysis and Bayes methods, we will explore parallels and lessons learned regarding the initial promise and ultimate realization of new analytical approaches. The unifying theme of this symposium is overcoming limitations to quality of life, in relation to cross cultural research in specific patient populations. The five papers will explore approaches to bridging some of the current gaps in our understanding of the quality of life of populations across the lifespan, with particular emphasis on the impact of disability and related health status. This symposium will highlight the relevance of cross-cultural methodology in identifying core aspects of quality of life across populations with diverse industrial, economic, social, and cultural profiles. These varied profiles incorporate different approaches to social welfare and the delivery of health care; the importance of understanding the impact of specific populations' attitudes to their quality of life; and the utility of involving target populations in the definition of their quality of life by addressing the dimensions of comprehensive measurement. Aims: To generate the set of facets required as the basis for a comprehensive measure of the quality of life of adults with intellectual disabilities, taking account of the views of the target population, their relatives, care staff and relevant professionals, as well as using conceptual and empirical literature. Methods: Focus group methodology was employed to generate quality of life themes relevant to the target population (adults with intellectual disabilities) and relate these to the set of facets contained in the World Health Organisation's quality of life measure the WHOQOL. Six focus groups were run, comprising adults with intellectual disabilities (2 groups, n=10), qualified staff with experience of working in intellectual disabilities services (2 groups, n=12), relatives of adults with intellectual disabilities (n=9), and staff in training contemporaneously on attachment in intellectual disabilities (n=15). From the focus group themes elicited, items were generated to reflect each quality of life facet. These were tested in a small pilot study, and a revised version of the scale was produced for subsequent testing in a study of approximately 400 adults with intellectual disabilities in health and social care settings. Results: Initial analysis of the psychometric properties of the scale suggested either a five or six domain model for the 36-item measure, including a number of dimensions supplementary to the base WHOQOL measure (WHOQOL-BREF version). Conclusions: The targeted focus groups produced a core facet set similar to that of the WHOQOL, but also generated additional themes in the context of the life experiences and quality of life of adults with intellectual disabilities. This more comprehensive, and population relevant, total facet set yielded psychometric properties sufficiently adequate and robust to warrant continued development in the context of a much larger cross-cultural field trial.

Aims: In the field of paediatric health care, self-report measures based on consequences of health conditions have been developed recently to screen for children with special health care needs. The current paper focuses on measures on health care needs and quality of life that have been developed in the DI-SABKIDS study with a particular emphasis on discrepancies between measures related to health services and to QoL. Methods: The DISABKIDS family of instruments were developed in a European project on instrument development in the field of pediatric health care. Focus groups, international instrument development, pilot and field testing assessed the perspective of children, aged 4-18, parents and clinicians. The following diagnoses were included: Asthma, arthritis, dermatitis, epilepsy, cystic fibrosis, diabetes and cerebral palsy, treated in hospitals across 7 European countries. The pilot study (n=546 children) and field study (n=1606 children) included a range of clinical measures to assess the severity of the conditions, functional status and a screener for identifying children with special health care needs in a population of children with chronic conditions diagnosed and treated in different European paediatric hospitals. Results: Results show that the association between different types of measures that might be used for health service research in pediatric care is weak. The prevalence of children identified positively with the CSHCN screener was 80%. There was no significant relationship between the screener classification and quality of life, however this varied across the type of diagnosis and its severity. Unmet needs and satisfaction of health care services showed some specific associations with QoL outcomes. Cross-cultural differences were stronger in health services-related measures. Aims: To develop a cross-cultural measure of attitudes to ageing based on the experiences and views of older adults themselves and to relate the measure to their self-reported quality of life. Methods: A pilot study and then a field trial test were set up to develop a measure of attitudes to ageing. The plan of development followed that of the recommended WHOQOL methodology and, in fact, the measure was developed alongside that of the WHOQOL module for older adults, the WHOQOL-OLD. In summary, focus groups were run in a number of centres worldwide with older adults, carers, and professionals working with older adults. Preliminary items were generated and tested in a pilot study, and a revised version of the scale tested in a field trial study of over 5000 participants from 15 centres worldwide. Results: A combination of classical and modern psychometric methods will be summarised that have resulted in a 24-item measure of Attitudes to Ageing. The scale consists of three subscales of 8-items each, the details of which will be presented in the talk. The relationship of the scale to quality of life, as measured by the WHOQOL-OLD, will also be briefly examined. Conclusions: This newly devised measure of Attitudes to Ageing is short and easy to use and has been devised drawing on the thoughts and experiences of older adults from a wide variety of cultures worldwide. It can be used both as a research and epidemiological tool and as a clinical tool with which to devise individual and group programmes for successful ageing.

S-5E/1095/THE QUALITY OF LIFE OF OLDER PEOPLE AND THEIR ATTITUDES TO AGING Ken Laidlaw, Clinical & Health Psychology, University of Edinburgh, UK Aims: The paper provides a conceptual background to the development of a new attitudes to ageing questionnaire that was developed in 23 countries worldwide. The Experiences and Attitudes to Ageing Questionnaire (EAAQ) is the first cross-cultural questionnaire that is designed for use with older people as a result of collaborative participatory research with older people. Method: The development of the scale makes reference to gerontological science and also reference to models of successful ageing in the selection and creation of scale items. Each item was examined for its utility and meaning across cultures and was included only after a careful iterative process that incorporated local country-specific focus groups and further modification through two delphi exercises prior to piloting in each center. The research project, coordinated by the Edinburgh Center, was funded as part of a European Union Framework 5 research grant 'The Measurement of Quality of Life in Older Adults and its Relationship to Healthy Ageing'. Results: In the pilot trial, 1414 older people (age range 60-99) completed question-naires, with 60% of the sample female and 72% considering themselves healthy even though many participants had more than one chronic illness. Currently we are analysing collected field trial data for over 5000 participants, with an age range of 60-100. Principal Components Exploratory Factor Analysis revealed an 8-factor model that was further refined using structural equation modelling resulting in a 3-factor model encompassing psychological growth, psychological loss, and physical change. The 3-factor model suggests a way of conceptualising and measuring successful ageing in individuals. Conclusions: The EAAQ provides researchers, clinicians and policy makers with a unique scale to measure the impact of successful ageing interventions. It also provides a vehicle for the measurement of how individuals age across cultures and under different economic, political and social circumstances.

S-5F/1094/CROSS CULTURAL APPROACHES TO MEASURING QUALITY OF LIFE ACROSS THE LIFE-SPAN Guus L. Van Heck, Psychology and Health, Tilburg University, Netherlands

Aims: (i) to describe how QOL develops over the individual's lifetime with an emphasis on the role of culture, personality, and age; (ii) to examine cross-cultural measurement of QOL, and (iii) to demonstrate the relevance of cross-cultural methodology in identifying core domains and facets of QOL. Methods: Review of the literature on cross-cultural QOL across the lifespan using the Web of Science and PsychLit databases. In addition, between-country differences in age-QOL relationships will be presented based on new analyses of WHOQOL-Bref data. The relative contributions to subjective ratings of general health and QOL of (i) cultural context, (ii) personality and temperament, and (iii) age will be compared using multiple regression analyses and structural equation modelling. Findings: In QOL, according to the WHOQOL Group conceptualisation, the context of the value systems in which individuals live play a major role. It is shown that these value systems are substantially influenced by age, culture, personality, as well as interpersonal and temporal contexts. Individuals' goals, expectations, standards and concerns change along the lifespan. These alterations reflect the health trajectories as well as the adaptive processes that develop over the individuals' lifetime. However, there are remarkable individual differences due to personality and temperament. Conclusions: The breadth and depth with which QOL instruments should address relevant health aspects requires in future a more holistic, dynamic and ecological approach. Four broad contexts should be weaved into research and theory on QOL: (i) the sociocultural context, including age and cultural background, (ii) the interpersonal context containing formal and informal social networks, (iii) the situational context of the situations that are encountered in daily life, and (iv) the temporal context featuring, for instance, the timing of illness in life and the different stages of disease. Cross-cultural measurement is necessary for such an ecological approach to QOL.

Aims: To examine the impact of Intensive Care (IC) stay on the health-related quality of life (HRQOL) of sepsis survivors, and to compare their HRQOL with the age matched normal Dutch population. Methods: Long-term prospective study on patients with sepsis identified by existing criteria and admitted to the IC for >48 hours. HRQOL was assessed with the SF-36. Patients or proxies completed the questionnaire in the first 48 hours of admission, to assess HRQOL in the pre-IC period. Patients also completed the SF-36 at IC discharge, hospital discharge, and 3 and 6 months after IC discharge. SF-36 data of survivors at 6 months were compared with normative data from the age matched Dutch normal population. Cross-sectional comparisons were done with t-tests, changes over time were assessed with repeated measures ANOVA. Results: Of the 170 patients with severe sepsis 95 could be evaluated 6 months after IC discharge. Eight patients were lost to follow up, 67 patients had died. HRQOL showed a sharp multidimensional decline from pre-admission to IC-discharge, and gradual improvement in 6 months after IC discharge in all dimensions of the SF-36 (all p<0.01). SF, VT, RE and MH returned to pre-admission level, but physical domains remained lower (all p<0.05). Compared to a normal population, the average SF-36 scores of survivors of severe sepsis were still lower in all dimensions at 6 months after IC discharge, with the exception of SF and BP (all p<0.05). Preadmission HRQOL of severe sepsis survivors was already lower in RP, MH and VT compared to the normal population (all p<0.01). Conclusions: In survivors of severe sepsis, HRQOL showed a sharp decline during IC stay, and gradual improvement during the 6 months after IC discharge. Sepsis survivors improved up to pre-admission values except for the physical domains, but remained lower than in the normal population in all domains except SF and BP. Aims: To develop treatment plans for patients with COPD it is important to evaluate the measurement properties of factors that influence HRQL and their relationship with overall HRQL. The objective of this study was to assess, using confirmatory factor analysis (CFA), the properties of latent factors that underlie HRQL in patients with COPD. A secondary objective was to evaluate the magnitude and directionality of the relationships between the latent factors. Methods: Baseline values on measures among 180 COPD outpatients participating in a randomized controlled trial were used to evaluate the properties of four underlying factors and their relationship with a HRQL factor in a second order CFA model. Variables measured included diseasespecific HRQL, generic HRQL, forced expiratory volume in one second, and dyspnea. Results: The second order CFA suggested that a HRQL factor in patients with COPD may be measured by four underlying factors (mental health, physical health, mastery, and energy). Fit indices for this model indicated acceptable fit as evaluated by the RMSEA (09; CI: 0.07, 0.1), and the Comparative Fit Index (0.9). The amount of variability in the above four factors, that was explained by HRQL, was greater than 50% except for the physical factor which was slightly lower (39%). HRQL had a direct and indirect effect on energy through the mental health factor. Conclusions: The results suggest that HRQL has a strong and significant relationship with areas that are targeted by pulmonary rehabilitation interventions. This is important for planning treatment regimens. Future work will evaluate predictive structural equation models to identify which areas may have the greatest impact on maintaining HRQL over time. As HRQL plays an important role in evaluating the effectiveness of interventions for patients with COPD, future work will also assess whether there are differences in this model between treatment and control groups which may indicate a potential response shift. Aims: Due to a shortage of donor organs candidates for liver transplantation are confronted with waiting times of considerable length. Among these patients and their relatives increasing psychological distress during the waiting period has been reported. Aims of this study were: (1) to investigate the course of quality of life (QOL) and psychological distress in liver transplant candidates during the waiting time; (2) investigate the course of psychological distress in their relatives; (3) identify predictors for deteriorating QOL and psychological distress in the patient sample. Methods: 47 patients (mean age 56.9 years, 21% female) eligible for the waiting list for liver transplantation were included in the study. Patients completed the EORTC QLQ-C30 and the Hospital Anxiety and Depression Scale (HADS) at the beginning of the waiting period and at each interim visit at 4-6 weeks intervals. Relatives filled in the HADS only. Short term (1-2 months) and medium term (3-5 months) changes from baseline were analysed. Results: On average, the patient sample did not show major changes in QOL or psychological distress during the waiting period. However, females and older aged patients showed a less favourable course in their QOL than the other patients. We found significant associations of older age with deteriorations in physical and cognitive functioning and of female gender with changes for the worse in role functioning and global QOL (p<0.05). Contrary to the patients, their relatives exhibited a significant increase in depression and anxiety levels. Conclusions: Among patients awaiting liver transplantation females and elderly patients may be subgroups requiring special psychological support. The same may apply for the relatives of the entire patient group. Aims: Zoster pain interference with patients' normal activities is often measured using general health-related quality of life instruments. We determined the reliability and validity of ADL items within two new zoster-specific measures, the Zoster Brief Pain Inventory (ZBPI) and the Zoster Impact Questionnaire (ZIQ). Methods: The ZBPI is a validated, modified version of the Brief Pain Inventory that measures zoster pain and its interference with 7 ADL items. The ZIQ contains 12 items measuring the interference of zoster pain with the patient's ability or desire to perform ADL. Data from a prospective validation study were used to assess reliability (internal consistency and test-retest), convergent validity (cross-sectional and longitudinal), and discriminant validity (known-groups analyses) of the ADL items in both measures. Zoster patients were enrolled from various US sites within 7 days of rash onset and followed for 6 months or until they reported no pain at 2 consecutive interviews. Other questionnaires included in the study were the McGill Pain Questionnaire, the EuroQoL VAS, and the SF-12. Results: Results are presented for day 14 after rash onset (n=149). Cronbach's alpha was greater than 0.70 for both instruments indicating that the ADL items were internally consistent. Testretest reliabilities were above the recommended cut point of 0.75. Correlations with other questionnaires were good (range 0.55-0.84) and statistically significant (p<0.001), demonstrating convergent validity. The two new measures differentiated between two pre-defined pain severity categories (p<0.05), supporting discriminant validity. Conclusions: The ADL items of the ZBPI and the ZIQ are valid and sensitive measures of the interference of zoster pain with patients' daily lives. Aims: To assess the Quality of Life (QoL) and psychosocial wellbeing of adolescent and adult children whose parent has either a chronic or acute neurological condition and make comparisons across conditions. Methods: Questionnaire batteries measuring QoL and psychosocial variables were postally administered to 438 family members where one parent has a diagnosis of Parkinson's disease (PD), Multiple Sclerosis (MS) or stroke. Responses were received from 336 participants (76%). Of these 173 were adolescent and adult children (age range 11-48), 98 were the affected parent, and 65 the non-affected parent. Results: Children living in the parental home report significantly greater impact than those who have left (t=)0.284, p<0.00), with female children reporting greater impact of parental illness in almost all variables measured. Comparing QoL between the three groups reveals children of parents with PD showing the least impact from their parent's condition, followed by children of parents with MS, with children of parents who have suffered a stroke reporting the greatest impact. 18% of children of parents with PD, 31% of children of parents with MS, and 36% of children of stroke victims reported mild-moderate depression. Correlations between child and parent variables suggest that it is the emotional manifestations of MS and the practical implications of PD and stroke that have the greatest impact on the child. Access to more information, and guidance in caring for their parent was identified as a priority by over 50% of children. Conclusions: Both chronic and acute neurological illness has a major psychosocial impact on children of sufferers. Longitudinal study is required to observe how the impact of parental illness develops over time and identify those families most 'at risk'. This will allow for the formulation of evidence based guidelines, and the development of interventions. Aims: CDH is a congenital malformation requiring surgical repair as a neonate. Survival has improved to >90% at Children's Hospital over the last 20 years. We examined parents' reports of HRQL and functional status (FS) of a cross-sectional cohort of survivors of CDH repair. Methods: Fifty-three CDH survivors' parents (65% of eligibles) completed (1) the 20-item Child Health Related Inventories (CHRIs), a HRQL measure, yielding physical, role, emotional functioning, and energy domain scores; (2) the 38-item Functional Status IIR (FSIIR), yielding general health and interpersonal functioning subscores and a total score; (3) a validated clinical severity score; and (4) a medical history form on their child's current condition. For both the CHRIs and FSIIR, lower scores connote worse functioning. Spearman's correlation coefficients and Wilcoxon's rank sum statistics were performed. Results: The average CDH survivor was 8 years old and in second grade; 57% were male. FSIIR and CHRIs scores were highly correlated (r=0.55-0.86, p<0.001). Sixty percent of CDH survivors had persistent medical problems. Children with ongoing medical problems had lower CHRIs (p<0.05) and FSIIR scores (p=0.01) compared to those without problems. Clinical severity ratings varied by presence or absence of current medical problems (p<0.001). CHRIs and FSIIR scores were strongly correlated with current clinical severity ratings (r=)0. 41 to )0.65, p<0.01). CHRIs scores did not vary by presence of cardiac anomalies (26%), genetic malformations (25%), or need for neonatal extracorporeal membrane oxygenation (ECMO) (36%). FSIIR scores varied by genetic malformations and need for ECMO, but not by cardiac anomalies. Conclusions: Despite striking improvements in overall survival, a subset of long-term CDH survivors continues to have ongoing medical problems, translating to lower FS and HRQL. These affected children and their families may benefit from prospective identification and ongoing interventions. Future confirmatory studies are needed. mode effect). This study investigated if mode effects exist when administering the Center for Epidemiologic Depression (CES-D) scale by a personal digital assistant (PDA) vs. the P&P format. Methods: Participants were patients, family members or friends in the genito-urinary, gastrointestinal, and psychiatry clinic waiting areas at M.D. Anderson. 64% agreed to participate. Participants took both versions and were randomly assigned to complete the PDA first (n=379) or the P&P first (n=377). An Item Response Theory method, the Differential Functioning of Items and Tests (DFIT) procedure, identified mode effects on both the test overall and the individual items. A classical approach using a mixed effects regression model summarized the effects in terms of the classical CES-D score. Results: Scores were on average higher (3 points) when the P&P was administered first than all other administrations (PDA 2nd, PDA 1st, and P&P 2nd). The mode-by-order effect was mediated by higher education and increased depressive symptoms; it was more pronounced in older people. 18 of 20 items exhibited a mode effect when the P&P was administered first. Conclusions: The mode-by-order effect affects interpretation of CES-D scores, especially when used to screen for depression. Results underscore the importance of checking validity of any psychosocial instrument prior to administering it in a different mode.

O-1-C03/1328/PSYCHOMETRIC ASSESSMENT OF HRQOL DIARIES: A SYSTEMATIC REVIEW OF METHODOLOG-ICAL QUALITY Joanne Greenhalgh, HCPRDU, University of Salford, Salford, Greater Manchester, UK Aims: This paper discusses the methodological challenges to the psychometric evaluation and interpretation of diaries that measure symptoms and their impact on HRQoL. It is based on a systematic review of studies that psychometrically evaluated symptom and HRQoL diaries in health. Methods: Six electronic health databases were searched using a combination of free text and thesaurus terms relating to diaries, HRQoL and psychometrics. Inclusion criteria were: diary was completed by the patient for at least one week, measured symptoms and their impact on HRQoL and at least one psychometric property had been assessed. Study quality was assessed using a combination of Lohr et al.'s (2002) and Greenhalgh et al.'s (1998) checklist. Results: 15 diaries met the inclusion criteria with asthma (6) and cancer (5) diaries being the most common. Validity was assessed in all 15 diaries but only 11 assessed responsiveness and 10 assessed reliability. For 12 diaries, their psychometric properties were assessed in one study only. A number of methodological weaknesses were identified. Reliability: score and test-retest periods varied widely, 6 did not report intraclass correlation co-efficients and 3 did not independently assess the stability of symptoms. Validity: 9 studies did not set a-priori hypotheses when assessing construct validity and 11 did not state the expected magnitude or direction of correlations between scales. Responsiveness: only two studies calculated an effect size or provided an independent assessment of change. Conclusions: There is less standardisation in the methods used to assess the psychometric properties of HRQoL diaries and the quality of psychometric assessment is generally poorer in comparison to single point HRQoL instruments. Few instruments are used or evaluated outside the original study for which they were designed. Further methodological research to address the challenges of the psychometric assessment of HRQoL diaries is needed. Aims: Although social well-being (SWB) is recognized as an integral component of health, it is rarely included in healthrelated quality of life (HRQL) instruments. Two SWB dimensions were identified by literature review: Role Participation (SWB-RP) and Social Support (SWB-SS). As part of a larger project to develop item response theory-derived item banks and computerized adaptive testing, we developed and tested SWB banks. Methods: Item ratings of three large (n>500) datasets was conducted by 15 reviewers. Rasch measurement analyses were conducted to initially define item hierarchies. Out of 83 total items, 8 were removed due to model misfit and 8 were removed because of overlapping item content. We then wrote 16 new SWB-RP and 11 new SWB-SS items to fill content gaps, and edited items to improve comprehension and consistency. A total of 94 items (29 SWB-RP, 65 SWB-SS) was administered by computer to 202 cancer patients. Exploratory factor analysis (EFA; principal components with promax rotation) and Rasch analyses were conducted. Results: Patients commented favorably on the content of the items and expressed appreciation for attention to this aspect of their HRQL. EFA and Rasch analyses supported two SWB-RP sub-domains: limitations and satisfaction. Specifically, 19 items had factor loadings>0.5 on limitations and 6 items loaded on satisfaction (64% and 8% variance explained, respectively). Each sub-domain demonstrated good internal consistency reliability (0.95, 0.86) and good fit to the Rasch model. The remaining four items were set aside. SWB-SS analyses are ongoing. Conclusions: Optimal care for patients with cancer or other chronic illnesses includes obtaining a complete picture of patients' physical and psychosocial health status. SWB measures are important since diseases like cancer and their treatment can affect quality of relationships, parental responsibilities, work abilities and social activities. With properly calibrated item banks, it will be possible to precisely and efficiently measure and monitor multiple HRQL dimensions in individual patients, and use their responses to inform care. Qualitative patient feedback and quantitative analyses suggest that it is possible and desirable to include SWB measures in HRQL assessment. Aims: To develop a leukemia-specific subscale for the Functional Assessment of Cancer Therapy (FACT) instrument and to assess the linguistic validity of the Czech, Korean, Portu-guese, Slovak and Spanish translations. Methods: The leukemia-specific subscale was developed through item generation from literature review and interviews with clinical experts and patients in 9 countries. The FACT-Leukemia (FACT-Leu) was then translated using the standard FACIT methodology: 2 forward translations, a reconciled version of the 2 forwards, back-translation of the reconciled version to English, and 3 independent reviews by bilingual experts. The linguistic validation sample included 85 patients (60% male, mean age 46 years) from 6 countries: Czech Republic=10, Brazil=15, Korea=15, Slovak Republic=15, Spain=15, and US (Spanish)=15. Patients diagnosed with leukemia completed the FACT-Leu in their native language and then offered structured input on problems with translation or content. Statistical and reliability analyses were performed, and the participant comments were analyzed qualitatively. Results: The Leu subscale had good reliability with the combined sample (alpha=0.88) and the Spanish sample (alpha=0.88). Low sample sizes precluded internal consistency testing of other language versions. There were no negative patient comments related to the FACT-Leu items in any of the languages. In addition, all leukemia subscale items had good item-total correlations in the combined sample. Item-total correlations in the individual language samples were reviewed to identify potential translation errors. Four items were flagged with this procedure: appetite, bleeding, bruising, and worry about infections. These translations were reviewed to ensure that the translation was not the cause of the relatively low item-total correlation. A minor revision was made to the Spanish version after testing. Conclusions: The FACT-Leu showed good reliability and linguistic validity in the five-language combined sample and in the Spanish sample, and good linguistic validity across the individual languages. These results contributed to a better understanding of how quality of life issues are perceived by leukemia patients in different countries and supported cross-cultural comparability of instrument scores. Aims: Assessment of Health-related Quality of Life (HRQoL) is becoming routine in oncology clinics (Wright et al., 2003) . Studies have shown that HRQoL assessment increases the frequency of communication of HRQoL in clinic consultations (Detmar et al., 2000) , and levels of self-reported HRQoL (Velikova et al., 2004) . However, evidence suggests that some physicians remain unsure how to respond to HRQoL scores (Greenhalgh and Meadows, 1999) , perhaps due to the relative complexity of the output from some HRQoL instruments. The aim of this study was to derive summary indices of the EORTC QLQ-c30 with a view to facilitating interpretation of HRQoL. Methods: A retrospective analysis was undertaken of data from 1620 cancer \patients (heterogeneous diagnoses) who had completed the EORTC QLQ-C30 on touchscreen. The data were analysed using a Principal Components Analysis (PCA, orthogonal rotation). Factors identified (eigenvalues>1) by the PCA were then analysed using Rasch analysis (Rating Scale model) to assess unidimensionality and item fit. Results: Four factors were identified from the PCA: (1) Physical Function Index (12 items: physical functioning, dyspnoea, fatigue and quality of life), (2) Social Function Index (7 items: role and social functioning, pain and finance), (3) Cognitive/Emotional Index (7 items), and (4) Gastro-intestinal Symptom Index (5 items). These four indices demonstrated good face validity. The Rasch analysis also identified an additional index corresponding to the total EORTC QLQ-c30. Furthermore, the Rasch analysis confirmed that these indices were unidimensional constructs with good item fit (fit statis-tics<1.4). Reliability for indices 1-3 and the total index was very good (range: 0.80-0.90), but moderate for the gastrointestinal index (0.53). Conclusions: The results of this study suggest that the scales from the EORTC QLQ-c30 may be combined either to form a single global quality of life measure or three or four individual summary indices. Additional work is required to explore convergent and discriminant validity of the summary indices, and whether these indices prove to be beneficial to clinicians. Aims: The goal of this pooled analysis was to study the agreement between clinically significant changes in QOL and patient reported AE. Methods: Data from 6 NCCTG lung cancer trials involving 358 patients who had completed baseline and at least one post-baseline assessment was used. All trials included the UNISCALE and either the Functional Assessment for Cancer Therapy-Lung, the Lung Cancer Symptom Scale, or the Symptom Distress Scale. Non-Hematologic AE (alopecia, anorexia, constipation, diarrhea, dyspnea, fatigue, nausea, neurosensory, vomiting) occurring in at least 2% of patients post-baseline were included in the AE analysis. A clinically significant decline (CSD) in QOL was defined as a 10-point drop from baseline. A CSD in AE was defined as a change from ‡grade 2 at baseline to ‡grade 3 post-baseline. The agreement between changes in QOL with changes in AE was determined. The Kaplan-Meier method was used to evaluate the time to the first occurrence of a severe (grade 3+) AE and first CSD in QOL. Results: A total of 183 (58%) and 125 (40%) patients reported at least one CSD in QOL as measured by UNISCALE and a multiitem tool respectively. At least one severe AE was reported in 33% of patients post-baseline. The percent agreement between experiencing a severe AE and a CSD in QOL was 49% and 59% for UNISCALE and the multi-item tools. For individual AE, the percent agreement between a CSD in AE and a CSD in QOL ranged from 37% to 50% for UNI-SCALE and 44% to 69% for the multi-item tools. Anorexia had the least percent agreement with both UNISCALE (37%) and the multi-item tools (44%). The median time to the first CSD in QOL for UNISCALE and multi-item tools was 67 and 142 days respectively. The median time to the first occurrence of a severe AE was 304 days. Conclusions: QOL and AE provide quantifiably different information. The multi-item tools are in better agreement with occurrence or changes in severe AE compared to the UNISCALE. There is some indication that a CSD in QOL occurs earlier than AE reporting. Further research to evaluate how AE impact QOL is needed. Aims: The question ''What is your overall quality of life during the past week?'' is vague and begs a response ''compared to what?'' We explored whether different patients have in mind different reference frames. We hypothesised that those comparing themselves against other patients would report better QOL than those contrasting against healthy peers or against their previous healthy state. Methods: An ongoing clinical trial of therapy for Paget's disease assessed HRQOL using SF36, EQ5D, a question about overall quality of life, and disease-specific items. Patients (n=1250) were asked whether they had a specific comparison group in mind while answering their QOL questionnaires -such as a healthy peer group (e.g. friends), other Paget's patients, or themselves before becoming ill. Results: The majority of patients identified a single peer group, the main groups (at baseline) being: 31% themselves a year or more ago, 24% other Paget's patients, 23% themselves before becoming ill. As hypothesised, there were substantial differences in the mean scores according to the reference frame employed. With a 5-point general health scale, those thinking about other patients scored on average 0.56 better than those who were considering their pre-illness state. For overall QoL the differences were even more pronounced: 0.75. Many patients shifted their reference frame at later assessments (1, 2 yr), but similar magnitude effects were observed. Conclusions: We observed a large shift equivalent to 15 points on a 0-100 scale, indicating that frame of reference makes a substantial impact on responses. This could have a major impact in trials such as those in which some patients attend clinics while others stay at home. It also implies that more consistent estimates of QOL might be obtained by explicitly specifying an intended reference. Aims: Individuals undergoing health state changes may change their internal standards, values, or conceptualization of QOL. Such 'response shifts' might affect or distort QOL measurement as well as the impact of medical or psychosocial interventions. We evaluated the clinical significance of response shift by doing a meta-analysis of published QOL articles on response shift. Methods: Extensive database searches and contact with researchers yielded 454 articles and 40 articles, respectively. Only published empirical longitudinal studies that measured response shift were retained, resulting in 29, of which 19 reported data requisite data for computing the effect size (ES) to determine clinical significance. ES for each study was calculated and compared with regard to the QOL domains measured, disease group investigated, sample size, and response shift method used. Articles were rated for quality of study design to allow ES weighting. Results: When ES absolute values were examined, ES magnitude ranged from small to moderate, with the largest ES detected for fatigue, followed by global QOL, physical role limitation, psychological well-being, and pain (mean |ES|=0.64, 0.47, 0.24, 0.21, and 0.20, respectively Aims: Identifying response shift at the individual level requires design features, such as a then-test or interview, to be built into the study. Statistical techniques such as change in the factor structure provide information only at the group level. A statistical method for identifying response shift at the individual level would help to interpret change scores when design options have or cannot be used. The purpose of this study is to explore whether a comparison between predicted and reported health status could be used as a method of identifying subjects who potentially experienced a response shift. Methods: Data from a longitudinal study of HRQL in 678 persons with stroke were used to develop a predictive growth model from physical, mental and cognitive ability. Fluctuations in change of the residuals over time were used to identify persons with a potential response shift. To validate this procedure, we compared its diagnostic accuracy against the then-test. This validation dataset arose from a randomized trial where 123 persons had a then-test administered specifically to evaluate response shift. Results: Fluctuations in residuals of the development model using all four times points ranged from )46 (a reported 46 point underestimate compared to the predicted score) to 73 (overestimate). The cut-point for identifying response shift via residuals was a 20-point change and 27 (22%) persons were identified. A then-test change of at least 25 units (out of 100) was recorded for 34 persons (28%); 30 scaled down and 4 scaled up their rating. A 2Â2 table analysis revealed the predictive value of a positive residual test to be 59%, the predictive value of a negative residual test was 80%; sensitivity was 46% and specificity was 88%. Conclusions: This type of residual analysis could be used to stratify a data set according to the potential for response shift. Persons with small residual fluctuations over time (negative test) were unlikely (<20%) to have a response shift and would form one strata while all others would form another strata. An examination of mean change and variability of change within these two strata would provide information for the purposes of interpretation of change for the whole group. Aims: A major challenge facing clinical researchers is the interpretation of the clinical significance of health-related quality of life (HRQoL) changes over time. A common anchorbased approach utilizes patient-generated transition ratings to classify domain-specific HRQoL change using a 15-point patient transition rating scale, ranging from )7 (a very great deal worse) to 0 (about the same) to 7 (a very great deal better). The Juniper et al. (1994) set traditional change levels on this scale at trivial (±1, 0), minimal (±2, ±3), moderate (±4, ±5) and large (±6, ±7) are subjectively defined and assume symmetry for improvement and decline. However, different cut points may yield dissimilar change categories, and therefore it is critical to evaluate the empirical premise behind the transition rating cut points. Methods: This study compared traditional and empirically derived Rasch and Graded Response Model (GRM) cut points and subsequent mean changes for assessing clinically important changes using data from interviews conducted with 1662 outpatients with asthma, heart or lung disease. These outpatients attended primary care clinics in the US, and completed a disease-specific HRQoL questionnaire (Asthma Quality of Life Questionnaire, Chronic Heart Questionnaire, or the Chronic Respiratory Disease Questionnaire) and corresponding domain-specific transition rating items every two months for one year. Results: Traditional, Rasch, and GRM classifications slightly differed and displayed a lack of symmetry between improvements and declines; however, nearly all mean changes across classification approach were equivalent. Conclusions: Traditional transition rating cut points, thereby, appear suitable to classify HRQoL clinical significance for these chronic health conditions. MIDs, and to present general MID guidelines for FACIT instruments. Methods: We review and highlight results from our prior work using a combination of distribution-and anchorbased approaches to determine MIDs. We discuss the importance of assessing the appropriateness of anchors prior to determining MIDs and offer suggestions for combining results into a single range of plausible MIDs. Results: Our review provides evidence that MIDs for FACIT scales and subscales are fairly stable across patient populations and settings. This stability justifies the development of general guidelines, which can be used to estimate MIDs for FACIT scales and subscales that currently lack established MIDs. The general MID guideline for individual subscales is 0.30-0.40 points per item or 7-11% of the total score. The general MID guideline for Trial Outcome Indices (i.e., disease-, treatment-, or condition-specific summary measures of physical function and well-being) is 0.20-0.30 points per item or 5-7% of the total score. The general MID guideline for FACIT Total scales (i.e., the FACT-G alone or FACT-G plus a disease-, treatment-, or condition-specific subscale) is 0.15-0.25 points per item or 4-6% of the total score. An application of these general MID guidelines will be presented. Conclusions: We derived general MID guidelines specific to the FACIT family of HRQL instruments. Although such guidelines are useful when available self-report data are limited, greater precision in MID estimation can always be obtained from more thorough analyses. Aims: Most health care providers are familiar with the term quality of life (QOL). The concept of QOL has been discussed in various medical settings. However, the definition of QOL still remains unclear. Furthermore, few attempts have been made to date to examine the levels of awareness about QOL among the general public. Methods: In October 2004, a nationwide survey was conducted to identify such levels among the general public. Two thousand individuals aged 20 years or older, who were selected by stratified two-stage random sampling, were interviewed. Results: A total of 1385 responses were collected. While approximately 25% of the subjects aged 20-59 were familiar with the term QOL, the awareness of QOL decreased dramatically in those subjects who were aged 60 or older. The level of awareness tended to be higher among those respondents who lived in large cities, those with college diploma and white-collar workers. Following is the list of words used by the respondents to describe the meaning of QOL: daily life, economic stability, meaningful life, health, mental stability, self-respect, sense of happiness, human life, and social roles. No association was found among the choice of words used to describe QOL, the respondents levels of awareness regarding QOL, and their perception of health. Conclusions: The present study suggests the levels of awareness about QOL among the general public is generally low, although such levels differ across individuals with different personal attributes. The health care providers who discuss QOL with patients should pay attention not only to the characteristics of the individual patients but also to the patients levels of understanding of QOL. Little is known about its relationship with pattern of alcohol consumption; in addition, gender differences in this relationship are still not well understood. The present study examined the relation between several patterns of alcohol use and subjective health in a general-population sample of men and women. Methods: The outcome measure, subjective health, was assessed using the two component summaries, physical and mental, of the Short Form-36 health survey questionnaire. Alcohol consumption measures refer to the 30 days before the interview. Analysis of covariance was used to compare gender-specific mean scores across multiple categories of drinking variables. Results: Former drinkers of both sexes consistently reported poorer physical and mental health than either lifetime abstainers or current drinkers. Among current drinkers, drinking pattern differentially affected the two components (i.e., physical and mental) of subjective health with evidence of gender-specific associations. In women, a moderate consumption of alcohol (1-1.9 drinks per day), daily drinking, and consuming alcohol with a mixed pattern in relation to beverage preference exerted more beneficial effects on perceived mental health than other drinking patterns. In men, a moderate consumption of alcohol (2-2.9 drinks per day), consuming alcohol only on week-ends, and drinking with a mixed pattern in relation to beverage preference were associated with more beneficial effects on perceived physical health than other drinking patterns. Conclusions: These findings indicate that aspects of drinking pattern may differentially affect physical and mental subjective health and that this association is genderspecific. Aims: To examine the predictive ability of 10 scales of Short Form 36 (SF-36) on mortality in a Taiwanese, community-based sample. Methods: A 6-year prospective, longitudinal study was conducted. Two waves of Health Survey were conducted in a representative sample of 1715 community residents in five counties of Taiwan in 1994 and 1995. The overall response rate of the first year was 72.13% and the rate of loss to follow-up for the second year was 14.87%. These subjects were drawn by a four-stage sampling design. Face-to-face interview by public health nurses was used for data collection. The SF-36 includes physical functioning (PF), social functioning (SF), role limitations due to physical problems (RP), role limitations due to emotional problems (RE), mental health (MH), vitality (VT), pain (BP), and general health (GH). In addition, SF-36 Physical (PCS) and Mental (MCS) Component Summary scales are derived. Mortality data were collected from the national mortality register. Multivariate Coxs proportional hazard models were used. Results: The overall 6-year mortality is 4.37%. Our study findings indicate that after adjusting for age, gender, and chronic conditions, only the change of MCS score was significantly associated with mortality. Subjects with higher score from the first wave to the second wave of survey had significantly decreased risk compared to those with no change or lower score (Relative Risk [RR]: 0.58, 95% CI: 0.35-0.95). When PCS and MCS were simultaneously taken into account, the change of MCS score still exerted significant effect on mortality (RR=0.57, 95% CI: 0.34-0.94). Conclusions: People who reported better health status between two consecutive health surveys had a lower risk of mortality over 6 years of follow-up. The unique contribution of health transition to mortality is independent of objective health measures such as chronic conditions. The objective was to examine whether there existed differences in health-related quality of life (HRQOL) between offshore islanders and Taiwanese residents of the same ethnic. Methods: The study design was a two-sample comparison at one point in time. A threestage sampling design was used to draw subjects, with sampling rate proportional to size within each stage of sampling for each sample. A total of 187 residents aged 18 years old and over from Feng Yuan city, located in central Taiwan, and 473 offshore islanders from Penghu island were randomly selected in 1997. Face-to-face interviews by public health nurses were used to ask questions regarding sociodemographic factors, chronic condition, smoking, alcohol drinking, and short form 36 (SF-36). The SF-36 physical (PCS) and mental (MCS) summary measures were derived. Regression models estimated the effects of offshore islanders on SF-36 by comparing the scores of offshore islanders with Feng Yuan residents. Results: The crude scores of physical functioning and general perception of health were lower in Penghu islanders than in Taiwan residents while the crude scores of role limitation due to emotional problem and MCS were higher in Penghu islanders than in Taiwan residents. After adjusting for the effect of age, gender, marriage, chronic condition, disability, ridden-bed status, smoking, and alcohol drinking, the estimated effects of offshore islanders on scales of role limitation due to physical and emotional problems, and social functioning were positive, often as much as a 5-8 point increases on the 100-point scale. Conclusions: Our findings indicate there existed better mental and psychosocial functioning in offshore islanders than Taiwan residents. Aims: Most findings confirm that quality of life of the elderly is strongly related to their health and age. On the other hand the elderly in East European countries that have experienced socioeconomic transition could be affected by number of other adversities such as poor financial situation or social exclusion. The aim of this study is to investigate whether the quality of life of the elderly in the Czech Republic is more related to biological, psychological or social factors. Methods: As a part of the WHOQOL-OLD Project, a representative sample of 325 older adults living in Prague completed the quality of life assessments (WHOQOL-BREF and WHOQOL-OLD), the Geriatric Depression Scale (GDS) and Socio-demographic form. The samples age and gender parameters corresponded to the structure of Prague population aged 60+. Results: Nearly 60% of the sample were women, mean age was 71, with range from 60 to 93 years. More than half of respondents (68% of men and 42% of women) lived with a partner; most of them (93%) lived in their own household. According to their report, financial situation was significantly worse in women than in men. Women also reported worse health condition than men. To identify determinants of various aspects of quality of life we ran regression analyses separately for each of the six WHO-QOL-OLD and four WHOQOL-BREF dimensions as well as for the overall quality of life item. The predictors included gender, age, marital status, education, number of children, previous employment position, financial situation, health status and depression score. Surprisingly the age and health status predicted only one dimension of the WHOQOL-BREF (Health) and two dimensions of the WHOQOL-OLD (Sensory Abilities and Social Participation). The strongest predictor across all dimensions was depression. Conclusions: The results confirm the importance of interventions and programmes promoting psychological wellbeing and reducing negative feelings. Supported by NO/7695-3. Aims: Due to better early detection and treatment interventions, there are increasing numbers of cancer survivors. Despite their expanded lifespan, many cancer survivors harbor concerns about their future health status, but no standardized questionnaire has been reported to measure such concerns. This study examined the validity of a newly-developed questionnaire to assess worries in cancer survivors. Methods: Cancer survivors were identified through a population-based tumor registry. Eligibility criteria included diagnosis with localized disease 18-30 or 60-72 months pre-assessment and currently disease free. Data were collected through mailed questionnaires and findings are based on total of 753 (57% response rate). Cancer worries were assessed through the ASC (Assessment of Survivor Concerns), a 6-item questionnaire with items developed based on input from survivors and the literature. It contains two 3-item subscales: cancer concerns and general health concerns. Questions are answered on 4-point scales. Reliability was assessed through coefficient alpha computation, convergent and divergent validity through correlations with other established scales (PANAS and CES-D) and construct validity through multiple group confirmatory factor analysis (CFA). Results: Responses were distributed across the scale, with a mean of 2.6 (SD=1.1) for cancer concerns and 1.6 (SD=0.6) for health concerns. Coefficient alphas for the cancer concern and health concern subscales were 0.93 and 0.63 respectively. Correlations between the two subscales and the corresponding PANAS and CES-D scales were in the predicted directions, significant (p=0.001) and moderate (0.24-0.45). CFA models showed strong support for the hypothesized structure of two distinct but correlated cancer worry subscales (CFI=0.99). Conclusions: The ASC cancer concerns subscale is recommended as an additional questionnaire in studies that assess quality of life in cancer survivors. It provides a valid assessment of worries important to this burgeoning group. Exploring interventions to address these concerns and their effects on other health outcomes is an important public health issue, given that cancer survivors now make up a substantial portion of the general population. (Novik et al., 2004) . Results: Patients who had 1-4 mild symptoms experienced no QoL impairment. Patients with 5 and more mild symptoms had mild QoL impairment (<25% decrease from population norm -PN). The presence of 1 moderate-to-severe symptom, if others were mild, resulted in mild QoL impairment; 2 symptoms -moderate QoL impairment (25-50% decrease from PN), 3 and 4 symptoms -severe QoL impairment (50-75% decrease from PN). Patients with 5 and more moderate-to-severe symptoms had critical QoL impairment (>75% decrease from PN). Conclusions: If all symptoms are mild, the presence of 5 and more symptoms in advanced cancer patients is crucial. If at least two symptoms become moderate-to-severe, significant QoL impairment takes place. The presence of 5 or more moderate-to-severe symptoms results in critical QoL impairment. The better understanding of the symptom interference phenomenon with QoL in cancer patients is helpful to determine the priorities in symptom management and QoL improvement. Aims: Describe the symptom experience of women following treatment from breast cancer and examine the relationship of symptom impact on functional status, quality of life (QOL), and future perspective. Methods: A population-based sample of women in two cities in the U.S. with Stages 0-II breast cancer (n=1555) completed a survey including the European Organization for Research and Treatment of Cancer QOL Questionnaire and the Breast Cancer-Specific QOL Questionnaire. Analyses were restricted to women who completed primary treatment (n=1372). We examined the presence and relative frequency of 11 symptoms scales/items and tested for associations with 7 QOL measures [5 functional status scales (physical, emotional, role, social, cognitive) , global QOL and future perspective]. Regression models controlled for sociodemographics, co-morbidities, treatment, and time from treatment initiation. Results: Mean time from initial treatment was 7.1 months (range 0.5-14.9). Over 90% of participants reported having >2 symptoms. The 5 most common symptoms/scales reported were systemic side effects (87%), fatigue (80%), breast symptoms (72%), insomnia (58%), and arm symptoms (53%). There was a statistically significant association between number of symptoms and QOL across all 7 dimensions, such that for each additional symptom, the average QOL measure decreased by at least 4 points (scale range 0-100) after adjusting for covariates (all p-values < 0.0001). Fatigue had the most impact on QOL with statistically significant association across all 7 dimensions (all p-values < 0.001). Systemic side effects, pain, insomnia, and breast symptoms were associated with QOL across three-five dimensions. Taken together, women's reported symptom experience accounted for between 23% and 54% of the variance of QOL across the 7 dimensions. Conclusions: More attention to the reduction and management of disease and treatment-related symptoms could improve functional status and QOL and potentially impact resource utilization for women with breast cancer. The results indicate that nurses, in spite of many obstacles, struggle for an ideal of promoting dignity for the dying. An important part of this ideal is to provide the physical frames for a peaceful death, which seems to be tentamount to a single room and an environment not dominated by technology. Furthermore, it implies ensuring optimal pain relief, providing high quality of physical care, taking good care of the patients' relatives and carrying out specific routines after death. However, these efforts seem to interfere with other concerns, which may be the source of ethical dilemmas. The nurses feel they are squeezed between palliation and treatment, between effectiveness and reflection, between patient and his relatives and between different relatives. Conclusions: The partly incompatible concerns mentioned above prevents the nurses from taking the individual patient's autonomy and preferences into consideration, which may tend to transform death into a stereotype event in the public health care system. Aims: Health-related quality of life (HRQOL) has been shown to be an important predictor of survival in patients with cancer. The aim of this study is to evaluate the role of HRQOL in predicting disease-free survival in newly diagnosed prostate cancer patients. Methods: Data from CaPSURE, a longitudinal disease registry of men with prostate cancer, was used to identify a group of men who were newly diagnosed with prostate cancer, underwent radical prostatectomy without any additional treatments, and had a pre-treatment QOL assessment. HRQOL outcomes were assessed at lower 10th percentile and the remainder of the patients. Univariate and multivariate survival analyses using Cox proportional hazards regression models were conducted to determine whether baseline HRQOL outcomes were independent predictors of disease-free survival for HRQOL alone and controlling for age at diagnosis, risk classification, race/ethnicity, and education. Results: 1103 patients met the study criteria, with 98 (8.7%) of patients having disease recurrence. Median disease-free survival times differed from 24.5 months in patients who did not have recurrence to 13.6 months in patients with recurrence. In univariate Cox proportional hazards regression models Physical Function, Vitality and Mental Health scales were associated with diseasefree survival (p-value < 0.05). A borderline significance was found for Role Emotional (p-value 0.06). In multivariate analyses controlling for clinical risk classification, age at diagnosis, race/ethnicity and education only Mental Health remained in the model with HR 1.79 (95% CI: 1.1-2.9). Conclusions: In addition to known clinical factors quality of life outcomes appears to be a significant prognostic factor for survival. Future evaluation of the mechanism underlying this relation is needed. Aims: Current studies assessing the impact of neo-adjuvant therapy for rectal carcinoma focus on toxicity and survival. Little is known about the impact on perceived QoL. Therefore, the aim of this study was to assess quality of life over time among patients receiving neo-adjuvant radiation and chemotherapy for rectal carcinoma and patients without neoadjuvant treatment at all. Methods: Since 1998, the EO-RTC-QLQ-C30 and a tumor specific ad-hoc module were prospectively administered to patients before surgery, at discharge, 3, 6, 12 and 24 months postoperatively. Altogether, 368 patients, who provided QoL data for at least one time point, were included in the analysis. Results: Out of the 368 patients, 183 were female and 185 male. Both groups were comparable in terms of age, tumor stage and surgical procedure. 16.8% of the male and 7.7% of the female patients received neo-adjuvant radiation and chemotherapy. In terms of general QoL (EORTC QLQ-C30) we found no significant differences preoperatively. At discharge, significant differences were found for Social Functioning, Global Health, Pain, Insomnia and Appetite Loss with patients following neoadjuvant treatment reporting poorer QoL. After 6, 12 and 24 months postoperatively there is no longer a difference. On the QoL module, patients following neo-adjuvant therapy reported significant higher levels of sexual impairment and strain. In a multivariate analysis we found out that women following neo-adjuvant therapy tend to report worse QoL scores (except for diarrhoea) than men, but these findings should be re-analysed with bigger sample size. Conclusions: Neo-adjuvant therapy seems to have positive effects in terms of downstaging. These positive effects however lead to a temporarily impairment of QoL especially in terms of sexuality: the sexual impairment remains stable over the time observed, while other aspects of QoL improve over time. This should be included in the preoperative counseling. Aims: The study aimed to finalize the Russian version of PedsQL by developing the forms for children aged 2-4 and 5-7. The results for the forms for children aged 8-12 and 13-18 were reported previously. Methods: The following samples were included in the study: children aged 5-7 with respiratory diseases (pneumonia, asthma, bronchitis, etc.) (n=146) and their parents (n=146); healthy children aged 5-7 (n=35) and their parents (n=35); parents of children aged 2-4 with respiratory diseases (n=130) and parents of healthy 2-4 yrs children (n=88). Results: After cultural and linguistic adaptation of the forms psychometric properties were tested. Cronbach's coefficient exceeded 0.7 for most of scales for child-and parent-reports. For Total Scale Score it exceeded 0.9 for both forms. The validity was tested and successful results were founded for all forms: QoL scores in healthy children were higher than in children with respiratory diseases (Student's T-test with Welch correction, p<.05); QoL scores in children moderately correlated with annual acute sickness rate (Spearman r=0.66, p<0.05). Additional test of validity for the forms for children 5-7 yrs with respiratory diseases revealed moderate correlations between scales of child-and parent-reports (Spearman r=0.56, p<.05). Sensitivity to changes was assessed by comparison of QoL scores for child-and parent-reports in acute disease period and in recovery for children with asthma (n=38) and pneumonia (n=62): QoL scores were higher in children after recovery than in acute disease period (Wilcoxon criterion, p<.05). Conclusions: Thus, the Russian forms of PedsQL (2-4) and (5-7) appear to be reliable, valid and sensitive to changes. Finally, the Russian version of PedsQL is now available for age groups (2-18) and may be used in research and clinical practice. Aims: Currently there are very few HRQOL instruments available, validated for Turkish children. The KINDL-R is a generic measure of children-and adolescents-(7-16 years) HRQoL, containing 24 5categorical items, assessing 6 dimensions of physical-, emotionaland social-wellbeing. The KINDL is available in many different languages. Within an elaborate translation procedure and cognitive focus group interviews, the KINDL was adapted into Turkish. This paper describes the first testing of the psychometric properties of the new Turkish KINDL-R in children. Methods: Within a school survey in Manisa 1918 children aged 8-12 years filled in the KINDL. Using the Multitrait/ Multi-Item Analysis Program (MAP) item-internal consistency and item discriminant validity was calculated to confirm the Instrument structure. Likert-scaling assumptions were tested. A confirmatory factor-analysis (CFA) was applied too. After the amendment of unsatisfactory items the KINDL was applied to a separate random selection of 104 children and the analyses were repeated. Results: The corrected item -total correlation for some item was below r=0.30. The wording of 2 items was changed according to suggestions from focus group interviews. After the amendment only 2 items displayed coefficients slightly below r=0.30. Cronbach a was 0.35-0.78 before and 0.54-0.78 after the amendment. MAP-scaling success was 60-100% before and 90-100% after the amendment. CFA confirmed the KINDL structure for the original version (RMSEA=0.08) less than for the modified one (RMSEA=0.06) although for the latter sample size was too low. Floor effects were negligible, ceiling effects were up to 19%.

The results indicate that the Turkish KINDL enables a reliable and factorial valid assessment of childrens HRQoL. The modifications have led to increased psychometric quality. Further research will focus on construct validity using additional information. Aims: Different IRT-models are available for constructing multiitem QOL scales. However the usage of only one model might be suboptimal. Recently some sequential applications of different IRT-models have been proposed which ideally should result in a set of items which are well-functioning with regard to every analysis. We investigated if the iterative application of different IRTmodels and CFA converges into stable results when applied to the KIDSCREEN psychological well-being scale. Methods: The pilot test of the KIDSCREEN project resulted in a data set of 3019 children and adolescents (8-18 years) from 7 European countries who answered the 11 items of the preliminary KIDSCREEN psychological well-being dimension. CFA was used to test unidimensionality and local independence (loading on common factor & residual correlation). The empirical tracelines were examined using TESTGRAF. Item parameters were estimated according to the generalized Partial Credit Model (GPCM). Items with low slope-parameter were discarded. A Partial Credit Model (PCM) analysis was conducted using WINMIRA. The Q-index was used for item exclusion. All analyses were repeated on the remaining items and then again on all initial items with the correlations between the different criteria across the items being calculated. Results: Some items were excluded during the sequential analyses. The remaining items fitted the data well according to all criterions (CFA-loading=0.78-0.85; residual correlation=0.02-0.19; GPCM-slope=1.36-2.20; Q-index p=0.117-0.761). The CFAloading, GPCM-slope and Q-index showed strong agreement. The maximum residual correlation contradicted the other criterions. Conclusions: The combination of the different models converged into stable results. Taking into account the different viewpoints of these methods, a scale with good psychometric properties for children could be constructed. Insight into how the items are actually functioning was enhanced.

O-4-B05/1733/ASSESSING PATIENT NEEDS AND VIEWS IN FAMILIES WITH CHILDREN AND ADOLESCENTS WITH CHRONIC HEALTH CONDITIONS ACROSS EUROPE Ute Thyen, Klinik fu¨r Kinder-und Jugendmedizin, Universi-ta¨tsklinikum Schleswig-Holstein Health care services for children with chronic conditions may vary across different systems of care and across conditions. We sought to describe the availability of services for children with a variety of chronic conditions in six European countries and measure satisfaction with care in a generic way across countries and conditions. We used clinical and sociodemographic data from the participants of a large European research project on health related quality of life in children with chronic conditions (DISABKIDS) and asked parents to resond to a newly developed, pilot-tested Inventory on Child Health Care, Satisfaction with Care and Unmet Needs (CHC-SUN), a 48 item tool with 21 items related to utilization and health service needs and 27 items related to satisfaction with processes of care 789 families agreed to answer the CHC-SUN. The sample includes families from Germany, the Netherlands, Sweden, France, United Kingdom and Austria. Conditions included asthma in all countries, and rheumatoid arthritis, diabetes, cerebral palsy, epilepsy, and skin disease. Patterns of care varied considerably among countries as regard to using regular primary care, emergency services, financial burden and satisfaction with care. Availability of school based services and integrated psychosocial care appeared to be the exception rather than the rule. Parents express moderate levels of unmet needs in those areas. Satisfaction with care varied across conditions with highest levels of satisfaction in parents of children with Asthma and Diabetes and lowest in Cerebral Palsy and Skin Disease. Parents reported highest satisfaction with doctors behavior and least satisfaction with the diagnostic process. Standardized parental reports may help to detect variations in health care delivery and impact on health care policies targeted at better care for children with chronic conditions. patients' QOL with that of the general U.S. population, and to identify factors associated with MS patients' QOL. Methods: Our study sample consists 2156 MS patients, identified via stratified random sampling of the National MS Society mailing list and via national outreach activities. The patients were from all regions of U.S., had various symptoms and all degrees of disability. QOL data were collected using SF-12. We compared the SF-12 data of MS patients to the U.S. norms and to the published data on patients with diabetes, congestive heart failure, myocardial infarction, hypertension or depression. To identify factors associated with QOL, we build regression models for SF12-derived mental and physical component scale scores (PCS and MCS), respectively. Results: 75% patients had MS for over 5 years, 30% had progressions or relapses in the past 12 months, and 72% needed help on Activities of Daily Living (ADL) or Instrumental Activities of Daily Living (IADL). The average PCS was 36 [standard error (SE): 0.27] in MS patients, 14-point lower than the US norms, and 4-12 points lower than patients with the five conditions; the average MCS was 49 [SE: 0.25], slimilar to all the comparison groups except for patients with depression, whose average MCS was 34 (SD: 0.7). After controlling for patients demographic variables, more severe pain, higher counts MS symptoms and more limitations on IADL were associated with worse self-reported physical and mental health; difficulty getting mental health care was associated with worse mental health. Conclusions: Patients' self-reported physical health is more adversely affected by MS than is self-reported mental health. To improve MS patients' QOL, health care providers should provide better pain management and better access to mental health care. Aims: Improving quality of life among the disabled is an important goal of Dutch health policy. Implicitly policymakers assume that social participation plays an important role and improves the subjective quality of life of the disabled. We have tested this assumption by relating objective and subjective aspects of the quality of life of both the disabled and non-disabled to a host of social and health related determinants. Questions: (1). How do the disabled and the non-disabled compare in both objective and subjective aspects of quality of life? (2). What are the roles that physical disabilities and socioeconomic factors are playing in quality of life differences? (3). How are social participation, perceived health and happiness related to each other among the disabled and the non-disabled? Methods: Linear structural equation modelling of data of a sample of 5,826 respondents, including 642 persons with physical disabilities, of the Dutch population. Among others the data include indicators of physical disabilities (OECD), and a range of objective (four types of participation) and subjective (SF-12, happiness) quality of life indicators. Results: In terms of the objective quality of life, the disabled are more at a disadvantage in labour participation than in social and cultural participation.

Regarding the subjective quality of life we find that the disabled are more likely to lag behind in perceived physical health than in mental health and happiness. Seriousness of the physical disabilities and the duration of illness are the main predictors of differences in labour participation and perceived physical health. This is much less true for perceived mental health and happiness, where a key role is played by socio-economic factors. In a multivariate model relating objective participation to subjective health and happiness, correlations are much lower than expected. Conclusions: It is striking to find that participation, perceived health and happiness are much less related than often assumed. Their determinants differ widely in nature and strength. The empirical model leads to rejection of the hypothesis that higher participation among the disabled is associated with higher subjective well-being. Aims: The increase in application of QoL research now requires more flexibility from patient-reported outcome measures (PROMs). Clinical trials focus on specific outcomes and need few subscales with good precision and range. Alternatively, observational studies and routine audit may require multiple scales and not need such precision. The aim of this study was to develop a new PROM for people with Friedreich's ataxia (FRDA) appropriate for all clinical research. FRDA is a progressive disabling neurological disorder affecting about 3000 people in the UK. Whilst clinician rated ataxia scales exist, none were developed using recognised methods nor incorporates patients' perspectives. Methods: There were two stages. First, patient interviews, expert opinion, and a literature review were used to generate a pool of items (n=160) forming the conceptual model of health impact of FRDA. Second, the item pool was pretested, modified accordingly, and then sent to and completed by 307 patients (65% response rate). Results: Rasch analysis of these data led to the construction of two PROMs: the long form Friedreich's Ataxia Impact Scale (FAIS-LF) and a short form version (FAIS-SF). Both include 8 scales (body movement, speech & swallowing, upper and lower limb functioning, ADL, isolation, mood, self perceptions). The former includes 117 items, the latter 60 items. Conclusions: The FAIS-LF allows for greater precision in studies of treatment effectiveness, whereas the FAIS-SF can be useful for general clinical studies. By using Rasch analysis, LF and SF scores are comparable on a common metric, thus helping to bridge the precision vs. pragmatism gap. Aims: Physical, mental and social factors are known to influence HRQL following a stroke. As part of a large 3-site inception cohort study on understanding the quality of life of individuals with stroke and their caregivers' the goal of this presentation is to characterize recovery in relation to HRQL after stroke and to determine predictors of HRQL that are amenable to intervention. Methods: In Montreal, 267 individuals with a confirmed stroke were recruited in the acute care hospital and interviewed by trained interviewers over the telephone at 1, 3, 6 and 12 months post-stroke using multiple measures including the Stroke Impact Scale (SIS) and the SF-36. Descriptive statistics and growth curve analyses were used to portray the results. Results: For the SF-36, recovery is most dramatic in role perceptions among survivors during the first 3 months but continues at a lower rate up to at least 6 months. Physical and social functions are resumed more slowly. For the SIS, mobility and daily living activities (ADL) improve most rapidly, while hand function and participation lag behind. The amount of recovery is influenced by three fixed (age, sex and stroke severity) and three potentially modifiable factors (comorbidity, social resources and depression). Mental health, age and stroke severity also influence the rate of recovery. Women score significantly lower on the physical component of the SF-36 over time, even after adjustment. Challenges at 6 months are still pervasive in mobility, ADL, communication, cognition and participation in usual activities. Conclusions: Treatment of depression, control or comorbidity and enhancement of social support have the potential of positively impacting the outcome of HRQL after stroke. Aims: To assess the content validity of the I-QOL, identify missing content for this incontinence sub-population, and assess the comprehension of the questions. This is particularly important in MS patients where cognitive deficits can be present. Methods: Two separate groups of in-depth interviews were conducted in MS patients and SCI patients with incontinence. The beginning of the interviews was designed to identify the major impacts of their disease on various aspects of quality of life. The second stage involved administration of the I-QOL followed by debriefing on individual questions, and detailed probing of terminology and content. Results: Nine patients with MS (7 women, 2 men) were interviewed with most patients identifying major impacts of their disease on quality of life already addressed within the I-QOL. Planning daily activities around bathrooms was common among patients. The MS patients reported that the questionnaire was relevant, thorough, clear, and there was no missing content which could be found from the probing prior to administration of the questionnaire as well as after completion. A total of eight SCI patients (4 women, 4 men) were interviewed with 50% of patients being wheelchair users and performing daily catheterization. This population also reported that the questionnaire was relevant, and there was no missing content identified. The SCI population also reported the questions were clear and easy to understand. While all patients understood the term 'incontinence' there was a consistent negative stigma with this term to old age. Conclusions: This initial qualitative research helped support the content validity of the I-QOL in neurogenic incontinence patients as represented by SCI and MS populations. We are investigating content validity in daily clean intermittent catheterization patients separately. Aims: Is agreement present between patient-and caregiverreported outcomes on 25 objective questions contained within the 51-item Schizophrenia Outcomes Assessment Project (SOAP) quality of life survey? Methods: 1500 communityresiding individuals with schizophrenia in 5 US states completed SOAP-51 survey at weeks 0, 4, 5, and 12. Previous factor analysis indicated 8 SOAP-51 factors with Cronbach a ranging from 0.73 to 0.94 and test/retest intraclass correlations at least 0.70 for all but 1 factor. An expert panel identified 25 SOAP items that could be objectively measured. This 25-item subset was given to each patient's primary caregiver concurrent with each patient's SOAP-51 administration. Caregivers were asked to answer each item in 2 ways:

(1) What is your objective response? and (2) Aims: A family of French generic self-administered HRQoL instruments for children and adolescents have been developed. Concurrently, proxy instruments have been built. The Vecu SanteṔ erc¸ue (VSP) project led to 3 different versions of proxy HRQoL questionnaires: a proxy-adolescent instrument, a proxy-children form and a common tool. The proxy questionnaires were based on the children/adolescent models. The aim of this study was to validate the 3 proxy versions of the VSP-A. Methods: This survey involved 1232 healthy and hospitalised children and adolescents and 1167 parents as proxies. Multidimensional construct has been assessed using MAP analysis, exploratory factor analyses (EFA) with procrustean rotations and confirmatory factor analyses (CFA). Rasch Rating Scale Model (RRSM) analyses allowed to test the scalability of each dimension. DIF by proxy (father or mother) has been tested relying on Zumbo's OLR. Cronbach's a and test-retest stability allowed to check the reliability. Comparisons with KIDSCREEN-52 proxy scores have been used to assess the convergent and divergent validity. Known group validity have been assess comparing scores between age, gender, clinical status, and social level. Results: EFA with procrustean rotation ensured the factorial validity of the 3 instruments. CFA indicated acceptable fit of the parent's data to the children or adolescent models (Adolescents: RMSEA=0.070, Children: 0.092, Common: 0.087). RRSM analyses results were satisfactory (INFIT: 0.80-1.30). No large DIF by proxy has been carried out (Delta-R 2 <0.4%). Reliability was good for the 3 tools (Cronbach's a: 0.70-0.92). Similar age, gender, clinical status, and social level as in the children and adolescent version were found in the proxies VSP-A domains. Convergent and divergent validity was supported by relevant correlations with KIDSCREEN proxy scores. Conclusions: These results support the reliability and validity of the 3 VSP-A proxy instruments. DIF study between proxy is a topic for future research. Aims: Building on a conceptual framework that proposed proxy health-related quality of life (HRQL) assessments may differ based on proxy viewpoint, i.e. as the patient would assess him/ herself (proxy-patient) or as the proxy views the patient (proxyproxy), this study in prostate cancer sought to determine whether HRQL assessments differed according to proxy perspective. Methods: Using a cross-sectional study design, prostate cancer patients and informal caregivers (dyads) were recruited from the Jesse Brown VA system in Chicago, Illinois. Dyads assessed HRQL using the EORTC QLQ-C30+3, and each caregiver provided an assessment both from the proxy-patient and proxyproxy perspectives. Congruence was evaluated using paired t-tests and intraclass correlation coefficients (ICCs). Results: Thirty-nine dyads were recruited as of April 2005. The majority of patients (74%) and caregivers (71%) were Afro-American, and 88% of caregivers were female. Proxy-patient and proxy-proxy assessments were significantly different [difference score(SD)] for emotional functioning [7.8 (13.4)], p<0.001, cognitive functioning [3.4 (10.9)], p<0.05, and sexual functioning [6.1 (13.9)], p<0.05. For the domains where significant differences were found, mean scores from the proxy-patient perspective were higher and closer to patient self-assessment than the proxy-proxy assessment. Conclusions: The perspective elicited from proxies can significantly affect the HRQL assessment they provide. In prostate cancer, caregivers perceived differences between the patient view and their own view of the patient's emotional, cognitive and sexual functioning. Further research that focuses on understanding the differences in proxy HRQL assessments based on perspective and implications for medical decision making is warranted.

O-5-A04/1808/COMPARISION OF CANCER PATIENTS QOL, MEASURED BY HYMESELFS AND PROXYS Francisco L. Pimentel, Health Sciences, University of Aveiro, Aveiro, Portugal; Alexandra Oliveira, Oncology, Hospital de Sao Sebastiao, Santa Maria da Feira, Portugal Aims: It is often difficult to determine the health related quality of life (HRQoL) of same cancer patients because psychological and congnitive problems limit their ability to complete questionnaires. A proxy, such as family member, may be able to give accurate information about patients HRQoL. Researchers and clinicians have been depending on proxies to assess HRQoL of cancer patients unable to provide the information. The aim of this study was to examine the agreement between cancer patients self reports and proxy-reports to evaluate quality of life (QoL). Methods: We use QLQ-C30 (Quality of life questionnaire of cancer) of the EORTC (European Organization for Research and Treatment of Cancer, Study Group on Quality of Life). Results: We found moderete to good agreement (r=0.49 for social function to r=0.79 for appetite loss) between responses from patients and those from their proxies for the QLQ-C30 scales. Statistically significant differences were found for physical functioning, emotional functioning and fatigue between patients and their proxies. Conclusions: In general, proxies rated the patients as having a more impaired QoL than the patients themselves. Significant other rated patients as having higher levels of disability and symptoms. The results suggest that proxies only provide valid and useful information about cancer patients QoL for QLQ-C30 domains where the correlations were high, however, should be used with caution particularly for psysical functioning, emotional functioning and fatigue, where de means difference were statistically significant. Aims: The Terri Shivo case highlighted the problems that can occur when proxies differ in their judgments about a patient's current or potential quality of life (QoL). This study assessed the accuracy of proxy judgments of individual QoL in healthy and chronically ill elderly couples. Methods: 80 elderly couples (65-90 years), married for an average of 42 years (range 23-61) completed the Schedule for Evaluation of Individual Quality of Life -DW. The sample, recruited through family practices in the Dublin area, consisted of: (A) 21 couples where both partners were healthy, (B) 20 couples where both partners had a chronic health disorder and (C) 39 couples where one partner was healthy and one had a chronic health disorder (cardiovascular, musculo-skeletal, cancer, respiratory etc.). In groups A and B, each partner independently rated their own QoL and then that of their spouse. In group C, the ill partner rated their own QoL which was then rated independently by their partner. Results: A two-factor ANOVA (gender and health status) on the absolute differences between self and proxy SEIQoL-DW index scores was not significant (F=1.5; df=5, 115; p=0.235). Male proxies significantly underestimated their wives' QoL in group C (healthy husband/ill wife: Mean absolute differ-ence=18, SD=13.9; t=2.6; df=9; p<.03). Wives accurately estimated their husband's QoL in all three conditions. Conclusions: We have proposed, elsewhere, that QoL is an individual phenomenon. Depsite this, couples who have been married for many years can generally provide accurate estimates of each other's QoL. The inability of healthy males accurately to estimate their wives' QoL points to the need for caution in some circumstances. While factors such as gender and differential adaptation may influence the validity and reliability of such judgements, proxies, if assisted by measures such as the SEI-QoL-DW, may be of greater assistance in clinical decision making than has previously been suggested in the research literature. Aims: To compare satisfaction and quality of life (Qol) in hemodialysis (HD) and peritoneal dialysis (PD) patients. Methods: The study sample consisted of all adult prevalent patients treated in 12 dialysis structures in Lorraine region on February 1st 2004. These patients were mailed a questionnaire built in 3 parts: (1) Dialysis satisfaction questionnaire: 41 items covering 7 dimensions: organization of care, medical problems management, information given by nephrologists, relation between nephrologist and general physicians, hospital environment, accessibility and overall satisfaction . Satisfaction scores were standardized from 0 (poor) to 100 (excellent). (2) Dialysis questionnaire: 68 items covering 5 chapters: medical care before regular dialysis, choosing the treatment for the kidney failure, transportation, employment, rehabilitation. (3) Generic and specific QoL questionnaires: SF12 with 12 items producing 2 summary scores (physical and mental) and KDQOL with 41 items covering 10 dimensions. Results: Respondents sample comprised 50 PD and 396 HD patients. Response rate was 44.3%. Mean age of patients was 65.9 years; 67% were men and 61% retired. Mean length of dialysis was 4.5 years. HD and PD patients were similar with respect to most demographic, clinical and dialysis variables except for age, length of dialysis, albumin concentration, physical activity: HD patients were significantly younger, longer on dialysis, had higher level of serum albumin and had more frequently activity (p<0.05). While both patient groups reported the same QoL in generic measure, effect of kidney disease on daily life was found to be more important in PD than HD patients with specific instrument (37.6 vs. 47.1, p<0.01). PD patients were overall more satisfied than HD patients as well as about care environment. These differences remains significant after adjustment for age, gender, length of dialysis, albumin concentration and physical activity. Conclusions: The fact that the PD patients rated their care better than HD patients suggests that information about the choice of dialysis modality is important. A specific instrument is useful while exploring differences between dialysis care modalities. Aims: The PCPT was a randomized, double-blind placebocontrolled study of the efficacy of finasteride in preventing prostate cancer. This abstract examines determinants of change in sexual function as measured by a Sexual Activities Scale [SAS] . Methods: Quality of life of 18,882 men was measured over a 7-year period. The SAS has four items with a total score ranging from ''0'' to ''100'' (higher, less positive). A mixed effects model was used to evaluate the effect of finasteride and baseline covariates, conditional on baseline sexual functioning: finasteride; baseline SAS score; age; race; SF-36 Mental Health Inventory (MHI), Physical Function (PF), and Vitality (VT) scores; body mass index (BMI); cigarette smoking status (never, former, current); presence of diabetes (DIA) & hypertension (HYPER); covariate time interactions. Results: There was no significant baseline-by-treatment interaction, allowing clear interpretation of the finasteride effect as an average baselineadjusted relative increase (worse sexual functioning) of 3.21 SAS points in the finasteride arm. This value is highly significant, but is not a large effect on a 0-100 scale. It is also small when compared to the maximum-likelihood estimate of withinperson change variability. Regarding other covariate effects, SAS would be expected to change 0.49 SAS points for each year of age from baseline; 0.53 points for former smokers (compared to lifetime non-smokers); 1.69 points for current smokers (compared to lifetime non-smokers); 3.06 points if have DIA; 1.14 points if have HYPER. Other covariates had negligible effects. Conclusions: In general, effects of finasteride on sexual dysfunction were small but additional analyses addressing clinical significance are in progress. Aims: Low HRQOL scores are associated with hospitalization and mortality in dialysis patients, but there is a lack of data about the association between HRQOL and mortality after succesful renal transplantation. The aim of our prospective cohort study was to asses HRQOL and its association with 2-year outcome in kidney transplanted patients. Methods: All patients 18 years or older (n=1067) who were followed at our university outpatient transplant clinic were approached to participate in the study. Baseline demographic characteristics and laboratory results were collected at enrollment. The patients completed a battery of validated, self-administered questionnaires including the Kidney Disease Quality of Life Questionnaire (KDQOL). Data on 2-year outcome (return to dialysis, mortality or combination of both) were extracted from the patients' charts. Statistical analysis was carried out using the SPSS software. Results: Mean age of patients was 49±12 years, 60% were male, the prevalence of diabetes was 17%. After 2 years the mortality was 5.2% and the prevalence of combined outcome (return to dialysis or mortality) was 9.6%. Patients who returned to dialysis or died during the 2 years had significantly lower initial QoL scores for most of the HRQOL domains (63±24 vs. 72±26 burden of kidney disease, 60±23 vs. 71±26 emotional wellbeing, 59±24 vs. 72±31 physical functioning; p<0.01, for all). In multivariate analysis burden of kidney disease were significantly and independently associated with combined outcome after adjusting for serum albumin, age and the number of comorbidities (p=0.04). Aims: Sexual dysfunction is a known complication of radical prostatectomy (RP). Normal sexual function (SF) encompasses libido, orgasm and erectile ability (quality and frequency). Our goal is to assess changes in global and specific domains of SF after RP. Methods: Validated quality of life surveys, administered to patients in CaPSURE, were assessed for changes in sexual domains in the following populations: all men, recipients of nerve-sparing surgery, and patients with high pre-operative SF. Results: Most men undergoing RP had compromised SF at baseline, most notably in frequency of intercourse and awakening with erections. After surgery, overall SF declined, sexual bother increased, and patients were most compromised in quality, frequency, and ability to get erections. Patients who received non-nerve sparing (NNS) RP had lower baseline pretreatment SF scores than patients who received nerve-sparing (NS) surgery. In a subset of patients who reported high pretreatment SF, patients who underwent NNS RP had greater percentage decline in SF item scores than those who had NS-RP. Conclusions: Men undergoing RP experience significant declines in all components of SF, not just limited to quality and frequency of erections. Such information may facilitate preoperative counseling with regard to SF outcomes and should be assessed for other managements of prostate cancer. Aims: Urinary incontinence (UI) has substantial and important impacts on health-related quality of life. The purpose of this research is to report the psychometric performance of 15 different language versions of the Incontinence-specific Quality of Life (I-QOL) measure, a patient-reported outcome measure specific to stress, urge and mixed urinary incontinence. Methods: The multinational dataset consisted of data from four clinical trials for stress incontinent females and from two additional population studies, enrolling women with stress, urge and mixed UI. All enrolled patients completed the I-QOL and comparative measures at baseline. The clinical trial populations had multiple administrations up to 12 weeks, and the two population studies included a shorter retest. Country-specific psychometric testing for validity, reliability, and responsiveness followed standardized procedures. Confirmatory factor analyses were performed to assess the I-QOL subscales. Results: The I-QOL measurement model was confirmed as three subscales. Summary and subscale scores for the 15 versions were internally consistent (alpha val-ues=0.91-0.96) and reproducible (ICC=0.72-0.97). Using changes in the independent measures of incontinence episode frequency standardized response means were predominantly strong (ranged 0.71-1.05) across 13 versions (out of 15) in association with these measures and effect sizes. These additional language versions of the I-QOL instrument demonstrate psychometric properties similar to the original version. Conclusions: The I-QOL has shown good results in both community studies and clinical trials with varying types and severity of urinary incontinence. It is a reliable and valid measure of HRQOL, suitable for use in a variety of international settings. Aims: To examine the existence of systematic differences in subjective health preferences compared to societal preferences for objective health status based on sociodemographic characteristics and chronic conditions in the US. Methods: Data were from the 2000 to 2002 Medical Expenditure Panel Survey (MEPS). Subjective health preference for current health status was measured using the visual analogue scale (VAS). Societal preference for current health status was measured using the US valuation of the EQ-5D. For comparability, individual EQ-5D scores were adjusted to the VAS scale using a linear transformation. Multivariate logistic regression was used to regress the dependent variable (EQ-5D>VAS scores=1; 0 otherwise) on sociodemographic characteristics (age, gender, race, ethnicity, education, income, insurance status and marital status) and 243 chronic conditions (a dichotomous variable for each ICD-9 code). Regression coefficients with p<.05 were interpreted as signifying a systematic difference in subjective preference for current health status compared to societal preference. Results: Individuals with higher levels of education were more likely and older, Hispanic and black individuals were less likely to have EQ-5D>VAS scores (i.e., the latter tend to subjectively rate their current health higher than society rates their objective health status). 33 of the chronic conditions were statistically significant in explaining the likelihood of having EQ-5D>VAS scores. For example, individuals with diabetes, hypertension, heart disease, multiple sclerosis and glaucoma were more likely to have EQ-5D>VAS scores (i.e., they tend to subjectively rate their health worse than society rates their objective health status). Conclusions: The results suggest that there are systematic differences in how individuals subjectively rate their current health compared to society that may be explained by age, race, ethnicity and the experience of certain chronic diseases. Aims: Utility values are often derived from average ratings from the general population based on direct elicitation from scenarios describing potential health states. In theory and some previous tests, people with experience of a condition provide higher utilities closer to optimal health values than those without such intimate knowledge. We hypothesized that people with disability (PWD) would follow this pattern, and that their higher values would be sensitive to distinctions of health status. Methods: We tested descriptive scenarios of 6 mobility impairment states with two groups: a general adult sample of people selected by randomdigit dialing (RDD), and volunteers who were adults with mobility impairments (PWD). 34 RDD and 54 PWD completed computer exercises using standard gamble (SG), time trade off (TTO), and a visual analogue (VA) preference rating technique. We also randomly varied health status (e.g., excellent versus good health). Results: Participants were 55% women, 21% African American, and had a mean age of 54.5 years. Across mobility scenarios, PWD gave preference ratings that were 0.02 (SG) to 0.09 (VAS) higher than RDD participants, although among 36 options of health status and rating technique, there were several comparisons that were reversed. Variation of health status resulted in counterintuitive results where RDD subjects were more consistent in rating the higher health state scenario with a higher rating compared to PWD. In an analysis that controlled for group demographic differences, utility differences persisted between PWD and the general population sample. Conclusions: Evaluations of disability states are substantially higher for PWD than in the general population. Applications of population values for resource allocations may need to consider differences based on PWD, especially for decisions that affect them directly. Aims: Generic preference-based measurements of health may not adequately cover important consequences for many diseases and conditions. The purpose of this study is to estimate a preference-based single index for the Asthma Quality of Life Questionnaire (AQLQ) from the UK general population for use in economic evaluations. Methods: AQLQ is a non-preference based condition specific measure of quality of life developed for people with asthma. A new classification was derived using AQLQ items selected mainly through Rasch analysis and other psychometric evidence, resulting in 5 attributes (concern for having asthma, shortness of breath, effect of weather and air pollution, sleep, and limitation in activities), each with 5 levels. A pilot study was undertaken on 100 respondents to compare different methods of explaining asthma to a general public sample: one was a 6-minute video explaining the mechanism of asthma and showing patients with asthma symptoms, and the other was a 200-word verbal description on asthma. In the main study a sample of 98 heath states was valued by 350 members of the UK general population. The time trade-off technique was used in an interview setting, and each participant evaluated 7 asthma states. Results: The pilot study found no significant differences between the two methods of presenting asthma. From the main study, a number of modeling approaches have been undertaken to predict health states valuations for all 3125 health states defined by the reduced classification. Models have been estimated using a random effects procedure on individual level data and OLS on mean health state. Alternative specifications of the model and transformations of the data will be examined. Final results will be presented. Conclusions: It is possible to estimate a preference-based scoring algorithms for conditions specific measures for use in economic evaluations. Aims: Racial and ethnic associated disparities have been observed in participation in randomized trials and adoption of medications in Human Immunodeficiency Virus (HIV) infection. The objective of this investigation is to determine if differences in preferences between ethnic groups exist that might partially explain apparent disparities in utilization. Methods: Using tablet computers, we measured patients' preferences for five different hypothetical life-courses consistent with HIV infection. Life courses had either a six-year or eleven-year duration. Health experiences changed over time and were described using the Corbin and Strauss model for illness impact. Patients reported their willingness to assume risk and trade off time to obtain perfect health for the duration of each life course. The population studied was a convenience sample of 407 English-speaking HIV+ infected patients recruited from university-affiliated HIV clinics in San Diego California. The sample was racially and ethnically diverse (57% Caucasian; 18% Hispanic), older (mean age 44 years) men (87%) and was well educated (72% with college education). Most were being treated with Highly Active Anti-Retroviral Therapy (84%). Results: African-American patients (AA's) were consistently less willing to take risk and less willing to trade time for improved health than other ethnic/racial groups. The median acceptable risk of death in the SG was typically 10% less in AA's than other groups and the median amount of time traded in the TTO was also about 10% less across all five states (all comparisons statistically significant by Sign-Rank test). The SG risk and TTO trade-off preferences of Asian patients were similar to those of Caucasians. Hispanic patients' preferences were similar to those of non-Hispanics. Conclusions: HIV infected AA's were risk-averse and less willing to trade time than patients from other racial groups. Such differences could contribute to observed disparities in the use of novel therapies and participation in randomized trials in AAs, as the AA's may perceive less benefit. Aims: Some quality of life instruments incorporate the respondents' values into their scores, using explicit ratings of importance, while others do not. It has been argued that the practice of weighting with importance ratings is cumbersome and unnecessary. The purpose of this paper is to examine whether weighting is justified, both from theoretical and psychometric perspectives. Methods: Two main questions will be addressed, based on the literature and data sets from patients with various chronic illnesses: (1) Does weighting have theoretical support; (2) Does weighting increase explanatory power. Results: (1) YES. Based on Campbell et al.'s (1976) theoretical framework and others, weighting is an implicit part of the process of making judgments about global QOL, which also has empirical support. (2) YES and NO -the answer hinges on the amount of variability in importance for the various aspects of life, which in turn depends on the following: (2a) Are satisfaction and importance so closely related that they give you essentially the same information? Examination of several large data sets has shown that importance and satisfaction ratings are NOT closely correlated. (2b) Are all the items very important, meaning we are multiplying by a constant, with virtually no effect on scores? Examination of 32 items aspects of QOL in a wellestablished measure revealed that that 73% of the items were rated ''not at all important'' by some people. Comparing weighted and unweighted scores showed a statistically significant difference that was also clinically meaningful (p<0.0001). (2c) Is there a different effect for group data vs. individual data? For individuals, weighting resulted in larger changes in scores, but for group data the effect of weighting was diminished. Conclusions: In some situations, weighting with importance is justified, on the basis of both theoretical and data-based considerations. In these cases, weighting will result in greater accuracy, because the values of the individual are reflected in the score. Aims: Local-item independence is a central assumption of item response theory (IRT). This assumption states that if the underlying latent trait is controlled for, item responses are independent and uncorrelated. However, item banks are typically constructed from responses to items that are presented in one fixed order; therefore, order effects between subsequent items might violate the independence assumption. In a prospective study, we investigated the effect of item order on item bank construction in general und item calibration in particular. Methods: 15 polytomous items similar to items taken from our computer adaptive test for anxiety were presented in one fixed order (group 1) or in a random order generated on the fly (group 2). A total of n=520 (n=239 group 1, n=281 group 2) psychosomatic patients participated in the study. Item analysis was conducted for each group along the line of our approach to item bank construction employed previously during our development of computer adaptive tests for anxiety, depression and stress. Results: Unidimensionality was tested by confirmatory factor analysis. One item in the fixed order group and one additional item in the random order group exhibited residual correlations greater 0.2 and were excluded from further analysis in both groups. Visual non-parametric analysis (Testgraf) of the remaining 13 items showed very similar item response curves between corresponding items of the two groups in almost all cases. Item calibration (generalized partial credit model) yielded only small differences between corresponding slope (absolute mean deviation for all 13 items 0.25±0.19) and location parameters. Conclusions: These results indicate that item order has little impact on item calibration and item bank construction. Moreover, these results provide evidence pertaining to possible concerns over the effect of item presentation in different orders in computer adaptive testing. Aims: To demonstrate use of a rapid hybrid approach to DIF detection by determining whether FACT items exhibit DIF. Methods: The FACT-G was administered to a convenience sample of 1615 patients with 5 diseases (273 HIV, 533 breast cancer, 258 colon cancer, 235 head and neck cancer, 316 lung cancer) in 3 cities (Chicago, Atlanta, and San Juan, Puerto Rico). To assess DIF we used a hybrid approach combining item response theory (IRT) using PARSCALE and ordinal logistic regression (OLR) using STATA. We analyzed the 4 domains of HRQL assessed by the FACT (physical well-being, PWB; social and family well-being, SFWB; emotional well-being, EWB; and functional well-being, FWB) for DIF related to language, race, Hispanic ethnicity, edu-cation, literacy, age, and gender. We determined the number of items flagged with DIF, the mean change in unadjusted and DIFadjusted IRT scores, and the 5 largest individual differences in unadjusted and DIF-adjusted IRT scores. Results: There were 28 analyses (4 domains times 7 covariates). DIF was found in 0 items in 17 analyses, 1 item in 8 analyses, 2 items in 1 analysis (DIF related to age in EWB), 3 items in 1 analysis (DIF related to Hispanic ethnicity in PWB), and 5 items in 1 analysis (DIF related to language in PWB). Mean differences between unadjusted and DIF-adjusted IRT scores ranged from negligible differences of 0.01 standard deviations (SDs) to modest differences of 0.16 (DIF related to language and education in FWB) or 0.17 (DIF related to race and literacy in SFWB) SDs. The largest individual changes due to adjusting for DIF ranged from only 0.10 SDs (DIF related to literacy in EWB) to 0.90 and 0.94 (DIF related to language and education in FWB) SDs. Conclusions: DIF analyses have been limited by labor intensive techniques. As the hypotheses we test regarding HRQL across different demographic groups become more complicated, the need for valid and rapid DIF detection increases. Our OLR-IRT method represents a viable approach to this need, with answers in hours rather than days. Aims: Cognitive decline is a common neurological problem. As such its measurement in clinical trials and practice is important. The Multiple Sclerosis Neuropsychological Screening Questionnaire (MSNQ) is a 15-item self-report scale for quantifying cognitive impairment. Patients must rate each item by a combination of its severity and frequency. This process is potentially complicated and confounding. Previous studies have shown high correlations between frequency and severity scores, suggesting it is unnecessary to measure both. However, correlations are based on person scores. It is also important to demonstrate that the instrument performs the same when measuring the two constructs. Using Rasch analysis, it is possible to compare the performance of the scale when measuring frequency and severity, separately from observed scores. Our aim was to test empirically whether it is legitimate and necessary to combine the measurement of frequency and severity. Methods: We modified the MSNQ to include separate response options for frequency and severity (15 items for frequency then same 15 items for severity) and administered this 30-item scale to 200 patients with MS. Data were Rasch analysed to determine whether items perform similarly when measuring the two constructs. We then tested whether frequency and severity measurements of people differed (agreement between scores). Results: All items performed the same in terms of item difficulty, differential item functioning, equating of frequency and severity subsets, when rated by frequency and severity. Estimates of peoples cognitive difficulties in terms of their frequency and severity were near identical. Conclusions: Results suggest it is unnecessary to combine frequency with severity, or measure the two separately. This implies that rewording of the MSNQ instructions and response options, to refer to either frequency or severity, would simplify the instructions, reduce measurement error, and improve validity. Aims: The SF-36V2 VT scale consists of four items concerning vitality and fatigue (V&F), hypothesized to measure a unidimensional construct. The purpose of this study was to evaluate the construct validity of the SF-36V2 VT scale in relation to 12 items from other commonly used V&F measures (Schwartz Cancer Fatigue Scale; Fatigue and Maastricht Questionnaires; Multidimensional Fatigue, Brief Fatigue, Multidimensional Fatigue Symptom, and Fatigue Assessment Inventories). Methods: We analyzed online responses from 900 subjects for whom 81% had a chronic disease. The sample was 48% male, 88% Caucasian, and evenly distributed through the ages of 18-65+. A cross-validation of exploratory and confirmatory factor analyses (FA) was used to evaluate dimensionality. Next, item response theory (IRT) models were analyzed to examine measurement properties of three item banks: fatigue only (12 items), SF-36V2 VT (4 items; 2 fatigue & 2 vitality), and V&F combined (all 16 items). Examining the three banks provided enhanced information on conceptualizing V&F as an appropriately combined construct. Results: FA confirmed three primary latent measures (factors) of V&F, with the SF-36V2 VT items loading on two of these factors. Of note, frequency and severity worded items loaded on the same factors. Each of the three primary factors had strong loadings (0.83-0.95) on a single secondary factor of V&F. All three item banks had appropriate fit to Rasch IRT models. For people with little fatigue, the SF-36 item bank and the combined V&F item bank provided more information than a fatigue only item bank. Conclusions: Findings reveal that SF-36V2 VT items load on two of the three primary factors of V&F, and are equally important in assessing overall V&F as the other 12 items from various fatiguespecific measures. Moreover, FA revealed that there is no difference in measuring fatigue with either severity or frequency worded items. Summarily, the SF36V2 VT domain demonstrated strong construct validity within the construct of V&F.

O-6-A05/1437/INVARIANT MEASUREMENT OF SUB-JECTIVE HEALTH OVER TIME -A RASCH ANALYSIS USING ADOLESCENT DATA Curt Hagquist, Social Sciences, Karlstad University, Karlstad, Sweden Aims: In monitoring population health across time possible instability of the psychometric properties of the survey instrument may be a cause for concern. Lack of invariance implies that comparisons in health between years of investigations are likely to be distorted. In order to meet the measurement requirement of invariance the operating characteristics of the items have to work in the same way across time. The purpose of this paper is to examine the psychometric properties of an eight-item scale of subjective health complaints, focusing on its ability to tap valid and reliable information about changes in health over time. Methods: The analysis is based on cross-sectional data collected at five points in time during 1988-2002. The sample comprises 14,000 students in 9th grade (15-16 years old) in a Swedish county. The data collection involved a self-administered questionnaire for completion in the classroom. The composite measure of subjective health consists of the following items: had difficulty in concentrating, had difficulty in sleeping, suffered from headaches, suffered from stomach aches, felt tense, had little appetite, felt low and felt giddy. The response categories for all of these items, which are in the form of questions, are: never, seldom, sometimes, often and always. The instrument is examined with the Rasch model in which the property of invariance comprises an essential integral part. Results: The analysis shows that the data conform to the Rasch model. On a general level of analysis the eight items show good fit across the latent trait, i.e. the measure works properly among students with good health as well as among those with bad health. Importantly, the items do also work invariantly across time, i.e. there is no differential item functioning with respect to year of investigation. Conclusions: This stability turns out to be a remarkable result since the time period of study is characterised by big changes in the Swedish society. It is concluded that the eight-item scale is an appropriate measure for re-current monitoring of subjective health complaints among adolescents. Aims: To assess the impact on health status of common mental disorders in the European general population. Methods: Data come from the European Study of Epidemiology of Mental Disorders (ESEMeD) project, a cross-sectional study of a representative random sample of 21,425 non-institutionalized inhabitants, aged 18 or older, from the general population of Belgium, France, Germany, Italy, the Netherlands and Spain (2001) (2002) (2003) . ESEMeD was designed to evaluate prevalence, comorbidity and quality of life in common mental disorders, among other issues. The WHO World Mental Health Survey Initiative version of the CIDI was used. The EQ-5D index, ranging from )0.54 (=worst impairment) to 1 (=perfect health), was calculated based on the UK tariffs developed by Dolan from time trade-off methods. All analyses took into account the complex sampling design of ESEMeD. Results: The prevalence of 12-month mental disorder/s was 9.6%. The mean EQ-5D index score for individuals with any 12-month mental disorder was 0.816 (95% CI 0.803-0.830), for those with any chronic physical condition (not exclusive of mental disorders) was 0.882 (95% CI 0.875-0.889) and for those with no metal nor physical disorders was 0.978 (95% CI 0.973-0.982). By type of mental disorders, mean scores were 0.769 (95% CI 0.746-0.792) for mood disorder/s, 0.808 (95% CI 0.790-0.826) for anxiety disorder/s and 0.893 (95% CI 0.866-0.920) for alcohol abuse/dependence. The five disorders with greatest impact on health status were dysthymia, generalized anxiety, panic, major depression episode and heart disease. Conclusions: Compared to chronic physical conditions, 12-month mental disorders have a greater health impact. Given their high prevalence (and relative low access to care), mental disorders emerge as a very relevant source of need for health care in Europe. Aims: The Hamilton Depression rating Scale (HDRS) is widely used to evaluate antidepressant treatment. However, research has demonstrated that the HDRS contains varied items proximally and distally associated with major depressive disorder (MDD). The objectives of this study were to develop and evaluate the psychometric characteristics of an item response theory (IRT) based short form using HDRS and Montogomery-Asberg Depression Rating Scale (MADRS) items. Methods: This was a secondary analysis of data from two antidepressant trials (n=1027). In both trials, patients with MDD were treated with either placebo, paroxetine or an experimental antidepressant. We performed exploratory factor analysis to examine unidimensionality and then IRT analyses, using a graded response model, on the combined baseline HDRS and MADRS items. Based on clinical recommendations, exploratory factor analyses and the IRT analysis, we developed 3 short form measures, version 1 (DV-1; HDRS items 1, 2, 7, 8, 10, 13 and MADRS item 8); version 2 (DV-2; MADRS items 1, 2, 3, 6, 7, 8, 9 and HDRS item 13); and version 3 (DV-3; HDRS items 1, 2, 8, 13 and MADRS items 1, 2, 3, 7, 8, 9) . We then examined internal consistency reliability, correlations among the instruments and responsiveness in comparing placebo and paroxetine. Results: The 3 short forms, HDRS and MADRS all had good internal consistency reliability (all >0.84). The short form measures were correlated with HDRS (p<0.001), MADRS (p<0.001) and Q-LES-Q scores (p<0.001). In study 1, neither the HDRS (p=0.62) nor any of the short forms (DV-1, p=0.50; DV-2, p=0.09; DV-3, p=0.08) detected statistically significant differences between the paroxetine and placebo groups. In study 2, statistically significant differences favoring paroxetine were seen for the HDRS (p=0.005), DV-1 (p=0.005), DV-2 (p=0.0003) and DV-3 (p<0.0001). Aims: Because depressed patients commonly experience reductions in motivation and energy, it is important to identify treatments that can restore their vitality. The Motivation and Energy Inventory -Short Form (MEI-SF), an 18-item acute variant of the 27-item MEI, was developed to facilitate such efforts. The current project was undertaken to build upon previous efforts to further evaluate the psychometric properties of the MEI-SF. Methods: Data from two 8-week double-blind, randomized clinical trials, comparing two antidepressants to placebo were utilized to evaluate the psychometric properties of the MEI-SF. In general, the data collected during the first and second trial were used for exploratory and confirmatory analyses, respectively. Psychometric analyses evaluated the structure, reliability, validity, and responsiveness of the instrument. Results: Factor analyses indicated a 3-dimensional structure best described the data:

(1) a global factor, including all 18 items, (2) a 6-item factor assessing cognitive or mental energy, and (3) a 5-item factor assessing social motivation. Item-level descriptive statistics, alphas, effect sizes, and validity correlations with other scales were all favorable. Most importantly, the MEI-SF demonstrated superior sensitivity to change: Guyatt's responsiveness statistics were very large (greater than 1) for the global and both subscale scores when used to compare HAM-D responders to non-responders. Conclusions: This study provides sound psychometric evidence for the MEI-SF. The content and responsiveness of this instrument make it an ideal measure to include in antidepressant trials of short duration. Future efforts will explore the use of this instrument in other therapeutic areas for which patient vitality is a primary concern.

O-6-B04 Aims: One important but often poorly captured area of functioning concerns an individual's ability to work. Several quantitative studies have now indicated that bipolar disorder (BD) can have a severe, and often enduring, negative impact upon occupational functioning. While this body of literature indicates that employment rates are low in people with BD, it throws little light on the specific ways in which this complex psychiatric condition can affect work, or upon how these effects are subjectively interpreted by individuals with the disorder. In order to further elucidate the relationship between BD and work, we report here on a series of qualitative interviews performed as part of the item-generation phase for a disease-specific scale, the QoL.BD. Methods: We conducted 52 interviews with people with BD (n=35), their caregivers (n=5) and healthcare professionals (n=12) identified by both convenience and purposive sampling. The affected sample came from a variety of employment situations, ranging between people with no employment history through to those in highly skilled professional positions. Interviews were tape recorded, transcribed verbatim and analyzed thematically. Results: Respondents described the different ways in which the symptoms of depression and hypo/mania impacted upon occupational functioning. Six themes emerged from the data relating to: lack of continuity in work history, loss of time, having too much or too little occupational routine, desire to engage in meaningful work, stigma and disclosure, and interpersonal problems at work. The majority (94%) of affected individuals talked about the role of work, or meaningful occupation, in determining QoL. Conclusions: The results highlight the complex, varied and intermittent effects of an episodic condition such as BD upon work functioning. A wide range of factors appear to contribute to ability to function in the workplace, pointing to the importance of developing more sophisticated and precise measures of occupational functioning for this population. Aims: The WHOQOL-Bref (Bref) is a generic quality of life measure, which has been developed simultaneously in many cultures and languages. It has 26 items covering four domains: Physical, Psychological, Social and Environment. Our main objective was to look at items exhibiting differential item functioning (DIF) by a depression factor. Methods: We investigated data from six countries (Australia, Brazil, Israel, Russia, Spain, and the USA) recently involved in a large international study, Longitudinal Investigation of Depression Outcomes (LIDO), assessing quality of life and economic aspects of undiagnosed depression among primary care patients. The sample consisted of 2359 subjects. The depression factor was defined by the diagnosis of depression using the Composite International Diagnostic Interview (CIDI) (n=1193). We have utilized the Rasch modeling as our method. Results: We found that 11 items of 26 items showed uniform DIF for depression factor and only one showing non-uniform DIF. To elucidate what is being measured by WHOQOL-Bref, quality of life (qol) or depressive symptoms, we divided arbitrarily the items into ''QOL items'' (15 items, related to overall quality of life, general health, pain, medication, safety, environment, body, finances, information, mobility, relationships, support, home, services and transport) and ''depression items'' (11 items, defined as those similar to DSM-IV criterion for depression and/or Hamilton Depression Scale items related to positive feelings, spirituality, think, energy, leisure, sleep, daily life activities, work, esteem, sex and negative feelings). Interestingly, only 3 of 11 ''depression items'' did not show DIF for depression, whereas only 5 of 15 ''QOL items'' showed DIF for depression. Conclusions: Our findings indicate a possible depression domain within the WHOQOL-Bref. In addition, as work continues including satisfaction within the quality of life construct, it may be affected by depressive symptoms. Methodologies were selected to test the underlying conceptual model that are tailored to the binary nature of the KENQOL data which comprises yes/no responses and the overlapping conceptual model; namely binary factor analysis with an oblique rotation. Results: The 112 question version tested among 447 community respondents was reduced to 83 items based on missing data, endorsement rates and item redundancy. A further 4 items were removed from the a priori model based on an assessment of Cronbach's a, removing each item in turn. The final Cronbach's a for the 79-item version a priori dimensions were Contentment 0.94, Corporeal Capacity 0.87, Co-operation 0.74, Completeness 0.69 and Cleanliness 0.68. Seven factor models were tested, and two alternative domain structures identified using binary factor analysis; namely a 1-factor and 3-factor model. The 1-factor model comprised 64-items, explained 44% of the variance, had a Cronbach's a of 0.96 and comprised items from each of the 5 conceptual a priori domains. The 3-factor model comprised 30 items and explained 22% of the variance. Factor One comprised 12 items (Cronbach's a 0.86), factor two 7 items (Cronbach's a 0.57), and factor three 11 items (Cronbach's a 0.84). The psychometric qualities of the two binary factor analysis models and the original a priori models were compared by assessing for construct and content validity, internal consistency and test-retest reliability. Each model was determined to have sufficient measurement qualities to warrant further testing using item response theory. Conclusions: We discuss the application of targeted binary factor analysis with oblique rotation by comparison to factor analysis models developed during earlier previous psychometric evaluation using standard factor analysis techniques with orthogonal rotation. Aims: Multiple quality of life (QOL) instruments are often used in longitudinal clinical intervention trials. Such instruments often include specific items that are not sensitive measures for the proposed interventions or the target populations. Resulting data may contribute to difficulties in detecting effects and patterns that are clinically important and otherwise, statistically significant. This paper describes an alternative technique (LIITE) that identifies highly sensitive items among the instruments. The research question was: does utilization of the LIITE technique result in the detection of significant effects compared with situations where utilization of all data does not detect such effects. Methods: The LIITE technique was developed and tested in a two-group Breast Cancer Education Intervention QOL trial (n=255). Three instruments were used: Brief Pain Inventory; Profile of Mood States, and QOL-Cancer Survivors. Instruments included 98 total items. We calculated and summed absolute change scores and determined item rankings. Thirty-six items common to both Experimental (EG) and Wait Control (WC) groups accounted for 50% of total change. Twelve items contributed to 50% of the variance for either the EG or WC. When combined, the total number of items increased to 48 and raised the percentage of variance accounted for to 63%. Between and within group analyses were performed at 3 and 6 months follow-up using the full data set and the LIITE scores. Results: At 3 months follow up, the results were equivocal regardless of which data set was analyzed. However, at 6 months, the LIITE technique identified continued improvement in the EG that would have been obscured if the full data set was used. Additionally, the LIITE detected significant change in the WC after receiving the intervention at month 6 that was undetected using the total number of items. Conclusions: The LIITE may have utility in increasing data sensitivity in QOL longitudinal trials. If future research confirms the efficacy of LIITE, smaller sample sizes may be indicated. LIITE may also be an important tool for efficient data mining in broad data sets, thus facilitating human intervention research. Aims: Because of its applicability to situations where patient preferences are unreliably reported, CA is aptly suited to systematic rating of patient treatment preferences in an AD caregiver population. By conducting a CA experiment of AD treatment scenarios with HUI-3 domains reflecting treatment attribute levels we assess subject preferences, trade-offs and willingness to pay for improvements in separate functional domains. Methods: After extensive pilot-testing, interviewers presented AD caregiver subjects with hypothetical AD treatment trade-off scenarios, along with background characteristics including caregiver burden, caregiver SF-36 and HUI-3 and (proxy) patient HUI-3. Because CA reflects a Random Utility choice theory rather than a Standard Gamble Expected Utility theory, the CA experiment independently tests the validity of the HUI scoring algorithm itself. Results: 74 AD caregivers enrolled in California AD Research Centers agreed to be interviewed. In a multivariate regression self-reported HUI was significantly related to SF-36 domains, ethnicity, caregiver burden, relationship to AD patient and working status, but not to age or education. In a multinomial logistic regression, CA treatment choice was strongly positively related to HUI score for the chosen intervention relative to alternatives, and negatively related to treatment costs (p<0.01). At the margin, caregivers were willing to spend an additional $5-$7 per month for an AD intervention that increased patient HUI utility scores by 1%. The treatment attributes with the strongest separate rankings include improvements in ambulation, emotion and cognition. The regression model based on the HUI multi-attribute score was superior to any single-attribute model and non-inferior to models allowing unconstrained parameter weights for each HUI attribute domain. Conclusions: CA provides a feasible and useful method for benchmarking the potential values for AD treatments and therapies trade-offs in terms of their costs and impacts on patient functioning. The results are consistent with utility theory, and provide an methodologically independent and innovative validation of the HUI utility scale as a strong predictor of subject health state preferences. Aims: Quality of life (QOL) has emerged as an important outcome of health care. Debate nevertheless continues regarding its clinical relevance. As shown in a previous cross-sectional study, symptom status has a significant impact on QOL in persons living with chronic illnesses. This paper reports the impact that symptom status has on QOL over time and discusses how symptom status can be used to translate the clinical relevance of QOL. Methods: A secondary data analysis using SEM tested the relationship between symptom status and QOL over time. The analysis included 367 HIV+ patients. The parallel process growth modeling approach was adopted for analyzing the data. To account for the non-normal properties of the indicators representing symptom status, analysis was conducted using WLSM chi-square statistic test in Mplus. Results: More than 83% of the patients had three or more completed responses at different times during the 13 years data collection period. There was a significant positive rate of change in symptom status and a significant negative rate of change in QOL over time. The rate of change in the symptom status had substantial negative impact on the rate of change in QOL across time. Also, patients who gained more than one symptom every 3 years had a faster declining rate of QOL than patients who gained less than one symptom every 3 years. Conclusions: In managing chronic illnesses, QOL is a critical outcome of patient care. Measuring QOL without reference to a conceptual model, however, hinders the development of a knowledge base for QOL research and subsequent translational application. The use of theoretically based models provide a mindset for interpreting the clinical relevance or impact that QOL has on patient care. Establishing an empirical relationship between symptom status and QOL begins to identify a means for translating the meaning of QOL for clinicians. This analysis demonstrated that symptom status had a direct relationship with QOL over time and therefore begins the dialogue about the usefulness of QOL for clinical decision-making and improved patient outcomes. Aims: Proxy HRQL assessment is often used as a substitute for patient self-assessment. Agreement studies between proxy-patient assessments often utilize Bland-Altman (B-A) plots. However, these plots are often misinterpreted when scales have fixed upper and lower bounds. We propose a new method to adjust for the agreement that occurs by chance alone. Methods: 10,000 randomly generated, independent pairs of HRQL responses bound between 0 and 1 were simulated and a B-A plot (average vs. difference) was generated. An empiric dataset of 96 proxy-stroke patient Health Utilities Index 3 (HUI3) assessments was examined. A new method was developed that considers the difference as a proportion of the maximum possible difference (PMPD) and thus adjusts for chance agreement. The results using this method were compared with other adjustment methods. Results: Due to range of scale limitations, B-A plots of rater concordance on HRQL assessment are subject to endof-scale bias as illustrated by the diamond shaped pattern from the simulated data. Similar patterns have led to erroneous reports of better agreement at the scale ends and worse agreement in the middle. Alternate adjustments (logarithms, etc.) did not adequately address this finding. By using the PMPD, the diamond shape is no longer apparent and the results no longer converge at the lower and upper scale ends. When applied to the stroke proxy-patient HUI3 dataset, the PMPD method provided a useful method to determine agreement beyond chance. Further graphical representations of our method will be presented. Conclusions: B-A plots are useful to examine agreement between proxy-patient pairs on HRQL assessments. However, due to the nature of HRQL assessment scales (a lower and upper bound), B-A plots are subject to misinterpretation. We have developed a method which adjusts for agreement by chance that is useful for HRQL assessments and could be generalized into other areas with similar data properties. Aims: In our previous work (Hawthorne & Richardson) we compared the AQoL, EQ5D, HUI3 and 15D in samples drawn from hospitals, outpatients and the community. This study extends this work through comparison of these instruments and the SF6D in a randomly sampled population study. Methods: A randomly sampled population cohort (n=3010) was administered the five leading utility instruments described above. Data were also collected on 17 health services used by participants. The results are analysed by instrument, health status, health service use and some co-morbidities. Results: In this large-scale head-to-head comparison, very clear differences between the utility instruments were observed. Some instruments performed very well, whilst others provided implausible utility estimates. The reasons for the differences and similarities are discussed. Conclusions: This study, the world's largest head-to-head comparison of utility instruments to date, suggests that although there are some similarities (thus giving confidence to utility scores) the differences, overall, in utility values from the five leading instruments are such that the cost-utility values for any particular study may depend more upon the choice of instrument than treatment efficacy. Aims: The target population for the US EQ-5D valuation study was the 210 million civilian, non-institutionalized adults who resided in the US in 2002. Hispanics and blacks were oversampled in the study, and sampling weights were developed to facilitate the unbiased estimation of population parameters. These weights yield estimates with excessive mean squared error for many socioeconomic characteristics when compared with other population surveys. The purpose of this research was to develop a new set of sampling weights that would produce more accurate estimates of population parameters. Methods: Design weights were developed as the inverse of the selection probabilities for households and persons. Adjustment factors based on Folsom and Singh's generalized exponential modeling procedure were applied to the weights to account for survey nonresponse, to post-stratify to known population control totals, and to control for extreme weights. Using the original and revised sampling weights, we compared weighted estimates to estimates from the US Census and other population surveys. Results: The original sampling weights yield reasonable estimates for age, sex, and the three largest racial/ethnic groups. However, the weighted estimates for other characteristics are less accurate. For instance, compared with the 2002 National Survey on Drug Use and Health, the US EQ-5D data underestimate the percentages of Asians (2.2 vs. 4.2%, p<0.001), college graduates (20.5 vs. 25.0%, p=0.012), and lifetime smokers (58.0 vs. 73.4%, p<0.001) and overestimate the percentages of unemployed (7.1 vs. 3.6%, p<0.001) and those with an annual household income <$20,000 (26.7 vs. 19.7%, p<0.001). Using the revised sampling weights, we observed substantial improvements in the accuracy of the weighted estimates. Conclusions: It is likely that researchers will seek to use the US EQ-5D valuation study data to examine relationships between socioeconomic characteristics and health state preferences. The work described herein will facilitate the more accurate estimation of these relationships. Aims: To compare health-related quality of life in Sweden and the UK using a standardised questionnaire. Methods: A standard questionnaire was mailed to randomly drawn samples of the general population in Sweden (S) and England & Wales (E+W). The questionnaire included EQ-5D plus items capturing comorbidities/health service usage. Returned questionnaires were received from 1945 and 1001 respondents respectively, representing comparable 65% response rates. Both datasets were weighted to be representative of their national populations. Results: Self-rated health status based on 0-100 VAS ratings (EQ-5Dvas)was generally higher for men than for women in both surveys. This position is reversed for women aged 70+ in the E+W survey. Differences in age/genderstandardised EQ-5Dvas between the 2 national surveys were small (typically <5). Whilst the gender gap remained constant with increasing age in Sweden, it increased in the E+W survey. Despite apparent convergence, the rates of reported problem on the 5 EQ-5D dimensions varied significantly both by gender and by country. The rates of problem on usual activities, for example, were 2.3 and 6.2% for men and women aged 20-44 in the Swedish survey but were 12.1 and 13.1% in the E+W survey . Within-survey regression models were constructed using EQ-5Dvas as the dependent variable and the recoded EQ-5D dimensions as 0/1 dummy variables. Both models fit the data reasonably well (r 2 > 0.450) with equivalent constants (87.9 and 89.5) however, the beta coefficients indicate large differences in the importance associated with each dimension. The highest decrements in the Swedish survey are for mobility (15.7) and pain/discomfort (12.0). The highest decrements in the E+W survey are for usual activities (11.4) and anxiety/ depression (9.5). Conclusions: The use of VAS ratings is helpful in demonstrating underlying differences in the value of dimensions of health. Intercountry comparisons of health status need to take account of systematic age/gender differences within national population surveys. Were similar findings to occur on a regional basis within countries, then this would have potentially serious consequences for the allocation of national healthcare budgets.

Aims: The validity of the HUI-3 and the SF-6D was determined in a sample of patients with latent and active tuberculosis (TB). We hypothesized that (1) latent patients would have higher scores than those with active TB, (2) patients with recent adverse drug reactions (ADRs) to TB drugs would have lower scores and (3) that agreement between the measures would be low. Methods: Patients attending an urban clinic with recently diagnosed latent or active TB (n=200) were given a selfadministered questionnaire that included the HUI-3, SF-6D, sociodemographics, and a drug toxicity index. Agreement between the utility instruments was assessed using the Intraclass Correlation Coefficient (ICC) with a 2-way mixed effect model (random subject effects and fixed instrument effects). Results: 170 respondents (96 female) completed the survey (85% response); 101 latent and 69 active TB patients. 40% of latent and 61% of active patients had a recent ADR to TB medications. The mean (SD) scores for latent and active patients for the HUI-3 were 0.90 (0.19) and 0.77 (0.30) and for the SF-6D were 0.82 (0.12) and 0.71 (0.16), respectively. Those active TB patients with recent ADRs had lower mean scores than those without ADRs (0.67 vs. 0.76 SF-6D; 0.73 vs. 0.80 HUI-3). The paired ICC between the HUI-3 and SF-6D for latent patients was 0.60 (95% CI 0.35-0.75) and 0.57 (95% CI 0.37-0.73) for active TB patients. Conclusions: These instruments appeared valid as patients with active TB had lower mean utility scores on both measures than latent patients, mean scores were lower in those patients with a recent ADR, and agreement was moderate between the two instruments. Aims: In this study, we compared VAS and 2 types of TTO valuations from parents for 10 Health Utilities Index mark 3 (HUI3) health states of hypothetical 8-year-old children. We estimated conversion (power) functions between the VAS and TTO valuations and compared the outcomes with Canadian HUI3 utility estimates. Methods: 153 parents of children aged 4-13 completed paper & pencil-VAS and Internet-based-TTO questionnaires. States were assumed to last 60 years. VAS anchors were most/ least-desirable; the relative preference for 'being dead' was assessed on a separate VAS. In 'TTO-dead', states were traded against perfect health/being dead; in 'TTO-ph/md' against perfect health/most disabled. Results: Mean age of respondents was 39 (SD 6); 88% were mothers. There were few missing answers (<3%). The VAS resulted in 1% inconsistent valuations; TTOdead in 6%; and TTO-ph/md in 5%; being dead in 15%. Interrater reliability as assessed by a G-study was 0.83 for VAS, 0.58 for TTO-dead, and 0.48 for TTO-ph/md ratings. 97% of the respondents scored most disabled £ dead; mean for most disabled was )0.32 (SD 18) if dead is set to 0. All Spearman rank correlations between the 3 methods and with Canadian utility estimates were >0.99 (p<0.01); person-mean TTO-dead scores corresponded best with the Canadian utilities (Intra Class Correlation 0.92; p<0.01). The power function between personmean VAS and person-mean TTO-ph/md was T=V (exponent 0.540) (adjusted R 2 =0.99); and TTO-dead T=V (exponent 1.512) (adjusted R 2 =0.98); unexpectedly, the TTO-dead ratings were somewhat lower than corresponding VAS ratings. Conclusions: In our setting, 2 types of TTO exercises were less feasible and yielded considerably less reliable results than the VAS. All methods, however, resulted in the same ranking order, and power functions could very well predict the person-mean TTO values from person-mean VAS scores. Aims: Most studies using the EQ-5D in the US have applied the UK preference weights, base on the assumption that preferences of these two populations differ minimally. We investigated the validity of a newly-developed US version of the EQ-5D preferences compared to the UK version. Methods: Data were collected from a randomized treatment trial for patients with HIV (n=1,126) in the US. The population-based preference weights for the US [Shaw JW. Medical Care 2005] and UK [Dolan P. Medical Care 1997] were used to score the EQ-5D utility index. We first compared the correlations of US/ UK versions of the EQ-5D index with the 10 subscale scores of the MOS-HIV, a HRQL measure for patients with HIV. Known-groups validity of the EQ-5D US/UK index scores were compared for clinical variables, including CD4 cell count (cutoff: 200 cells/mm 3 ) and viral load (cutoff: 5.8 log copies/ ml), and the MOS-HIV physical (PHS) and mental summary scores (MHS) (cutoff: median). Responsiveness of the EQ-5D index scores from baseline to Week 24 and 50 were compared for both versions using the standardized effect size (ES). Results: The mean US EQ-5D index score was slightly higher than that of the UK version (0.87 vs. 0.84 respectively). The correlation between the two versions was 0.98. Correlations of the US version with the MOS-HIV scores were slightly lower than those with the UK version, ranging from 0.49-0.64 vs. 0.50-0.69, respectively. Together the 10 subscales of the MOS-HIV explained 54% of variance in the US version of the EQ-5D utility and 58% of the UK version. Patients with lower viral load, higher CD4, higher PHS and MHS scores had higher EQ-5D utility scores than sicker patients for both versions (p<0.05). The changes in EQ-5D index scores from baseline to Week 24 and Week 50 were similar for the two versions (ES: 0.25 vs. 0.22 at Week 24, and 0.21 vs. 0.22 at Week 50 for US and UK, respectively). Conclusions: In this study comparing UK and new US weights for the EQ-5D utility index, we found that using the two versions had similar validity. The EQ-5D index scores generated using either set of weights can be used interchangeably to measure health utility in the US. Aims: An European trans-cultural HRQoL questionnaire for children and adolescents is being developed during the KID-SCREEN project. The aim of this study is the validation of the KIDSCREEN-27. Methods: The field survey involved 22481 children and adolescents across 13 countries (Ge, Sp, Nl, Au, UK, Fr, CH, Hu, Gr, Cz, Ir, Po and Sw). Construct has been assessed with MAP ananlysis, EFA (procrustean rotations), CFA and Rasch analyses. Zumbo's OLR method was used to assess the DIF across countries, by age groups and gender. Trait validity has been explored using CFA to test a Multi-Trait Multi Methods (MTMM) comparing children and proxy HRQOL reports. Discriminant validity has been assessed by score comparisons between age groups, gender, countries and social level. Reliability of the instrument has been assessed with Cronbach's a and test-retest score comparisons. Results: Results of EFA with procrustean rotations and MAP analyses are satisfactory. CFA indicates satisfactory fit (RMSEA=0.068, CFI=0.960). The unidimensionnality of every dimension has been confirmed (INFIT: 0,81-1,15). No DIF effect was found (DR 2 <0.4%) except for 4 items across countries. Reliability was good (Cronbach's a: 0,80-0,84; ICC: 0.61-0.74). MTMM model test is satisfactory (RMSEA=0.027, CFI=0.998). Discriminant validity was supported by significant difference between age, gender groups, and social level categories. Conclusions: This first European trans-cultural instrument to assess children and adolescents HRQOL is a new promising tool. O-7-03/1680/USING ITEM RESPONSE THEORY TO CROSS-CALIBRATE THREE MENTAL HEALTH SCALES Anastasia Raczek, Jakob B. Bjorner, QualityMetric Inc., Lincoln, RI, USA; David Cella, Center on Outcomes, Research and Education, Northwestern University, Evanston, IL, USA; John E. Ware, QualityMetric Inc., Lincoln, RI, USA Aims: To cross-calibrate three widely-used Mental Health scales (the EORTC QLQ-C30 Emotional Function, the FACT Emotional Well-being and SF-36 Mental Health scales) using item response theory (IRT) methodology and to evaluate to what extent original and calibrated scores lead to the same results in group-level analyses. Methods: We analysed responses from 1533 Cancer and HIV patients with complete scores on the three scales. We used confirmatory factor analysis for categorical data to evaluate dimensionality and used the graded response model for IRT analyses. For each scale, we calculated the expected scale score at each IRT score level and developed scale-to-scale correspondence tables. We generated calibrated and original scale scores for each respondent. To evaluate criterion validity, we selected two questions from the CARES-SF Psychosocial scale (''I frequently feel anxious'' and ''I have difficulty sleeping'') and compared group mean scores at each level of the items. Results: Factor analyses and IRT fit tests supported the hypothesis that the three scales generally measure the same concept and can be cross-calibrated. A few items did not fit the standard IRT model well, but could be fitted by adjusting the standard approach. IRT analyses showed that the EORTC covered a narrower range of Mental Health, but provided more precision within that range. The FACT and the SF36 performed similarly. SF36 scores calibrated to the metric of the FACT and the EORTC agreed well with the observed FACT and EORTC scores in criterion validity analyses. Conclusions: IRT based methods enables comparison of results from studies that used different instruments measuring the same underlying concept. Cross-calibration seems to work fine, even when some items do not show good fit to standard IRT models. Aims: US Spanish translations require a balanced input from each of the major US Spanish language groups (Mexican/Chicano, Central American, the Caribbean, and South American). Once the reconciliation, back translation, and patient interviews are complete, the resulting version demonstrates both equivalence with the original and appropriate cognitive equivalence to the relevant patient and language groups within the diverse US Spanish population. Subsequent steps often include additional reviews by developers, clinicians and proofreaders. Suggestions are frequently made by single reviewers who have their own personal language styles. If suggested changes are based on preference and do not represent an improvement on the existing translation, the quality of translations that have already been confirmed by multiple input can be undone. Therefore, we aim to propose the following criteria for assisting project managers in making decisions about these types of review suggestions. Methods: We reviewed criteria of the standardized translation methods, and identified four criteria to assist in decision making about the appropriateness of implementing suggestions made late in the process. Results: Criterion 1: confirm that the suggestion conveys the concept of the original questionnaire. Criterion 2: confirm that the suggestion is grammatically correct and is consistent with other items in the questionnaire. Criterion 3: confirm that the change will be universally comprehensible for the diverse US Spanish population by checking with reconciliation panel consultants. Criterion 4: determine whether the suggestion represents a significant improvement to the existing translation (clarity, style, common language) without adding ambiguity, controversial terms, or deviation from the intended content of the original item. Conclusions: If suggestions made by single reviewers late in the standardized translation process meet these criteria better than the original translation, then the change can be accepted. If there are doubts against any of these criteria, or if the existing translation offers no improvement in any of these categories, then the original version should be retained.

Aims: The purpose of this study was to describe health-related quality of life (HRQL) of Venezuelan pharmacy students and to explore the association among quality of life, health behaviors, and demographics. Methods: A random sample 171 of pharmacy students, ranging in age from 18 to 35 years were surveyed using a written questionnaire. HRQL was determinate using the Medical Outcome Study Short Form 36 (MOS SF-36).The associations among HRQL, demographics, and health behaviors were examined using both bivariate and multivariate models. Results: The sample consisted of 127 females and 44 males. The sample had a mean age of 22.3 years (SD 2.71 years). As expected the sample was healthy: only 11 subjects (6.47%) evaluated their health as poor, 20 subjects (11.7%) reported health deterioration during the past year, and 30 subjects (17.5%) reported to suffer from an illness. Forty subjects (23.4%) reported current medication use. The prevalence of alcohol consumption during the previous month was 65.8% and for smoking it was 15.5%. One third of the sample reported no exercising during the previous month. Multiple regression analyses were used to model HRQL score as a function of age, sex, income, illness, lack of regular exercise, alcohol consumption, and smoking. The regression model explained approximately 20% of the variance in HRQL. Controlling for other variables in the model, low income, illness, and smoking had a significant negative impact on HRQL. Lack of regular exercise and age were not associated with HRQL. Controlling for other variables male students had significantly higher scores in HRQL than female students. Alcohol consumption was associated with HRQL in bivariate but not in multivariate models. Conclusions: The overall quality of life of pharmacy students in Venezuela is good. This exploratory study demonstrates sex differences in perceived quality of life of college students. Quality of life is associated to certain predictors of future health status, including health behaviors such as smoking. Aims: Questionnaires within the same therapeutic area are usually translated at different times and by different vendors, resulting in translations inconsistent with one another. When used in clinical trials or studies, inconsistently translated questionnaires may generate confusion in respondents, especially when key terminology and response scales are affected. Our objective is to propose the implementation of translation memories (TMs), for use as an active reference to maintain consistency across similar key concepts, relevant terminology, and response scales. Methods: TMs are created by storing source text segments and the corresponding translations as source-target pairs into a database. At the beginning of a new project, the TM software compares the new source language text with this database and, where a match is found, retrieves previously validated translations. These are then provided to the in-country consultants to be incorporated into a standardized forward translation step. When the translation is finalized, TMs are updated to incorporate into the database any changes that may have occurred during back translation evaluation, debriefing, harmonization, and proofreading steps. Results: A single English word or phrase can generate multiple correct translations in the target language. For instance, the response scale option somewhat could result in multiple, equally acceptable Italian translations: un po', piuttosto, and abbastanza that reveal slight variations in meaning when back translated (a little, rather and enough.) Implementation of this consistency control system within the cross-cultural adaptation process preempts the need for discussions over which alternative to use in similar contexts. Conclusions: TMs facilitate consistent terminology for translations in the same therapeutic area and will consequently help reduce discussion and revision time, thus decreasing project management and consultant costs. Aims: In a Norwegian investigation among adolescents, where the overall aim was to develop methods to promote their quality of life, an adequate generic health-related quality-of-life (HR-QOL) measure was needed. A limited number of well validated instruments that measure HR-QOL in adolescents exist, and to date only a few has been translated into Norwegian. The purpose of this study was therefore to examine some psychometric properties of the first Norwegian version of a simple, generic, German HR-QOL questionnaire for adolescents, KINDL. Methods: The instrument consists of 24 Likert-scaled items, distributed into six subscales, which correspond to six domains of adolescents' health-related quality of life; Physical, Emotional, Self-esteem, Family, Friends and School. 239 healthy adolescents were enrolled in the survey. Reliability is expressed by Cronbach's a, which was computed for each subscale and for the total scale. Furthermore, the distributional properties were calculated for the items, the subscales and the total scale. Exploratory factor analysis, with oblimin rotation, was performed in order to evaluate the construct validity of the instrument. Results: The internal consistency reliability is acceptable for both the total scale and the subscales of Self-esteem and Family, fairly good for the Emotional subscale, but lower for the subscales Physical, Friends, and School. The distributional properties of the scales are satisfactory. Finally, factor analyses yielded interpretable solutions. The factor solutions at item level were interpreted to be in line with the original subscales, while factor analysis at subscale level indicated that a common quality-of-life core is involved. Conclusions: The Norwegian version of KINDL appears, in general, to be a psychometrically acceptable method of measuring HR-QOL in healthy adolescents. However, the alpha values of some of the subscales are not optimal, and these scales should be used with caution in research and profession. Still KINDL-N is considered suitable for screening purposes in the public health area and especially within school-health care. Aims: This study is part of a larger Norwegian study among adolescents, where the overall aim is to develop methods to promote their quality of life and to discover risk factors or threats to adolescents' well-being. The specific purpose of this paper is to describe the health-related quality of life (HR-QOL) in a group of healthy, Norwegian adolescents, and discuss possible implications for further research and clinical work. Methods: In this survey, the sample consisted of 229 adolescents (104 boys and 125 girls). HR-QOL was measured with the Norwegian version of the Child Health Questionaire (CHQ-CF87), and in addition demographic data were collected. Descriptive statistics were performed to describe the demographic conditions. Mean and standard deviations were computed for the total sample and for boys and girls separately. Further, t-test for independent samples was performed in order to describe differences between boys and girls. Finally, multiple linear regressions were performed in order to illuminate variations in the adolescents' quality of life. Results: The results show that HR-QOL among Norwegian adolescents is relatively high on average. However, the findings also reveal several areas of quality of life where the average measures are lower and where the adolescents seem to be more vulnerable, i.e. mental health, self-esteem, global health and bodily pain. There are also significant gender differences with girls scoring lower than boys on several of the scales of CHQ, ie bodily pain, behaviour, mental health and self-esteem. The background variables, parents' marital status, the adolescents' membership in an organisation and the presence/absence of serious illness, in statistical interaction with the adolescents' gender, was able to explain some variations in the adolescents' quality of life. Conclusions: This study shows that Norwegian adolescents in most cases have a fairly good quality of life, but in some areas an effort is needed to reach an optimal level of QOL in adolescents. Future research and clinical interventions should focus on the adolescents' psychosocial health and in addition be gender specific. Aims: Previously mandated to be the sole language in China, Mandarin Chinese was chosen over all other languages and dialects as a way to ensure cultural unification. Under the influence of history, politics, geography and religion, this unified language began to fragment, resulting in a wide array of culture-specific variations of Chinese languages and dialects. Our objective is to describe the impact of this language fragmentation on the development of Asian language versions of PRO measures. Methods: We evaluated primary differences in versions of Chinese spoken across various Asian countries (Singapore, Malaysia, Hong Kong, Taiwan and China.). Results: Several influences can be cited to explain the diversity of Mandarin Chinese in these countries, many related to the uniqueness of their cultural development. While cultures more homogeneous in terms of religion, stability of government, and international exposure (i.e. Taiwan, Mainland China) are usually paired with steadier and more linguistically consistent versions of Chinese, heterogeneous environments tend to develop a different style of language, as well as a multiplicity of coexisting languages. For example, language use in the former British colony of Singapore tends to be stratified by age cohorts, with the older generation speaking Chinese, the middle generation more comfortable with English, and a younger cohort of Chinese speakers now emerging from the schools. Countries with more international traffic tend to develop a variety of parallel languages (Malay, English, Tamil, etc.) which also add to the unique character of the Chinese spoken locally. Conclusions: A careful analysis of the specific population targeted for clinical trials should be the starting point to identify the choice of languages to be adapted in multicultural countries, in order to ensure a valid study outcome. Aims: Behc¸et's Disease (BD) is a multisystematic disorder characterized by oral and genital ulcers as well as cutaneous, ocular, arthritic, vascular, central nervous system and gastrointestinal involvement. Poor oral health due to recurrent painful oral ulcerations has been reported in patients with BD. This study aimed to investigate oral health in BD patients and assess the impact of recurrent aphthous stomatitis (RAS) on quality of life (QOL). Methods: A cross-sectional study was conducted on all members of the Japan Association of patients with BD. The questionnaire included QOL instruments such as the Short Form 8 (SF-8) and the General Oral Health Assessment Index (GOHAI). Development of the Japanese version of the GOHAI was previously reported at the ISOQOL's 11th Annual Conference. Oral status and symptoms as well as the disease severity in BD were also self reported. Results: A total of 883 patients with BD aged 17-91 years (mean 56.1, SD=13.7) completed the questionnaire (response rate=64%). 50% were female. Half of the patients reported poor oral health. 70% indicated that their mouth had affected their pattern of daily living in the past year. 83% reported having RAS in the past year. More cases of RAS occurred in patients younger than 40-years-old. The mean number of RAS occurrences per year was 11.7 (SD=28.0). The mean of total RAS duration per year was 58.9 days (SD=87.2). Tongue was the most commonly affected. The physical component summary (PCS) and mental component summary (MCS) of SF-8 were lower in the patients than Japanese norms. There were significant correlations between the frequency of RAS and the MCS or the PCS (p<0.05). After adjusting for age, sex, treatments and disease stage of BD, the PCS of patients with RAS in the past year was significantly lower than that in patients without RAS (p<0.05). The RAS occurrence also had a negative effect on the GOHAI scores (p<0.001). Conclusions: The present study showed that QOL was impaired in patients with BD and RAS could increase negative impacts on physical health related to QOL. Aims: Tremor, rigidity, abnormally slow body movements and postural instability characterize Parkinson's disease (PD). As the disease progresses the physical symptoms may affect other aspects of daily life and create additional psychological and social problems. PDQ-39 is a specific quality of life scale for evaluating the dysfunction of daily living and quality of life for PD patients. The objective of this study was to validate a Korean translation of PDQ-39. Methods: The cross-cultural adaptation included the instrument translation (forward and backward), an expert consensus meeting and a pilot test to produce a conceptually equivalent Korean version of the PDQ-39. We compared the quality of life 102 PD patients with 90 normal aged matched controls using Korean consensus version of the PDQ-39. To validate the Korean consensus version, psychometric properties were examined: reliability was assessed by Cronbach's a. For construct validity, convergent validity was checked by Spearman correlation coefficients and known-group validation was performed by t-test. In addition, quality of life according to modified Hoehn & Yahr stage was examined as well. Results: The PDQ-39 Korean version is reliable (Cronbach's a=0.98). Convergent validity is fair to good (Spearman correlation coefficient of eight subscales with each other, rho's between 0.23 and 0.86, all p's <0.05). Also PD patients had significantly lower quality of life than normal controls in all the eight scales of PDQ-39. PD patients in stage 4 had the worst quality of life in scales of mobility, activities of daily living, emotional well-being, cognition, communication, stigma and bodily pain. Conclusions: The findings confirm the Korean adaptation and validation through excellent reliability and validity Aims: The information given to cancer patients is crucial to patients' wellbeing and support patients can receive. We report the development of an instrument to measure the level of information received by cancer patients. Methods: The development followed the EORTC guidelines for developing Quality of Life (QOL) questionnaire modules which consist of a four phase process. In Phase I and II a literature review was conducted and items formulated. In Phase III, the 26 item questionnaire was tested in order to identify and solve potential problems in its administration. To determine the need for additional questions or the elimination of others, structured interviews with each patient were conducted after completion of the questionnaire. Results: 133 cancer patients from eight countries were interviewed. These countries represented northern, central and southern Europe and Korea. Clinical and demographic data were recorded. The resulting module (EO-RTC QLQ-INFO 26) consisted of 26 items grouped into four scales (information about the disease, the medical tests, the treatment and other services) and other single items such as satisfaction with information and desire to receive more information. There were no differences among subgroups of the whole sample for disease stage, treatment modality, level of information or gender. Conclusions: These preliminary results suggest that the EORTC QLQ-INFO 26 is a robust instrument to be used in cross-cultural studies. Cronbach's as ranged from 0.74 to 0.92. Test retest reliability was ensured by intraclass correlation coefficients ranged from 0.63 to 0.89. Several items presented a differential item functioning across countries. Correlation coefficients between SF-36 and MUSIQOL scores were high for those assessing similar constructs (r=0.42 to 0.65). All MUSIQOL dimensions showed a gradient according to the severity of the MS defind by physicians. Responsiveness was good for physical domains (effect-size >0.20). Low differences were carried out between the different methods to compute HRQoL norms. Conclusions: The results demonstrate that the MUSIQOL questionnaire is a well-validated measure of HRQoL assessment in MS patients. Aims: Assess the psychometric properties of the 28-item version of the Esprint Questionnaire (EsQ) a disease-specific QoL scale for allergic rhinitis (AR). Methods: An observational, prospective study was performed. EsQ-28 overall and dimension scores (nasal symptoms -NS-, others symptoms -OS-, daily life activities -DLA-, environment -E-, vitality -V-, sleep -SL-, psychological impact -PI-) were obtained. Patient evaluation included: a symptom severity scale (RASP), the SF-12 (physical -PC-and mental -MC-scores), a disease-specific QoL questionnaire (RQLQ), and a measure of global change. Average time spend for administration, ease of questionnaire use and usefulness of the scale for clinical practice, ceiling/floor effects, reliability, validity and sensitivity to change were assessed. Results: 206 patients were recruited: 33.7 years, 65% women, 68% medium cultural level, 66% in treatment for AR, 51% with persistent AR. Cronbach's a (CA) for the EsQ-28: 0.95 for overall score and 0.78 (NS) to 0.94 (V) for dimensions. Intraclass correlation coefficient (ICC) in patients reporting a stability in their health status was 0.86 for overall score, and 0.63 (NS))0.89 (V and SL) for dimensions. The correlations between the RASP scale and EsQ-28 was consistent and ranged from 0.63 (NS) to 0.41 (V). The correlations between the EsQ-28 overall score and SF-12 were: 0.12 (MC) and 0.54 (PC). The pattern correlations with the RQLQ was consistent with dimensions and was 0.85 with overall score. The overall effect size in patients reporting an improvement in their health status was 1.63 and ranged from 0.97 (V) to 1.8 (NS) by dimensions. 87% of patients spent £10 min for answering the EsQ-28, 82% of patients considered the scale ease to answer, 87% of investigators stated that they would use it in daily clinical practice. Conclusions: The EsQ-28 is a new questionnaire which has proved to be reliable, valid and sensible to change for measuring QoL in patients with AR, complementing disease-specific scales previously developed in this area. and Affect Balance (0.43). In a regression equation to predict happiness, the LBI was a significant predictor along with stress, income and sense of coherence. (R=0.50, p<0.001) As expected, age was correlated significantly with the LBI (n=25, r=0.19, p<0.001). Conclusions: The study provided support for our model and measure of lifestyle balance. Lifestyle balance is positively correlated with measures of well-being, life satisfaction and quality of life and negatively correlated with stress. The Lifestyle Balance Inventory appears to be measuring a unique variable from other positive states yet shows the expected associations with other measures. Aims: HRQoL is an important topic in paediatric practice. However, there is a lack of HRQoL studies and suitable instruments in the scientific literature. The aim of the study was to develop a new disease specific module for children with recurrent urinary tract infections in addition to the generic HRQoL ques-tionnaire KINDL. Methods: The development of the instrument was conducted in 2 phases. In the first phase, issues that are important for children with a bladder dysfunction were identified. Therefore, interviews with 20 children and their parents were performed. Open-ended questions were used to identify the main areas of concern and allow maximum input from the children. In the second phase, these areas were further discussed with 10 children and their parents based on a semi-structured interviewer guide, the key issues were summarized and a pool of relevant items was developed. Results: The following domains were found to be relevant influencing the HRQoL of children with chronic cystitis: physical wellbeing (burning on urination, pain in the bladder and kidney area, incontinence, reduced general state of health, reduced immunological state), psychological wellbeing (fear of recurrent urinary tract infections, frustration, psychological distress due to frequent visits to the doctor) and social wellbeing (limitations of social activities, sense of shame towards other children). Conclusions: This new disease-specific module of the KINDLR for children with recurrent urinary tract infections is expected to become a useful tool for HRQoL research in pediatric urological populations. Aims: Chronic obstructive pulmonary disease (COPD) is characterized by progressive breathlessness. In the advanced stage, patients suffer from dyspnea in association with minor exertion or even at rest, resulting in a gradual impairment of their physical ability, activities of daily living (ADL) and affect their quality of life (QOL). Lung volume reduction surgery (LVRS) has been reported to improve pulmonary function, lung mechanics, exercise tolerance, and dyspnea. However, the longterm effects of LVRS on psychological states have not been fully elucidated. In this study, we evaluated QOL and psychological state before and after LVRS in patients with severe COPD for 60 months. Methods: We studied 11 patients with severe COPD who underwent LVRS following pulmonary rehabilitation. QOL was assessed by Sickness Impact Profile (SIP). Psychological state (anxiety, depression) was measured by using the State Trait Anxiety (STAI) and the Self-Rating Questionnaire for Depression (SRQ-D). Serial measurements of QOL and psychological states were done before and 3, 12, 24, 36 and 60 months after LVRS. Results: The best scores in SIP were obtained at 36 months after LVRS. Especially, physical-SIP scores at 12, 24, 36 months reached to normal range. However, psychosocial-SIP score at 60 months was worse than pre-LVRS. As for psychological state scores, STAI scores did not show significant changes during 36 months, and indicated in the range of high anxiety at 60 months after LVRS. The SRQ-D scores have been in the suspected level of depression for 60 months. Conclusions: It is difficult to maintain the beneficial effect of LVRS on psychosocial QOL and psychological state up to 60 months. Aims: Background: Breast cancer is an important disease and one where health care services have the potential to improve the quality and quantity of life. This demographic imperative, coupled with the dramatic increases in rates of breast cancer with advancing age, is expected to translate into a large absolute increase in the number of older women treated for and surviving breast cancer. In Iran, breast cancer is one of the most growing and important women's health Problems women diagnosed with breast cancer today have many different treatment options from which to choose. While most women chosen treatments that maximize survival, information about quality of life can be an important component in decision-making in several clinical situations We discuss special considerations in measuring outcomes and quality of life in this group, and make some recommendations for future research. The overall aim of the study was to describe the Quality Of Life of (BCS)women and to examine the role of ethnicity in influencing their wellbeing. Methods: The authors conducted an interview survey to examine QOL, health perceptions, and life stress among BCS. Subjects diagnosed between May 2001 and May 2003 identified among register at oncology center of Shohada University Hospital (Tumor Registry). This survey instrument were include standard measures of QOL (e.g., the RAND SF-36 Health Perceptions Scale, the Cancer Rehabilitation and Evaluation Survey-Short Form [CARES-SF] , and the Ladder of Life) and new items. Results: The result showed that the quality of Iranian women with breast cancer is not so good. Conclusions: Discussion: .It seems The lack of knowledge about breast cancer is an important factor in Iran and there is a need for public educational programs especially for less educated women. However, in Iran social values and moral considerations limit the use of mass media for publicizing breast cancer and quality of life awareness. in the terminal cancer patients can be highly meaningful. The instrument used for this study was the Korean Version of the McMaster Quality of Life Scale (K-MQLS), which developed by Kim, et al and its validation has been confirmed in Korea in 2004. Thus, our research was targeted at evaluating a K-MQLS to provide more qualitative care in clinical practice particularly in Korean end-stage cancer patients. Methods: The study was undertaken at 4 Korean university hospitals. The eligible study subjects were 104 adult end-stage cancer patients and family members. Discriminant validity was examined by correlation between K-MQLS subscales. Construct validity was evaluated by ANOVA and Tukey's multiple comparison analysis. The reliability of the K-MQLS was evaluated by ICC and Cronbach's a. HrQOL evaluation of the K-MQLS was evaluated by Mean, Standard Deviation, t-test, ANOVA. Results: 1. K-MQLS showed good discriminant and construct validity, ICC, internal consistency. 2. There was statistically significant difference of total K-MQLS mean score which lower among those within 3 weeks of death than among longer-term survivors. 3. In the patients characteristics, mean scores of K-MQLS showed significant differences with gender, age, family living together, analgesics, ECOG PSR and No. of symptoms. 4. Among the variables affecting quality of lives, Number of symptoms accounted for 20% of HrQOL in end-stage cancer patients by linear regression analysis. Conclusions: K-MQLS was a valid HrQOL instrument in the end-stage cancer patients who received hospital care in Korea. The findings showed that HrQOL score was significantly higher in the patients with the less number of symptoms than in the patients with the more. Aims: To evaluate the responsiveness of 10 scales of short form 36 (SF-36) in a Taiwanese, community-based sample. Methods: Data came from a two-year follow-up study, conducted in a representative sample 1715 community residents of five counties of Taiwan in 1994 and 1995. The overall response rate of the first year was 72.13% and the rate of loss to follow-up for the second year was 14.87%. These study subjects were drawn by a four-stage sampling design. Face-to-face interview by public health nurses was used for data collection. The SF-36 measure eight multiitem variables: physical functioning (PF), social functioning (SF), role limitations due to physical problems (RP), role limitations due to emotional problems (RE), mental health (MH), vitality (VT), pain (BP), and general perception of health (GH). In addition, SF-36 Physical (PCS) and Mental (MCS) Component Summary scales are derived. Two responsiveness indices were calculated: effect size (ES) and standardized response mean (SRM). The variable for external criteria was a single item on changes in respondents' health over the past year. The ''unchanged'' group included individuals with ''no change'' (n=977). The ''change'' group was defined as those who answer ''better'' or ''much better'' (n=123). Results: Between baseline and one-year later, GH scale showed the highest change among scales of SF-36, while RE scale showed the lowest change. Aims: Surgery results in a decline in performance of activities, followed by recovery. However, little is known about the functional form of this recovery. We examined the trajectory of health-related quality of life in patients who underwent abdominal surgery. Methods: The primary outcome measure was a revised version of the previously developed Surgical Recovery Index (SRI). The revision incorporated an individualized approach to assessment in which patients nominated their 5 most important activities. Patients then reported the number of hours spent on each activity each week and the week post-operatively when they fully recovered each activity following abdominal surgery. Resolution of pain was also ascertained on this questionnaire. Results: A total of 47 patients (63% female, mean age 51.2) were surveyed, at a mean post-operative day of 31.2 (SD 18.5, range 12-126). Most patients reported resolution of pain by week 4 (data not shown), but significant activity limitations remained. The mean activity recovery for weeks 1 to 6 were plotted, with x axis representing the time in weeks and the y axis representing the percent of total hours spent on their top 5 activities at each week. While a linear trend line fitted to this curve (y=0.15x+0.038) has a r 2 of 0.94, a logarithmic trend line fitted to this curve (y=0.43ln(x)+0/087) has a r 2 of 0.98, implying a better fit. The recovery curve for elderly patients (age > 65) trends towards more rapid recovery than younger patients. This is likely not significant due to small sample size of elderly patients (n=7). Conclusions: Patient-reported post-operative recovery from abdominal surgery follows a logarithmic course. Results such as these are important for patient education and for helping calibrate expectations regarding post-operative recovery. Further refinement and validation of the SRI will be used to explore potential differences in recovery among varied surgical procedures as well as in patient subgroups (e.g., elderly vs. young). Aims: Several instruments for measuring functional status in rheumatic diseases have been developed and validated. However, none is yet available in Mandarin for use in Taiwan. To translate the Health Assessment Questionnaire (HAQ) from English to Mandarin and to test the metrological properties, reliability, and validity of the Mandarin version. Methods: The study included a total of 225 patients (87% female with a mean age of 54 years) with rheumatoid arthritis (RA), seen at the Taipei Veterans General Hospital. The HAQ was translated from English to Mandarin by translators who were aware of the intended use of the questionnaire. The translated questionnaire was administered to a subset of 119 patients on two occasions 4 weeks apart. Test-retest reliability was assessed using the intraclass correlation coefficient. Functional status was assessed with the HAQ, Short Form 36 (SF-36) and Disease Activity Score (DAS-28). Internal consistency of individual items with the overall score was assessed using Cronbach's a coefficient. Construct validity was evaluated by determining Pearson's correlation between the subscale scores on the Mandarin HAQ and scores from the mental component summary and physical component summary of the SF-36. Results: Test-retest reliability was 0.76. As predicted, a comparison of subscale scores on the HAQ, SF-36 and DAS-28 indicated strong correlations between the physical component of the SF-36 (r=0.7) and the DAS-28 (r=0.6). Internal consistency was range from 0.7$0.85, and construct validity outcomes were statistically significant (p<0.01). Conclusions: Taiwan's Mandarin version of the HAQ retains the characteristics of the original English version, and is a reliable and valid instrument for measuring functional disability in Taiwanese Aims: Physical dysfunction, experienced by all palliative care patients, affects most aspects of life and is among their most distressing concerns. Thus, Physical Function (PF) is a crucial domain for quality of life (Qol) assessments. This study aimed to examine the content of the existing PF assessments developed for palliative care. Methods: To identify instruments a Medline search was performed using the terms (palliative care OR end of life care OR terminal care) AND quality of life AND (assessment OR instrument OR questionnaire). Of 1189 hits, 878 were excluded as irrelevant. Among the remaining 311, named instruments were detected in 203 abstracts and included in this study. Results: Of 133 instruments detected, 35 were identified as developed for palliative care. The content of these was explored. Seven tools assessed symptoms, 2 assessed satisfaction with care, 5 assessed existential/emotional issues and 1 was a global uniscale. Five multidimensional instruments had no PF items, and for 3, only brief descriptions indicating no PF assessment were obtained. Twelve instruments included PF assessment. Two of these were comprehensive palliative performance scales for staff's assessment, and 10 were multidimensional instruments including 2 to 8 PF items. The PF items' content was non-consistent across the instruments. Two instruments used the ECOG/ WHO performance scale, the others included various aspects of self-care, mobility, help needs or participation in activities (work, hobbies, transportation). Conclusions: Conclusions: PF assessment seems a minor part of palliative care instruments and only 12/35 instruments included PF. The content and conceptualisation of PF varied considerably. Clearer definitions and conceptualisation of PF are needed, as well as a consensus on relevant aspects. Improvement of instruments is needed and existing performance scales should be further tested.

Aims: The PGI's reliability, validity and responsiveness studied by several groups but no review of its psychometric properties exists. This paper presents a graded review using criteria adapted for individualised assessments to assess both findings and quality of the studies. Methods: A systematic review found 16 studies included a formal assessment of one or more of reliability, validity or responsiveness of the PGI. Grading criteria to examine the quality of the studies were developed from those used with outcome measures for elderly people. For example, a top grade was given to studies of validity assessing both content and construct aspects, with a lower grade if only content validity was examined. Overall conclusions were based on both study results and quality. Results: The PGI has been variously adapted in terms of instructions and rating scales, in some cases for specific populations. The quality of the studies was generally high, although there was some variability. Overall psychometric properties are favourable. Only one study found PGI reliable for individual comparisons, however, five studies of high quality reported reliability adequate for group comparisons. Two studies reported that the PGI was not reliable, however in one case the correlation coefficient approached that required for low-level group comparisons and the other discussed potential effects of response shift. Of the eight papers directly addressing quantitative aspects of validity, all eight included some results that suggested at least low validity. Responsiveness was confirmed in four out of five studies. Conclusions: Results indicate the psychometrics properties of the PGI to be good and for reliability to be adequate for group comparisons in most cases. However, the different versions of PGI render conclusions problematic. The graded review criteria represent a way to understand the quality of measures and are applicable and workable. Further work to take into account the potential effect of response shift on psychometric properties is required for all quality of life instruments. Aims: Dysphagia is a common disorder associated with impairment complications and even death. There is no standardized method to follow impairment complications and assess dysphagia specific interventions. A new measure is needed that reliably and accurately captures impairment complications both for identification and prognosis. Any new measure needs a conceptual foundation to clearly define target population, purpose and concept, and to guide instrument development. This research established the conceptual framework for development of a new tool measuring impairment complications in adult patients with dysphagia Methods: Medical impairment was defined according to the model by Verbrugge (2001) and includes anatomical, physiological, mental and emotional impairments secondary to dysphagia, regardless of etiology. Focus group methodology elicited perspectives of clinicians and patients on the impact of dysphagia and factors influencing severity of impact (personal and external). The measure is designed for clinical utilization; hence, it requires high levels of reliability and clinical sensibility. Results: Development of a conceptual framework permitted probing three identified domains of impairments secondary to dysphagia in great detail including pulmonary, nutrition-hydration and psychological impairments. The need for two subscales in each domain was identified; one for current severity and one for future risk of impairment. Further, we recognized the need to have a scale capable of assessing status at one point in time as well as changes over time in a way that reflects the perspective of both the clinician and the patient. Conclusions: This conceptual framework was the first established to guide development of dysphagia specific item inventory for medical impairment. consultation with the rheumatology multi-disciplinary team and patient interviews (n=30). From this, a measure comprising 67 items was generated. Twenty patients completed this initial version providing critical feedback regarding face validity and structure, content of the items and response scales, leading to further revision of the LupusQoL (63 items) which was completed by 322 patients. Results: Principal components analysis and Cronbach a coefficients highlighted eight domains. The LupusQoL was further revised (42 items) based on factor analysis, clinical decision and patient feedback. Principal components analysis was performed on the data from the second version of the LupusQoL, completed by 213 patients. This confirmed the factor structure of the Lup-usQoL. The Aims: Neuro-QoL is a NIH-project (National Institute for Neurological Disease and Stroke) funded project. This 5-year study is designed to construct a clinically relevant and useful health-related quality of life (HRQOL) measurement system for major neurological conditions. The 2 phases of this project are instrument development and field testing/instrument refinement. Both phases include multiple steps.

Step 1 of phase 1, the process of identifying target neurological conditions that will guide the rest of the instrument development is reported here. Methods: While selection of target conditions is the research team's responsibility, we sought opinions of thought leaders using 4 methods: we conducted (1) a literature review of major neurological conditions and HRQOL impact; (2) individual interviews with 44 neurology experts who recommended the 5 neurological disorders that have HROQL impact; (3) a consensus meeting with an expert panel of 13 adult and pediatric specialists. They were advised of results from the previous steps. Using a modified nominal group technique the panel developed the criteria that would guide their condition selection then generated their condition recommendations; (4) we sought recommendations from the American Academy of Neurology Practice Committee and consulted with our NIH-NINDS project officer. Results: The 5 adult conditions that will serve as the basis for our project are stroke, Parkinson's disease, multiple sclerosis, epilepsy and neuro-muscular diseases. The pediatric conditions are epilepsy and muscular dystrophies. Conclusions: Selected conditions include adult and pediatric conditions. There are common HRQOL consequences across these conditions and aspects that are unique to each. The Neuro-QoL measurement system will be designed to assess all of these factors. Aims: Quality Metric Incorporated has recently changed the methodology for scoring their SF tools (SF8, SF12, and SF36). This study compares results from the prior scoring methods and the new IRT methodology. Methods: A demonstration disk provided by Quality Metric Incorporated contained test data for each of the SF tools. Each of these test datasets contained 300 observations. These datasets were scored on the test dataset using the new methodology and were rescored using the prior methodology. Comparisons were made between the scores obtained from the new and old methods. Methods of comparison included correlation analyses and Bland-Altman analyses. Results: Correlations between the two scoring methods were extremely high. Correlations between subscores ranged from 0.9996 to 1.0000. Bland-Altman comparisons showed that old and new scores were highly related on the SF8 and SF12. The new physical composite score on the SF36 was 1 to 4 points lower than the score using the old method. Mental composite scores tended to be about 0.5 point higher. Most of the other subscales did not change dramatically, except for the role emotional score. Conclusions: The new scoring algorithms for the SF tools do not drastically change the subscale scores on the SF8, SF12, or SF36. However, the new algorithm creates subscales more often if there was missing data. There does not appear to be a need to rescore data from prior studies. A next step would be to look at some clinical trials to see if the choice of scoring algorithm would change the results of the trials. Aims: To assess the construct validity and internal consistency of a 12-item patient satisfaction with treatment questionnaire in overactive bladder. Methods: A cross-sectional sample of 891 patients participating in a National Family Opinion web-based survey were assessed. Reliability of the questionnaire was assessed using Cronbachs a. Factor validity was assessed using the variance extracted by the factor, maximal reliability (coefficient H), and reliability of the construct. Polychoric, polyserial, and pearson product moment correlations were also conducted on the OAB-PSTQ (transformed score) with incontinence severity on life, incontinence episodes, time with symptoms of OAB, global assessment of satisfaction with treatment, adverse event severity, and frequency of urinations per day. Hypotheses were pre-specified with higher expected correlations with the global assessment and high correlations with incontinence episodes and severity on life. Lower correlations were expected with frequency and the other variables. Results: The Cronbach's a was 0.97 for the 12 item scale. The variance extracted from the unidimensional questionnaire (via factor analysis) was 0.73. Coefficient H was 0.98 with a maximal reliability of the construct of 0.97 indicating good support for factor validity. A transformed score (100 point scale) was utilized and correlations with the other outcome measures in the study further supported the hypothesis consistent with construct validity. The highest correlation existed between the OAB-PSTQ and the global assessment of satisfaction (0.823). There were also moderate correlations with severity of side effects ( Aims: Current rehabilitation outcome instruments for children are narrowly focused on selected functional tasks. Our objective was to design broader, clinically relevant measures appropriate for assessing the status of children following brain injury. Methods: A prospective observational study was conducted with 77 children and adolescents diagnosed with brain injury admitted to 15 acute care hospitals and rehabilitation facilities. Surveys were also completed by children's caregivers after hospital discharge. Item response theory and classical test theory were used to select items measuring pre-defined domains of functioning. Results: Functioning After Brain Injury (FABI) scales were developed for the domains of Daily Functioning (14 items), Readiness for Community Functioning (7 items), and for Caregivers (7 items). The Daily Functioning scale taps a wide array of physical activity, cognitive, and behavioral skills. All three scales were highly reliable (Cronbach's a greater than 0.90), valid in comparison to disability scales and known patient groups, and responsive to clinician perceptions of improvement by discharge. Conclusions: The FABI scales provide brief, psychometrically sound measures that may be used by clinicians to monitor children's functioning following acquired brain injury. Aims: This study examines relationships between quality of life (QoL) and clinical outcomes, specifically A1C, in Type 2 diabetes patients. Methods: Type 2 diabetes patients at the outpatient clinics of a university hospital completed a generic QoL measure (SF-12) and a diabetes-specific QoL measure (Audit of Diabetes Dependent Quality of Life (ADDQoL)). A cover letter signed by the patient's provider was sent with a mail out questionnaire. Patient reported data was merged with a retrospective collection of his/her clinical and utilization data, including A1C, from electronic medical records. A Charlson comorbidity score, diabetes complications score, BMI, and number of ER and hospital visits were also calculated. Results: Usable response rate was 44.3% (n=385). Respondents were 57.1% female, 93.8% Caucasian, and 64.1% were in the 40-69 age range; 49.1% were on oral medications only, 31.7% were on oral medications and insulin, and 9.4%, on insulin only. Mean A1C of respondents was 7.2 (+1.4), mean diabetes duration was 10.2 (+9.1), and 62.1% were obese (BMI>30). Patients were dichotomized into glycemic control levels based on the ADA recommended A1C level <7.0, versus>7.0. Cronbach's as for the ADDQoL, PCS (SF-12 Physical Component Score), and MCS (SF-12 Mental Component Score) were 0.92, 0.92, and 0.88, respectively. Only ADDQoL scores showed significant correlation with A1C (r=)0.19, p<0.000). Hierarchical regression models were used to separately explain ADDQoL, PCS, and MCS scores, using A1C as the primary independent variable, and controlling for demographics and clinical variables including comorbidities and complications. A1C was not a significant predictor in any regression model, although univariate analyses indicated significantly lower AD-DQoL and SF-12 PCS in the group with A1C>7.0, and in the group with one or more diabetes complications. Obesity was a significant predictor in models explaining only the PCS and MCS scores. Conclusions: These results support complementary use of generic and disease-specific QoL measures in Type 2 diabetes populations. Patient's perceptions of their own life may differ from what a biomedical measure like A1C suggests. Aims: The aim of this study was to validate a scale for functional bother specific for nail psoriasis. The questionnaire will have to be adapted both in the case of toe or finger nail psoriasis. The measured criterion will be unidimensional and related to the bother caused by nail psoriasis in daily life. Methods: The scale was developed according to the international recommendations on quality of life research. In October 2004, in collaboration with the French Association for the fight against Psoriasis (APLCP), a vast survey was carried out in France by sending to the 4000 members, selected by drawing lots, a questionnaire that included several items. The first step of the process has led to the selection of 10 items related to functional bother induced by nail psoriasis. Results: In total, 795 questionnaires concerning individuals affected by nail psoriasis were usable. Validation analyses included the 10 selected items. Questionnaire's contents were coherent with the a-Cronbach coefficient equaling 0.88. The unidimensional feature of the questionnaire was verified: the analysis in principal components (ACP) revealed that 49% of the total variance was explained by one component. The DLQI (Dermatology Quality of Life Index) specific to dermatological pathologies was also given and enabled a comparison with the scale to validate. Pearsons correlation coefficient between both scales was 0.48. The severity of the affection assessed through the DLQI evolved in the same way as the evaluation for the ''Nail Psoriasis'' scale. A test-retest performed on a sample of 15 individuals showed that the scale could be reproduced with an intra-class correlation coefficient of 0.82 between 2 administrations. Conclusions: The ''Nail Psoriasis'' scale is simple to use and easy to give to the patient. The qualitative features which must be found in a quality of life scale have been checked: comprehensibility, reliability and validity. The scale will have to be used during clinical trials in order to demonstrate its ability in measuring change in condition (before and after treatment). Aims: to determine a concept of spirituality, which would be used in the WHOQOL instruments to assess the effect of spirituality among Japanese people Methods: After conducting 7 focus groups of 44 participants, the WHOQOL SRPB questionnaire of a 118 question items, consisting of a scale that measures the constructs of spirituality Questions on the importance of each facet, Socio-demographic were delivered to 226 males and females living in Tokyo, Shizuoka, Kobe, Nagasaki, Kyoto, and Toyama with mean age of 38.8 years±12.8. Among them, 128 subjects (56.6%) had no specific religion. Among the 98 subjects with a religion, 50 Buddhists, 23 Christians, 9 Islam (16 unknown). Results: By single factor analysis, nine facets were found suitable for the scale; kindness to others/selflessness, acceptance of others, inner strength, inner peace/serenity, harmony, death and dying, meaning of life, connectedness to a spiritual being or force, faith and divine love. Chromatic's fscoefficient alpha for each facet 0.71^0.923. Three of the four domains proposed by the WHO-QOL-SRPB projects, ''personal human relations'', ''code to live by'', ''transcendence'' were confirmed as valid to Japanese people. Conclusions: The results indicated that based on the focus group results, ''connectedness to ancestors'', and ''awe to nature'' were found very important factors for concept of spirituality, and by the survey, personal relationships and inner world were more important than religiousness. Present survey alone has completed the task of elucidating Japanese spirituality, partly due to the sampling bias within the survey, and inadequate sampling size. Since the concept of spirituality is greatly affected by cultural factors, the job remaining is to obtain a larger sample encompassing a wide range of geographical and cultural backgrounds. Aims: The GERD-HRQL symptom severity instrument was developed 10 years ago to assess symptoms of gastroesophageal reflux disease (GERD). Since then, investigations provided evidence of its validity. The purpose of this study was to review existing and new data assessing the validity of the GERD-HRQL. Methods: The GERD-HRQL is a 10 item questionnaire based on symptom severity anchors. The best score, 0; worst, 50. Existing studies were reviewed assessing its validity. In addition, 372 patients with GERD and 50 patients without symptoms of GERD completed the questionnaire. Assessment of the types of validity were completed. Results: Content Validity: Items for the instrument were generated from textbook descriptions and author's experience with GERD patients. GERD patients reviewed the instrument and 85% felt that it accurately reflected their symptoms. Face Validity: It has been assessed by mulitple investigators and felt to adequately assess GERD symptoms. Criterion Validity: It correlates to objective physiologic measurements of GERD. Concurrent Validity: There is a statistically significant correlation of esophagitis grade with GERD-HRQL scores. Predictive Validity: The satisfied patients with medical management had lower scores than patients who proceeded to surgical treatment. Construct, Known-Groups Validity: Patients without GERD had a median score of 3 compared to 32 in GERD patients (p<0.0001). Patients who underwent GERD treatment had statistically significant lower scores. Construct, Convergent Validity: The GERD-HRQL has statistically significant correlations with other instruments which measure GERD symptoms. Construct, Divergent Validity: Compared to the SF-36 (not designed to measure GERD symptoms), there was no correlation with the GERD-HRQL scores. Conclusions: Ten years of experience with the GERD-HRQL has demonstrated its validity in measuring symptom severity of GERD. Aims: Theory of Planned Behavior (TPB) has explained and predicted human health behavior but it has rarely been applied to the issue of overweight and obesity. The present study developed the TPB-based questionnaire and determined the explanatory power of TPB for weight control behavior. Methods: Elicitation review of weight control behavior was obtained from 32 females and content analysis was performed. The TPB questionnaire was administered via face-to-face interview to 397 overweighing and obese women living in rural communities of Mahasarakham Province, Thailand. Confirmatory factor analysis (CFA) was used to determine how well TPB explained diet control behavior. Results: An initial model was developed based on TPB and the statistical diagnostic of the model fit were as follows: A preliminary list of issues was checked for redundancy and adequacy and condensed to a list of 70 items. Thirty-six experts (physicians, nurses, psychologists) from 8 different European countries checked the preliminary list of 70 items. They rated the relevance, importance and the breadth of coverage and assessed if the items are easily translatable into their native language. This resulted in 43 items that were checked for meaning and translation problems. Those items difficult to translate and all items >2 of relevance were excluded from the resulting questionnaire. The revised item list were checked with the EORTC Item Bank Database for consistency of item translation. Additionally, a global rating of interference with daily activities was included within the questionnaire. Conclusions: The preliminary phase II fatigue module is now available in two versions: FA-R15 (reduced version with 15 items) to be used in combination with EORTC QLQ c30; FA25 (with 25 items) to be used without QLQ C30. The phase II module is ready to be tested with larger patient samples. Following psychometric considerations phase III testing will be started with the FA-R15. Aims: The aims of this study were to develop and validate an instrument for the measurement of health-related quality of life (HRQoL) based on Traditional Chinese Medicine (TCM) perspective. Methods: Six groups (scales) of questions, five questions for each scale, were developed from literature review and professional consultation: general health, and health of major five aspects according to TCM theory (heart, liver, spleen, lung, and kidney) to measure the HRQoL of each aspect. One hundred and thirty-six patients (age 49±15 years) visiting the TCM clinic of a medical center in northern Taiwan were interviewed. Health conditions of the five aspects for each patient were rated by a TCM physician. A telephone interview was conducted 1 week later as a retest. Test-retest reliability (intraclass correlation coefficient, ICCR), internal consistency (Cronbach's a coefficient), and the ability to differentiate the health conditions of the patients for of each scale were assessed. Results: Five of the test-retest reliability and internal consistency coefficients of the six scales were fair to good (0.43-0.75 and 0.51-0.75, respectively), except that of spleen. Both the test-retest reliability and internal consistency coefficients for the liver scale for male were higher than the counterpart for female. The scores of spleen, lung and kidney scales were significantly different in patients with different health conditions of the same aspect. Conclusions: Most scales in the questionnaire had fair to good test-retest reliability and internal consistency. Scores in three scales could differentiate patients' health conditions. Aims: OAB is highly prevalent with significant QoL burden. We assessed the impact of long-term darifenacin (dari) treatment (tmt) on QoL. Methods: This 2-year, multicenter, open-label extension study enrolled patients (pts) from two 12-week, placebo-controlled, double-blind (feeder) studies of dari CR 3.75, 7.5 or 15 mg qd in pts with OAB for ‡6 months 1, 2 . After the first 2 weeks of the extension (when all pts received dari 7.5 mg), pts could titrate between dari 15 and 7.5 mg. OAB symptoms and QoL (King's Health Questionnaire [KHQ] ) were assessed at baseline and during tmt. This analysis examines pts on dari 7.5 or 15 mg during feeder studies with a gap of 3 tmt days or less before continuing into the extension. Results: Three hundred and three pts on dari 7.5 or 15 mg during the feeder studies entered the extension (22-89 years; 86.5% female) and 199 pts (65.7%) completed. Dari 7.5/15 mg produced significant improvements in 8 of 9 KHQ domains from baseline to end of the feeder studies, which were maintained/further improved Emotions: )13.7 and )12.8; Personal Relationships )7.9 and )8.7; Sleep/Energy: )9.3 and )11.6; General Health Perceptions: 0.0 and 2.1; all domains p<0.001 vs. feeder study baseline at both timepoints (Student's t-test) except General Health Perceptions. A decrease from baseline=improvement. Dari was well tolerated and the overall long-term safety profile was consistent to that observed in the 12 week phase III studies 1, 2 . Conclusions: Dari is associated with significant improvements in QoL in double-blind, 12-week studies, which are maintained for 2 years during open-label tmt Aims: The objectives of this study were to compare the factor structure of the Hungarian version of the Center for Epidemiological Studies-Depression (CES-D) scale to the factor structure of the original English version and also to assess the reliability and validity of the Hungarian scale. Methods: The Hungarian scale was prepared according to current recommendations. Following pilot testing, 308 patients on maintenance haemodialysis from five dialysis centers, and 308 randomly selected kidney transplanted patients completed the CES-D scale and the Kidney Disease Quality of Life (KDQoL-SF36) questionnaire at same time. The data obtained was compared to pooled data from 308 North-American end-stage renal disease (ESRD) patients. Simultaneous confirmatory factor analysis (SCFA) was performed using data from both the North-American and the Hungarian ESRD patients. Subsequently internal consistency and testretest reliability was assessed. Finally, to test construct validity, correlation between data obtained with the Hungarian CES-D scale and the emotional well-being domain of KDQoL-SF36 questionnaire was analyzed. Results: SCFA indicated a good fit between the originally established fourfactor model (Depressed affect, Somatic component, Positive affect, Interpersonal dimension) of the original English version and data obtained with the Hungarian translation both in the Hungarian hemodialysis and the kidney transplant groups. Internal consistency of the Hungarian scale was very good (Cronbach's a=0.87) and test-retest reliability was also high (r=0.78). The CES-D score showed a moderate-strong correlation with the emotional well-being domain of KDQoL-SF 36 questionnaire (r=)0.716; p<0.001), respectively. Conclusions: The Hungarian version of the CES-D scale showed remarkable factorial invariance to the original English version both in dialysis and in kidney transplanted patients. Furthermore, we demonstrated that the instrument is a reliable tool to assess psychological distress in different chronic kidney disease populations. Aims: As cancer care becomes more complex, satisfaction with care has emerged as a key issue to address in many countries, and is increasingly seen as an integral part of cancer care. This study provides an analysis of key characteristics regarding inpatients satisfaction with cancer care in a large cross-cultural sample. Methods: Cancer patients (pts) discharged from a surgery or medical ward of oncology hospitals in 9 countries (Europe-Asia) rated their level of satisfaction using the EO-RTC IN-PATSAT32 at home and returned it in a self-addressed envelope. Results: Of 762 pts recruited, 647 (85%) returned the questionnaire. Mean age was 57 (range: 19-91) and 438 (58%) being female. Patient satisfaction levels differed significantly according to country: patients in Northern countries being more satisfied with doctors' information provision than in Southern countries (p=0.005). Patients undergoing chemotherapy and surgery were more satisfied with the exchange of information between caregivers (p=0.004), other hospital personnel (p=0.005), and overall care (p=0.009); the larger the hospital size the less patients were satisfied with the exchange of information between caregivers (p=0.04), other hospital personnel (p=0.001), hospital comfort (p<0.001) and overall care (p=0.002). Age and education level were only related to hospital comfort as well as to doctors' information provision and availability. Time since cancer diagnosis and disease status was not related to patient satisfaction. Conclusions: These findings help interpret patient satisfaction scores when comparing hospitals within and across nations. They also provide insights into the needs of patients from various cultural backgrounds. Aims: In quality of life research, many terms, such as quality of life, health related quality of life, health status, well-being, happiness, life satisfaction, etc., are used interchangeably. However, these terms are much or less different from each other. This study attempted to distinguish these terms by conducting multidimensional scaling (MDS) analysis. Methods: One hundred and fifty undergraduate students were asked to rate 66 paired comparisons on a seven-point similar/dissimilar response scale with 12 QOL-related terms as the stimuli. Moreover, they rated the stimuli on each of the five bipolar adjective scales (short-long; easy-difficult; abstract-concrete; material-spiritual; subjective-objective). Multidimensional scaling analysis was conducted to recover the latent structures on judging these terms. To enhance the interpretation of the MDS result, the five property fitting vectors were also plotted on the MDS configuration. Results: Two dimensions, i.e., physical and psychological, were found by the MDS analysis. Property fitting vectors were helpful to interpret the MDS result. Conclusions: Participants judged the 12 QOL-related terms differently. Interchangeable use on the QOL-related terms may not be appropriate. To clarify the confusion, we suggest that QOL researchers should define the terms before using them in their studies. Aims: It has been proposed to determine MID for a patient reported outcome (PRO) in a phase III study based on blinded data combined across treatments, e.g., using global rating of change (GRC) as an anchor. Anchoring works only if the anchor and the PRO are correlated. We model a hypothetical study to illustrate that the condition of non-zero correlation can be violated and consequently MID determination is not meaningful without this being evident from the combined blinded data. Methods: The hypothetical study was modeled as a parallel group study with test drug and placebo. Change from baseline (ChgBl) in PRO and a 3-point GRC (worse, unchanged, better) are recorded (assumptions see Table) . MID of PRO is determined based on combined blinded data as average of the absolute mean ChgBl in PRO for patients with GRC=better and patients with GRC=worse. Within a treatment group, we model ChgBl in PRO (normally distributed) as uncorrelated with GRC. Results: Based on the model, the means of PRO ChgBl for patients with GRC=worse, unchanged, better are )0.66, 0, +0.66 and MID=0.66 which is <treatment difference=2, and the correlation between GRC and ChgBl in PRO is 0.5. Clinical relevance would be concluded, although GRC, uncorrelated with ChgBl in PRO within the treatment group, is not suited to assess MID. Conclusions: Determination of MID based solely on blinded data of one study might not always be appropriate. One should also assess the impact of treatment on the correlation structure.

Test drug Pearson's correlation coefficients between our questionnaire and LSI were r=0.78 (p<0.01) before the treatment and 0.80 (p<0.01) after the one. After the pool therapy 25 patients with rheumatoid arthritis showed the improved total QOLs, but 4 patients with one indicated the deteriorated total QOLs. Compared with before treatment, there was significant improvement of total QOL in after treatment (p<0.01). Significant improvements were shown in strata of well-being, dietary problems, sleep pattern, psychological problems, mental function, medical service, social participation, environmental problems, work performance (p<0.01), physical function and sexual problems (p<0.05) after treatment. Conclusions: As the recent progress of both internal and surgical treatments play very important roles to improve the QOLs of patients with rheumatoid arthritis, the water training therapy in pool showed the remarkably significant improvements on various categories on QOLs. Our questionnaire had high enough reliability and potency of varidity to use for the patients with rheumatoid arthritis. These findings indicate that the water training therapy is very effective, and we must pay more attention to the evidence based efficacy on water training therapy for the patients with rheumatoid arthritis. Of the subjects who received this information, 42% (n=50) reported using the information and 58% (n=69) reported never using the information. Content analysis showed incidental findings that many subjects believed they were not at risk for developing lymphedema. Women reported that their physician told them they were not at risk or that lymphedema was not a concern for them. Some women stated they were unaware of lymphedema as a potential side effect. Conclusions: Lymphedema can have a significant impact on quality of life in breast cancer survivors with an incidence of 6 to 62.5%. While well-established prevention and management guidelines exist, it is concerning that the majority of women did not use them. These findings present future opportunities in quality of life research. These include the development of tools to quantify risk perception among breast cancer survivors and the exploration of non-adherence to current lymphedema guidelines in women at risk Aims: In this study Quality of Life (QoL) was evaluated in patients who underwent pancreaticoduodenectomy for ductal adenocarcinoma of the pancreatic head. Due to ongoing controversy, special attention was given to the influence of extended lymphadenectomy and classic Whipple procedure (PPD) vs. pylorus-preserving pancreaticoduodenectomy (PPPD). Methods: From 1993 to 2004 QoL was analyzed in a prospective single center study in 91 of 169 patients. All patients received either PPD (n=34) or comparable PPPD (n=57). The majority of patients had additional extended lymphadenectomy (ELA, n=70); remaining patients underwent regional lymphadenectomy (RLA; n=21). QoL was assessed by the EORTC QLQ-C30 and a disease specific module. Data were collected pre-operatively (baseline), before discharge, 3, 6, 12, and 24 months after surgery. Results: No significant difference was found in long-term survival between the PPD and PPPD groups or between the ELA and the RLA groups. QoL significantly decreased after surgery and slowly recovered to almost preoperative levels within 24 months. This was true for functionalscales, symptom-scales and disease specific items. Compared to the RLA group, patients with ELA reported clinically significant higher levels of diarrhea, pain and poorer global health. Comparisons of QoL between PPD and PPPD groups showed disadvantages particularly concerning pain-scales in the PPPD patients. However, the PPPD-group included a significantly higher proportion of ELA patients. Conclusions: QoL in patients with cancer of the pancreatic head is significantly impaired after surgery and it recovered slowly to almost preoperative levels. The two surgical reconstruction techniques did not demonstrate significant differences in terms of QoL, but an ELA seemed to impair the patients' QoL. The aim was to determine the minimum important difference (MID) for the outcome measurement of 'satisfactory sexual events' (SSEs) using three methods: anchor-based, distribution-based, and interviews with women. Methods: Anchor-and distribution-based methods: Data from two double blind, placebo-controlled, parallel group trials for the treatment of Hypoactive Sexual Desire Disorder (HSDD) were pooled (n=726) for assessment of MID for SSEs. The anchor used was taken from a questionnaire administered at baseline and end of treatment allowing both a cross-sectional and a longitudinal anchor-based approach. Distribution-based methods used three pooled estimates of variance of the SSEs (baseline standard deviation [SD], change from baseline SD, and SD of the change from baseline among stable subjects), and the empirical rule (1/6 of the range of SSEs at baseline), for effective sizes of 0.2 and 0.5. Interview approach: 77 women with HSDD were interviewed to determine what they considered to be a 'minimum' improvement over what they had experienced in the previous 6 weeks. Conclusions: Using a variety of methods for estimating an MID for an outcome measure has routinely been recommended. However, how to then determine the estimate has not been addressed. We recommend that a single estimate be selected that is in the middle of the set of estimates formed by the limits of the 95% CIs and which provides a reasonable, readily interpretable value. Aims: To evaluate the association between differences in SF-36 Vitality (VT) scale scores with clinical conditions and outcomes. Methods: Analyses were performed on data from the Medical Outcomes Study (n=3445) and the National Survey of Functional Health Status (n=2909). The strategy consisted of two components: (1) comparing VT scores in patients with and without clinical conditions with well-understood effects on VT/fatigue; and (2) investigating whether VT score differences of 5-and 10-points predict outcomes. The first analysis used multivariate regression to evaluate the association between chronic disease conditions and VT. To approximate the impact of within patient VT differences, the second set of analyses used logistic regression models to examine the relationship between baseline VT scores and four outcomes: (1) inability to work due to health; (2) subsequent job loss; (3) subsequent hospitalization, all at 1 year; and (4) mortality at 7 years. Results: VT scores were significantly reduced in patients with anemia (5-6 points), CHF (6 points), and COPD (7-10 points). Both 5-and 10-point decreases in VT score were significantly associated with increased odds of negative outcomes, including inability to work due to health (OR(5)=1.28, OR(10)=1.62), job loss (OR(5)=1.13, OR(10)=1.28), hospitalization (OR(5)=1.08, OR(10)=1.17), and mortality (OR(5)=1.08-1.28, OR(10)=1.14-1.54, depending on VT level). Absolute risk for these outcomes increased non-linearly with decreasing VT. Conclusions: VT scores are reduced in patients with serious medical diseases. Differences of 5-10 points in the VT score were associated with significant increased risk of negative outcomes. A 10-point lower VT score was consistent with 62, 28, and 17% increased likelihood of inability to work, job loss, and hospitalization, respectively, by 1 year. Similarly, a lower VT score was associated with an increased mortality. weekly during and post-therapy up to 12 weeks. Results: Fatigue, pain, lack of appetite, sad, drowsiness, sleep disturbance, and short of breath were the most severe symptoms in the study sample. Uncontrolled fatigue was the most severe symptom. Changes in disease and treatment-related symptom severity were estimated during early, mid, late and post-therapy by mixed effects models with linear splines. Four symptom patterns were identified, with symptoms peaking at the end of therapy in three of the four patterns. Fatigue explained most of the variation in total symptom interference during each of the four periods (r 2 =0.6-0.7). Conclusions: The major symptom burden in patients with NSCLC under chemoradiation starts in the early phase of therapy and peaks at the end of the therapy. The longitudinal symptom study approach is a new way to state symptoms and side effects from aggressive cancer treatment over time. The results provide the valuable information for targeted symptom management for the study sample.

Aims: New healthcare technology can often improve patient quality-of-life (QoL). However, the clinical relevance of QoL benefits may be difficult to interpret. The objective of this study was to assess the clinical importance of QoL differences between apodized diffractive multifocal intraocular lenses (AD-IOLs) and conventional monofocal intraocular lenses (C-IOLs). Methods: In a large prospective multi-center US clinical trial, 495 cataract patients were bilaterally implanted with either AD-IOLs (N=339) or C-IOLs (N=156). Patient QoL data were collected using the Cataract TyPE questionnaire (TyPE). The TyPE is a reliable, validated instrument designed to measure vision related QoL. The TyPE was administered pre-operatively and 6 months after the second eye surgery. Effect size (ES) scores were calculated for all significantly different TyPE dimensions. The ES scores were then compared to clinically important difference (CID) thresholds found in the medical literature. Results: AD-IOL subjects indicated significantly better QoL in 12 TyPE dimensions. ES scores for 11 of the 12 significant TyPE dimensions exceeded CID thresholds. Specifically, AD-IOLs provided small but clinically important differences in reduced social limitation, reduced trouble with night-vision, reduced trouble with day-vision, higher rating of vision, reduced distance-vision frequency-of-spectacle-wear, higher satisfaction with night-vision and higher satisfaction with day-vision. Additionally, AD-IOLs provided moderate but clinically important differences in overall visual status, overall satisfaction with vision, and reduced near-vision frequency-ofspectacle-wear. Lastly, AD-IOLs provided a large clinically important difference in reduced near-vision limitation. Conclusions: AD-IOLs provide greater improvement in cataract patients QoL compared to C-IOLs. The Aims: The Multitrait-Multimethod (MTMM) approach proposed by Campbell and Fiske is a common method for examining convergent and discriminant validity of constructs. However, the full implementation of the MTMM procedure is not easy to put into practice in QOL filed, especially for the QOL measurement designed for certain diseased populations who have not enough capacity to complete many measures with limited time. In this study, two alternative methods on examining convergent and discriminant validity proposed by Fornell and Larcker, and Anderson and Gerbing were conducted. Methods: These two methods examine the convergent and discriminant validity of constructs from confirmatory factor analysis. Specifically, these two methods were applied to examine the convergent and discriminant validity of the four QOL domains (physical health domain, psychological domain, social relationships domain, and environmental domain) in the WHOQOL-BREF using the data from the 2001 National Health Interview Survey (NHIS) in Taiwan. The NHIS was conducted by stratified multistage systematic sampling resulting in responses from 13,012 participants aged 20$65 who completed the WHOQOL-BREF. Results: According to the results, the four domains of the WHOQOL-BREF had excellent convergent validity within each domain. However, the four domains had weak discriminant validity. That is, although the four domains claimed in the WHOQOL-BREF were not the same, they were not distinguished clearly by the items in the WHOQOL-BREF. Conclusions: Researchers can use the same procedure to examine the convergent and discriminant validity of a multidimensional measurement. Further discussion on the two methods will be provided. Aims: To investigate the association between Quality of Life, and perceived parental rearing behavior. Methods: Three hundred and three college students enrolled in Social Medicine classes participate in a cross sectional study. The data was examined through step-by-step regression analysis (SPSS12); the variables entered in a chronological order: (1) Gender and social class. (2) s-EMBU factors for mother and father (rejec-tion, emotional warmth and overprotection). (3) Aims: The EQ-5D proxy version may be used by a respondent other than a patient to evaluate the patient's health status. The study objective was to determine the factors that contribute to patient health status when the EQ-5D proxy is used as a physician assessment. Methods: Physicians participating in a study of the treatment of patients with type 2 diabetes (n=1058) were administered the EQ-5D and visual analogue scale (VAS). Patient demographics, lab values, and diagnostic information were obtained through medical chart review. Physicians were asked about frequency of patient visits and patient compliance. Correlations between patient demographics, lab values, physician ratings, and EQ-5D utility and VAS scores were calculated. All variables significantly correlated (p<0.5) with patient health status were entered into a stepwise linear regression using EQ-5D or VAS as the dependent variable. Results: Mean EQ-5D utility and VAS scores were 0.83 and 61.3, respectively. Correlation between EQ-5D and VAS was 0.47. Utility scores were weakly, but significantly (p<0.05) associated with demographics, lab values, and patient compliance. Stronger associations were found for total numbers of diabetes-associated conditions (e.g., hypertension), diabetes symptoms, foot problems, and medications. Similar associations were found between VAS scores but stronger associations were seen between the VAS and A1C and compliance. Significant contributors to EQ-5D scores were diabetic symptoms, diabetesassociated conditions, gender, frequency of physician visits, foot problems, compliance, patient age, and body mass index (BMI) (r 2 =0.23). Significant contributors to VAS scores were diabetes-associated conditions, A1C, BMI, diabetic symptoms, compliance, medications, and cholesterol (r 2 =0.25). Conclusions: Physicians tend to use comorbidities and demographics as the basis of their evaluations of patient health status using the EQ-5D, but base VAS scores on comorbidities and lab values. Aims: Generic QOL instruments have not been used widely in studies of assaulted victims of non-domestic violence. The aim of this study was to investigate distress reactions as Post-traumatic stress disorder (PTSD) and anxiety/depression, overall QOL and overall health perception to non-domestic violence in the acute post-traumatic phase. Methods: A Norwegian sample of 138 physically assaulted victims answered a questionnaire containing the Impact of Event Scale-15 (IES-15), the Hopkins Symptom Check List-25 (HSCL-25) and WHOQOL-Bref. Results: In addition to the physical injury, which in some situations nearly caused death, the traumatic event also caused psychological distress and pain. Thirty-four per cent of the victims scored as probable PTSD cases according to the IES-15 score and 44% of the participants scored as cases with probable anxiety/depression, according to the HSCL-25. One way ANOVA showed statistically significant differences (all p-values < 0.001) between overall QOL and IES-15 categories (mean probable PTSD cases: 2.95, mean not PTSD cases: 4.05) and between overall QOL and HSCL-25 categories (mean probable anxiety/depression cases: 2.88, mean not anxiety/depression cases: 4.14). One way Anova also showed statistically significant differences (all p-values < 0.001) between overall health perception and IES-15 categories (mean probable PTSD cases: 3.00, mean not PTSD cases: 4.03) and between overall health perception and HSCL-25 categories (mean probable anxiety/ depression cases: 2.95, mean not anxiety/depression cases: 3.99). Conclusions: Overall QOL and overall health perception are connected with the acute psychological impairment that results from experiencing assault violence. Aims: health care assessment system in Nepal is still take long time and high cost. Collection of data about prioritization of community health problems by the community is essential for planning and monitoring of programs and interventions for improving community health status. Scientific Rapid Community Health Assessment Methodology needs to be validated in rural community developing country like Nepal which is economic and less time consuming. Objective: To validate the Rapid Community Health Assessment Methodology (RCHA) for prioritization of community health problems in a rural community. Methods: Study area: Fourteen Villages of rural areas. Study population: Thirty-four Primary School Teachers from 10 randomly selected Primary Schools in above villages. (One Teacher each from Class I to V, in each School). Validation: Heads of households from 500 Households (50 per village, systematically sampled). Data collection technique:

Self-Administered Questionnaires for Teachers, Interview Schedules for Heads of Households. Data analysis: With the help of EPI info program. Results: There was significant correlation between the responses of the school teachers and heads of households on community health problems viz., (a) the prioritization of 10 village problems (r=+0.77, p<0.02), (b) prioritization of utilization of services of various health functionaries for treatment of and advice for children's illnesses (r=+0.75, p<0.05), and (c) prioritization of households using water from different sources (r=+0.975, p<0.02). The method was also found to be more rapid (3.3 times) and less costly (6.3 times) compared to the traditional household survey method. Conclusions: Rapid Community Health Assessment Methodology for prioritization of community health problems in a rural community is validated. The information thus obtained can be utilized for purposes of health policy and program planning, monitoring and evaluation. This is especially relevant for micro planning of child health services in developing countries. Repeated use of questionnaires for monitoring disease control programs must be carefully considered. Further studies to confirm and reconfirm the results of this study may be done before wider application of above methodology. Aims: ISAT has revealed significantly better clinical outcomes for patients randomised to endovascular treatment as measured using the modified Rankin scale (mRS). ISAT compared the proportion of patients with a mRS grade of 3 or over (indicates higher level of impairment and dependency) following endovascular or neurosurgical treatment. The HRQL data were explored to determine whether there are differences in patients ranked 0-2 on the mRS as well as 3 or over. Methods: A subsample of ISAT patients from eight UK centres completed a thorough assessment of HRQL (SF-36 and Functional Limitations Profile) and cognitive function at 12 months following treatment. HRQL data are reported here. Differences were tested using one-way ANOVA with post-hoc comparisons (LSD) between mRS grades. Results: Every domain of the SF-36 declines significantly with deteriorating mRS grade. The post-hoc comparisons reveal that each grade is significantly worse than the next lower grade, apart from grades 4 and 5. Patients within the 0-2 mRS range still report substantial declines in HRQL (PF is 90.1 at grade 0 (n=137), to 66.6 at grade 2 (n=144); VT is 71.6 at grade 0 and 39.9 at grade 2; BP is 93.7 at grade 0 and 64.0 at grade 2). Sample sizes for grades 3 and over are small, but the data indicate some sensitivity to differences in HRQL outcomes in patients at these higher mRS grades. Conclusions: The HRQL data greatly elucidates the differences in health status for patients at different mRS grades. The HRQL data demonstrate that simply categorising patients using a single cut-off score on the mRS is a very crude way of measuring outcomes. Most of the decline in HRQL scores occurs in the 0-2 range. Aims: We evaluated whether there is equal performance of the widely used health-related quality of life measure CHQ-CF87 across Paper and Internet modes of administration in a large school-based adolescent population by assessing the equivalence of (a) mean scale scores, (b) the internal reliability of multi-item scales, and (c) the ability of the CHQ-CF87 to discriminate between groups with/without self-reported chronic conditions (discriminative validity). Methods: One thousand and seventyfour students (13-17 years) were assigned at random to either the standard Paper or an Internet mode of administration of the CHQ-CF87 and items on chronic conditions (at school, under supervision). Results: Participation rate was 87%. Four CHQ scales showed statistically significant differences in mean scores between the Paper and Internet modes of administration (p<0.05). Two scales showed Cohens's effect sizes above 0.20: for the CHQ scale 'General behavior' the Paper mode showed better health (ES=0.22) and for 'Family activities' the Internet mode showed better health (ES=)0.26). For each mode of administration, all scales showed adequate internal reliability (Cronbach's a > 0.70). All CHQ-scales, in both modes of administration, discriminated clearly between adolescents with no and with three or more chronic conditions (p<0.01). However, in the subgroup students that completed the Internet mode of the CHQ, except for one scale, all CHQ scales indicated a smaller negative impact of chronic conditions on health-related quality of life than in the subgroup that completed the Paper mode of administration. Conclusions: Both the Paper and Internet modes of administration of the CHQ-CF87 showed adequate internal reliability of multi-item scales and supported its discriminative validity. Yet, score equivalence between modes of administration was absent for four CHQ scales and the magnitude of impact on HR-QOL of chronic conditions differed per mode. We have no explanations for these findings and recommend further quantitative and qualitative studies since researchers increasingly apply combinations of these modes of questionnaire administration. Aims: To assess the feasibility and validity of the computeradministrated version of the SEIQoL-DW. Methods: Gastrointestinal (GI)-cancer patients (n=20; mean age 65; 50% women) filled out the touch screen version of the SEIQoL-DW. Pharmacy students at Uppsala University (n=40; mean age 25; 92% women) were administrated the touch screen version and the paper and pen version in a cross-over study. Both patients and students were interviewed following their completion of SEIQoL-DW to assess their views of using the instrument. Results: A majority of patients (85%) found the instrument easy to use and understand. The mean time for patients to complete the computer test was 11.5 min. The student data showed that there was no significant difference between the mean QoL scores obtained from the computer and paper and pen versions (73.7 vs. 74.9, p=0.5). The ICC comparing the formats was 0.77 (CI: 0.57-0.88). The time for students to complete the computer version was half the time for completing the paper and pen version (4.7 vs. 9.9 min, p<0.05). The majority of the students (65%) preferred the computer version, while almost a third (27%) preferred the paper and pen version. Most respondents, (82% students; 95% patients), judged their QoL as being equivalent to the mean SEIQoL-DW score. Conclusions: The computer version of SEIQoL-DW seems to suit both older persons, with no prevoius experience of working with computers and younger persons with extensive experience of computers. Most respondents perceived the computer test to be easy to understand and to use. There was no difference in mean SEIQoL-DW scores when comparing the two administration forms. The computeradministrated version of SEIQoL-DW seems to be feasable and a valid alternative to the paper and pen version. In addition this version was timesaving, preferred over the paper and pen version and appear to make data capturing easier.

Aims: To study the relative impact of pain-related variables, demographic variables and co-morbidities on SF-12 Mental and Physical Summary scores. Methods: This was a cross-sectional study using data from the baseline visit of a study to validate a pain treatment satisfaction questionnaire. Study variables used in this sub-analysis were: patient age, gender, education, presence of co-morbidities, pain intensity, pain type, pain frequency, pain treatments and HRQoL. Pain intensity was measured using a 10 cm horizontal Visual Analogue Scale, with endpoints of 0 (no pain) and 10 (unbearable pain), and HRQOL was measured using the SF-12. Data were obtained from 1119 patients, who were over 18 and receiving treatment for pain in Primary Care centers. A multiple linear regression analysis was performed to analyze the impact of study variables on the SF-12 Mental Component Summary (MCS-12) and Physical Component Summary (PCS-12) scores. Results: According to the Beta standardized coefficients, study variables showing a statistically significant association with the MCS-12 were: pain intensity (BETA=)0.163 p<0.000), pain frequency (b=)0.150, p<0.000), gender (b=)0.127, p<0.000), presence of co-morbidities (b=)0.118, p<0.001), pain drug treatment (b=)0.113, p<0.002) and educational level (b=0.103 p<0.009). Variables showing a statistically significant association with the PCS-12: age (b=)0.284, p<0.000), pain intensity (b=)0.206, p<0.000), pain drug treatment (b=)0.172, p<0.000), pain frequency (b=)0.124, p<0.001) and educational level (b=0.084, p<0.022). Rheumatology outpatients using qualitative & quantitative methods. Results: The ST was well accepted by subjects (n=66, 76% female, mean age: 47.3±11.93 years, 73% with >6 years of education) with high (n=43) or low (n=23) literacy, 98% of whom found it easy or very easy to use; 85% found the VTS feature useful. By virtue of design, there were no missing HRQoL data. In low literacy subjects who had never used a computer (30%), none reported any difficulties using the ST. The mean (SD) time spent to complete the ST (2 instruction & 2 practice screens, 24 questions, 1 VAS) for high & low literacy groups was 16.2 (7.58) & 21.0 (7.63) min, respectively. Subjects who expressed a preference for 1 (n=42) or more (n=5) modes of administration favored the ST (n=26) over interviewer (n=16) or self (n=10) administration. Conclusions: The Smiling Touchscreen is well-accepted by subjects with varying levels of literacy, including those who had never used a computer before. It is thus a promising new approach for HRQoL assessment among subjects with varying levels of literacy. P-62/1313/THE EFFECTS OF AGING TO HRQOL OF DIABETIC OUTPATIENTS Hisako Adachi, Nursing, University of Gifu, Gifu, Gifu, Japan; Takaaki Oyamada, Early Childhood Studies, Gifu Women's University, Gifu, Gifu, Japan

Aims: This study aimed to examine the effects of aging to HRQOL of diabetic outpatients by using SF-36. And diabetics were compared with no disease persons. Methods: One hundred and eleven diabetics (Diadetic G.) were recruited the outpatient clinic. The average age of Diabetic G. was 55.6±14.3 years old. Ninety-five persons with no disease (Health G.) were recruited from some volunteer groups. THe average age of Health G. was 48.1±10.8 years old. The subjects of 2 groups were divided into 3 subgroups, respectively by the range of age: younger age group: 19-39, middle age group: 40-64, elder age group: 65 over. Results: (1) Effects of aging: (i) Diabetic G. showed the significant increment of mean score on general health with aging (p<0.05), but the significant decrement of mean scores on vitality and mental health with aging (p<0.05). Middle and elder age groups of diabetics showed significantly lower mean score on physical functioning (p<0.05), however, showed significantly higher mean score on general health in comparison with that of younger age group of diabetics (p<0.05). (ii) Health G. showed the significant increment of mean scores on vitality and mental health with aging (p<0.01), but showed the significant decrement of mean score on physical functioning with aging(p<0.1). (2) Differences between Diabetic G.and Health G.: (i) On the physical functioning, role physical functioning and general health, mean scores of Diabetic G. were significantly lower than those of Health G. (p<0.05). Mean score of younger age group of Diabetic G. was significantly lower than the same age group of Health G. on general health (p<0.05), however, mean scores of middle age group of Diabetic G. were significantly lower than those of the same age group of Health G. on general health and physical functioning (p<0.05). Conclusions: On the effects of aging to HRQOL of diabetic outpatients, general health, physical functioning, vitality and mental health of SF-36 subscales were important facors. Aims: Hypoglycaemia can occur suddenly and is often associated with unpleasant physical and psychological symptoms. If blood glucose continues to fall, severe hypoglycaemia can occur which is associated with delirium, unconsciousness and even death. As a result, it is not surprising that many people with type I diabetes or insulin treated type II diabetes have a significant fear of developing hypoglycaemia. The aim of this literature review was to explore the concept of fear of hypoglycaemia and examine how this concept has been applied in medical and psychological research. Methods: A literature search was conducted on Medline, Embase, Psychinfo, and Cinahl. The search was limited to journal articles published in English from 1985 to 2005 inclusive. Results: A total of 35 papers were identified for inclusion in the review. These studies focus on a number of areas, including the impact of previous hypoglycaemic episodes and anxiety on fear of hypoglycaemia. Qualitative accounts of fear of hypoglycaemia, the negative impact of fear of hypoglycaemia on quality of life and blood glucose self-management were also reviewed and the efficacy of interventions in reducing fear of hypoglycaemia were also considered. Conclusions: Overall, it was seen that fear of hypoglycaemia has a negative impact on many aspects of quality of life. Anticipation of hypoglycaemia can prove to be a large psychological burden and can lead individuals to engage in unsafe avoidance behaviours such as deliberately maintaining a high level of blood glucose. There is some evidence to suggest that various interventions can reduce levels of fear which will have a significant impact on the health related quality of life of a person with type I diabetes. Aims: The number of people diagnosed with diabetes is increasing and the disease has considerable effect on people's quality of life. This study was carried out to measure quality of life in patients with diabetes and compare QOL among Iranian with and without diabetes. Methods: A cross sectional study was conducted in Tehran, Iran. Quality of life data were collected for a convenience sample of 131 patients with diabetes and then it was compared to data for an age matched random sample of the community (age range between 14 and 69 years, n=4004). Quality of life was measured using the SF-36. Results:

The study results are summarized in the following Table. In all measures people with diabetes showed significant lower scores (all p values < 0.01). These differences were more profound for physical functioning and bodily pain. Aims: To explore the patient experience of change in visual symptoms related to diabetic retinopathy in pre-and post-laser treated patients. Methods: This qualitative interview study was implemented at four eye clinics in the UK. One hundred and twenty-eight diabetics with proliferative diabetic retinopathy and 118 diabetics with macular edema were recruited. Patients were grouped into three groups based on stage of laser treatment: (1) first laser treatment, (2) first follow-up and (3) multitreatment. Multi-treatment patients were required to have clinically documented loss of visual function in at least one eye (VA £ 6/12). Patients in groups 1 and 3 received follow-up telephone interviews post-treatment. Group 2 patients were asked to recall symptoms before their first laser treatment as well as after their last treatment. Symptom differences between groups were evaluated by Wilcoxon-rank sum test. Results: Patient expressions of visual symptoms were elicited. The most frequently reported symptom prior to initial treatment was blurred vision (55%); the least frequently reported symptom was micropsia. Self-reported blurred vision showed a marked decrease (27%) following initial laser treatment. First laser treatment patients reported fewer symptoms than multi-treatment patients. Median visual symptoms in group 1 decreased post-laser treatment (2 vs. 1; p=0.07), while median visual symptoms in group 2 did not change (2 vs. 2; p=0.59). Group 3 patients reported a statistically significant increase in the number of symptoms experienced post-treatment (p<0.001). Conclusions: The relationship between the symptoms resulting from the progression of diabetes and those from multiple laser treatments is complex. The measurement of symptoms may not translate to improvements in quality of life due to the underlying progression of diabetes. Language subscale showed significant improvement in the VPA+MEL group after adding melatonin (p=0.005). However, not much significant difference was seen across all the four groups in the language subscale. In the depression subscale, significant difference was seen among all the four subgroups post-treatment. In the selfesteem subscale, significant difference was seen among all the four subgroups post-treatment (p=0.0006). In the social interaction subscale, significant difference was seen across all the four groups in the post-treatment score (p=0.01). No significant difference was seen among the four groups in the other subscales. Conclusions: To our knowledge, this is the first study assessing quality of life in epileptic children with add-on melatonin administration in form of a randomized double blind placebo controlled trial. Add-on melatonin, which has a wide safety margin, can be of promise in pharmacotherapy of pediatric epilepsy by improving quality of life. (2) Significant differences were observed between T1 and T2 in a STAI score and some scale scores of disease-specific QOL, i.e., speech, eating, communication and worry/concern (SOHSI), social aspects of dentofacial deformity, facial aesthetics and oral function (OQLQ), recognition and satisfaction.

(3) T2 showed significantly higher score on satisfaction of facial esthetics, and worry/concern about teeth and mouth than those of CONT.

The results of this study suggested that QOL of orthodontic patients who required jaw surgery generally improved following treatment. Particularly they showed higher satisfaction on their facial esthetics, and were more interested in their teeth and mouth post-operatively. Aims: Health profiles, such as the MOS-HIV, have been frequently used to measure health-related quality of life. However, the MOS-HIV in its current form cannot be used to measure individuals' utility of health states. The purpose of this study is to predict EQ-5D utility scores using the MOS-HIV. We applied different statistical methods to address ceiling effects seen with the EQ-5D scores. Methods: Data were collected from a treatment trial for HIV (n=1126) in the US. We developed four models to predict EQ-5D utility scores using the MOS-HIV: Model 1 includes physical and mental summary score (PHS/MHS); Model 2 includes the 10 subscale scores; Model 3 adds age, gender, CD4 count and viral load into Model 1; Model 4 adds age, gender, CD4 and viral load into Model 2. We applied three statistical methods: Method 1 is ordinary least square (OLS) approach; Method 2 is censored least absolute deviations approach (CLAD) which emphasizes medians and is more robust to ceiling effects; Method 3 is two-part model (TPM) predicting EQ-5D with and without ceiling effects separately. We used 10% random samples for model development and the remaining 90% for model validation. We used R-square (R 2 ) and mean absolute error (MAE) to evaluate the model predictive validity. Bootstrap sampling (n=1000) was used to estimate 95% confidence intervals of the predictive validity. Results: The distribution of EQ-5D utility scores was highly skewed with a mean of 0.87 and 40% ceiling effect. Significant predictors of EQ-5D scores across different models were PHS, MHS, physical function, pain and health perceptions (p<0.01). For predictive validity, models that included PHS and MHS were more powerful than those included the 10 subscales. Aims: Uncorrected refractive errors is one of the major causes of preventable blindness in rural communitees in developing Nations like Nigeria with 60-70% of its population living in the rural areas. The aim of this study was to determine the impact of vision correction on quality of life of recipients in five rural communitees in south eastern Nigeria. Methods: The National Eye Institute Refractive Error Quality of Life (NEI-RQL-42) instrument was administered to 1216 subjects with refractive errors before vision screening exercises in five rural communities in south eastern Nigeria. Questionnaires were readministered to subjects in a follow up visit to the same communities after 2 months. Only 223 subjects who were corrected with glasses during the previous screening programmes were seen in the following categories: myopia (n=65), hyperopia (n=48), presbyopia (n=110). The principal outcomes assessed were visual acuity and changes in NEI-RQL scores. Paired student's t test and Pearson's product momment coefficient were used to analyse results. Results: The mean QOL scores on the 0 to 100 possible score range before and after vision correction were 26 and 88 respectively. Statistically significant difference was observed between baseline and after vision correction in the 13 scales on the NEI-RQL (p<0.05). The QOL scores were closely correlated with presenting visual acuity before and after vision correction (r=0.51-0.62). Presbyopes reported the largest improvements in 9 of the 13 scales. There was no significant difference in scores between myopes and hyperopes before and after correction.

Conclusions: These results demonstrate that preventable blindness are avoidable, and quality of life can also be improved in rural communitees in developing nations with vision correction. Aims: The purpose of the paper is to describe a qualitative method for determining factors women consider to be instrumental in defining their quality of life in making decisions for breast cancer as well as living with the outcomes of those decisions. Methods: A series of qualitative studies have identified attributes of choices facing women diagnosed with early stage breast cancer. An initial sample of 48 women were interviewed at the time of diagnosis and a series of follow-on studies have refined the decision making experience as well as the attributes of treatment alternatives (mastectomy or lumpectomy) in an effort to determine women's values in selecting an appropriate treatment. The interviews were transcribed and analyzed using the constant comparative method of Glaser and Strauss (1967) . Results: The attributes of the treat-ment alternatives (expedience, health, safety, survival, and body integrity) provide a means for women to determine their relative importance in making a treatment decision that is most likely consistent with her values. Further, the personal selection and ranking of the most important attributes, provides a unique preference structure that helps define individual profiles of quality of life. Qualitative analysis indicates that an individual's personal assessment of quality of life is shaped by her satisfaction with obtaining the most valued attributes of breast cancer treatment.

Conclusions: The conclusions for this paper are: (1) the richness of qualitative data yields important information regarding how individuals assess their personal quality of life in making treatment decisions, (2) that the values expressed in the selection of a treatment option and the same as those values that shape post-decision quality of life, and (3) the attributes described here remained consistent across a number of studies and will be useful in future measurement studies examining quality of life outcomes specific to women with early stage breast cancer. Aims: Among the CDC's battery of HRQOL items are two addressing the number of physical or mental unhealthy days one has experienced in the last month. These items are commonly combined to form an unhealthy days index, and many studies have examined the relationship between a specific disease or a few predictors and unhealthy days. We believe that creating such an index may obscure important differences in characteristics that affect physical and mental health and also, that more complete models that go beyond the single disease approach are necessary to understand the complexities of HRQOL. This study AIMS to develop separate, comprehensive predictive models for physical and mental unhealthy days. Various demographic, chronic disease, physical activity, nutrition, health care access, and alcohol and tobacco use variables will be examined. Methods: The Behavioral Risk Factor Surveillance System (BRFSS) is a cross-sectional telephone survey conducted yearly under the guidance of the CDC. Using standard instruments, samples of adults are interviewed by each state and sent to the CDC for consolidation, weighting, and reporting. In this study, 2003 BRFSS data (n=264,684) were used. Potential predictor variables and relevant interactions were tested using the regression procedures in SUDAAN 9.0. Results: In addition to many significant first-order effects, several disease by disease and disease by behavior interactions emerged, particularly in the prediction of physical unhealthy days. For example, both diabetes and arthritis are associated with more physical unhealthy days. These variables also interact such that individuals who are both diabetic and arthritic expect an additional unhealthy day per month. Interesting differences also emerged through creating separate models for physical and mental unhealthy days. For example, females experience more mental unhealthy days but there was no gender effect for physical unhealthy days. Conclusions: The prediction of HRQOL is a complex phenomenon that lends itself most appropriately to a person-centered rather than disease-centered approach. In addition, although related, there are advantages to modeling physical and mental unhealthy days as distinct phenomena. Aims: The relationship between health-related quality of life (HRQoL) and leisure time physical activity (LTPA) in the general population has mainly been investigated in cross-sectional studies. This study aimed to assess whether the usual level of LTPA and its variation over a 3-year period is associated with changes in HRQoL. Methods: 3891 participants of the SU.VI.MAX study, a nutritional trial conducted in France, filled out the 36-item Short Form (SF-36) and the Modifiable Activity Questionnaire (MAQ) in 1998 and 2001, in order to assess HRQoL and LTPA, respectively. Linear regressions were performed to determine the association between changes in LTPA and changes in HRQoL. Results: Cross-sectional results showed that LTPA was associated with all dimensions of HRQoL, in both sexes. Over the 3-year follow-up, changes in LTPA were associated only with changes in the mental component of HRQoL in women. Independently of baseline LTPA and HRQoL, an increase of one hour per week in LTPA was associated with a 0.23-point increase in this component in women, but was not significant in men. Conclusions: Although mean changes over time in HRQoL components in this study were of limited magnitude, results suggest that, especially in women, increasing LTPA may help maintenance of HRQoL mental component over time in middle-age subjects. Aims: To our knowledge, no study has examined whether perceived communication between breast cancer patients and health care providers could be a determining factor in explaining the quality of life of these women at various phases of the disease. The main objective is to ascertain whether communication between breast cancer patients and health care providers (physicians, nurses, technicians in radiotherapy) influences the quality of life of these women at diagnosis and during radiotherapy. Methods: 120 French-speaking women aged 35-78 (mean=55 years) with early breast cancer diagnosis (stage 1 or 2) who underwent lumpectomy and completed adjuvant therapy were recruited at two Montreal hospitals. They were asked to complete, around the diagnostic period and during half-way of radiation therapy, either at the hospital or at home, a questionnaire comprising demographic and medical information, the MOS Social Support Survey, two quality of life questionnaires: EORTC QLQ-C30/BR23 and FACT-B, the Medical Communication Competence Scale, the Perceived Physician's Communication Style Scale, and a Visual Analog Scale on perceived communication. Results: Multilevel analyses demonstrate the manner in which physicians, nurses and radiotherapy technicians interact with breast cancer patients (information provided, style of communication and socioaffective aspect of communication) influences the quality of life of these women at diagnosis and during radiotherapy. Conclusions: By better understanding how communication affects the quality of life of these women at diagnosis and during radiotherapy, it should enable health care providers to change their approach towards their patients by using better communication strategies in order to improve the quality of care and quality of life of women afflicted with early breast cancer. Aims: MPM is a relatively rare, yet aggressive and often fatal cancer. The survival rates are poor with a median of 7-11 months after diagnosis. However, until recently very little reference data exist to understand the symptoms or QOL of MPM patients, particularly for the EORTC QLQ-C30 and EORTC LC13 module. Such information is critical to design/calculate future studies, and allow for comparisons between groups. Methods: These data were collected prospectively from an international, multi-centre study (EORTC trial 08983/NCIC BR17). Only baseline QOL data were reported. The trial used the EORTC QLQ-C30/EORTC LC-13. Data were compared to a normal population and a reference NSCLC sample. Results: During March 2000 -January 2003, 250 patients were recruited, with 229 providing a valid baseline QOL forms (91.6%). The mean age was 58 years, with 80% being males. Baseline scores, compared to a normal population sample showed a large clinically significant impairment (>10 points) on all scales with the exception of cognitive and emotional functioning, nausea/vomiting and diarrhea which were close to the normal population. Symptomatic problems were also seen in the LC 13 for various pain scales (>19), along with dyspnoea (32.7) and coughing (30.2). Compared to NSCLC reference data MPM patients had clinical significantly more pain and a greater impairment of role functioning. Conclusions: When compared to reference data from NSCLC and normal reference populations, these data provide clear evidence of the impairment in QOL in MPM patients. Our results also highlight the major symptom difficulties in untreated MPM patients. These data should be used to help design future trials, and understand the clinical improvements from new and promising therapies.

Aims: The WHOQOL-BREF is a commonly used measurement in QOL research. Although previous psychometric studies have showed that the WHOQOL-BREF has adequate validity regarding to factor analysis, criterion-related correlations and discriminant ability on health-disease groups, however, the content validity of the WHOQOL-BREF was rarely examined. Content validity, the extent to which a measurement reflects the specific intended domain of content, is a basic type of validity of a valid measurement. The purpose of this study was to examine the content validity of the WHOQOL-BREF. Methods: Contrary to the commonly used method by experts' judgments, 120 undergraduate students were asked to rate the appropriateness of each item with the four domains of the WHOQOL-BREF, respectively. Then, three quantitative methods on examining content validity were applied to analyze the rating data to give a quantitative criterion of the content validity, including (1) substantive validity (Anderson and Gerbing 1991), (2) content adequacy (Schrriesherim et al. 1993) , and (3) ANOVA approach (Timothy & Tracy 1999) . In addition to the appropriateness rating, participants were also asked to sort the items into the four domains of the WHOQOL-BREF (physical health domain, psychological domain, social relationships domain, and environmental domain). The rating and sorting results were used to compare with the original content structure of the WHOQOL-BREF, to see if the original structure is consistent with the structure of subjects' judgments. Results: The results showed that the content structure gained from appropriateness rating and item sorting were generally consistent with the original content structure of the WHOQOL-BREF. Conclusions: The finding of this study indicated that the WHOQOL-BREF has adequate content validity. However, a few items were not consistent with the original content structure. More discussion on these items and the issue of content validity will be provided. Aims: This prospective cohort study compared the levels of psychological distress, quality of life and patient satisfaction, between women who underwent breast reconstruction with either expander/ implant (Implant) or transverse rectus abdominis musculocutaneous flaps (TRAM). Methods: Patients scheduled to undergo postmastectomy reconstruction at MD Anderson Cancer Center were prospectively recruited to complete a questionnaire, which included assessment of anxiety (IES), depressive symptoms (CES-D), body image, and quality of life (SF-36). Visual analogue scales were also used to compare patient satisfaction with surgical outcomes. These questionnaires were completed before surgery, and 1, 6, and 12 months after surgery. Ninety-one patients (55 TRAM, and 36 implant) included in this study completed questionnaires on at least 2 time points (pre-surgery and one of the follow-ups). Eighty-two percent completed all 4 of the questionnaires. Results: Prior to surgery, both groups had no significant difference in physical component scale (PCS), body image, or satisfaction with their chest area. However, TRAM group had significantly better mental component scale (MCS), less anxiety and fewer depressive symptoms compared to women in the implant group. Using Generalized Estimating Equations, both groups showed no difference in their PCS over time. After an initial postoperative small drop in body image and satisfaction in their chest area, both groups showed gradual improvement. However, only TRAM group showed significantly higher improvement to the pre-surgery level compared to the implant group (b=)1.4, p=0.001, and )0.76, p=0.001 respectively). Women who underwent the TRAM procedure also showed consistently better levels of MCS (b=)2.1, p=0.01), anxiety (b=3.3, p=0.04) and depression (b=1.7, p=0.03). Conclusions: Women who undergo TRAM reconstruction have a better chance to return to their pre-surgery level of body image and satisfaction with their chest area within 1 year compared to women who have implants. Overall physical activity (as measured by the SF-36) did not differ by reconstruction type. Aims: Self-administered measures are widely used to measure quality of life and health status. Research demonstrates that format and layout affect respondent comprehension and impact ability to successfully complete measures, thus influencing data quality. The Measure Format Checklist is a systematic, comprehensive tool for evaluating measure format. Methods: After reviewing a diverse literature, including cognitive science, typography, survey research, health literacy, and clear language, we developed a 60-point checklist encompassing six formatting domains: overall layout, measure organization, text formatting, use of symbols, images, and visual relationships between questions, instructions, and responses. The checklist was tested on a diverse sample of self-administered mental health measures. Results: No single formatting feature was problematic for every measure. Recurring features that proved central and problematic in a number of mental health measures included: ambiguous navigational path, small font size, confusing skip patterns, and inconsistent presentation of response formats. The most serious formatting problems were detected in multiple domains of the checklist. A single feature, such as prominent irrelevant information, can lead to problems in other domains, such as confusion about the navigational path and a generally difficult-to-read appearance that may discourage a respondent from completing the measure. Moreover, we realized that format attributes were not uniformly good or bad. Characteristics can only be evaluated in context, while also considering respondent demographics and measurement goals. Conclusions: The Measure Format Checklist provides a structured evaluation of subjective visual elements and can be a useful tool for evaluating the format of self-administered measures, offering systematic assessment that can lead to format refinement, enhanced comprehension, and improved data quality. Aims: To examine consistency in parents assessments of health status and HRQL of their child with severe vision impairment (SVI) using two methods: a generic instrument and a semistructured interview. METHODS: Parents of children with SVI aged 3 to 8 were contacted through statutory services and voluntary organisations in England. The main carer was invited to complete a questionnaire and/or to take part in a semi-structured interview. The questionnaire included the Health Utilities Index as a measure of their child's health status and HRQL. The interview included sections on their child's daily life and activities. Information was also collected on the child's ophthalmic condition and other disorders. Results: Parents of 48 children completed both a questionnaire and an interview. The mean age of children was 5.8 years. For 24 children, their SVI derived from an eye condition (eg retinopathy), for 19 from a visual pathway condition (e.g. optic nerve hypoplasia) and for 5 from nystagmus alone. 32 children had an additional disorder. Parents' qualitative accounts of their child's daily life conveyed assessments of the level of functional ability similar to those identified by the HUI2 health status classification system. The exception was Mobility, where lack of confidence was seen as significantly limiting ability to get around independently. Parents also focused on the social consequences of limited functional ability (stigma and social exclusion) in describing its significance. In contrast to the HUI2 overall HRQL scores, most parents described their child's overall HRQL as good on the grounds that they were happy and generally free from discomfort. The difference between parents' qualitative assessments of HRQL and the HUI2 HRQL scores was greatest for children with visual pathway conditions and children with additional disorders. Conclusions: Parents make judgements about specific domains of their child's health consistently across methods. However, global assessment of HRQL differs between qualitative interviews and the HUI2. Complementary methods may be needed to evaluate services. Aims: Collecting and monitoring longitudinal QOL data in patients with PBT is not the standard of care for many health care institutions. In this study, we evaluated the feasibility and utility of collecting and monitoring longitudinal QOL data for up to 2 years in patients with BTP. Methods: With IRB approval, newly diagnosed patients with PBT referred to the IHC Neuro-Oncology Service were prospectively enrolled from January 2003 to December 2004. Validated QOL measures (EORTC QLQ-C30, EORTC QLQ-BCM20, Spitzer QOL Index, Schwartz Fatigue Scale, HADS) and standard clinical measures (KPS, MMSE, Neurological Function Status, Neurological exam and brain imaging) were collected from enrolled patients in 3-month intervals. No proxies were allowed. A database was developed to collect and display over time changes in global and subset QOL scores. Repeated Measures Analysis of Variance was performed to determine if patients experienced changes in QOL factors over time. Results: We enrolled 54 patients: 48% (26/54) were female, median age=43 years (range 18-71 years). 91% (49/54) completed the baseline and follow up questionnaires without assistance. QOL measures, global health status (p=0.0028), role functioning (p £ 0.0001), emotional functioning (p=0.0116), and social functioning (p<0.0001), significantly improved over time in these patients. The presence of pain (p=0.0425) and insomnia (p=0.0008) decreased over time. However, fatigue (p=0.0193) and future uncertainty (p=0.0001) fluctuated over time, and constipation problems (p=0.0437) increased over time.

Conclusions: Collecting longitudinal QOL data is feasible, and more importantly, it improves the care of patients with PBT. This process is facilitated by the use of a database as a monitoring tool to collect and display changes in QOL scores prior to office visits. Aims: Self-management educational interventions commonly link health-related quality of life (HRQL) outcomes to changes in behavioral skills. Measurement of cognitive changes will guide development of intervention models with components that motivate patients to adopt better self-management strategies as well as improve health behaviors. Methods: This meta-analysis evaluates the impact of chronic illness self-management educational interventions on cognitive function compared with other patient teaching and learning activities in adults. Reports are drawn from a key word search of electronic databases and indexed bibliographies, review of peer-reviewed journal articles and other reports, and search of references of studies and review articles. Data are from English language reports of randomized and other controlled trials with separate pre-post comparison controls, involve structured group interventions in adults with chronic conditions, measure change in at least one cognitive variable, and have adequate statistical reporting for effect size (ES) calculation. Coding includes characteristics of the reports (n=33), research quality and method of data collection, intervention components, intervention fidelity, and standardized mean differences of outcome variables. Results: Overall results indicate (p<0.05) a modest ES (0.43) that remains consistent over self-management performance (ES=0.51) and cognitive (ES=0.34) domains as well as over mediator variables. Groups show significant homogeneity of effect. Data quality supports a high validity framework. Conclusions: Understanding cognitive processes and outcomes in self-management educational interventions will contribute to understanding of the developmental learning experience affecting HRQL in the chronic illness setting. Use of detailed experimental study designs will enhance outcome validity and generalizability. Interim analyses are limited to 1473 peri-and postmenopausal women. Final analyses will utilize the full sample of nearly 5000 women. In addition to reporting VMS frequency, severity, and related healthcare, respondents completed the Menopause-Specific Quality of Life Questionnaire (MEN-QOL), a 29-item self-report measure assessing the presence and bothersomeness of menopausal symptoms. Subjects were classified into three groups based on VMS in the past month: (1) met FDA criteria (7 or more moderate-to-severe daily hot flashes/night sweats); (2) hot flashes/night sweats but less than FDA criteria; or (3) no hot flashes/night sweats. Results: Mean age of the preliminary sample was 55.3 years. Over half (54.4%) reported hot flashes and 46.4% reported night sweats within the past month. While nearly 40% reported experiencing both types of VMS, only 7% met FDA criteria. MENQOL scores between the peri-and post-menopausal groups did not significantly differ. However, significant differences in MENQOL total and subscale (Vasomotor, Psychosocial, Physical, Sexual) scores were observed among the three VMS groups, with both VMS groups reporting significantly diminished QOL (p<0.05). Conclusions: The current study provides much-needed prevalence estimates of VMS and their QOL impact among menopausal women. US regulatory guidelines focus on women with moderate to severe menopausal symptoms and support a treatment indication targeted only at this segment of the symptomatic population. However, VMS significantly impact QOL for a much larger portion of the peri-and post-menopausal population.

Aims: To understand if and how the systematic use of HRQOL information is used in oncology practice in the US and to propose ways in which such measurement can become more widely used and effective in helping with clinical decisionmaking. Methods: Systematic review of peer-reviewed and other literature in English after December, 1989 and interviews with oncologists and quality-of-life researchers. Results: Factors affecting the use of HRQOL in routine practice are distinct from those in clinical trials. Several studies have reported the feasibility and acceptability of HRQOL measurement in clinical practice, but few examples in routine oncology practice could be identified. In addition to well-recognized instrument-related issues, challenges include clinician, patient, and health system issues. For clinicians, the effect on clinic work flow, resource constraints for data collection and management, added value in clinical decisionmaking, and reimbursement for measurement and services provided in response to patient need are noteworthy. For patients, acceptability, response burden, data confidentiality and use may be important. Patients providing information will expect it to affect their care. Health care organizations will need to deploy resources for data management, timely review of data, and to respond to identified patient needs. Conclusions: Even with well-accepted measures, effective use of HRQOL information is not likely to occur without using principles of active dissemination and exploring new forms of organizing clinical practice rather than adding data collection tasks to a patient visit. This may involve new approaches to medical records including the use a variety of electronic platforms for data collection and new ways of organizing patientclinician interactions. To respond in a timely way to patients' needs and to improve clinical decisionmaking, for example, one promising approach is for patients to enter data between, and perhaps instead of, face-to-face visits using the internet, hand held devices, cell phones, and other means. Aims: Over the past 20 years, an increasing number of PRO instruments have been developed and are being used in multinational research and clinical practice. To help researchers in the choice of the most appropriate instruments for a given application, the web-based ProQolid database provides structured, synthesized, reliable, and updated information on hundreds of PRO instruments to all those involved in health care evaluation. Methods: To provide accurate and current information to ProQolid's 800+ daily users, regular updates of the information on each instrument are performed in close collaboration with the instruments' developers. In addition, more than 50 new instruments are added to the database each year. Results: Since the launch of the ProQolid project (initially known as QOLID) in January 2002, 4 updates of the entire database have been completed and over 150 new instruments have been added. The database, which started with 300 instruments 3 years ago, currently describes over 480 PRO instruments. To be eligible for inclusion, an instrument must be the subject of a publication that describes its development and/ or validation. During each update, the instruments' developers review and validate the existing information and complete new information fields for their instruments. The four previous updates have allowed the inclusion of information on the domains covered by the instruments, the main scoring rules, the methodology of development, and the internal consistency reliability for each instrument (when relevant). The 5th update of ProQolid began in March 2005 and will include the review of the reproducibility (test-retest reliability) and clinical validity of each instrument (explores logical relations that should exist with other clinical, biological, and level-of-severity measures). Conclusions: Its facilitated access, broad coverage, and regular updates ensure that ProQolid is a major Internet source of reliable and current information on existing PRO instruments.

Aims: The specificAIMS of this study were to: (1) compare the quality of life of midlife men and women in Australian and Taiwan; and (2) to explore the impact of country of residence and gender on quality of life. Methods: A cross-sectional, supervised self-administered survey was conducted using census data and a probability proportional sampling (PPS) strategy on a general population of men and women aged 40-59 years old who live permanently in Taiwan and Australia. The sample consisted of 278 Australians and 398 Taiwanese men and women. Results: Analysis showed that country of residence does have an impact on quality of life with significant differences seen between midlife Australian and Taiwanese men and wo-men (F=59.31, p<0.001). The results suggest that midlife Australian men and women have a better quality of life than midlife Taiwanese men and women. No significant differences were seen, however, between gender and quality of life (F=1.91, p=0.107), and between gender, nationality and quality of life (F =0.18, p=0.948). Conclusions: This study has addressed and contributed to the assessment of multi-cultural quality of life research and has important implications for both health providers and policy makers of both countries. Consideration of the areas where quality of life is enhanced in Australian midlife men and women could help to formulate, enhance and utilize strategies to improve standards of living and quality of life now and in the future. Aims: To investigate changes in the Health Related Quality of Life (HRQoL) of patients receiving different treatments for prostate cancer during the first 6 months of treatment. Methods: A longitudinal, prospective, multicenter study was performed in patients with a diagnosis of prostate cancer. Data on clinical variables was collected and HRQOL was measured using the RAND 36-Item Health Survey and UCLA Prostate Cancer Index (UCLA-PCI). In this intermediate analysis, changes in global and dimension scores on the HRQOL after 6 months of treatment were compared across different treatment groups. Results: Useable data was collected from a total of 1426 patients: mean age was 74.4 years; 10% received surgical treatment for prostate cancer during the study period, 3.9% received radiotherapy, 70% hormone therapy, 1.4% chemotherapy, 13.5% combination treatment, and 1.1% other types of treatment. When analyzing the overall sample, significant differences were found after 6 months of treatment on 9 out of 14 HRQOL dimensions (3 on the SF-36, and 6 on the UCLA PCI). Statistically significant differences 6 months from baseline were seen on nine dimensions in the hormone therapy group (with statistically significant improvements on six dimensions); on eight dimensions in patients receiving combination therapy (with statistically significant improvements on two dimensions); on six dimensions in the surgery group (with improvements on two dimensions); and on two dimensions in patients receiving radiotherapy and chemotherapy (with improvements on one dimension in each group). The dimensions showing improvement in the highest number of groups were urinary function and sexual bother, which improved in 5 and 4 of the different treatment groups, respectively. Conclusions: Of the five treatment groups studied, those which showed improvement on the highest number of HRQOL dimensions were hormone therapy, surgery, combination treatment, radiotherapy and chemotherapy, in that order. Aims: There is little evidence on agreement between patient and proxy reports on change over time in health status and healthrelated quality of life. In a longitudinal study of a communitybased cohort of patients admitted for treatment of hip fracture, we examined the agreement for changes in the HUI Mark 2 (HUI2) and Mark 3 (HUI3), Functional Independence Measure (FIM) and hip pain over 6 months. Methods: 137 patients who were 65 aged and older along with their respective primary caregivers were enrolled within 3-5 days of surgery. Baseline assessments included the FIM, an 11-point Likert-type pain scale, and the HUI2/HUI3 questionnaire. Follow-up assessments were completed at 1, 3, and 6 months. Agreement of patient-proxy pairs was evaluated using intraclass correlation (ICC) for four intervals: baseline to 1 mon; 1 to 3 mons; 3 to 6 mons and baseline to 6 mons. Results: Vision, hearing, speech and dexterity, attributes which were not directly affected by hip fracture, showed little change over the 6 months of recovery. Large changes from baseline to 6 months with reasonable agreement were seen with ambulation (ICC=0.79), mobility (0.65), and self-care (0.48). Agreement for change scores for overall HUI2 scores was poor to moderate; ICCs ranged from 0.17 to 0.58. Agreement for change scores for overall HUI3 scores was poor to good; ICCs ranged from 0.36 to 0.65. No group differences were seen (p>0.05) except for the change between baseline and 1 month. Conclusions: Physically based observable attributes such as mobility, self-care and ambulation, for which change scores were large during recovery from hip fractures, had a higher agreement for change scores than for less observable attributes such as cognition and emotion. Agreement for change in scores is lower than agreement between patient and proxy respondents observed cross-sectionally. Aims: To explore patients' and oncologists' perceptions of the use of a computer administered individualized QoL instrument to support the consultation. Methods: A purposive sample of patients with gastro intestinal cancer (50% females; mean age 60 years) at Uppsala University Hospital and Karolinska University Hospital, Stockholm, Sweden, completed a computer-administered version of the SEIQoL-DW before their medical consultation. Both the patient and the doctor received a copy of the results to support the consultation. All patients were interviewed directly after the consultation to identify their perceptions of the QoL test, while six doctors participating in the intervention, were interviewed at the end of the study. All the interviews were audio taped, transcribed verbatim and analysed according to the phenomenographic approach. Results: Most patients and doctors were optimistic about the instrument, since it allows for systematic monitoring of the patient's QoL and detection of the patient's areas of concern. It was deemed to empower the patient, give the doctor a broader picture of the patient, and inform clinical decision making. Still, QoL results were only brought up to discussion in a minority of the consultations, because doctors were short of time or unsure of how to use the instrument. Conclusions: There is a potential for using individualized QoL instruments in clinical practice, since it provides a holistic and patient centred means of follow up of care and treatment. However, for the intervention to be effective, it is vital to insure that the doctors have internalised this new way of working. Randomized and controlled trials assessing the outcomes of the intervention have to be conducted before wider implementation into routine clinical practice can be recommended. Aims: The EQ-VAS, a measure of self-reported health status, has been operationalized in ways that depart from the original format. This study examines the equivalence of the original paper-based vertical format with a touch screen-based horizontal format. Methods: Non-probability sampling was employed to recruit 314 subjects intended to reflect the general adult population. A two part questionnaire was administered in a randomized crossover design. One part was the original paper-based 20 cm vertical EQ-VAS; the other part was touch screen computer-based (designed by Assist Technologies) and included, among other items/scales, a horizontal EQ-VAS, the SF-36, and socio-demographic items. The two EQ-VAS formats were completed roughly 10 min apart. To test for minimally important differences (MID) between EQ-VAS scores, a difference of half a standard deviation ($ 8 points on the 100 point scale) was used as the equivalence threshold. Results: The mean (SD) EQ-VAS score was 81.0 (15.4) on the paper and 79.6 (15.2) on the touch-screen. The mean (CI) difference between scores on the two formats was 1.4 (0.19-2.58) points and the mean absolute difference was 5.3 (4.22-6.44) points. The intraclass correlation coefficient (ICC) was 0.75, indicating good agreement between the two scores. Almost a third (30.1%) of the respondents reported identical scores on both formats and 80.1% of the respondents had difference scores within ±8points. Using nonparametric bootstrap techniques, both the mean difference and the mean absolute difference between scores on the two formats were significantly less (p<0.001) than the equivalence threshold. Conclusions: These results provide evidence for the measurement equivalence of this EQ-VAS touch screen format with the original paper format. Aims: For many years, outcomes research in schizophrenia was limited to the measurement of symptoms, usually rated by a clinician during patient interviews. More recently, research into Patient Reported Outcomes (PROs) has become increasingly common and recognised as important. Subjective self-report measures place the patient at the centre of the assessment process providing unique and valuable insights, complementing clinicians evaluation of symptoms. This, however, assumes that the patient is sufficiently self-aware to make such assessments independent of symptomatology. The aim of this review was to provide an overview of the use of PRO measures in symptomatic schizophrenia. Methods: An electronic search (MED-Line and PsycINFO) was conducted using the terms: acute schizophrenia, symptomatic, psychosis, questionnaire, quality of life, self report. Results: Of the 83 relevant papers, only 31 were empirical studies measuring PROs in patients with symptomatic schizophrenia. While several PRO measures have been used, most studies used clinician-administered interviews. Although comparison between studies is difficult due to varying measurement and lack of consensual definition of symptomatic schizophrenia, where self-reported questionnaires were used (mainly to assess symptoms), most patients completed them with little difficulty and reasonable reliability. Conclusions: Successful use of self-reported questionnaires infers that full assessment of PROs might be possible. However, quality of life instruments are generally more cognitively complex than symptom measures. The question of whether or not patients with symptomatic schizophrenia are capable of providing reliable subjective report could not be comprehensively answered by this review. Therefore an empirical study has been initiated with the aim of addressing this specific question. Aims: This study aims at analysing the relationship between attitude to drugs and drug treatment, general health, psychological well-being, and adherence to the prescribed drug treatment among elderly patients recently discharged from hospital. Methods: Two-hundred patients over 65 years of age who had been admitted to an emergency medicine ward and had been treated with at least one drug participated in an interview study 1 week after discharge from the hospital. The patient provided information about his/her current drug therapy. This information was compared with the prescriptions documented in his/ her medical record in order to evaluate concurrence. The patient's drug use was compared by multinomial logistic regression with the patient's attitude to the prescribed drug therapy as well as with three instruments that measure general health: Health Index (HI), Well-Being Questionnaire (WBQ-12), and the individual's capacity for coping with stressful events, Sense of Coherence (SOC). Results: Thirty percent of the study population reported a drug use in agreement with the medical record, 28% used less drugs than prescribed, and 42% more. The analysis showed a significant association between the number of drugs, and both under-use and over-use. The multinomial logistic regression analysis showed a high probability of non-accordance with prescriptions among the patients who reported poor health. No association was found with well-being or sense of coherence, and drug use. Conclusions: This study indicates that the number of drugs prescribed and consumed is the main factor for concordance between the self-reported drug use and the medical record. The results showed that the probability of being adherent with prescriptions of drugs increases with good self-reported health. Aims: People with advanced lung cancer often experience multiple debilitating symptoms, many of which emerge between clinic appointments. Unrecognized symptoms can result in reduced health-related quality of life (HRQL), decreased treatment efficacy, ER visits and hospitalizations. Barriers to adequate symptom management on patient, physician, and system levels make routine symptom assessment challenging. To address some of these barriers, we piloted the Advanced Lung Cancer Management Program (ALCaMP) over 2 years to assess the feasibility and acceptability of a computerized telephone assessment program to monitor symptoms in advanced lung cancer patients. Methods: Patients telephoned the system weekly for 12 weeks to complete a 14-item symptom measure. A nurse monitored weekly symptom scores and contacted patients and physicians when scores reached a pre-defined threshold warranting clinical attention (''alert''). A graph of patients' cumulative symptom experience was generated for discussion during clinic visits. Results: All patients (n=90) identified at least one alert over 12 weeks (range 1-64, M=5.5); 78% of patients reported an alert in the first week. Patient compliance with weekly assessments was 92%. Fatigue, HRQL, and appetite generated the largest number of alerts. Interventions in response to alerts were categorized as clinical care (20%), education (36%), support (39%), and coordination of care (4%). Acuity of the interventions was rated as critical (3%), important (16%), standard of care (59%), nonessential (15%) or system-related (6%). Patients found the system easy to use, helped focus the discussion with their physician, and they wanted to continue to receive the reports. Physicians reported improved communication and greater awareness of patients' symptom experience. Conclusions: A technology-based system for targeted monitoring of lung cancer symptoms is feasible, well accepted, and promotes symptom-focused patientphysician communication. The system's role in reducing symptom burden remains to be tested in a future randomized trial. Given the broadness of the search terms, abstracts were reviewed to screen for inclusion (e.g., pain as term used in MedLine included acute pain which was excluded). Full-text was reviewed for screening when abstracts were not available, the focus on adults with chronic pain was not clear, or method of evaluating QOL was not described. Finally, articles were reviewed. Results: Reaching a conclusion about the nature of the relationship between QOL and chronic pain was not straightforward. The focus of articles ranges from theoretical to evaluation of chronic pain program that offer diverse interventions (e.g., medication, relaxation, cognitive-behavioural strategies). Interventions reduced pain which was associated with improved QOL. Measurement of QOL ranged from symptoms and their management to health-related QOL or a more global assessment of over overall satisfaction with life. The association between pain and QOL was clearer when the focus was on symptoms or health-related QOL. When other aspects of QOL or a broader perspective was used this relationship was not as clear. was used in a randomized trial to examine the effectiveness of a multidimensional, interdisciplinary low back pain education program for improving patients' health-related quality of life. This is a brief outpatient regimen that utilizes a health education approach to empower participants through a process of assessment, education and skill building. Quality of life was assessed in two points time: at baseline and at 3 months followup. Patients (n=102) were randomly allocated into two groups, matched in terms of age, weight, education, socioeconomic level, occupation and some aspects of risk behavior. Group 1 (n=50) but not group 2 (n=52) received the back school program. Results: The back school program was effective in improving patients' quality of life; significant differences were found on all eight subscales of the SF-36 for group 1. In the clinic group (group 2), improvement was observed on three scales (bodily pain, vitality and mental health) but these improvements were less than in group 1. In addition, the mean improvement over all eight subscales of the SF-36 was significantly better for the back school program group. Conclusions:

The back school program is an effective intervention and might improve the quality of life over a period of 3 months in patients who experience chronic low back pain.

Aims: It is widely believed that pain has a profound effect on health related quality of life. This study was conducted to assess quality of life in patients suffering from different intensity of chronic low back pain. Methods: This was a cross sectional study of 102 patients with chronic low back pain attending to a rheumatology center in Tehran between July and September 2003. All patients were female, married, aged 18 years and over and underwent rheumatologic clinical examination. Data were collected by face-to-face interviews using the Iranian version of the Short Form Health Survey (SF-36). Relative to mean score of bodily pain, patients were divided into two groups: mild pain group (group 1) and severe pain group (group 2). Then, quality of life scores was compared between these two groups. Results: There were significant differences between quality of life scores among people with different intensity of low back pain in all dimensions but the role emotional and social functioning scales.

The results are shown in the Aims: Described in the 90s' in American and later in french guidelines of good clinical practice, the concept of optimal pre end-stage renal disease (ESRD) care involves interventions (medication or not) to retard the progression of renal failure, to prevent or attenuate comorbidies, to adequately prepare ESRD therapy, and to timely initiate dialysis. The aim was to determine the impact of pre-ESRD (i.e. during the year before dialysis initiation) therapeutic care on health-related quality of life (HRQoL) at dialysis start. Methods: We designed a prospective study in the Lorraine region, in which all consecutive incident ESRD patients are included in a registry. Pre-ESRD treatments (erythropoietin, angiotensin converting enzyme inhibitors, phosphate binders, calcium, iron, vitamin D supplementation) and pre-ESRD biological data are collected from the patients' medical records. HRQoL is assessed at dialysis initiation using a specific instrument, the kidney disease quality of life questionnaire (KDQoL-36), from which two generic scores (physical and mental summary scores) and three specific scores (symptoms, burden of kidney disease, effect of kidney disease) can be calculated. Results: Inclusion began at January 1, 2005. About 20 patients are included monthly. Characteristics and HRQoL scores of patients will be described. Planned analyses consist for each HRQoL dimension, on a comparison of scores between subjects with and subjects without pre-ESRD therapeutic measures, performed with analysis of variance models, adjusted for potential confounding factors. The awaited major result is the demonstration of a positive impact of a pre-ESRD therapeutic care on HRQoL. Conclusions: This study is designed in an attempt to improve the preventive measures care delivered by the nephrologists to ESRD patients. Aims: Generalized outcome expectancies and coping strategies are considered to be prognostic of the quality of life (QoL) of patients undergoing treatment. We investigated the mediating function of the individual's coping style (active/problem-oriented, passive/depressive) and the moderating function of disease activity in patients with inflammatory bowel disease (IBD).

Methods: A total of 129 out-patients with IBD at age 18 to 75 participated in the study. At time 1, they completed established questionnaires to assess ways of coping, optimism, and expected self-efficacy. At time 2, they completed a generic questionnaire of life satisfaction, a disease specific QoL questionnaire and a visual analogue scale of QoL in intraindividual temporal comparison (ACSA). Results:

The assumption that an active, problem-oriented coping modus was able to predict QoL was not supported. However, a passive coping modus had adverse effects. Optimism and self-efficacy did not prove to be significant predictors of QoL. Mediating effects can only be shown regarding passive coping. Conclusions: Overall, the results suggest the necessity to consider the processing character of coping with regard to QoL in further studies. Aims: Studies on deep-vein thrombosis (DVT) mainly focus on mortality and recurrent disease. These outcomes are limited in their ability to quantify the impact of the disease on patients' lives. We examined the effects of socio-demographic, diseaserelated and treatment-related variables on the quality of life (QL) of patients with DVT. Methods: The study sample comprised 400 patients participating in a multi-center clinical trial. QL was assessed at study entry, after 2 weeks, 3 and 6 months with the MOS-SF-20 and a revised version of the RSCL. Effects of socio-demographic, disease-related and treatment-related variables on patients' overall QL, physical functioning, pain and thrombosis complaints were investigated through mixed modeling. Results: Overall, patients' QL improved over time.

QL was most impaired in women, older patients, and patients with extensive DVT (p<0.05 for each QL scale). Co-morbidity and the quality of treatment of patients treated for 6 months with oral anticoagulants affected overall QL, physical functioning and health perception adversely (p<0.05), but were not associated with pain and overall QL. Symptom duration, type of initial treatment and duration of treatment with oral anticoagulants were not associated with QL. Conclusions: The relationship between extensive DVT and impaired QL might indicate that it takes more time to recover from proximal DVT than from DVT in the poplitea only. 5) . It was significantly related to gender with a score of 7.6 (SD=6.2) for men and 9 (SD=6.8) for women (p=0.0333). It was also significantly related to age, the younger the individuals, the more quality of life was affected: 9.4 (SD=6.6) for the group <39 years old and under>, 9.2 (SD=6.7) for the group < 40-54 years >, 7.8 (SD=6.6) for the group < 55-64 years > and 6.7 (SD=5.6) for the group < 65 years old and above > (p<0.0001). Quality of life was also linked to the localization of the psoriasis. Conclusions: Quality of life assessed through the DLQI shows an important alteration in the study population with a mean score of 8.3. In comparison with other studies using the DLQI ; a score of 8.9 was found for severe psoriasis, 12.5 for atopic dermatitis and 4.3 for acne. Scores of 13 or more is considered to be a significant case of postnatal depression, while scores of 9-12 represents borderline and 0 to 8 normal. Results: The mean age of the study samples were 24.5 (SD=2.9) and 25.0 (SD=3.9) and most had secondary educational level (60 and 56%). Severe postnatal depression in both groups was reduced from time1 to time 2. Overall 19% of women at time 1 and 16% at time 2 scored 13 or higher on the EPDS. However at both assessment points a higher proportion of the women who had caesarean section had a postnatal depression score of 13 or above as compared to the normal delivery group (at time 1: 22% vs. 16% and at time 2: 18% vs. 14%). Conclusions: In general the findings suggest that normal delivery might lead to a better postnatal psychological status. Indeed the results not only do not support the believe that caesarean section protect women from postnatal depression but also indicate that it is associated with an increased risk of severe postnatal depression. Aims: Patients with aspirin-sensitive rhinosinusitis (ASRS) frequently present with symptoms of nasal congestion, sinusassociated headache, fatigue, daytime somnolence and associated with decreased cognitive performance and impaired quality of life. Methods: Twenty nine patients with ASRS (mean age 45.9±7.5 yrs), of about 7 years duration have been included in study. All patients were receiving nasal decongestant. Clinical outcome measurements (nasal inspiratory peak flow -NIPF, questionnaires) before and after chronic rhinosinusitis treatment were analyzed for the treatment group (19 patients received topical corticosteroids (TCS) 4 weeks 400 mkg daily) and the control group (10 patients received only decongestant therapy). The patients were daily recording 24-hour severity evaluation scores of rhinosinusitis, needs in drops of decongestant. Results: Before treatment the patients in both groups had very high score of nasal congestion -3.5 (a scale from 0 (no) to 4 (max)). NIPF in these patients has been reduced -up 45 to 130 (mean -82) l/min. The impairment of QOL in them was associated with decreased cognitive performance, restriction of contacts to colleagues, the friends, poor concentration at work. The patients had daytime somnolence and more half of patients often had headache. Daily use of decongestant was significantly lower during TCS treatment compared with DT, 1.6 vs. 4.5 times a day, (p<0.05). There was significant reduction nasal congestion (p<0.05) and increase NIPF to 120 (in mean) in TCS group. Quality of life statistically significant increased in this group. It has been connected with decrease the needs in decongestant at night and improvement sleep quality, social and work related activities, reduction of sinus -associated headache, nighttime awakenings and daytime sleepiness.

Conclusions: Nasal topical corticosteroids significant relief of nasal congestion and have considerably improved quality of life in patients with ASRS. Means of the different subscales were determined, and the impact of the demographic variables on the dimensions was explored using student's t-test or one-way ANOVA. Results: Demographic profile of the respondents shows, males (52%), singles (45%), married (44%). A majority were civil servants (32%), had attained post-secondary educational level (56%), and age group of 20-39 years (72%). Cronbach alpha coefficient of reliability ranged from 0.60 to 0.91. The quality of life of the subscales measured on a scale that ranged from 0 to 100 were: Overall function (61) (3) the purpose of the review is to identify all measurements assessing (an aspect of) health status and to report on the clinimetric properties of these measurements. Two independent reviewers selected the articles. A standardised data-extraction form was used. Results: Based on a first search in PubMed, at least 75 systematic reviews were included. We will present an overview of how many systematic reviews used explicit criteria for assessing measurement properties or for defining good measurement properties. Next, we will give an overview of which criteria were used and if evidence for this criteria was presented. Conclusions:

We expect a lot of variability in definitions of measurement properties and in criteria used for evaluating these properties, with little evidence. Without consensus about explicit criteria for good measurement properties, instrument selection is strongly hampered. This review can serve as basis for reaching consensus on explicit criteria for good measurement properties. (1994), was first examined to determine whether a possible methods effect associated with the reverse-scored items significantly contributed to better model fit. The preferred measurement model was then examined for measurement invariance using multigroup confirmatory factor analysis (MG-CFA) to compare adolescents with poor perceived health to adolescents with good perceived health. Perceived health was determined based on adolescents' aggregated responses to items measuring both physical and mental health. Results: The findings provided support for a methods effect to account for covariation among the 10 reverse-scored items. Subsequent MG-CFA for measurement invariance provided preliminary support for configural, metric, and scalar invariance. However, the findings for invariance of item reliability estimates were inconclusive. Conclusions: The measurement invariance analyses provide tentative support for the use of the MSLSS in the context of health when the purpose is to compare mean life satisfaction scores across differences in perceived health status. The analyses further point toward the importance of accounting for the possibility of a methods effect associated with the reverse-scored items. Ongoing analysis using item response theory is recommended to further examine differential item responses associated with differences in various aspects of perceived and observed health. . We compared the QOL, the physical and psychological status in EP with YP. The physical status was assessed by the exercise tolerance and serum lipid profiles. The psychological status was assessed by the Spielberger state-trait anxiety inventory questionnaire (STAI) and self-rating questionnaire for depression (SRQ-D). QOL was assessed using the QOL score of the Ministry of Health and Welfare in Japan, which established and validated QOL scale for the Japanese Patients with Cardiovascular Disease. These were evaluated before participation in the program, and 1, 6 and 12 months after. Results: In the both groups, serum lipid profiles and STAI anxiety score were improved significantly, at the 1, 6 and 12 months follow-up. In YP, the QOL total score and the social attitude and subjective well-being score were improved significantly at the 6 months follow-up. In EP, the QOL scores tended to improve and were maintained for the 12 months follow-up. For several years BlueCross BlueShield of Tennessee has been collaborating with their healthcare providers to develop both medical policies and utilization guidelines. This collaboration allows BlueCross BlueShield of Tennessee to fill the gaps in medical evidence with provider experience, thus supporting the concept of Evidence-Based Medicine. This experience led to the next logical step, which was to expand the collaborative process to include our research and academic community. Methods: This study is intended to determine whether a cognitive-behavioral prevention program has a significant advantage over usual care in reducing the prospective incidence of mood disorders and depressive symptoms. Results: (1) The primary hypothesis is that participants in the cognitive-behavioral preventive intervention group will have a significantly lower incidence of first and repeated episodes of depressive disorders and symptoms compared to adolescents in the usual care group. In addition, it is anticipated there will be improved global functioning in the treatment group.

(2) The cognitive-behavioral preventive intervention will be effective across geographic locations, managed care settings, and adolescent characteristics (e.g., gender, socioeconomic status).

(3) The cost per outcome (level of functional improvement) achieved will be lower for the cognitive-behavioral preventive intervention group. Conclusions: The ultimate goal is to determine if cognitive-behavioral preventive intervention is an approach that should be adopted as part of routine care for offspring of depressed parents. Aims: Lifestyle is associated with health-related quality of life (HRQOL) in adults. However, there is little information about relation between lifestyle and HRQOL in children. This study aims to examine the correlation between lifestyle and HRQOL in Japanese elementary school children. Methods: A total of 4000 children (7-12 years old) from the 10 elementary schools, in Ohsaki City in Japan participated in the study. The questionnaire was designed to measure lifestyles including questions on eating, physical activity, sleep, and other factors. For example, the frequency of breakfast, it was rated on a fourpoint scale (1=everyday eating, 2=skip 2-3 time/week, 3=skip 3-4 time/day, and 4=skip 5-7 time/day). Sleeping hours was categorized into the following: <6, 6 to <8, 8 to <10, and 10> hours/day. The QOL was measured a selfadministered questionnaire, and health conditions including the frequency of headache and abdominal pain during the past 6 months. In all, 3892 children answered their questionnaires, with a response rate 97.3%. The data from 3852 children (boys 49.9%, girls 50.1%) were available and logistic regression analyses were used. Results: Twenty percent of them had poor QOL. Nine percent of them answered that they do not eat breakfast everyday and five percent of them answered that sleeping time is less than 6 hours. Children with undesirable lifestyles, such as skipping breakfast, less participation in physical activity, and later bedtime, were more likely to have poor QOL. Conclusions: Undesirable lifestyle is positively associated with impaired QOL among Japanese elementary school children. Further understanding of these relationships may facilitate the development of interventions to help Japanese elementary school children from poor QOL. Aims: To evidence gender differences in perceived health during childhood and adolescence and discuss theoretical implications. Methods: Several samples and generic perceived health questionnaires were used: (a) the Spanish versions of the CHIP, administered in representative school-based samples in Catalonia (Spain) to children aged 6-11 (n=971) and adolescents aged 12-19 (n=1.453); (b) the Spanish versions of the KINDL-R and VSP-A, administered in a representative non-institutionalized sample of Spanish children and adolescents aged 8-18 (n=882). Analysis of the variance were carried out to assess gender differences in perceived health in different age groups. The percent distribution of girls and boys with high and low scores in the Spanish CHIP was described. A literature review complemented information of studies that have evidenced gender differences in health during childhood and adolescence. Results: A total of 3.306 Spanish children and adolescents were included in this study, collected between 1999 and 2003. Girls presented lower scores (worst) in Satisfaction, Comfort and Resilience in the Spanish CHIP in all age groups (6-11, 12-15 and 16-19) . Boys presented lower scores (worst) in Risks domain and lower scores (worse) in Achievement compared to girls (p<0.05). Girls presented lower scores (worse) in comparable domains of the Spanish KINDL-R and VSP-A such as Physical and Psychological well being, Self-esteem and Global HRQL at ages 12-18 (p<0.05). No statistical differences between girls and boys were found in age groups 8-11 in the Spanish KINDL-R. Conclusions: Gender differences in perceived health in children and adolescents are consistently evidenced in Spain and international studies. From available data, differences can be detected from 6 onwards. Gender differences in perceived health exist partly because girls and boys are educated differently, and society emphases these differences considering values or behaviors as genderly adequate. P-149/1610/LAUGHING ALL THE WAY TO THE ITEM BANK: HOW TO IMPROVE YOUR TREATMENT EF-FECT BY IMPROVING YOUR OUTCOME MEASURE-MENT Rachel Baron, Jeremy Hobart, Clinical Neuroscience, Peninsula Medical School, Plymouth, Devon, UK Aims: Patient completed rating scales are increasingly used as outcome measures in clinical trials. As such, there should be no compromise on their scientific quality. Scales typically suffer from one of two problems. Some measure a wide range of functioning, but with poor precision and poor responsiveness. Others have superior precision and responsiveness, but over a limited range with inevitable floor and ceiling effects. Both cases risk treatment effects being underestimated. The goal of scale measurement is good precision and responsiveness across a wide range. This requires large numbers of items and might appear impractical for use. However, new psychometric methods allow the development of item banks that achieve that goal.

Moreover, patients need only answer a few items. We are developing an item bank to measure physical functioning in Multiple Sclerosis (MS). Methods: Items were generated from a comprehensive literature review, examination of existing scales, and semi-structured interviews with 20 MS patients. Rasch analysis of data from multiple surveys will guide item selection, item modification, scale development and validation. Results: From the qualitative data and literature review, we identified three related, but clinically distinct and clinically meaningful, dimensions of physical functioning. We then selected items that attempted to operationalise these variables as rating scales for: mobility (51 items); upper limb functioning (57 items); tasks involving upper and lower limb function (41 items). The first survey (n=100) is underway. We will report the results of the initial and validating (n=300) surveys. Conclusions: Item banks enable clinicians to solve the main problems of existing scales without increasing patient burden, and will be invaluable tools for measuring patient outcomes in future clinical trials. However, the variables they aim to operationalise need to be clearly defined in advance Conclusions: Item banks enable clinicians to solve the main problems of existing scales without increasing patient burden, and will be invaluable tools for measuring patient outcomes in clinical trials. However, the variables they aim to measure need to be clearly defined in advance. Aims: The aim of this study was to develop and validate a brief self-report tool to assess pain severity and pain impact on individual's health-related quality of life (HRQOL). Conventional pain questionnaires are developed on the basis of the classical test theory (CTT) and often too long to be feasible in everyday clinical practice. The Pain Impact Survey was constructed using CTT and item response theory (IRT) methods. The new tool aims at meeting clinicians needs for a short, precise and practical tool. Methods: The Pain Impact Survey was developed using an initial item pool of 65 items from 10 of the most commonly used pain tools. Those items were webbased administered to a sample of 782 people. On the basis of statistical analyses using CTT-and IRT-methods, 8 items were selected, because they had the best item characteristics and spanned the desired content. In a following validation study the pilot survey, two pain severity rating scale items and the SF-8 were web-based administered to a representative sample of 830 people and 350 chronic pain patients. Results: The data analyses revealed the Pain Impact Survey to be unidimensional (CFI=0.96, TLI=0.98) with items showing no differential item functioning for age, gender, ethnicity and level of education. The final survey consists of 6 items showing good item characteristics, i.e. a wide measurement coverage, favorable item information functions and discriminative item response curves. Construct validity was supported by strong correlations with pain severity rating scale items (r=0.81-0.84, p<0.001). Correlations with the SF-8 physical and mental component scores were r=0.74 (PCS) and r=0.45 (MCS), respectively (p<0.001). Discriminant validity was suggested by significant mean score differences (p<0.001) between groups of patients differing in pain severity and medical condition. Conclusions: We conclude that the Pain Impact Survey being available as paper-and-pencil and computerized version is an efficient and precise tool. We develop item oriented fit tests for polytomous items by generalizing the S-X2 and the S-G2 tests proposed by Orlando and Thissen (2000) . The tests compare the observed and expected distribution of item responses conditioned on the simple sum of all other items in the scale. The fit statistics can be evaluated using parametric bootstrap or through chi-square approximations. Plots of observed and expected values can be used to visualize misfit. For these and other polytomous item IRT fit tests the observed/expected tables are sparse with many cells having low expected scores. We suggest a strategy for collapsing cells to deal with this problem. Results: In simulation studies, the new statistics showed satisfactory type I error rates. At sample sizes of 1000+ respondents, the test statistics had sufficient power to identity misfit of a type and magnitude we have encountered in empirical studies of mental health. The bootstrap approach provided more power to identify misfit, in particular for long scales, but the chi-square approximations seemed adequate for initial item evaluations. Conclusions: With research initiatives such as the Patient-Reported Outcomes Measurement Information System (PROMIS) project, IRT methodology is increasingly applied to patient-reported outcomes data. These new statistics are promising and crucial ingredients in the statistical evaluation of this approach. Aims: We examined levels and predictors of psychological distress in wives of men treated for early-stage prostate cancer (PCa). Methods: PCa patients (n=165) and spouses were interviewed to assess general and cancer-specific distress. Social and intra-personal factors of spouses as well as clinical characteristics and quality of life of patients were assessed as potential predictors of spouses' distress. Results: Spouses reported more cancer-specific distress than did patients (p<0.001), but did not differ from patients in general distress. Several spousereported factors predicted higher spouses' distress including: less education (p<0.005), worse marriage quality and less social support (ps<0.005), more negative social interaction with the patient (ps<0.001), lower self-esteem (ps<0.001), less positive coping (ps<0.005), searching for meaning (p<0.001), not finding meaning (p<0.005) and greater illness uncertainty (ps<0.001). Patients' bowel function and mental health also predicted greater spouses' distress (ps<0.005). Conclusions: Our findings indicate that overall distress in spouses of early-stage PCa patients is modest, and it is more likely to be predicted by psychosocial than medical factors. These factors may help to identify spouses at risk for distress. Aims: Delirium is common in palliative care, but it is frequently undetected and therefore under-treated. A short and simple questionnaire for screening cognitive impairment in these patients is required, to identify those who deserve more comprehensive testing. Methods: Two samples of patients completed the Mini-Mental State Examination (MMSE). Sample A included 290 general medical patients with mean age 80, of whom 142 had delirium (MMSE score <24); sample B was 217 palliative patients with mean age 62. Stepwise logistic regression was used to identify the items that best discriminated for cognitive impairment in the sample A patients. IRT was used to determine a brief but sufficient subset of screening items, and these were validated in Sample B. Results: Most MMSE items were too easy for screening purposes. In sample A, four items sufficed to predict delirium as defined by ICD-10 research criteria, with 92% sensitivity: ''current year'', ''date'', ''backward spelling'', ''copy a design''. These items correctly identified 99% MMSE-cases (score <24) in sample A and 96% in the validation sample B. Adding the two items ''recall three objects'' and ''repetition of a proverb'' further reduced the standard errors of the estimated MMSE scores. Conclusions: Many of the MMSE items performed poorly when used as a screening tool. IRT methods enabled us to identify a subset of four items that provided fairly accurate determination of cases, while using a total of six items resulted in greatly improved discrimination. Aims: The overall aim is to describe Health-related Quality of Life (HRQL) reported by adult survivors of autologous and allogeneic HCT and risk factors associated with poor HRQL. Methods: Eligible subjects had an HCT at City of Hope Cancer Center or University of Minnesota between 1974 and 1998, and survived two or more years after HCT. We analyzed data from 1003 adult HCT survivors (older than 21 years) enrolled in a retrospective cohort study who had completed a validated City of Hope HCT-QOL tool. Descriptive statistics were used to describe the population, and ANOVA was used to compare QOL scores with tool norms for physical, psychological, social, spiritual domains and overall QOL. Multivariate analysis was used to identify risk factors for those reporting poor QOL. Results: Results revealed 54.5% males, 80.2% Caucasians, and 55% allogeneic HCT. The primary diagnoses included CML (n=233), AML (n=241), NHL (n=200), ALL (n=99), HD (n=88), and other (n=142). The median age at HCT was 35.4, the median age at study participation was 43.4 years, and the median length of follow-up was 9.4 (2-28 years). Poor QOL (defined as QOL scores less than two standard deviations below that reported for the tool norms) was found in 9.2% of this cohort for overall QOL; 13.7% for poor Physical; 10% for poor Psychological; 6.7% for poor Social; and 4.3% for poor Spiritual well-being. Significant risk factors identified for poor QOL by multivariate analysis included being Non-white, a low annual income (<$20,000), unable to work, experiencing pain, or anxiety. Age, sex, type of HCT, CGVHD and time since HCT were not associated with poor QOL. Conclusions: Certain vulnerable sub-populations exist, specifically, low household income, difficulty in obtaining health insurance, inability to return to work, and the presence of pain and anxiety that impact upon HRQL, and require additional attention and possible tested research intervention. Aims: We developed an adaptive, Item Response Theory (IRT) based questionnaire to measure health-related quality of life (HRQL) in arthritis (ISOQOL 2003 (ISOQOL , 2004 . The instrument produces IRT-based scores for five domains of HRQL plus an overall utility score. The aim of this study was to validate this measure in patients with rheumatoid arthritis (RA). Methods: Data were obtained by a self-administered mailed questionnaire at baseline and 2 years later from patients with RA treated at a rheumatology clinic. Outcome measures at Time 2 included the adaptive questionnaire, HAQ, SF-8, EQ-5D, pain, self-rated disease severity and control. We evaluated a priori hypotheses about correlations between HRQL measures, group differences, effect sizes (ES) for patients who were expected to improve, and minimal clinically important difference (MCID). Results: There were 137 patients who responded at Time 2. Overall utility scores on the adaptive questionnaire correlated with HAQ (r=)0.72), pain (r=)0.57), SF-8 Physical (r=0.70), and Mental Component (r=0.39), and EQ-5D (r=0.70). However, they were higher than EQ-5D utilities. Sores varied as expected according to self-reported severity and control. In longitudinal data, scores in patients receiving etanercept or infliximab (n=39) improved significantly over time (ES=0.23, p=0.03), while other patients did not change (ES=0.11, p=0.64). Changes in utility scores correlated with changes on a transition question (p<0.0001). The MCID was 0.02 using an anchorbased approach and 0.04 using a distributional approach. Conclusions: The adaptive questionnaire is a valid and responsive measure of HRQL in patients with RA. Aims: The asthma quality of life questionnaire (AQLQ) is a 32item instrument that was designed to assess quality of life in patients with asthma. However, it cannot be used in economic evaluations to measure health states, as in its current form, it is not preference based. Therefore, this study aims to use Rasch analysis to develop from the AQLQ a health state classification that is amenable to valuation studies where selected health states can be valued in time-trade off or standard gamble exercises, using a representative sample of the general population. Methods: A Rasch model was fitted to a dataset consisting of 482 patients who have answered the AQLQ with the aim of (i) reducing the number of levels for each item from the seven presented in the original form to a more manageable number, (ii) reducing the number of items to a manageable number in such a way as to ensure that as much of the severity range of the latent variable as possible is covered for each of the four AQLQ domains (symptoms, activities, environmental stimuli and emotion). Conventional psychometric criteria for item selection (distribution of responses across item levels, missing data, correlation, discriminant validity, responsiveness between two time points and regression against overall health) were also explored. Results: Rasch analysis and psychometric methods were used to reduce the number of levels per item from seven to five and they identified a total of five items from the AQLQ for the reduced health state classification: one from each of the four AQLQ domains and a further question about problems with sleep due to asthma symptoms. The five items all did well under the conventional psychometric criteria. Conclusions: The study showed Rasch analysis to be a useful tool in the development of a health state classification system amenable to valuation exercises for deriving preference weights for different asthma states identified by AQLQ. Aims: CDH is a congenital malformation requiring surgical repair as a neonate. CDH in-hospital survival has increased between 1981 and 2000 from 53% to >90% at Children's Hospital Boston. We examined the impact of CDH on families and hypothesized that families who have children with ongoing medical morbidity will experience greater financial, social, and personal strain compared to CDH families without medical morbidity. Methods: A cross-sectional cohort of 53 (65% of eligible) CDH families participated in this study. The following measures were completed by self-administration: Impact on Family Scale (IFS) [lower scores connote less impact]; Child Health Related Inventories Parent Report (CHRIs-PR) [lower scores connote worse functioning]; items from the National Survey on Children with Special Health Care Needs, including a severity rating of the child's condition, and a parent medical history form. Spearman's correlation coefficients and Wilcoxon's rank sum statistics were performed. Results: CDH survivors had a mean age of 8; 57% were male. Study respondents were predominantly female (83%), well-educated (81% with > high school), and employed (40% full-time, 36% part-time). Clinical severity was highly correlated with IFS total score (r=0.71; p<0.001) and moderately correlated with CHRIs-PR physical (r=)0.43; p=0.01) and role (r=)0.33; p=0.05) scores. Family impact, as measured by the IFS, and parental emotional functioning (EF) were worse among the 48% of families with children with ongoing medical morbidity (IFS, p=0.001; EF, p=0.01) and among the 22% with public insurance (IFS, p=0.01; EF, p=0.003). Further, family impact was worse for those with a higher level of out of pocket expenses (>$500 vs. <$500, p=0.002). Conclusions: There is significant functional and financial impact on families of CDH survivors. The child's current medical morbidity influenced the differences found in these families. Long-term follow-up of the family as well as the child is essential to understand the quality of life of the CDH survivor.

Martha A. Grootenhuis, Heleen Stam, Pediatric Psychology; Jaap W. Groothoff, Pediatric Nephrology, Emma Children's Hospital, Amsterdam, Holland Aims: As renal replacement therapy in children with end-stage renal disease has become routine treatment, concern has arisen about its implications on adult life. The objectives of the study are to assess the course of life of End Stage Renal Disease (ESRD) patients, and to compare their course of life with the general population. Furthermore, we explored how course of life is associated with QoL in young adulthood, while controlling for age, gender, and medical determinants. Methods: A total of 76 adults with ESRD since childhood, aged between 20 and 30 years, completed the RAND-36 Health Survey and the Course of life questionnaire, which retrospectively assesses the achievement of developmental milestones and examines socio-demographic outcomes in young adulthood. Medical data were obtained from the long-term follow-up. Multiple linear and regression analyses were performed to investigate differences for course of life and prediction of QoL. Results: Adults with ESRD since childhood had a delayed course of life, they achieved less mile-stones than their peers with respect to Autonomy, Social, and Psycho-sexual development. Furthermore, it is shown that fulfilling developmental tasks while growing up has a predictive value for QoL later in life. Especially a relation between social development and QoL was shown. Those patients who were not able to continue peer related activities during childhood reported a lower mental QoL later in life. Conclusions: As more and more pediatric patients with ESRD will reach adulthood, more and more physicians will be confronted with these young adult patients. Optimal transition from pediatric to adult health care requires physicians with knowledge of the psychosocial history of growing up with ESRD. Aims: Generic QOL instruments have not been used widely in studies of people with increased night time voiding frequency, even though the symptom affects wellbeing and therefore often is a key driver for seeking health care. The aim of the study was to investigate whether the short version (WHOQOL-bref) differentiates between of men with/without increased night time voiding frequency. Methods: The sample consisted of men waiting for urologic evaluation. Based on the general practitioner information in the referral letter to an urologist, BPH was tentatively set for 612 patients who entered a waiting list for urologic evaluation. Fourhundred and eighty patients gave their consent to participate in the study and filled in a questionnaire containing the item on night time voiding in the International Prostate Symptom Score(I-PSS) (range 0-6) and WHOQOL-bref. Patients reporting 2 or more voiding episodes per night were categorized as having increased frequency. The WHOQOL-bref results in a profile of 4 domains (range 4-20), and two single items (range 1-5). Results: Of the 480 men, 133 were defined as having no increased frequency of night time voiding (Group I), while increased frequency was defined in 333 (Group II). Fourteen did not respond to this item. The two groups had no significant mean differences in the domain Social Relationships, and in the single item Overall health perception. For the single item Overall QOL, Group I/Group II had means 3.8/3.6 with Confidence Intervals (CI) for the means' difference at 0.0-0.4. In the remaining domains: Physical Health 15.2/14.2 (CI 0.3-1.6), Psychological 15.5/14.9 (CI 0.1-1.1), and Environment 15.6/15.1 (CI 0.1-1.0). Conclusions: The WHOQOL differentiates between the responders with/without increased night time voiding frequency. Why there were no differences in General Health Perception and Social Relationships domain need further discussion. Aims: Although recommended in several countries, only a minority of patients with an acute myocardial infarction (AMI) are offered systematic psychosocial care after discharge from hospitals. The purpose of the present study was to evaluate the effect of a post-discharge nurse-initiated telephone follow-up intervention on anxiety, depression and health related quality of life (HRQOL), in a context where existing follow-up services are poorly developed. Methods: All patients with a diagnosis of AMI admitted to the Department of Heart Disease at Haukeland University Hospital from September 2001 until December 2003 were potentially eligible and invited to participate in the study if they fulfilled the inclusion criteria. Two-hundred-andeighty-eight patients consented to participate and were randomised to an intervention or a control group. The intervention group received nurse initiated telephone follow-up in addition to the standard post-discharge follow-up which the control group received. Data are collected at baseline, and three and six months after discharge. The Hospital Anxiety and Depression Scale was used to measure emotional distress, and HRQOL was measured using the Mental and Physical Component Summary Scales in the SF-36 ÒHealth Survey Results: Although HRQOL improved significantly over time in both groups, a statistical significant difference, in favour of the intervention group, was found on the Physical Health Component Summary Scale (PCS) (p=0.034), but not on the Mental Health Component Summary Scale. No statistical significant differences were found between the control and the intervention groups for levels of depression and anxiety three and six months after discharge.

Conclusions: An effect of the intervention was found in the PCS reflecting improvement in the physical dimension of HRQOL. The intervention had no effect on anxiety and depression; however the potential for improvement was smaller than previously indicated in observational studies. Aims: Disease-free period and improvement of patient's quality of life (QoL) are the major treatment outcomes for MS patients. According to our QoL treatment response concept the following grades of QoL response exist: minimal (less than 25% improvement compared to the baseline), moderate (25-50% improvement), good (50-75% improvement), and excellent (more than 75% improvement). The aim of this research was to study clinical and QoL response in MS patients after HDCT+ASCT. Methods: Thirteen patients with MS were included in this study. The following time-points were used for clinical examination and QoL assessment: at baseline, at discharge, at 3, 6, 9, 12 months, and then every 6 months. FACT-BMT and FAMS were used for QoL evaluation. Integral QoL index was assessed by the method of integral profiles. The median follow-up duration was 36.3 months (range 8-66 months). Results: According to the data of clinical examination and MRI eleven patients had disease-free period during the time of follow-up. In two patients disease progression was observed. All patients with clinical response exhibited improved QoL at the end of follow-up. At one year after HDCT+ASCT all patients exhibited either good or excellent QoL response. These levels of QoL response were preserved throughout the follow-up period. Conclusions: HDCT+ASCT appears to be an effective treatment for MS patients both in terms of clinical and QoL response. The concept of QoL treatment response allows us to gain valuable information about MS patients after HDCT+ASCT. Aims: Little is known about the quality of life (QOL) of colostomates with stoma performed large intestine changes after surgery. We evaluated the generic and disease-specific QOL in Japanese colostomates in 2001. Moreover, we evaluated changes of those QOL in Japanese colostomates in 2004. Methods: A prospective evaluation of 120 colostomates, who had answered the questionnaire and had been registered their names and addresses in 2001 was undertaken. The generic QOL and disease-specific QOL were evaluated by SF-36 and Olbrisch ostomy adjustment scale (OAS), respectively in 2001. Those questionnaires were sent by mail. We also evaluate changes the generic and disease-specific QOL of Japanese colostomates in 3 years. Data analysis was performed by Mann-Whitney U-test. Results: One hundred and twenty colostomates (age: 67±11 years, 5.0±4.8 years after surgery) replied in 2001. Fifty colostomates out of 120 colostomates (age: 69±8 years, 7.6±4.6 yrs after surgery) replied in 2004. In 2001 data, the SF-36 scores of colostomates were below the average scores of Japanese general population except 'bodily pain', and the mean total score of OAS were lower than that reported by Olblisch (1983) , Burckhardt (1990) and Brydolf (1994) , 130.6 compared to 155.7, 168 and 178.6. In SF-36 scores and the mean total score of OAS, there were no significant differences between the data of 2001 and 2004. Conclusions: Both generic and disease-specific QOL of Japanese colostomates were lower than those of Japanese general population and Western countries, respectively. There were no changes in QOL of them during 3 years observation. The longitudinal further studies of large samples of patients may contribute to improvement of QOL in Japanese colostomates. Aims: The aim was to explore gender differences among spouses in perceived psychological well-being and general life situation, during the first year after the patients' stroke event. Methods: Eighty female and 20 male spouses of stroke patients admitted to a stroke unit in Stockholm, Sweden, during November 2000-July 2002 accepted the invitation to participate. The spouses were followed with assessments at baseline and after six and 12 months, with assessments regarding psychological wellbeing and general life situation. Data were analysed using Analyse of variance. Internal dropouts over time consisted of eight women and one man. Results: Female spouses of stroke patients have a negative impact on their psychological well-being, while male spouses have a lower occurrence of emotional contacts in their social network. Already at baseline, and consistently over time, the female spouses reported lower quality of life and well-being than the male spouses. Conclusions: Nurses must ensure that the informal caregivers are assisted in the transition of care from the structured hospital care to home settings and to provide support, so that the caregivers can adapt to their new role and be comfortable and effective as informal caregivers. It is important to take the gender differences under consideration when designing a nursing intervention, focusing on spouses of stroke patients, to meet the different needs and demands of male and female caregivers. Nursing interventions are needed to reduce these negative health effects, and to inform the caregivers on how to care for the stroke patient as well as for themselves. Further studies are needed to develop and evaluate this type of intervention. Aims: While the emphasis of epilepsy treatment is shifting towards health-related quality of life (HRQoL) outcomes, patients too need to develop specific coping skills to live with the condition due to potential adverse drug reactions and disease-related psychosocial implications. This study aims to (1) investigate overall HRQoL status in Malaysian epilepsy patients and (2) compare the types of coping strategies used by groups with different Overall Score. Methods: Upon written consents, 96 patients from nine government hospitals in Sabah completed the Malay versions of Quality of Life In Epilepsy-30 (QOLIE-30) (7 domains; 30 items) and Brief COPE instruments (14 domains; 27 items). Results: Respondents were on average 32 years old (range:18-76 years; 57 males; 54 unmarried; 44 jobless). Overall, patients generally scored highest in the Medication Effects domain (mean=68) and lowest in the Seizure Worry domain (mean=46). A total of 49 patients (Group A) scored below average for the Overall Score (mean=61) while the rest reported above average level (Group B). No significant difference was found in their demographic characters. Independent samples t-test revealed that Group A patients employed the Denial (p=0.027), Behavioural Disengagement (p<0.001) and Self Blame (p<0.001) strategies significantly more frequent than Group B respondents. Conclusions: The results generally indicate that patients were more concerned with seizure attacks than medication side effects, hence generating lower HRQoL in the former domain. It also demonstrates that the three coping styles were less effective in maintaining an above average HRQoL level in this sample. Such outcomes show that coping skills play an important role with regard to HRQoL and could guide healthcare professionals in the successful management of epilepsy control. Aims: Both full spectrum lighting and negative ion environments have independently demonstrated effectiveness in improving mood, increasing energy and reducing stress. However, the combined impact of these two modalities has not been evaluated. This presentation describes the combined effects of full spectrum light and a negative ion environment on a group of patients with mild levels of stress and mood difficulties. Methods: A total of 94 onehour visits were made to the controlled environment by two different groups: office workers under mild stress (n=73 visits) and participants in a patient education class for depressed mood (n=21 visits). Negative ions were produced naturally by falling water. Full spectrum lighting provided 30,000 lumen at the ceiling level. A pre/post-evaluation measure was developed based on qualitative identification of attributes where affects were noted (mood, stress, energy, sensation around the eyes, and overall ability to think and reason). Subjective responses were made for each attribute on a 10 cm true VAS which is scored on a 0-100 point scale. Results: All pre-post change values were positive and significant (p<0.01). Improvements in self-rated depression, degree of stress, and energy levels were greater for the depressed group (24.4, 29.5, 19.4, respectively) than for the office workers (11.4, 23.9, 15.1, respectively). Similar improvements were demonstrated in both groups for levels of anxiety, sensation around the eyes and ability to think and reason (16.0, 21.6, 13.3 and 16.5, 24.8, 17 .9 respectively). Conclusions: The combined effects of full spectrum lighting and negative ion exposure for a one-hour period contributes to significant improvements in mood, stress and energy. When combined with positive gains from the other three variables, all participants experienced a noticeable increase in general well being. level of independence (LID); social relations (SRD); environment (ED); and spirituality, religion and personal beliefs (SRPBD). Domain and facets scores range is 4-20,being 20 the best score. Data collected included social-demographic and clinical data Results: the mean age was 37.7 years (range 18-72),the majority (51.9%) were male, most (61%) were single, 55% (405) had more than 8 years of education, 28.9% (213) were unemployed, 48.8% (359) were catholics, 29.2% protestants and 59.6% (439) referred going to church. The average income was US$ 263.89 and family income US$ 133.67 per capita. About 76.1% (560) were classified as AIDS cases, the mean duration of disease was 30.59 (range 1-191) months. 82.6% (608) acquired HIV through sexual relations. The mean duration of antiretroviral therapy was 40.20 months. About 65.5% (482) classified their health as good or very good. General QOL score was 13.5 (SD=3.1). QOL had (p<0.05) better scores in the SRPBD (15.93, SD=3.1). The worst scores were for ED (12.41, SD=1.9) and the SRD (13.54, SD=2.6). AIDS patients had worse mean scores (p<0.05) in PYD and LID than HIV patients. There was correlation between: General QOL (r=0.08, p<0.05) and LID (r=0.14, p<0.01) with CD4 count; ED and income (r=0.39, p<0.01); DYP and duration of disease (r=)0.09, p<0.01) Conclusions: the SRPBD showed the best score and general QOL, SRD and LID the worst. Aims: Caesarean section deliveries might increase the incidence of surgical interventions and problems resulting from hospitalization and thus affecting quality of life in women who receive such deliveries. This study aimed to compare quality of life in women with normal and caesarean section deliveries. Methods: A sample of 50 women with normal delivery and 50 matched cases with caesarean section delivery from 5 health care centers in Esfahan, Iran were entered into the study. Quality of life was measured using the SF-36 at two points in time (Time 1: 6-8 weeks after delivery; Time 2: 12-14 weeks after delivery). Results: The mean age of the study samples were 24.5 (SD=2.9) and 25.0 (SD=3.9) and most had secondary educational level (60 and 56%). Postnatal quality of life in both groups was improved from time1 to time 2. However, comparing the mean scores between the normal delivery and caesarean section delivery groups the results showed that in general the normal group had a better quality of life for almost all subscales in both assessment times. The difference was significant for vitality (mean score 62.9 vs. 54.4) and mental health subscales (mean score 75.1 vs. 66.7) in the first assessment (p=0.03, and p=0.03, respectively) and for physical functioning (mean score 88.4 vs. 81.5) in the second assessment (p=0.03). Conclusions: The findings suggest that normal delivery might lead to a better quality of life especially resulting in a superior physical and mental health in women and therefore it is better to be considered as the first priority in elective termination of pregnancy. Aims: Relatively little is known about the interrelations between different patient-reported outcomes, such as patient satisfaction, health-related quality of life (HRQOL) and coping. In this paper, we test the hypothesis that positive patient experiences with information contribute to improved health, more frequent use of approach coping and less avoidance based coping. Methods: A prospective study of 556 people with chronic illness. Patient experiences with information provided by hospital staff were assessed approximately 6 weeks after hospital stay in 1998. Coping behaviour was assessed 2 years after baseline hospitalization in 1998. HRQOL was assessed 6 weeks and 2 years after baseline hospitalization. We used chi-square statistics and t-test for independent samples to analyse the extent of selective attrition. Structural equation modelling was used to analyse the structural relations between patient experiences with information, (change in) HRQOL and subsequent coping behaviour. Results: Positive experiences with information were cross-sectionally but not longitudinally associated with improved HRQOL. Patient experiences with information are not related to avoidance coping, but positive experiences contribute to more frequent use of supportant coping. Conclusions:

The findings in the present study indicate that measures of patient experiences with information are not merely a reflection of patients HRQOL or of their coping behaviour. Although related, these patient-reported outcomes may be considered separate outcomes of health care delivery. More theoretical work is required to describe the relationships between different patient-reported outcomes.

Feng Xie, Pharmacy, National University of Singapore, Singapore; Julian Thumboo, Rheumatology and Immunology, SGH, Singapore; Shu-Chuen Li, Pharmacy, NUS, Singapore Aims: (1)To determine the extent to which the important items identified from perspective of patients with knee osteoarthritis (OA) using focus group methodology can be covered by the International Classification of Functioning, Disability and Health (ICF); (2) to determine the structure underlying these items in terms of the ICF; and (3) to critically evaluate the ICF Core Sets for OA. Methods: All 74 items of importance reported by patients with knee OA were independently linked with the ICF by two researchers on the basis of 10 linking rules which have been tested to be reliable. The percentage agreement between two researchers was calculated at each level of the ICF. Any disagreements were resolved by reaching a consensus through a discussion between two researchers led by a third researcher who is also trained in these linking rules. Results: Of 74 items, 11 (15%) can be connected with the ICF at the second level, while 62 (84%) and 1 (1%) at the third and fourth level, respectively. Concurrently, all items within the domain of pain can also be connected at the fourth level. The agreements between two researchers were 83 and 77% at the second and third level, respectively. The ICF Core Sets for OA can cover the majority of the items except several items within mental and social health domain. Conclusions: Most of the impacts of importance to patients with knee OA can be expressed in a global language at a higher level (i.e. third level). This study can specifically complement to the development of the ICF Core Sets for OA and render the possibility to compare the HRQoL measurement worldwide. Aims: To describe health related quality of life among women who have experienced intimate partner violence. Knowledge about these facts will be useful in evaluating the need for health services among women in this situation. Methods: It is a self administrated questionnaire survey, a cross-sectional study. All womens shelters in Norway were asked to participate. Eightyseven women who meet the inclusion criteria from October 2002 to April 2003 participated. Description of domestic violence was done by use of SVAWS (The Severity of Violence against Women Scale) about physical violence and PMWI (Psychological Maltreatment of Women Index) and health related quality of life by scores of SF-36. Results: Women staying median two weeks at womens shelters in Norway had experienced a multitude of threats and actual physical and psychological violence during their partnership. Their health related quality of life was extremely low and significant (p<0,001) below the norm for Norwegian people in all dimension. Conclusions: Women at Norwegian shelters who had experienced domestic violence have very low and clinically significant reduced health related quality of life scores. Health care workers must give priority for developing intervention programs for victims of domestic violence. Aims: The purpose of this paper is to introduce a model and measure of lifestyle balance and to report preliminary research on its validity. The model distinguishes the construct of lifestyle balance from other global positive states such as quality of life, subjective well-being and life satisfaction, and proposes principles that enable the model to be used as a guide for comparing an individual's customary life activities with a configuration of activities known to be related to well-being. Our research will answer the following questions: (1) Do people who lead balanced lifestyles experience less perceived stress? (2) How does lifestyle balance relate to other concepts of positive state? (3) Does the model contribute unique variance when combined with other predictors of well being? (4) Do demographic variables such as age, income, gender and education correlate with lifestyle balance in expected ways? Methods: Subjects included convenience samples of adults living in the Minnesota and New York State. Participants filled out the Occupational Balance: Lifestyle Inventory, Life Regard Index, Quality of Life Scale, Attitude and Behavioral Scale, Sense of Coherence, and the Satisfaction with Life Scale. Pearson correlations and regres-sion analysis were completed. Results: With over 330 subjects, we found: (1) a significant inverse correlation between perceived stress and measured balance. (r=)0.27, p<.001) (2) lifestyle balance correlated significantly with Sense of Coherence, (0.22) Quality of Life (0.43), Happiness (0.37), Life Meaning (0.33) and Affect Balance (0.43). (3) In a regression equation to predict happiness, the LBI was a significant predictor along with stress, income and sense of coherence. (r=0.50, p<.001) (4) Because lifestyle balance is a function of meeting needs in the context of regular activity patterns, one might expect that only experience (age) should be correlated significantly with the LBI. This is what we found. Aims: Subsyndromal depression has been found to affect quality of life, result in a significant degree of clinical and functional impairments. The objective of this study is assess the quality of life in individuals with subsyndromal depression who seek treatment at primary care services in Porto Alegre in sought of Brazil. Methods: This was a cross sectional study which the cases were divided into three groups according to the severity of depressive symptoms. One group consisted of subjects who were major depressive disorder, the second group patients with subsyndromal depression and the other group of subjects without depressive symptoms -controls. The participants completed the World Health Organization Instrument to Assess Quality of Life (WHOQOL -BREF), Centers for Epidemiologic Studies' Depression instrument (CES-D) to screening depression and Composite International Diagnostic Interview Instrument (CIDI) to confirm the diagnostic Results: The sample consisted of 438 primary care users where 35.2% had subsyndromal depression. Assessing the quality of life using the WHOQOL instrument, the patients with a subsyndromal depression reported significant higher scores in all dimensions than patients with major depressive disorder. The effect size of the perception of quality of life between these groups was for the most of domains moderate. About social and demographic variables , there was a direct association between quality of life and the schooling and a inverse association with physical comorbidity at all domains. For a more detailed analysis of variables correlated to quality of life a linear multiple regression was performed with demographic variables and level of depressive symptoms. This model showed that the depression was the variable most correlated with all dimensions of quality of life. Conclusions: The results suggest that subsyndromal depression has a impact in quality of life and worthy of clinical consideration mainly in primary care patients. Aims: The association between quality of life and depression is very complex. The literature data are insufficient for the recognition of an integrated or distinct concept. Quality of life and depression can be understood as opposing phenomena of wellbeing , they have a causal association or sometimes that depression can be seen as component of quality of life . The objective is evaluate the association between a generic measure of quality of life and levels of depressive symptoms to verify if these concepts are similar and can be considered as only model . Methods: Principal Components Analysis is a multivariate statistical technique for the purpose of examining the interdependence among variables. It has been applied to this study of the Factorial analysis based on Kaiser Meyer Olkin Model . Centers will be Epidemiologic Studies -Depression instrument (CES-D) instrument with 20-items that evaluate different levels of depressive symptoms and the World Health Organization Instrument to Assess Quality of Life (WHOQOL -BREF) with 26 items were analyzed . Results: Of the 438 patients , 39% presented symptoms suggestive of major depressive disorder and 26% no symptoms. Five factors were found to explain 50% of the total variance of this association: Factor I -negatives depressive symptoms (eigenvalue=7.85, 17%); factor II -Psychological dimension (eigenvalue=4.96, 10.78%); factor III -Physical dimension (eigenvalue=3.91, 8.51%); factor IVpositive feelings in the CES-D and a independence level dimension (eigenvalue=3.56, 7.76%); and factor V -social and Environment Dimensions (eigenvalue=2.77, 6.03%). The factor I components was only CES-D items, while in the other factors the WHOQOLs Items were most of them , being correlated in accordance with theirs domains. Conclusions: These data suggest that depression and quality of life are distinct concepts, although correlated. Aims: Twelve percent of patients after trauma of lumbar spine are suffering chronically of posttraumatic stress disorder's (PTSD). Aim of the study is to explore the impact of PTSD on Health related Quality of life. Methods: We examined 71 patients at average 50 months after acute trauma of lumbar and thoracolumbar spine (Th12-L5). The diagnosis of chronic PTSD was done by the revised form of the Impact of Event Questionnaire (IES-R). In addition to sociological and medical variables the Health Related Quality of Life (HRQoL) was evaluated by means of the revised Aachen Longterm Score (reALOS) and the Short Form 36 (SF-36). Depression was assessed using the Beck Depression Inventory (BDI). Results: Pain and mobility are the most important factors involving PTSD and HRQoL. According to this we found that patients with PTSD have more pain (T=)2.56; p=0,03) and poorer mobility (fingertip-to-floor distance T=)2.41; p=0,04; subjective mobility T=)2.74; p=0.02). PTSD patients had poorer values in all HRQoL scales (T>3.00; p<0.001) except the SF-36 social functioning scale (T=2.16; p=0.06). According to this, the IES-R scores correlate significantly high with all HRQoL Scales ()0.69<Spearman r< )0.47; p<0.001). Conclusions: PTSD symptoms do highly effect the HRQoL after spine trauma. Clinical Physicians do well if they try to prevent PTSD by reducing pain and preventing restricted mobility. Furthermore there seems to be need in early psychological support for the patients. Aims: The aim was to describe predictors of major depression, and how stroke patients' mental health and physical disability affect the life situation of their significant others. Methods: Seventy-one dyads consisting of patients with a diagnosed post stroke-depression and their significant others were included. The patients were included in a multi-center study of poststroke depression, and were assessed for depression with the Montgomery-Å sberg Depression Rating Scale and activities of daily living, according to Barthel. Depression diagnosis was according to DSM-IV with exclusion of the D-and time criterion. The assessments of significant others included their own life situation and proxy assessments of the patients' state of depression, anger, change of personality and need of assistance. Statistical analysis were performed using Mann-Whitney Utest, Chi square and stepwise multiple regression. Results: The multiple regression analysis showed that female gender (p=0.002), younger age (p=0.002), lower mood state (p=0.004), less perceived change in personality (p=0.036) and previous treated depression (p=0.010) predicted increased severity of depression among these patients. There was a significant relation between the patient's level of depression and Barthel score, with those with less impaired ADL having more major depression than those with minor depression (p=0.007). Significant others of male stroke patients reported a more negative impact on their life situation, than did significant others of female stroke patients (p=0.04). Conclusions: Being younger, female patient with lower mood state increases the risk of major depression for the patient. For the significant others of male stroke patients, life situation was negatively affected. Aims: Although the prevalence of mental health symptoms among peacekeepers is lower than that of combat veterans, little is known of the effect of these symptoms. This study of Australian peacekeepers who were recently awarded compensation for a mental health condition is the first study to explicitly examine these issues with respect to veteran peacekeepers' quality of life, health status and health service use consumption. Methods: As part of a veteran health service use study, a random sample of 187 peacekeepers who had been recently compensated for a mental health condition was drawn. Participants completed a questionnaire and follow-up telephone interview, including the AQoL and WHOQOL-Bref.

Results: The quality of life of compensated peacekeepers was poor; typically 50% worse than that of comparable civilians and about the same as Vietnam veterans seeking treatment for posttraumatic stress disorder. When compared with Vietnam War and World War II veterans, peacekeepers were less likely to have seen a general practitioner (GP) or medical specialist, but were more likely to have seen a psychiatrist, psychologist or allied health professional. However, their health service use was the highest, including use of mental health services. Although the numbers were small, peacekeepers were more likely to report that treatment had not helped them when compared with other veterans; 13% reported treatment had made their condition worse. Forty per cent reported stopping their treatment, usually because of treatment failure. Although peacekeepers were more satisfied with GP treatment for a mental health condition, they were less satisfied with their overall health care, possibly because of access difficulties. Conclusions: Peacekeepers who receive compensation for mental health conditions suffer a very poor quality of life, use excess health services and are in poor health. That 40% stop their treatment because of treatment failure suggests greater efforts are needed in health literacy and understanding of the mental health consequences of peacekeeping. Aims: The effect of salutogenic therapy has not been fully explored among people with mental health problems. Accordingly the aim of this study is to investigate the effect of salutogenic therapy principles in talk-therapy groups on coping with mental health problems. Methods: The study is a randomised controlled trial. The participants, residents in the community, were allocated to a coping-enhancing experimental group (n=59) and a control group (n=47) receiving standard care. Coping was measured using the Sense of Coherence (SOC) Questionnaire. The SCL-90-R (Symptom Checklist-90-Revised) was used as a single one-time assessment of the participants' mental health status at baseline. Results: Over 80% of the sample was encumbered with a number and intensity of symptoms that may be disturbing and prevent activities of daily living. Coping improved significantly in the experiment group (+6 points) compared with the control group ()2 points), p=0.03. Conclusions: The results of this study show that talk-therapy groups based on salutogenic therapy principles improve coping among people with mental health problems. This should encourage mental health professionals to offer a group treatment model and individual treatment model based on salutogenic therapy principles in the rehabilitation process among people with various mental health problems. However, the likelihood of sustaining control of schizophrenic symptoms may depend on treatment persistence. In this study, we compared treatment persistence between patients who were initiated on risperidone or olanzapine, the two most widely prescribed atypical antipsychotic agents. Methods: We identified patients with schizophrenia by ICD-9-CM codes (>1 inpatient or >2 outpatient ICD-9-CM codes >7 days apart) between July 1, 1998 and June 30, 1999. We further selected those who were prescribed the target drug during 4/1/1999 through 3/31/2000 provided that they were not on any antipsychotic agents during the prior 6 months. Using event history analysis, we compared treatment persistence in terms of hazard ratio between olanzapine and risperidone initiators, adjusting for patient sociodemographic and clinical characteristics. Results: Following the initiation of the target drug, more patients switched from risperidone to olanzapine than visa versa. Olanzapine initiators had decreased hazards of discontinuation by 14% (unadjusted; p<0.001) and 12% (adjusted; p=0.002), respectively, than risperidone initiators. Conclusions: Compared with risperidone, olanzapine seems to be better tolerated by patients as indicated by better treatment persistence. The initiation of olanzapine may thus increase the likelihood of sustaining control of symptoms of schizophrenia. Future research needs to provide a more comprehensive assessment of treatment persistence by considering other factors, such as formulary decision, and other antipsychotic agents in the study and developing models to assess treatment persistence and switching as two interdependent competing risks.

Mohammad Taghi Sheykhi, Social Science, Al-Zahra University, Tehran, Iran Aims: The paper explores the problems increasingly faced by the widows in Tehran. The paper also reflects network relationships as caused by widowhood leading to isolation and loneliness. Health-related problems of the widows are among the main objectives of the ''study population'' to be tackled. Widowhood transition in life, followed by change in relations, is examined from various dimensions. Finding a new perspective of the widows will help in improving their quality of life, their health conditions and so on. However, widowhood involves loss of roles for the women. Methods: The technique used in the present study to yield the required data is a combinedmethod approach of investigation. While the main method of research was administering questionnaires, the author used interview method also as per required. In the theoretical section of the work, relevant sociological literature on women was probed into, before entering the empirical part. However, while so far little attention has been paid to the widowhood studies in Iran, in the theoretical section, the author has referred to various theories and approaches. In completing the research survey, over 250 widows were randomly selected and for whom the questionnaires were filled in. Eventually 250 questionnaires were elicited, extracted and analyzed. Results: The data collected in the present survey reflect different dimensions of quality of life of the widows in Tehran, will help in planning a better life for them. Conclusions: The widows studied, acknowledged that many changes had occurred in their lives. Widows studied, expressed sufferings from diseases, health issues, financial constraints etc. Aims: Successful kidney transplantation improves quality of life and increases life expectancy in haemodialysis patients. While the number of chronic dialysis patients is constantly increasing in Hungary, the number of patients on the transplantation waiting list has not changed during the recent years. In this study we tested the hypothesis that direct or indirect negative experiences with transplantation, lack of information and also social status predicts the odds of being registered on the waiting list. Methods: Four-hundred and eighty-one patients from 7

Budapest dialysis units (N=481, mean age=58.6±15.1 yrs , median of dialysis=42 months [3-255]) completed a battery of questionnaires. Results: Twenty-seven percent of all patients were enrolled on the waiting list. Patients with at least a high school degree had a greater chance of transplantation waiting list registration then those with a lower degree of education (odds ratio: 3.33 compared to patients with primary education or less, p<0.01). In accordance with our preliminary studies, patients who encountered unsuccessful transplantation cases were less likely to register for the waiting list then others. Sixty-two percent of all patients reported not to have received sufficient information about transplantation. This opinion was significantly more prevalent in patients with lesser education (p<0.05, chi-square test). Conclusions: In summary we found that socio-cultural characteristics and the potential communication barrier between staff and patients could both play a role in forming negative attitudes towards renal transplantation in patients on chronic dialysis. Aims: Since the 1980s, community-based services have become the mainstay for mental health consumers in Hong Kong. With the recent surge of young consumers due to early detection of schizophrenia, community-based services must tailor their services to fit with the needs of varied consumers. Given that recent consumers may differ in their illness perception and experience with the service system from their more chronic counterparts, the present study compared the quality of life between recent and long-term mental health consumers and investigated a quality of life model by examining the effects of macro factors (i.e., continuity of service, service needs) and micro factors (i.e., self-stigma, insight of illness, perceived treatment coercion) on multiple indices of quality of life (i.e., physical functioning, mental functioning, life satisfaction), after consumers symptom severity and medical side effects were controlled. Methods: Using a cross-sectional design, 162 mental health consumers, with broad diagnosis of schizophrenia from various community-based psychiatric rehabilitation centers were interviewed using structured questionnaires. Results: Based on hierarchical regression analyses, differential effects of macro-micro factors on dimensions of quality of life were found. Whereas unmet service need was negatively related to physical health, both unmet need and poorer continuity of service were negatively related to mental health. Poorer continuity of service and higher perceived treatment coercion were negatively related to life satisfaction. T-tests showed no significant difference in physical health and mental health between recent and chronic consumers although life satisfaction was higher among the chronic group. Conclusions: Macro factors were found to be better predictors of quality of life than micro factors. Results highlighted the importance of enhancing the organization and scope of mental health care services in the improvement of consumer quality of life. Aims: Several studies have addressed the issue of the impact of cancer and its treatments on sexual functioning, especially for patients with breast and prostate cancer. Our study is part of a more extensive research project on the quality of life of cancer patients in Uruguay. We aimed to evaluate the performance of the FACT-G satisfaction with sex life item (GS7: ''I am satisfied with my sex life'') included in the Social/Family Well-being subscale. Methods: The study sample consisted of 309 cancer patients with various tumor sites at different stages of disease, and undergoing different treatments. Since the FACT-G offers the patients the option of not responding to this item, we first looked for differences in clinical and sociodemographic features of respondents vs. non-respondents. Second, in those patients who responded, the differences in sex life satisfaction were studied in relation with tumor location, tumor stage and other clinical and psychosocial variables. For statistical analysis Chi square, t-test, ANOVA statistics, and Spearman correlation coefficients were used. Results: The results showed that those who responded (n=261, 84.5%) tended to be younger (p<.002), married (p<.05), and more educated (p<.02); while patients still on treatment (p<.007) and with lower income (p<.001) were more reluctant to answer the item. Among those who responded, significant correlations were observed between the satisfaction with sex life item and the ECOG-PSR, total FACT-G, Functional Well-being subscale and TOI scores. Significant differences in GS7 mean scores were observed only for prostate cancer vs. all other tumors (p<.01), and self-administration of the questionnaire (p<.004). Conclusions: We concluded that the assessment of sexual satisfaction using the FACT-G GS7 item might be biased due to a high rate of patients refusing to answer. However, among those responding, we found that the satisfaction with sex life of patients with prostate cancer appeared to be significantly impaired when compared with patients with other tumor locations, pointing us to an important issue for further research.

Aims: Curative treatment (CT) of prostate cancer (PC) was reported to affect health-related quality of life (HRQOL). We hypothesize that HRQOL following CT is related to the longterm state of continence. Methods: Self-administered HRQOL questionnaire [Urology 57: 112-116, 2001 ] was used. Higher score means worse HRQOL. Continence score composed of self-assessment of wetness, number of pads and fecal incontinence. Included are consecutive patients, CT Candidates, who viewed a video presentation about CT options and chose either surgery, radiation or watchful waiting (WW). First questionnaire filled before CT and latest available questionnaire at least 12 months after CT were used. Urinary incontinence defined as either use of at least one pad a day or self report wetness more often than once a day. Results: CT continent patients main-tained their HRQOL as those on WW. Incontinent patients had significant reduction in their HRQOL. CT patients were also divided into 2 groups on follow up based on self reported no change vs. deterioration in sexual activity. We found no difference in HRQOL measured on follow up between the 2 groups. Fecal and urinary incontinence had same effect on HRQOL. Conclusions: Incontinence following CT for PC causes deterioration in HRQOL while self-reported reduction in level of sexual activity did not. Patients on WW seem to preserve HRQOL despite increase in the PSA level. Aims: This study aimed to analyze current levels of health-related quality of life in individuals with leukemia and lymphoma and compare them with existing normative data. Methods: This was a cross-sectional study of quality of life in patients attending to a hematological clinic at Imam Hospital in Tehran, Iran. Quality of life was measured using the SF-36 questionnaire. Then the results were compared with quality of lifer data for the normal population (n=4163). Results: In all 94 patients were interviewed. The mean survival time for patients was 40.1 months (SEM=4.3). The study findings are shown in the Table. Except for the physical functioning and role physical patients recorded poorer level of HRQOL. Conclusions: Apart from descriptive findings, the study results suggest that the SF-36 can be easily administered to the patients with leukemia and lymphoma. However, to have a better insight into the quality of life issues in these patients the use of specific questionnaires are recommended. Aims: Clinically relevant deficits in quality of life (QL) of tumor patients can be diagnosed by questionnaires, visual QL profiles and expert reports to coordinating practitioners (GPs), but patients require effective therapy. Quality and effectiveness of QL-enhancing treatment options, however, are uncertain and therefore were assessed and improved by methods of PDCA management in a regional tumor centre. Methods: One hundred and seventy (age 58 ) were enrolled in a prospective study between 12/02 and 06/04. Continuous PDCA (=plan, do, check, act) circles were applied to five QL-specific therapies (physio-, psycho-, social-, painrelief-therapy and strengthening fitness). According to expert opinion (5 specialists in oncology, psychology and trial methodology/QL) therapies were indicated if the patient showed deficits <50 on the respective score of a EORTC QL-profile (0=bad to 100=good). Thirty-eight coordinating GPs evaluated baseline situation of QL-specific therapies in the region. Two rounds of PDCA were performed:

(1) Identifying therapeutic deficits in the region and creating a list of specialists.

(2) Classifying qualified specialists, developing quality indicators, check quality of specialists, get advice and input from external experts. Results: 105 (62%) patients showed clinically relevant QL-deficits in at least 1 treatment option (max. 4 deficits). Their GPs, however, had no specialist available for physiotherapy (22% of cases), psychotherapy (19%), social rehabilitation (48%), pain therapy (19%), and fitness (42%). The first PDCA circle offered numerous specialists to all GPs for corresponding treatment options: 25, 10, 7, 11 and 22 (in total: 75 specialists). The second PDCA circle developed quality indicators by barrier analysis and interactive mode in quality circle, specifying levels of professional education and experience for each therapy option (e.g.,>400 lymph drainage/ year for physiotherapy). Conclusions: A system for treatment of clinically relevant QL deficits was achieved in breast cancer patients by applying PDCA methods. Results are promising, but must be confirmed in an ongoing randomised trial. . These 22 undetected cases had a median of 4 (range 1-7) additional, specific QL-deficits, particularly in emotional functioning (n=15), fatigue (n=13), and pain (n=12). Conclusions: Patients' self-reported clinically relevant QL deficits have to be regarded as a separate part of the disease that is overlooked and consequently not treated by physicians. Only the combination of objective health status as determined by the physician and subjective suffering as expressed by patients portray a complete picture of the patient and direct to the way of cure and care. Aims: The purpose of this study was to determine the levels of menopausal symptoms, sexual functioning, psychological distress and generic QL among younger breast cancer (BC) patients treated with adjuvant chemotherapy, and their interest in symptom management. Methods: Cross-sectional data (FACT-ES, SAQ, HADS, SF-36) were available on 214 patients (response 78%) who had been premenopausal at the time of BC diagnosis. The study sample consisted of: 93 women (<45 years; group 1) treated with chemotherapy, and 35 women (<45 years; group 2) and 40 women (>50 years; group 3) who had undergone surgery + radiotherapy only. Results: In group 1, the most frequently reported symptoms were: hot flushes (54%, sleep problems (53%), decreased libido (44%) and mood swings (43%). Controlling for age, group 1 reported significantly more menopausal problems (FACT-ES; p<.05) than did naturally menopausal women (group 3). Women in group 1 reported significantly worse sexual functioning than those in group 2 (p<.01). A quarter of women in group 1 were classified as clinically depressed (HADS), vs. 11% and 12% in groups 2 and 3. Group 1 reported significantly lower vitality and worse general health (SF-36), than groups 2 and 3 (p<.05). Patients with current menopausal symptoms expressed greatest interest in physical activity (78%), relaxation (64%) and psychosocial interventions (60%). Only a small minority of women (11%) had utilized psychosocial care for menopausal symptoms in the past. Conclusions: A substantial percentage of younger women with BC treated with chemotherapy experience menopausal and sexual problems, depression and impaired generic QL. There is considerable interest in psychosocial and/or behavioral approaches to symptom management. Aims: Considerable negative impact on health-related quality of life (HRQoL) has been reported in the acute phase of diagnosis and initial treatments for breast cancer, with improvements occurring with the passage of time in women remaining diseasefree. Standard adjuvant tamoxifen, when appropriate, has traditionally been prescribed for 5 years post-diagnosis. Although associated with side effects, little is known about tamoxifen's long-term effects on HRQoL. These were evaluated in this study. Methods: Breast cancer survivors (n=1373) were the Canadian participants in a trial of letrozole or placebo for 5 years following 5 years of standard adjuvant tamoxifen (NCIC CTG MA.17). Their quality of life data were obtained before randomization to letrozole/placebo. Controls (n=6266) were similarly aged well women drawn from a Canadian population-based cohort [Canadian Multicentre Osteoporosis Study (CaMos): Hopman et al.: CMAJ 2000 ]. The SF-36 was used to measure HRQoL. Results: More survivors than controls reported ‡3 comorbid diseases (29% vs. 16%). After adjusting for age, geographic region, race and comorbid diseases, survivors reported better HRQoL on each of the 8 SF-36 scales than controls (p<0.0001). However, differences were unlikely to be clinically important: adjusted effect sizes for each scale were small (range: 0.18-0.32) and between-group differences were smaller than the smallest increment of change, this latter representing the difference discernable by an individual when completing the SF-36. Analyses stratified by age (<50, 50-69, and ‡70 years) yielded similar results. Conclusions: HRQoL of breast cancer survivors after completing the standard 5 years of adjuvant tamoxifen was at least as good as that of similarly aged women who had never had breast cancer. Our results are reassuring because these survivors, by virtue of their interest in 5 more years of adjuvant treatment in a clinical trial, were likely to have been very compliant in taking the full 5 years of tamoxifen and thus at risk of any associated negative qualityof-life effects. (1) To cure a patient and to maintain his/her QoL (when a cure is possible),

(2) To improve survival and QoL (when a cure is not possible, but we have a chance to improve the patient's life expectancy) and (3) To improve a patient's QoL (when we have no chance to improve the patient's life expectancy). II. Treatment approach is chosen depending on treatment goal(s). The goal and the approach are specified by the type and the stage of cancer, and the patient's individual characteristics. -In the initial stages the first two goals (1 and 2) dominate and the approach of choice is an attempt to cure the patient. -In the advanced stages, the possibility of a cure decreases and the importance of QoL issues increases. -For terminally ill patients the major goal is to support QoL and, in this case, palliative treatment dominates. III. QoL is an important end-point at any stage of the disease and is to be measured before, during treatment, and at follow-up using standard QoL instruments. IV. Palliative care is to be started from the early stages with its proportion increasing as the disease progresses. Conclusions: The paradigm gives grounds for the unified approach to cancer treatment. It is of value in identifying end-points and outcomes in clinical trials and might be used by physicians in their routine practice.

Aims: The proposed prospective study aims to assess the way in which certain personality traits of women influence their HRQoL after breast cancer diagnosis. More specifically, the study will assess the amount of which sense of control and control modes influence the level of HRQoL among women after the diagnosis of breast cancer. The central hypothesis in the present study is that the pre-diagnosis sense of control and control modes of the women will predict HRQoL after the diagnosis. In other words, women whose prior personality involves more positive modes of control and a stronger sense of control, will show better adjust-ment after the diagnosis. This relationship will be stronger as the disease stage is higher. The emphasis is on the longitudinal design in which causality can be assessed. Methods: In this interdisciplinary study, medical variables (the disease stage, histological variables, the treatment plan etc.) and their moderating effect on the HRQoL of the women will be examined. The participants will be 4000 women between the ages of 45 and 75 who arrive to the Breast Health Center in Soroka University medical center -Beer-Sheva, in order to do a routine breast examination with no history of prior cancer and/or chronic disease. All women that receive a diagnosis of cancer will be invited to participate in the follow-up.

It is estimated that about at least 50 women will be eligible for this group. A control group (matched for age, education and marital status) of 50 women whose diagnosis is negative will be asked to participate. The participation will be strictly voluntarily and the women will be explained the purpose of the study. They will all sign an informed consent. Results: initial results of the study indicates that the more positive modes of control and a stronger sense of control is, the higher adjustment after the diagnosis.

The study is unique in its prospective design. To the best of our knowledge no prospective study examining the influence of predisease personality factor on later coping has been performed. The implications are developing intervention programs to help potential women to increase the sense of control and to develop more adaptive control modes. To date, few interventions have been tested specifically to assist cervical cancer survivors manage illness-related stress and improve health behaviors. Effective interventions could also influence disease endpoints through cross-talk between the neurologic, neuroendocrine and immune systems, the psychoneuroimmunologic (PNI) axis. We propose as our central hypothesis that a psychosocial telephone counseling (PTC) intervention can positively influence QOL and associated stress-related biomarkers. Methods: Fifty women with Stage I-III cervical cancer was randomized to receive either six sessions of PTC or usual care. Baseline and follow-up assessments included QOL measurement (e.g., FACT-Cx), and neuroendocrine and immunologic parameters (e.g., Th1:Th2 bias). To examine potential PTC benefits over time we conducted a repeated measures analysis of variance, controlling for baseline scores. Preliminary correlations between Th1:Th2 bias and QOL have been conducted. Results: PTC participants report significant benefit from baseline to follow-up on overall QOL (p=0.05), whereas control participants report a slight QOL decline over time. In addition, a promising association exists between longitudinal improvement in QOL and specific biomarkers. Conclusions: The use of PTC has the potential to positively influence QOL and modulate the PNI axis for cervical cancer survivors. -Hep] ; QLQ-HCC18) were used or developed in HCC pts. The SF-36 and NHP showed appropriate discriminant validity, but there was a lack of data to support their responsiveness. The FACT-G and FACT-Hep were validated in a cross-sectional survey; however, in an interventional study, the FACT-G was responsive to change, while the FACT-Hep was not. The QLQ-C30 was responsive to change in a small study; however, another larger study only found a tendency. The face validity of the QLQ-HCC18 is more evident in HCC than the FACT-Hep, but the psychometric characteristics of both instruments need to be further investigated. Conclusions: HCC is associated with a significant QOL burden and research of new treatments should capture QOL as an endpoint. Only a few studies used validated QOL instruments in HCC pts. Future research is warranted to further explore psychometric properties of QOL instruments in HCC. Aims: As parents life can be negatively affected by the asthma of their children, we aimed to develop a specific questionnaire to measure how parents live with children with asthma. We assume that as parents improve their QoL, they can better care their children. Methods: After focus group meetings with dozens of mothers, an early 48-item questionnaire was prepared.

To increase its content validity and sensitivity, we carried out further focus group meetings that leaded to a new 60-item questionnaire. It was administered to mothers recruited by paediatricians throughout Italy. Items were reduced by using combined criteria: ceiling-effect, effect-size (by asthma severity); factor analysis (main component oblimin rotated) structure. Results: One hundred and sixty five mothers (24-52 y.o.) filled in the questionnaire. Their children were 4-12 y.o., suffering from intermittent (N=66), mild (N=45), moderate (N=38) and severe asthma (N=16). The first data reduction excluded 30 items (those with ceiling effect >35%). Factor analysis produced 5 sub-scales (mother's worries, perception of children symptoms, comparison to others, social life, avoidance). To preserve the conceptual model, items sensitive to asthma severity were retained after a further reduction, so that each original factor was represented. In the final 16-item solution, items showed three factors underlying, though the first one explained 44.628 of variance and second order factor analysis showed their belonging to only one macro factor. This brought to aim at a one factor scale. Its performances are very good, with no ceiling effect, moderate skewness and a=0.92. Effect size against children asthma severity was 0.770 (severe-intermittent) and 0.480 (mild-intermittent). Conclusions: The AST-PQ16 (ASThma Parents Questionnaire, 16 items) showed to be sensitive to asthma severity, and able to detect parents' overunder estimation of children clinical condition. This estimation could lead to excessive paediatric examinations or underestimation of the risk. It does not replace children QoL measurement, but it can measure a wider range of age than direct QoL questionnaires can do. reporting an improvement in their health status was 1.5 ranging from 0.9 (sleep) to 1.8 (symptoms) for dimensions. Correspondingly for the m-R were:1.8, ranging from 0.9 (eye sympt.) to 1.8 (nasal sympt.). Conclusions: Despite differences in questionnaire development and content, psychometric properties of these two questionnaires are similar. Given that, the instruments may be selected according to which adapted best to the target population. Aims: The factor structure of a 12-item patient satisfaction with treatment questionnaire was explored with the goal of supporting a single factor solution and simplified scoring.

Methods: A large cross-sectional survey of patients with incontinence taking medication for overactive bladder were given the OAB-PSTQ via on-line survey. Two random samples were utilized to perform exploratory and confirmatory factor analysis. Exploratory factor analysis was conducted in SAS and LISREL. Confirmatory factor analysis was done using LISREL. Results: A total of 891 patients with completed surveys were analyzed. A sample of 445 patients was used to perform exploratory factor analysis. A one factor solutions resulted with scree plot from SAS showing an eigenvalue of 33.09 for factor 1 and only 1 other factor yielding an eigenvalue over 1 (1.12). All items loaded very highly on factor 1 (range 0.63-0.93), and much higher then other possible factors. The total weighted communality was 38.64 with unweighted of 8.5 for the 12-item questionnaire. Quartimax rotation supported a one factor solution. Promax and varimax rotations also resulted in high factor correlations (>0.68), further supporting these results. The confirmatory factor analysis in LISREL (sample=446) demonstrated high loadings on the latent variable (0.68 or greater).Good incremental fit indices were reported (NFI, NNFI, and CFI=0.96 or greater), absolute fit indices were good (SRMR=0.037, GFI >0.8), however parsimonious fit indices were not as good (AGFI=0.72, RMSEA=0.16). When the modification indices were examined and error variances were correlated between several of the items the fit indices improved. Conclusions: The exploratory and confirmatory factor analysis support using the OAB-PSTQ as a unidimensional questionnaire. Polytomous IRT analyses are being conducted to examine item and scale properties. Local dependence between items where error correlations optimized fit will be explored further. Aims: Within the EORTC, we developed a 32-item satisfaction with care questionnaire to measure patients' appraisal of hospital doctors and nurses, as well as aspects of services. This study assessed the psychometric characteristics of the questionnaire, the EORTC IN-PATSAT32, in a large, international sample of patients with cancer. Methods: Patients discharged from a surgery or medical oncology ward in 9 countries were invited to complete at home the EORTC IN-PATSAT32 as well as other instruments for psychometric testing. Results: Of 762 eligible patients recruited, 15% failed to return the questionnaires. Of the 647 compliant patients, 63% completed the questionnaires within 15 minutes and 82% required no help in its completion. Multitrait scaling analysis revealed excellent internal consistency and convergent validity, although some scales within the IN-PATSAT32 were relatively highly correlated. Test-retest data on 113 patients showed high reliability for most scales. Scales of the IN-PATSAT32 and of the QLQ-C30 were not significantly correlated, suggesting that the two questionnaires are assessing quite distinct concepts. The scales of the IN-PATSAT32 were able to discriminate clearly between patients with differing care expectations and differing intentions to recommend their hospital to others. Conclusions: This study supports the acceptability to patients, and the psychometric robustness of the EORTC IN-PATSAT32 questionnaire. Further studies are needed to assess the responsiveness of the questionnaire. Aims: This study validated the SF-8, a generic measure of health status, in Chinese with cervical and breast cancer in Taiwan. Methods: The study used a cross-sectional design. Patients with cervical or breast cancer recruited from 4 geographically representative oncology hospitals in Taiwan were interviewed using a questionnaire containing the SF-8, the EORTC QLQ-C30 (a validated cancer-specific quality-oflife instrument in Taiwan) and questions assessing sociodemographic and clinical status. Construct validity was examined using a priori hypotheses. Patients' SF-8 scores were factor analyzed and compared with US population norms. Results: A total of 1010 patients (breast cancer: 530; cervical cancer: 480; response rate: 98.3%) were interviewed. Pearson's correlations between SF-8 and QLQ-C30 scales measuring similar domains of HRQOL were strong (range: 0.54-0.73) and generally higher than correlations between scales measuring different domains. Patients with higher Karnofsky Performance Scale scores had better SF-8 scores than those with lower KPS scores (p<0.001 for all, ANO-VA). A factor analysis yielded a one-factor solution, which was different from the hypothesized factor structure for constructing physical component summary (PCS-8) and mental component summary (MCS-8). The mean PCS-8 and MCS-8 scores of the patients were lower and higher than age-adjusted US population norms, respectively. Conclusions: The SF-8 demonstrates satisfactory acceptability and construct validity in patients with cervical or breast cancer in Taiwan. However, the interpretation of the SF-8 scores (calculated using the standard US scoring method) in Taiwan needs to be further studied. Aims: The prevalence of diabetes is increasing worldwide. The purpose of this study was to assess the psychometric properties of a newly developed instrument for measuring quality of life in people with diabetes in Iran. Methods: Literature review and interviews were conducted with 14 people with diabetes to design the DQLQ in Persian (the Iranian language). Then 131 people with diabetes completed the questionnaire and the psychometric properties of the DQLQ were assessed. Results: The mean age of the respondents was 47.3 (SD=12.7) years and mostly were male (60%) suffering from type 2 diabetes (85%). In all, 70 items were identified for inclusion in the questionnaire. Each item was rated on a 5-point Likert scale. After performing face and content validity the number of items were reduced to 61. Then item correlations and principal component analysis was carried out to delete unsatisfactory items and to specify the factor structure of the questionnaire. Only 45 items achieved the satisfactory criteria resulting in five subscales namely: physical (9 items), psychological (11 items), social (7 items), financial (5 items), and impact of diabetes and its treatment (13 items). The internal consistency reliability of the five subscales was satisfactory (Cronbach's a coefficients varied between 0.88 and 0.93) and test-retest results revealed good scale stability (correlations varied between 0.86 and 0.90). Conclusions: The findings suggest that the DQLQ-45 is a valid and reliable measure of quality of life in people with diabetes. It is hoped that the instrument and our data would provide the basis for future outcome studies among diabetes patients. Indeed conducting larger studies are essential to lend support for the inclusion of the questionnaire in such investigations. Aims: Cardiac rehabilitation programs in the USA are required to demonstrate their effectiveness in terms of QOL. A previous head-to-head comparison of the SF-36 and the QLI showed the QLI to be more sensitive to change in patients with cardiac arrythmias. In addition, 11 published studies of cardiac interventions have shown the QLI to be sensitive enough to detect change in QOL. The purpose of this study was to establish the psychometric properties of Ferrans and Powers Quality of Life Index -Cardiac Version when used with cardiac rehabilitation patients. Methods: Questionnaires were mailed to patients at least 6 months after treatment for cardiac disease (cardiac bypass surgery, angioplasty, or implantable defibrillator). Participants (n=576) were drawn from cardiology practices throughout Illinois, including Chicago. Results: Evidence of construct validity was provided by the strength and direction of association between QLI scores and variables expected to be related to quality of life (r=0.33-0.79). Consistent with predictions, better health status and life satisfaction were significantly related to better QOL as measured by the QLI. Greater depression, poorer coping with stress, and greater pain were significantly related to poorer QOL as measured by the QLI. Alphas above 0.70 provided support for internal consistency reliability for the total QLI (as=0.94-0.96) and for the four domain subscales (health and functioning, psychological/spiritual, social and economic, and family). Comparisons of data for men and women showed no significant gender differences, indicating the QLI was equally appropriate for both. Conclusions: The findings supported the reliability and validity of the QLI when used with these cardiac patients. The QLI provides an additional choice for an instrument for documentation of effectiveness of cardiac rehabilitation programs. Aims: To determine if improvements in the SF-36v2 demonstrated in the US source version are observed in the Hong Kong Chinese translation. Methods: Data were collected via telephone from 2410 Chinese adults randomly selected from the general population in HK (response rate=85%, mean age=42.9, 52% female). Respondents completed the full Chinese (HK) SF-36 Version 1 (v1) and the Version 2 (v2) Role Physical (RP) and Role Emotional (RE) scales, the primary scales that differ in the v1 and v2 HK forms. SF-36v2 RP and RE scales use 5 responses (all of the time-none of the time) vs. yes/no responses in v1. Role functioning scales were compared in terms of reliability (Cronbach's a), range covered, and floor and ceiling effects. The factor structure of the eight SF-36 scales was compared using v1 and v2 role scales with principal components analysis and varimax rotation. Discriminant validity of the scales and the v1 and v2 physical (PCS) and mental (MCS) component summary measures was evaluated using a 4-group test (self-reported major medical only, psychiatric only, medical and psychiatric, no reported condition). Results: Reliability increased for the v2 role scales (RPv1=0.83, RPv2=0.89; REv1=0.82, REv2=0.86) , while the range of measurement approximately doubled with v2. Floor effects were below 1% for both v2 role scales, compared to 7% (RP) and 16% (RE) for v1. Ceiling effects also declined for both scales. The 8-scale factor structure was similar if either v1 or v2 role scales were used. V2 role scales and v2 PCS and MCS measures had 5-85% greater precision in discriminating between medical and psychiatric groups than corresponding v1 measures. Aims: The purpose of this study was to develop a desirable QOL questionnaire for child care. Methods: 205 mothers who take care of children under 5 years participated in this study. Our new original self-administered questionnaire consists of 42 questions divided into 10 categories. Results: The Cronbach's a coefficients of our questionnaire were high enough to accept for clinical use: 0.91 in financial circumstances and eating habits, 0.88 in sleeping habits and living environments, mother-child interaction, 0.82 in social circumstances, 0.78 in everyday life, etc., respectively. Our questionnaire contained 11 main factors which matched the 10 categories quite precisely. There were significantly positive correlations between mother-child inter-action and well-being (p<0.01), mother-child interaction and psychological conditions (p<0.01), mother-child interaction and everyday life (p<0.01), everyday life and well-being (p<0.01), everyday life and psychological conditions (p<0.01) etc., respectively. For the high QOL child care, mother-child interaction and everyday life are most important for the childrearing mothers from the viewpoint of their psychological conditions and well-being. The QOL evaluation scores were best for mothers communicating well with their peers (p<0.01). Conclusions: We had believed the QOL of child-rearing mothers had been too difficult to evaluate, but these findings indicate that our new QOL questionnaire has a sufficient reliability and validity to use for child-rearing mothers. The most important points of this study are intimacy between mother and child and support for the child-rearing mothers. We think it's important for child-rearing mothers to relax, since this attitude itself will have a positive effect on their child's development. We must pay more attention to many kinds of problems child-rearing mothers have at the various developmental stages of children. Item pools were derived from: (1) literature review to identify issues relating to HRQOL measurement in these groups, existing HRQOL instruments for use in HIV + adults, and instruments for use in non-HIV + children and adolescents; (2) Metaplan consensus session with clinical experts and HRQOL measurement experts to define quest, characteristics, outline possible content and identify areas of particular difficulty in measuring HRQOL in Ped-HIV+; (3) 6 focus groups (FGs) with a total of 38 HIV+ children and adolescents and their parents. Content analysis of FG transcripts was performed by two independent reviewers. Preliminary item reduction was performed in 3 separate expert review sessions, in which redundant, ambiguous or irrelevant items were eliminated. Results: Initial item pools for children aged 6-12 included 76 items in 10 dimensions and 115 items in 13 dimensions for adolescents (87 items and 110 items for parents, respectively). Various items were common to the child and adolescent versions to allow continuity of instrument use across age groups. After expert review, item pools were reduced to 51 items (children) and 65 items (adolescents). Challenges arising in the construction of these item pools include the fact that most children under 12 are not aware they are HIV+, and the differing degree of experience and characteristics of children and adolescents participating in the FGs. Conclusions: Measuring HRQOL in this population entails several specific challenges. Nevertheless, it was possible to construct item pools by appropriate adaptation of recommended procedures. Further item reduction will take place in larger samples of the target population. Aims: The purpose of this study was to develop the desirable QOL care note questionnaire for the elderly persons supported by mandatory health care system. Methods: 53 elderly persons supported by mandatory health care system participated in this study. Our new original self-administered questionnaire consists of 21 questions divided into tripartite parties (i.e. patients, family and care giver specialist). Results: Pearson's correlation coefficient between our new QOL care note questionnaire and the established Japanese Society of QOL Research Questionnaire (JSQOLRQ) was r=0.82 (p<0.01). Cronbach's a coefficients of JSQOLRQ were high enough to accept for clinical use: 0.90 in well-being, 0.88 in mental function, 0.76 in economical problems, etc., respectively. JSQOLRQ contained 6 main factors and cumulative contribution was 0.98. The QOL score by care giver specialists demonstrated the significantly more excellent than the one by patients by themselves (p<0.05) and the one by family (p<0.01). Significantly positive correlation (r=0.91, p<0.01) was indicated between the total QOL score by patients and the one by family. QOL improvements evaluated by care giver specialists due to health care service were more excellent than the ones evaluated by patients. Conclusions: The dissociation of QOL evaluation among patients by themselves, family and care giver specialists, especially overestimation by care giver specialist which was demonstrated in our study is most dangerous factor to happen the trouble on health care practice for the elderly. It is very important to communize the objective QOL evaluation data by three different parties openly. We must pay more attention to prevent the dissociation of QOL evaluation, and must practice the evidence-based health care service for the elderly in future. Aims: The Health Utilities Index Mark 3 (HUI3) is a widely used utility-based generic, preference-scored, comprehensive system for measuring health status and health-related quality of life; however, its Korean version has not been validated. We therefore present its Korean adaptation and validation. Methods: Subjects were consecutive sample of patients who were admitted to the emergency center of the Bundang CHA general hospital near in Seoul, Korea. For validation we communicated with the Health Utilities Inc., and complied with their guidelines. Cultural adaptation included the instrument translation (forward and backward), an expert consensus meeting and a pilot test with a few patients (n=10) in order to produce a conceptually equivalent Korean version of the HUI3. To validate the Korean consensus version, psychometric properties were examined: reliability was assessed by Cronbach's a and test-retest intraclass correlation coefficient (ICC), and validity was checked against the SF-36 self-rated health transition item (SF-1, scale of 1 to 5; 1=perfect health and 5=worst health) by examining Spearman correlation coefficients. Results: Two hundred patients were participated (mean age=48.6, SD=19.1; 51.5% male). When patients were admitted they were 'walk-in by themselves' (52%), 'walk-in by holding guidance's arms' (30%), or 'unable to walk at all' (18% Aims: For the WHOQOL-BREF, three intermediate scale descriptors on the 5-point Likert-type scale was originally selected using rating method, in which participants were asked to rate the level of descriptors on a visual analogue scale (VAS) given the two end anchors. However, this direct evaluation (absolute judgment) assumed that participants can provide a precise scale value for each descriptor. This assumption might not be tenable in practice. In this study, a relative evaluation (relative judgment) on the descriptors by full-ranking method was proposed to investigate the scale value of each descriptor Methods: Eighty undergraduate students were asked to rank all the descriptors at a time according to the level within each of the four categories (capacity, frequency, intensity, and evaluation). In each category, there were 17-24 descriptors including the extreme ones. Thurstonian ranking model based on the Gibbs sampler [Yao and Bockenholt, 1999] was used to analyze the ranking data for calculating the scale value of each descriptor. As a result, five descriptors with two end points and three intermediates were selected according to the scale values derived from the ranking method and they were compared with the selected scale descriptors from the rating method. Results: The results showed that the orders of the scale descriptors by ranking method and rating method were slightly different. The interval values of scale descriptors calculating from the two methods were slightly different as well. Conclusions: Generally, results of ranking method were generally consistent with the findings gained from the rating method, but ranking was better than rating because end descriptors were not necessarily fixed. In addition, ranking method is an easier approach for respondents than rating method. Consequently, ranking method was proposed as a better method to find the scale values of descriptors than rating method. Aims: Although assessment of health-related quality of life (HRQoL) is increasingly used alongside biological measures in the management epilepsy patients, its utilisation in Malaysia remains to be explored. One reason could be due to lack of established HRQoL instruments in Malay language. This study aims to test the reliability and validity of the Malay version of Quality of Life In Epilepsy instrument (QOLIE-30) which contains 30 items tapping on 7 domains (seizure worry, overall QoL, emotional well-being, energy/fatigue, cognitive functioning, medication effects and social functioning). Methods: The original QOLIE-31 was translated, modified and cross-culturally adapted into Malay to generate the final Malay QOLIE-30 and administered to a group of patients with epilepsy. Descriptive statistics, reliability test (Cronbach's a), Pearson's correlation coefficient and principal factor analysis were carried out using SPSS 12. Results: A total of 96 out-patients from 9 public hospitals in Sabah completed the QOLIE-30 (mean age=32; range:18-76 years; 57 males; 54 unmarried; 44 jobless). Internal consistency reliability ranged from moderate to high for all domains including Overall Score (Cronbach's a=0. 495-0.906) . Correlations between domain and Overall Score were all strong and significant (Pearson's r=0.495-0.884; p<0.01). Eight domains explaining 70% of the total variance were produced by factor analysis, indicating evidence of construct validity. Conclusions: These outcomes show early evidence of reliability and validity of the Malay QOLIE-30 for assessing HRQoL in Malaysian patients with epilepsy. In view of this, such population stands to gain valuable benefits if treatment is focused at improving disease condition, managing adverse drug reactions as well as enhancing HRQoL status. Aims: To compare the responsiveness of two global quality of life (QOL) scales in heart surgery patients. Methods: In a clinical trial, 114 patients completed a single-item scale (gQOL) and the 16-item Quality of Life Survey (QOLS-N, sum score range 16-112) before and 12 months after surgery. All items were measured on 7-point Likert response scales. The Wilson & Cleary causal pathway model guided data analysis. Convergent validity, and discriminant validity related to physiological and functional measures, were evaluated with Pearson's r. Distri-bution-based (using standardized response means) and anchorbased responsiveness was assessed. For the anchor-based approach, a binary factor identified patients with clinically significant improvement in the SF-36 Physical as well as Mental Component Summaries. The predictive value of this factor was tested in analysis of covariance models. Results: At 12 months after surgery, both scales demonstrated an expected low correlation to a physiological variable (preoperative creatinine, gQOL r=0.02, QOLS-N r=0.11) and moderate correlation to a functional measure (sleep difficulties, gQOL r=0.29, . Anchor based responsiveness was demonstrated by both scales (gQOL, FANCOVA=5.41, p=0.022; FANCOVA=12.87, p=0.001) . Standardized response means at follow-up were 0.53 (gQOL) and 0.17 (QOLS-N). Moderate (baseline r=0.44) and high (follow-up r=0.52) correlations were found between the gQOL and the QOLS-N. Conclusions: Despite similarities in discriminant and responsiveness properties, the two instruments do not appear to reflect identical content regarding global QOL in this patient population. Prior to expressing a preference for either instrument, further analysis is needed to clarify relationships between the gQOL and QOLS-N at the construct level, and to separate differences in content from scaling effects. Aims: The Edinburgh Depression Scale (EPDS) is a widely used instrument to measure postnatal depression in women. The EPDS contains 10 items and each item is rated on a 4-point scale, giving maximum scores of 30. Scores of 13 or more is considered to be a significant case of postnatal depression, while scores of 9-12 represents borderline and 0-8 normal. This study aimed to translate and test the reliability and validity of the Iranian version of the EPDS. Methods: The English language version of the EPDS was translated into Persian (Iranian language) and was used in this study. The questionnaire was administered to a consecutive sample of 100 women with normal (n=50) and caesarean section (n=50) deliveries at two points in time: 6-8 weeks and 12-14 weeks after delivery. Statistical analysis was performed to test the reliability and validity of the EPDS. Results: In general the Iranian version of the EPDS found to be acceptable to all women. Overall 19% of women at time 1 and 16% at time 2 scored 13 or higher on the EPDS. Cronbach's a coefficient (to test reliability) has been found to be 0.77 at time 1 and 0.86 at time 2. Validity as performed using known groups comparison analysis showed satisfactory results. The questionnaire discriminated well between sub-groups of women differing in mode of delivery in the expected direction. Conclusions: This validation study of the Iranian version of the EPDS proved that it is an acceptable, a reliable and valid measure of postnatal depression among women. Aims: Patients with advanced cancer have an impaired quality of life (QoL) due to a variety of symptoms. We have developed a model of symptom interference with QoL based on a 5-grade scale of QoL impairment as compared to a population norm (PN): no impairment (no changes as compared to a PN), mild (25% decrease from a PN), moderate (25-50% decrease), severe (50-75% decrease) and critical (>75% decrease). Understanding of symptom interference with QoL provides the basis for better cancer outcome measurement. The goal of research was to study the structure of advanced cancer patient population based on the model of symptom interference with QoL. Methods: 201 advanced cancer patients (male/female 80/121; mean age 61.5) were enrolled in the study. To compare with controls a gender-and age-adjusted sample from the St.Petersburg PN was used. SF-36 and M.D. Anderson Symptom Inventory were used for patientreported outcome assessment. The distribution of patients according to the grades was provided using the Integral QoL Index (IQLI). Results: The following distribution of patient population into 5 groups according to the model was observed: critical impairment-42%, severe impairment-21%, moderate-10%, mild )11%, and no impairment-16%. The values of IQLI differed significantly between the groups (p=0.00002). The most severe symptoms were: fatigue (95%), sleep disturbance (75%), and pain (65%). Moderate-to-severe symptoms led to critical and severe QoL impairment. Conclusions: The 5-grade model of symptom interference with QoL allows to identify 5 groups of advanced cancer patients with different levels of QoL impairment. The majority of patients exhibit critical and severe QoL impairment. Accurate grading of QoL impairment and analysis of symptom burden is a valuable approach to improve adequate management of advanced cancer. (DLTV) has been used to assess functional impairment among patients with age-related macular degeneration (AMD). The instrument is comprised of 24 items; an overall score and four domain scores can be calculated. This study evaluated the psychometric properties of the DLTV using patient-reported rates of use of services and devices. Methods: The DLTV was included in a survey mailed to members of an AMD patient advocacy organization. Included in the survey were demographic and clinical items and questions about use of services, visual assistive devices, and caregiving. Respondents provided informed consent.

Responses were analyzed by visual impairment category based on publications relating DLTV scores to visual acuity. Data were analyzed in SAS. Results: Four of 803 respondents did not complete the DLTV; missing data were uncommon. Most (56%) respondents were male; mean age was 73 years. Estimated visual acuity was 20/50 or better for 60% of the respondents. Internal consistency reliability of the DLTV overall score was 0.97; for each of the four domains, it ranged from 0.72 to 0.95. Construct validity was demonstrated with moderate correlations between the DLTV and use of services and devices. The DLTV demonstrated discriminant validity and could distinguish between respondents with different levels of impairment, i.e., those who reported regular caregiving compared to those without regular caregiving (mean DTLV overall score 49.8 vs. 79.0, p<0.0001). Conclusions: Although this study was limited by the lack of validated visual acuity data, in this population, the DLTV demonstrated reliability and validity. This analysis complements existing research and supports the use of the DLTV in AMD patients. Future research should seek to confirm that the relationships between DLTV scores and use of services, devices, and caregiving are similar in the general AMD population. Aims: The objective of this study was to estimate influence of a cranium irradiation (CI) on quality of life (QoL) of adolescents with long remissions of acute lymphoblastic leukemia (ALL). Methods: 60 adolescents (m-41; f-19) had been treated at children's age concerning ALL during 1984-1998 enrolled. Twenty patients received no program therapy and 40 patients had been treated under the protocol of ALL-BFM-90m. The questionnaire of EORTC QLQ C-30 used for estimation of QoL. The age median of patients at moment of filling a questionnaire was 19.0±2.0 years. CI and risk group used as factors determining the intensity of all therapy. Results: In group treated according to ALL-BFM-90m 9/40 (22.5%) patients a CI did not received (standard risk) and 16/40 (40.0%) and 15/40 (37.5%) patients had received 12 Gy and 18 Gy accordingly (intermediate risk). The analysis of QoL did not reveal distinctions depending on irradiation doses or risk group. In group treated before introduction in practice of ALL-BFM-90m to 5/20 (25.0%) patients a CI did not use and 8/20 (40.0%); 3/20 (15.0%) and 4/20 (20.0%) patients had received 12 Gy; 18 Gy and 24 Gy accordingly. Deterioration of QoL of patients received 18 and 24 Gy had revealed. Reduction of emotional, social and especially cognitive functioning determined ( Aims: Treatment centres (dedicated units offering efficient, elective day and short-stay surgery) have recently been introduced in the UK. Clinicians, patients, and policy makers are concerned that efficient service does not compromise quality. We were commissioned by the UK Department of Health to conduct a systematic review and scientific appraisal of patient-reported outcome measures (PROMS) in five areas of surgery (cataract surgery, hip replacement, knee replacement, procedures for varicose veins, hernia repair) and to provide recommendations for their routine use in treatment centres. Methods: We developed a systematic search strategy that combined clinical and methodological terms and searched all relevant databases: MEDLINE (via PUBMED), EMBASE, CINAHL, Psychinfo, Web of Knowledge, IBSS (BIDS), and The Cochrane Library. To identify unpublished literature, we also searched SIGLE and the New York Academy of Medicine Grey Literature Report. We used clearly specified inclusion/exclusion criteria and standard data extraction methods to identify PROMS for review, and gold standard psychometric methods to evaluate measures. Results: Our searches identified 9157 articles, which were reduced to 927 potentially relevant articles during initial screening. Final screening, carried out by two members of the team trained in psychometrics, identified 415 articles (71 for cataracts, 281 for hip and knee replacement, 39 for varicose veins and 24 for hernia). Extraction, analysis and synthesis of these articles led to recommendations of at least one disease-specific and one generic measure in each of the five areas of surgery. Conclusions: Findings from this study provide a scientifically rigorous evidence-base for a national study to pilot the routine use of recommended PROMS in evaluating UK treatment centres. Aims: Evidence has shown a decrease in QoL during the menopausal transition. Since menopause experience and QoL perception are affected by sociocultural factors, there is the need to create a measure which is specific to both premenopausal women and the culture they live in. The aim of this study was therefore to develop a menopause-specific QoL instrument based on Thai women's perspectives. This paper reports conceptualization and development of QoL constructs for the new measure. Methods: Using semistructured interviews, the study described midlife Thai women's perceptions of QoL. Fifty women aged 45-54 years were selected to cover the broadest range of views possible. Participants were invited to explicate the meaning of QoL and components of good QoL, nominate 5 most important things in life in order of importance, and evaluate their current QoL as a score between 1 and 100 or a choice of 6 words. Results: A qualitative content analysis of responses revealed 4 characteristics of a person with good QoL: healthy; happy family and good social relationship; enough money to live comfortably; and doing something worthwhile. Attributes of life determined as most important were health, social relationship, and financial concern, respectively. Interpersonal relationship appeared to be as valued a component of a good life as health status. On average, this group of women perceived their QoL as good with a score 80 out of 100. Women with higher QoL tended to have higher education and income than those with lower QoL. The high QoL women also valued social relationships more than health while the opposite was true for women with lower QoL. Of note was that some women rated their QoL as high despite giving their health a low score; the opposite also happened. Conclusions: In conclusion, the present findings support the notion that QoL is a multidimensional construct, including but not limited to health. The results of the study provide some foundation for QoL assessment in menopause population. Aims: Studies of the higher-order factor structure underlying SF-36 scale scores in Japan and other Asian countries suggest that the hypothesized two-component model confirmed in many Western countries may not be robust in Asian countries. In this study, we investigated the goodness-of-fit of a threecomponent model for the Japanese-language version of the SF-36 Health Survey. Methods: Data (n=2963) came from a random sample of all registered residents of Japan between 20 and 80 years old. In addition to the SF-36, the survey included questions about medical conditions, perceived stress, and days of absence from work. With the hypothesis of 3 components, we used structural equation modeling and computed the Goodness-of-Fit Index (GFI) and the adjusted GFI. Scores for each of the three summary scales were computed with the bestfitting model, and were evaluated with external criteria. Results: The GFI and AGFI were 0.982 and 0.918 for the best-fitting 3-component model, and 0.946 and 0.860 for the 2-component model. The Physical Component Summary (PCS) was strongly associated (factor loading >0.6) with PF; the Mental Component Summary (MCS) was strongly associated with GH, VT, and MH; and the Role Component Summary (RCS) was strongly associated with RP, SF, and RE. The correlation between PCS and MCS was 0.405, that between PCS and RCS was 0.516, and that between MCS and RCS was 0.467. The three summary scores were correlated with external criteria (the number of medical conditions, perceived stress, and days of absence from work, respectively), as follows: correlations with PCS were )0. Aims: To report the normative scores of physical and mental health for Thai population and to assess the known-groups validity of the multi-item health measure used in a recent national household survey. Methods: A national representative sample of 37,202 adults aged 15-98 years were interviewed for health status using a 9-item questionnaire (called 9-THAI). A confirmatory factor analysis (CFA) of the 9-THAI supported its construct validity of two health domains: physical and mental. Based on CFA, the final scores were calculated by a summation of the items and transformed to 0 (worst health)-100 (best health) scale. Appropriate tests according to data distribution were used to assess known-groups validity. Results: The average (SD) scores of physical health and mental health were 91.5 (15.6) and 92.0 (13.7), respectively. The normative scores were derived for subgroups based on age, gender, region, and comorbidity of the surveyed population. As hypothesized, the physical and mental health scores for the worse health groups were significantly lower than those for the better health groups, suggesting the known-groups validity of the measure. The groups with worse health status were relatively older and those who reported prior year hospitalization or had chronic diseases. Conclusions: The 9-THAI possesses known-groups validity. The measure could be applied in evaluative research because the normative scores for the general population have been determined. Aims: Safe fast-acting anxiety medications are important to patients and society. While the HAM-A is the standard for assessing response to anxiolytic treatment, it requires a clinician interview and has not been validated for use sooner than 1 week. This study develops and evaluates the psychometric properties of a new patient-reported instrument designed to detect reduction of anxiety symptoms during the first week of treatment. Methods: Qualitative: clinician interviews and patient focus groups were conducted to identify relevant constructs; discussions focused on early symptom improvement and meaningful changes in GAD symptoms. The draft questionnaire underwent iterative sets of cognitive interviews to inform item reduction and revision. Quantitative: A double-blind, randomized, placebo-controlled study of paroxetine and lorazepam assessed the performance of the DAS-A in GAD patients. Analyses evaluated the structure, reliability, validity, and utility of the DAS-A. Results: Qualitative: There was consistency across focus groups and clinicians in the description of symptoms that improve first (e.g., feelings of anxiety and irritability). The final item set was easily understood by interview participants. Quantitative: Factor analyses indicated a unidimensional structure best described the data. Itemlevel descriptive statistics, alphas, effect sizes, and validity correlations with other scales were favorable. Most importantly, the DAS-A detected separation from placebo 24 h post-dose for lorazepam and correlated with other anxiety measures. Conclusions: This study resulted in the development of a reliable and valid instrument addressing the DSM-IV dimensions of GAD. The DAS-A is capable of detecting reduction in anxiety symptoms within 24 h, making it a desirable measure to include in future trials of fast-acting anxiety medications. P-220/1453/DO TAIWANESE SELECT THE SAME 26 CORE ITEMS FOR THE WHOQOL-BREF AS IN THE STANDARD WHOQOL-BREF? Ching-hua Wu, Grace Yao, Chia-huei Wu, Psychology, National Taiwan University, Taipei, Taiwan

Aims: The items of the standard WHOQOL-BREF were selected from the standard WHOQOL-100 based on the data from 18 countries, 11,275 subjects. However, the appropriateness of the items for Taiwanese was questionable. The purpose of this study was to select 26 core items for the WHOQOL-BREF (called Taiwan version) based on Taiwanese responses on the WHOQOL-100. Methods: 1068 participants were randomly selected from 17 hospitals throughout Taiwan completed the WHOQOL-100. Following the psychometric criteria suggested by the WHOQOL, we selected 26 items from the WHOQOL-100 completed by these participants. The psychometric qualities of the 26 items, in terms of reliability, validity, and predictability, were compared between the standard and Taiwan versions. Results: The 26 items selected from Taiwanese were quite different from the items of the standard version; however, both versions have similar reliability, validity, and prediction on overall quality of life. Conclusions: Although the individual items selected were different from the standard version, we can still conclude that the standard WHOQOL-BREF is applicable to Taiwanese on predicting peoples quality of life. Aims: To develop a spiritual wellbeing questionnaire, focusing on the ''function'' rather than the ''substance'' of patients' beliefs and practices. Methods: EORTC Guidelines for module development were used with modifications. A literature search identified potential issues in three areas: relationships with self and others, existential concerns and religious or spiritual beliefs and practices. These, translated as necessary from English into six other European languages, were shown to 22 patients and 22 health professionals from seven European countries for comments on content and relevance. Highest ranked issues were rewritten as items (i.e. in statement form), and shown to 17 patients from three countries. Items from this phase were retained on the basis of patient rankings, translatability, and correlations between items. Results: Simple questions for assessing importance of issues were inadequate for complex issues, so were rephrased. Translation was difficult for items relating to ''spirituality,'' which has culturally-specific meanings. EORTC Guidelines were modified for earlier deletion of items. Eighty-four issues were reduced to 40 items for pilottesting. For some patients, responding to the items became a spiritual intervention as they considered previously unaddressed issues. Conclusions: The EORTC Guidelines were a useful framework for development but required modification for this complex project. The next study phase will involve a larger sample, and debriefing interviews investigating the implications of self-completing the questionnaire. Aims: Factor analysis is widely used to evaluate whether questionnaire items can be grouped into clusters representing different dimensions of the construct under study. This review focuses on the appropriate use of factor analysis to examine the construct validity of health status questionnaires, using the Medical Outcomes Study Short Form-36 (SF-36) as an example. Methods: Articles were systematically searched in Medline and assessed according to a number of criteria: choice and justification of the methods used, quality of performance of exploratory and/or confirmatory factor analysis, sample size and quality of data, and reporting of statistical entities. Results: 26 Studies were identified: an exploratory factor analysis was performed in 20 studies, a confirmatory factor analysis was performed in five studies and in one study both were performed. Substantial shortcomings were found in the reporting and justification of the methods applied. In 15 of the 20 studies in which exploratory factor analysis was performed, confirmatory factor analysis would have been more appropriate. Confirmatory factor analysis is more powerful in terms of quantifying the fit of the hypothesized data structure with the data. Cross-validation was rarely performed. Presentation of the results and conclusions was often incomplete. Conclusions: Some of our results are specific for the SF-36, but the finding that both the application and the reporting of factor analysis leaves much room for improvement probably applies to other questionnaires as well. Optimal reporting and justification of methods is crucial for correct interpretation of the results and verification of the conclusions. Our list of criteria may be useful for journal editors, reviewers and researchers who have to assess publications in which factor analysis is applied.

Hitomi Hayashi, Michiko Kobayashi, Siebold University of Nagasaki, Nagayo-cho, Nagasaki, Japan; Yoshihiko Yamasaki, The University of Tokyo, Bunkyo-ku, Tokyo, Japan; Junko Sakano, Okayama Prefectural University, Soja, Okayama, Japan Aims: We examined the SOC and QOL in adolescence in order to evaluate the composing and improving factors besides school life, family circumstance and subjective well-being. SOC is the scale of ability for the adaptation to difficulty. SOC is not the hereditary ability but the ability of acquired immunity. These background indicate that childhood is important for the elevation of SOC. Methods: In order to evaluate the effect of SOC on QOL and subjective well-being, we compared the mean scores between the good SOC group (over average score) and poor SOC group (under average score). We also examined the QOLs on the categories of school life, family life, physical condition, self evaluation, sexual acceptation, sexual information, the other sex, sexual rule by the comparison between high and low SOC groups. Results: SOC scores were 47.01±5.64 (mean±SD), maximum 63, minimum 25, full marks 70, and QOL scores were 278.22±36.59 (mean±SD), maximum 356, minimum 184, full marks 369, respectively. There were high correlations between 5 scale of SOC and all QOL domains (p < 0.05). Compared with low SOC group, there were significantly excellent health degrees in domains of subjective wellbeing, QOL score and school life, family life, physical condition, and self evaluation in high SOC group. There were signifiant correlations between SOC and subjective wellbeing(p < 0.05), QOL score (p < 0.001), school life (p < 0.01), family life (p < 0.001), physical condition (p < 0.001), self evaluation (p < 0.001), respectively. Conclusions: In order to improve the adolescent QOL and SOC we need to make supports for these two factors, because these two factors have a mutual relation ship each other. However, it is said that SOC is established from childhood to young adult and makes no changes. Thereafter we would like to find the hint of care to elevate the both SOC and QOL in adolescence. Aims: The KDQOL-SF could be used to assess HRQOL across treatment modalities for renal failure(dialyis or kidney transplant-Tx). The instrument has only been validated in the HD population. Our aim was to determine the basic psychometric properties, reliability and validity of the KDQOL-SF in Tx population. Methods: A rigorous translation process was utilized to produce the Hungarian version of the tool followed by cognitive debriefing in a small convenience sample. Test-retest reliability of the final version was evaluated in a 63 HD patients in Budapest. Subsequently reliability and validity of the instrument was determined in dialysis patients (n = 418) and 418 Tx patients from Budapest.Patients completed the KDQOL-SF and the Center for Epidemiologic Studies-Depression (CES-D) scales. Results: The overall reliability pattern was similar to the original US pattern. In the Tx group internal consistency of all SF-36 domains was very good: Cronbach's a values were above 0.70 for all domains, except 3 disease specific sub-scales (quality of social interaction 0.52; work 0.65; cognitive function 0.68). We found significant, moderate to strong negative correlations between most of the KDQOL-SF domains and the CES-D scores. Older patients had worse scores along the dimensions related to physical health or functioning but less difference was seen along the mental health domains and some of the disease specific subscales. Impaired renal function was associated with significantly lower KDQOL-SF scores for all the generic and for most of the disease targeted dimensions. Conclusions: The KDQOL-SF is a reliable and valid tool to assess HRQOL in kidney transplant patients. Aims: Choosing the most appropriate measurement instrument depends on the measurement properties (e.g. internal consistency, reproducibility, validity, responsiveness, and interpretability) and on practical issues such as requirements, costs and burden to the subjects. Available standards for the evaluation of measurement properties often consist of general guidelines. But explicit criteria for proper evaluation methods that have been agreed by a substantial group of experts are mostly lacking. Without such criteria instrument selection is strongly hampered. The COSMIN study aims to develop consensusbased standards for the selection of health measurement instruments, including explicit criteria for good measurement properties. Methods: A Delphi-study in an international expert panel will be carried out using four written Delphi rounds and a face-to-face-meeting. As a basis for the Delphi study, we will identify all available evaluation criteria for measurements based on a systematic review of the literature. Results: Consensusbased criteria will be developed for which and how measurement properties should be assessed, and for what constitutes good measurement properties. The criteria will be field-tested in systematic reviews and disseminated through a website and publications. Conclusions: The challenge is to find all existing criteria, and to reach consensus in this complex field with little evidence. Many experts from different backgrounds and countries in the field of health status measurement will co-operate in this study, in order to guarantee a broad and sound foundation of the standards. This will enhance the credibility of the standards, their adoption and implementation. P-227/1333/QUALITY OF LIFE AND SPIRITUAL-RELI-GIOUS COPING RELATIONS Raquel G. Panzini, Denise R. Bandeira, Psychology, Federal University of Rio Grande do Sul, Porto Alegre, RS, Brazil Aims: Investigate the expected relationship between quality of life (QoL) and spiritual-religious coping (SRC) in a Southern state of Brazil by examining scores on WHOQOL-BREF and SRCOPE Scale, a Brazilian adaptation of RCOPE Scale. Methods: As part of the SRCOPE Scale validation, 616 participants (65% female), 13-82 years old (l = 41.38; SD = 18.44), were accessed at one of the following locations they attended: religious institutions or spiritual groups (74.4%), universities (13.5%), health care institutes (9.1%), and web mail (2.9%). Participants also completed an informal consent form and demographic, socioeconomic, religious and health questionnaire. Results: QoL and SRC total scores were positively and significantly correlated. Also, negative SRC (NSRC) was negatively correlated to QoL to a greater degree than the positive SRC (PSRC) was positively correlated with QoL. Using Student's t-test, those who had high SRC total scores showed higher overall levels of QoL and in all domains of WHOQOL-BREF, plus higher objective health classification. Furthermore, those who had higher levels of QoL demonstrated higher use of PSRC and lower use of NSRC. Conclusions: QoL and SRC are correlated constructs. A minimal 2PSRC:1NSRC proportion was proposed to obtain positive results of SRC on QoL. Evidences suggest that interventions focused on SRC processes, could be effective and beneficial to policy makers because of their potential to reduce interventions costs and their significant impact on population's health and QoL. Next steps are tests of proportion and causation (direction of correlation effect). Aims: Palliative care is defined by WHO as the type of care delivered to patients with an active and progressive disease and with short life expectancy. Care should be focused on symptoms relief, on prevention and mainly on quality of life. Methods: 104 patients with breast and lung cancers with life expectancy less than a year and aged between 40 and 85 years old were selected in a specialized oncology hospital. A Portuguese version of the Palliative Care Outcome Scale (POS) was administered to them along with other variables including pathology, treatment and socio-demographics. Results: Construct validity of POS evidenced five factors: emotional well-being, real fife consequences, information and support, anxiety and burden of disease. A moderate a strong correlation was shows when POS scores were compared to EORTC QLQ-C30 ones. Test-retest reliability and internal consistency were very good. Sensitivity after 1 month was also tested with good results. Comparing pathologies we observed that the burden on the patient is sensitive to his/her pathology. Conclusions: This study showed the good performance of the Portuguese version of the Palliative Care Outcome Scale. Aims: Governance can mean many different things to different people. However, several authors see good health governance as the enabled participation of those concerned in the formulation and deployment of policies, programmes, and practices leading to equitable health systems. Among the main elements of this concept we include people-centeredness, accountability, transparency, broad participation and consultation, and regular monitoring and evaluation. The purpose of this presentation is to evidence the usefulness of the Europep instrument as a tool to implement health governance in the primary care context. Methods: The set of the indicators included in the Europep core items measure five dimensions through 23 questions: patientdoctor interaction, provision of medical care, information and support to patients, continuity of care and cooperation, and organization of services. All questions are associated to 5-point Lickert scales from ''bad quality'' to ''excellent quality''. This instrument was implemented in a large study in Portugal focused on an age-gender representative sample of all the existing 358 health centers. The final sample size was 11,166 obtained after a proportional random sample of from the original population formed by 1,575,061 users. Results: We created a classification system for the health centres with five categories. Associated to each category, it was designed a quality improvement procedure and a set of recommendations to guarantee that any health centre user has the same change to be delighted with the service provided. Conclusions: Europep satisfaction questionnaire is a valid, feasible and useful tool to promote the empowerment of patients and to involve them in the ways health care is provided. It also seems to be a good instrument to continuously assess the quality of care and to prioritize actions for quality improvement. Aims: Glaucoma is one of the most important causes of low vision leading to blindness in Thailand. No disease-specific instrument on health-related quality of life (HRQOL) is available for Thai patients with glaucoma. This study developed and validated a HRQOL questionnaire applicable for such population. Methods: A multi-disciplinary research group (ophthalmologist, nurse, pharmacist and HRQOL researcher) was involved in development. Twenty-five patients were asked to indicate factors relevant to their HRQOL. Items were generated from the patients' list along with adaptation of the 25-item National Eye Institute Visual Function Questionnaire, using forward and backward translation. Final version of Thai Visual Function Questionnaire 28 (VFQ-28) consists of 28 items (11 domains, general eye dis-ease) plus 10 symptoms (glaucoma specific). Content validity was tested by the experts (n = 4) and patients (n = 20). The VFQ-28 was tested in 153 Thai glaucoma patients. Known-group validation was used to compare mean HRQOL scores between patients with a range of disease severity using Mann-Whitney Utest. Correlation of domains with global visual scores was analyzed using Spearman-rank correlation. Cronbach's a was used for internal consistency reliability. Results: Of 153 patients recruited, 89 were female (58.2%) with mean age 63 years (range 25-86) and highest education of primary school (73.9%). VFQ-28 Thai showed good internal consistency with alpha range of 0.69-0.90. Patients with normal visual function (N = 88) had significantly better HRQOL scores than those with visual impairment (N = 33) in all domains. Patients with mild to moderate defect of visual field (N = 36) also showed significantly better HRQOL scores than those with severe defect (N = 22) in 9 of 11 domains (p < 0.05). All domains showed moderate to high correlation with global visual score, except for role difficulty and driving. Conclusions: VFQ-28 Thai is shown to have good reliability and validity. The ongoing study with larger sample size and wider socioeconomic background should confirm construct validity. Aims: COPD is a debilitating condition with high morbidity, mortality and cost. Evidence is emerging on the association of anemia in COPD with increased breathlessness and reduced functional status. The purpose of this analysis was to quantify the impact of anemia on HRQL in subjects with COPD. Methods: A web-based survey was administered to a random sample of subjects with COPD. Seattle Obstructive Lung Disease Questionnaire (SOLQ) 1 was incorporated as a validated HRQL tool. Only subjects with COPD were eligible, with limits of $50 anemic (A+) and $850 non-anemic (A)) subjects for a total sample of $1000. Results: 2114 consecutive responders were screened to enroll the planned number of subjects. Of the 1006 eligible enrollees, 152 reported awareness of a history of anemia within the previous 6 months. Although the A+ group included a higher proportion of females than A) (65% vs. 56%, p = 0.046), there was no difference in age, body-mass index (BMI) or COPD severity between A+ and A) groups. A+ subjects were more likely to report poor/ frail health (43% vs. 24%, p < 0.001) and to have a history of renal disease (9% vs. 2%, p < 0.001), while the groups did not differ in the history of cardiovascular or oncologic diseases. The total SOLQ score was significantly lower in A+ subjects compared to A) (45.3±21.22 and 49.9±20.01, respectively, p = 0.01), indicating further impairment of HRQL in COPD subjects with anemia. Significantly lower scores in the A+ subjects were found on three (physical function, emotional function and coping skills) of the four domains (p < 0.05); the only domain to show no betweengroup difference was satisfaction with care. In addition, in a regression model controlling for age, gender, BMI and disease severity, anemia was independently associated with impaired HRQL as measured by the total SOLQ score. Conclusions: COPD patients reporting anemia are more likely to be female, be in poor/ frail health and have a greater likelihood of a history of renal disease. Importantly, anemia is independently associated with impaired HRQL as measured by SOLQ. Further work is needed to better understand the impact of anemia severity on HRQL in patients with COPD Changes were of small to medium effect size. By controlling for age, sex, and chemo-therapy in time-lagged regression analyses, higher finding of meaning predicted worse role and emotional functioning; higher positive reframing predicted healthier scores on role and cognitive functioning, fatigue and diarrhea; higher acceptance predicted better emotional and role functioning. Positive reframing was also predictive of the residualized changes of cognitive functioning and diarrhea, whilst accommodation was predictive of the residualized change of fatigue. Conclusions: Results suggest that emotion-focused/accommodative coping subjectively improved participants' HRQOL in several domains. It is a task for future research to investigate if emotion-focused/ accomodative coping also objectively ('truly') changes health by promoting for example health-related behavior. Considering pre-diagnosis health the average thentest scores were better than the observed average scores, while considering pre-treatment health the average then-test scores were worse than the observed scores. Conclusions: In hindsight patients considered their health before diagnosis as better, but their health 1 month post-diagnosis as worse than when asked at that time, suggesting that 'true' change between pre-and post-diagnosis is bigger than observed scores disclosed. These results indicate the presence of response shift and confirm earlier findings of others and of ourselves, e.g. in this same group of patients whose ratings of prostate cancer specific dysfunction changed after they had received their diagnosis. Aims: Response shift (RS) has been described as a change in the meaning of one's self-evaluation (X) of a target construct (T) as a result of a change in one's frame of reference (internal standards, values, and conceptualization of T). In an attempt to clarify and resolve differences in understanding of RS that confuse present research, we describe two perspectives on RS and propose two formal definitions. Methods: In a measurement perspective, RS can be defined as a violation of measurement invariance, f(X|T)=g(X|T,V), where V are possible violators of invariance. If this invariance does not hold, the measurement of T by X is biased by V, and test scores can not be validly compared. In a conceptual perspective, RS can be defined as a violation of explanatory invariance, f(T|C) = g(T|C,V), where C are generally acknowledged causes of T. In this perspective we do not distinguish between T and its measurement, yet if the explanatory invariance does not hold, the explanation of T by C is biased by V. Results: In HRQL research, the RS definitions can be used with, e.g., T = ''true'' HRQL, X = HRQL test scores, and C = health state change.

V may be measurement occasion, group membership, frame of reference, or any other patient or environmental characteristic of interest. E.g., instead of trying to operationalize V as a patient's change in frame of reference, it may be sufficient to operationalize possible causes of a change in frame of reference (e.g., adaptation, coping, social comparison). After operationalizing the variables that feature in the preferred RS definition, RS can be detected by testing the associated invariance. Aims: There is increasing interest in the use of ordinal methods for eliciting utilities as alternatives to conventional methods such as standard gamble and time trade-off. The aim of the study was to compare ranking, time trade off (TTO) and a discrete choice experiment (DCE in the estimation of a preference-based index for a condition specific quality of life instrument (the sexual quality of life questionnaire. Methods: A representative sample of the UK general population (n = 200) were interviewed and asked to rank a series of health states derived from the condition instrument followed by TTO tasks to elicit utilities for each state. Four weeks following the interview, consenting individuals received a postal self-completion questionnaire containing a DCE. Respondents were asked to make a series of pair-wise comparisons of a selection of states. TTO data was analysed by random effects and OLS models, rank data by rank ordered logit (entering dead as a dummy) and DCE data by random effects probit. Results: In total, there were 207 respondents to the TTO interview survey and 102 of these returned a follow-up DCE. Mean TTO health state values ranged from 0.643 to 0.966 (where full health is 1.00 and zero for states equivalent to being dead). The RE and mean TTO models produced significant decrements that were broadly consistent with the health state classification, with mean absolute errors (MAE)of 0.037 and 0.022, respectively. The rescaled rank and DCE models also produced significant and consistent decrements but with worse MAEs of 0.08 and 0.077, respectively. Conclusions: The rank and DCE models were not able to predict mean TTO health state values as well as the TTO models. This result contrasts with findings from applying the rank method to generic preference-based measures (Solomon, 2004; McCabe et al. 2005 Mean HUI3 overall scores were compared for the two countries. Linear regressions using normalized weights were conducted. Overall HUI3 scores were regressed on a set of potential confounders with a country dummy variable. Results: Canadians (mean HUI3 = 0.88, median = 0.97, SD = 0.20, min = )0.24, max = 1.00) are slightly healthier than Americans (mean HUI3 = 0.87, median = 0.97, sd = 0.21, min = )0.26, max = 1.00). Income is associated with health in both countries. Canadians in the lowest income level (mean HUI3 = 0.81, median = 0.91, SD = 0.26, min = )0.24, max = 1.00) are healthier than Americans(mean HUI3=0.77, median=0.91, SD=0.29, min=)0.26, max=1.00). Across education levels Canadians(less than high school HUI3=0.81, high school=0.89, vocational college=0.89) are also healthier than Americans(less than high school HUI3=0.74, high school=0.86, vocational college=0.87). Canadians in the survey are slightly younger with more males (mean age 45.11 and 49% males) compared to the U.S. (mean age 45.35 and 48% males). Results from regressions are similar. Differences in health are explained by age, gender, education, health insurance, marital status, and country of residence. HUI3 scores are lower for respondents who are single, older, less educated, and/ or have no insurance. Conclusions: On average, health status in the two countries is approximately equal. The HUI3 is a useful measure of health status. It is able to identify and distinguish differences in general health status among groups and between countries. The regression models are useful in explaining country differences in health status. Aims: To assess measurement properties of HUI Mark 2 (HUI2) and Mark 3 (HUI3) for health-related quality of life (HRQL) and cost-effectiveness studies of breast reduction mammoplasty. Methods: Consecutive patients from a surgeon's practice were enrolled. The longitudinal study collected patient reported outcome (PRO) measures at one week pre-surgery (T1), one day pre-surgery (T2), and six months post surgery (T3). PROs were HUI2, HUI3, SF-36, Multidimensional Body Self Relations Questionnaire Appearance Assessment (MBSRQ-AS) and Breast Related Symptoms Questionnaire (BRSQ). Reliability was assessed from T1 to T2 by intra-class correlation coefficient (ICC), responsiveness from T2 to T3 by Guyatt's responsiveness statistic (GRS) and t-test, and construct validity by Pearson's correlation coefficient (r) between PROs of T3-T2 change scores. Differences in mean HRQL scores of >0.03 are important, ICC>0.80 represents good agreement and r>0.40 represents moderate or better association. Results: 52 women were eligible. Three patients were not enrolled because of language problems or refusal. Patients had a mean age of 38 years and mean resection weight of 790 g per breast. HUI2 and HUI3 scores of HRQL had test-retest ICCs of 0.86 and 0.84, GRSs of 0.89 and 1.09, and increases in means of 0.07 (p<0.05) and 0.13 (p<0.01). As predicted, moderate or better associations were observed for HUI2 emotion with SF-36 mental (r = 0.489, p = 0.003) and MBSRQ-AS (r = 0.618, p<0.001), for HUI3 emotion with SF-36 mental (r = 0.501, p = 0.002) and for HUI3 pain with SF-36 physical (r = 0.412, p = 0.013). Conclusions: HUI2 and HUI3 measures of HRQL have good reliability, detect clinically important size improvements, and have construct validity. HUI should be used to assess HRQL and evaluate costeffectiveness of breast reduction mammoplasty, including such studies of obese patients not currently insured for surgery in some jurisdictions. Aims: Our study focused on the measurement bias of the visual analogue scale (VAS) in the context of the quantification of health states. Three prominent biases that are expected to be associated with the VAS were studied: anchor-point bias, context bias and end-aversion bias. These biases may distort the assumed interval measurement property of VAS values, which is a requirement for valid application (e.g., QALYs). Methods: Respondents placed 18 cards with descriptions of EQ-5D health states and a card with 'death' on a 20 cm EuroQol VAS and adapted versions of it, in such a way that the intervals between the positions of the health states corresponded with the differences in strength of preference perceived by the respondent. Anchor-point bias was examined by applying the standard EuroQol VAS (n = 212) and an adapted version (n = 97) with a different lower anchor: 'death' instead of 'worst imaginable health state'. Context bias was examined on the basis of another group of respondents (n = 112) who valued three different sets of EQ-5D health states. End-aversion bias was studied by yet another group of respondents (n = 100) who placed the same set of states as that in the first experiment on the standard EuroQol VAS and on an ('true') VAS without any anchors, categories (the numbers 10, 20, ..., 90) or markers. To test for statistically significant effects in or between the experimental groups, multivariate (MANOVA) and t-tests were carried out. Results: Anchor-point nor end-aversion bias was detected. However, the values were undoubtedly affected by the context of the set of health states, showing changes up to 20% in VAS values for identical states. Conclusions: We recommend a broad and balanced set of states when the VAS is used to quantify health states. Aims: For patients receiving treatment for prostate cancer, the predominant (disease specific) symptoms in these patients are erectile dysfunctions and urinary problems. These symptoms may impair the patients' overall QoL but especially their sexual life and their spouses' sexual life as well. The aim of this study was to compare QoL in prostate cancer patients to their spouses' QoL using the EQ5D. Methods: We mailed the questionnaires to 33 patients that underwent prostatectomy 6-12 months before. A short version of the same set of questionnaires was sent to the spouses. The questionnaires used for the patients were the EORTC QLQ-C30, a disease specific module, a questionnaire on sexuality and partnership, and the EQ5D (EU-ROQOL). The spouses' questionnaires consisted of EQ5D and the questionnaire on sexuality and partnership. The response rate was 71.9% for the patients and 54.8% for the spouses. Altogether, QoL data from 16 complete pairs were included in the statistical analysis. Results: Mean age was 64.9 (SD 5.3) years for the patients and 62.7 (SD 6.0) years for the spouses. Eight of the patients included had tumor stage T2, the other 8 had T3. No statistical significant difference between patients and spouses was found for QoL (EQ5D). The spouses rated their sexuality and partnership for all items more positive than their partners. Statistical significant differences were found for satisfaction with sexuality. Conclusions: The EQ5D -which is one of the standard instruments, used for health economic evaluationswas not able to discriminate between prostate cancer patients and their (healthy) spouses. A possible explanation for this is the fact, that the domains that are highly impaired in patients with prostate cancer (erectile dysfunction), are not covered by the EQ5D. From the author's point of view the usefulness of the EQ5D in health economics especially resource allocations is questionable. Since the design of this current study was a historical cohort study, these findings should be re-analysed in a prospective sample with a bigger sample size, which is on the way. Aims: The management of insulin therapy and diet for patients with type I diabetes can be very complex, and so many patients struggle to keep their blood sugars in the correct range. Under-standing the preferences of patients regarding insulin therapy may well increase patient adherence to therapy, decrease complications from disease and improve quality of life for patients. This study aimed to collect qualitative data from patients regarding their preferences for different aspects of insulin therapy. Focus groups were also designed to inform the development of quantitative patient preference and willingness to pay study. Methods: Patients with type 1 diabetes were recruited through advertisements and all completed a telephone screener designed to identify type 1 patients. Two focus groups were organised (n = 5 and n = 6). A discussion guide was developed based upon literature review and discussion amongst study authors and clinical experts. Exploratory discussion focussed on whether participants could identify a monetary value for improvements in insulin therapy. The sessions were taped and analysed for content and themes. Results: The main themes from the discussion centred on the impact of diabetes on patient's lifestyle, their fear of hypoglycaemic events and ability to maintain blood sugars within their target range, views regarding injection devices, and worries regarding side effects. In discussing the value of a hypothetical, superior insulin therapy (compared to their current therapy) participants said that they would be willing to pay a range of different amounts from UK£ 20-30 per month to UK£ 5 per day, including several participants who were not willing to pay anything at all. Conclusions: A wide range of views regarding the impact of diabetes, and its management emerged from the focus groups. This information is powerful in its own right, but will also be very informative for the development of the quantitative patient preference survey. Aims: The consequences of inactivity on the health status of elderly individuals with chronic conditions are not well documented. We hypothesize that the consequences of physical inactivity are particularly detrimental. Methods: The CCHS is a multi-cycle, cross-sectional, representative health-related survey of the Canadian population.This survey contains information pertaining to chronic conditions, health related quality of life (as assessed by the HUI3), health resource utilization,sociodemographics, and physical activity (kcal/week). We utilized responses from respondents 765 years old from cycle 1.1 (September 2000) . The relationship between the HUI3 (dependent variable) and physical activity and chronic diseases (independent variables) was assessed using multiple linear regression. Other variables in the model included age, sex, alcohol use, smoking, marital status, income and body mass index (BMI). Two-way interactions between specific chronic diseases and physical activity were explored. Results: Over 24,000 respondents were 7 65 years old, 60% were female, 76% had 71 chronic condition, the mean HUI3 score was 0.78 (SD 0.26), and the average physical activity was 711 kcal/week (SD 877). In an unadjusted model (adj. R2 = 0.19), the HUI3 was significantly associated with both the number of chronic diseases (ß=)0.070; 95% CI )0.073 to )0.068), physical activity per 100 kcal/week (ß=0.0051; 95% CI )0.0047 to )0.0054). In the fully adjusted model, there was minimal change in the beta estimates and explained variance. Conclusions: On average, older Canadians are not achieving target physical activity levels. Chronic diseases and physical activity appear to influence health related quality of life in opposite directions. However, physical activity appears to have minimal influence on HRQL. Aims: Medical cost-effectiveness analyses (CEA) traditionally use utility values that average across the whole population of persons who might be appropriate for the treatment. However, most treatments are used only when the patient accepts them, suggesting that the preferences of patients who would choose a treatment if it were offered are the relevant ones for CEA. This study uses the example of intensive therapy for diabetes to study the effect of accounting for patient choice in incorporating patient preferences into CEA and to examine the implications of this perspective for preference elicitation. Methods: We interviewed 394 adults with type 2 diabetes to determine their utility values for health states relating to common diabetic complications and treatments and their choice of therapy. We then used a simulation model developed previously by the CDC to assess the cost-effectiveness of intensive therapy for the population as a whole and for the set of patients who chose intensive therapy. Results: The mean age of subjects was 62 years. 58% were women. 43% were African-American, 30% White, and 27% Latino. 31% of patients reported using intensive therapy. The mean utility for intensive glucose control was 0.67 (SD 0.33) and the mean value for conventional control was 0.75 (0.30). Utilities for complications were similar to prior estimates, and were similar between patients on intensive and conventional therapies. However, patients on intensive therapy had higher utilities for intensive therapy than did patients on conventional therapy. When patient utilities were used in the costeffectiveness model, intensive therapy for diabetes was harmful on average for the full population of patients with diabetes (mean delta costs $8007, mean delta QALYs )0.35) but beneficial and cost-effective for the patients who select it (mean delta costs $7777, mean delta QALYs 0.18, ICER $43 K/QALY). Conclusions: Costeffectiveness results for diabetes care are dramatically altered by accounting for preference heterogeneity and patient selection of treatment choice. Utility elicitation for cost-effectiveness studies should also collect data on treatment choices to inform analyses that reflect the effects of self-selection. Aims: It would be useful to be able to compare the costs and health value associated with alternative quality improvement targets. Health value in this context refers to the effects of an intervention on the quantity and quality of life. The RAND-UCLA expert panel method is an efficient way to obtain ratings of care for a broad range of health interventions and types of patients. This study employs expert panel methodology to elicit quantitative ratings of the utility of usual patterns of treatment for two medical conditions. Methods: The expert panel approach consisted of development of clinical profiles and derivation of estimates of treatment benefit via elicitation of utility ratings before and after treatment. Clinical profiles specified all information regarding characteristics of patient groups, treatments to be rated, and their interactions, that might affect utility estimates. A panel composed of 13 asthma and depression experts used a web interface to make a series of utility ratings (before any new treatment, 1 or 3 months later with no treatment, 1 or 3 months after initiating various common treatments) for adult patient groups with depression or asthma. The panel convened to discuss discrepancies, and subsequently made final ratings. Final ratings were examined for face validity and variability, and t-tests and ANOVAs on the pre-and postmeeting ratings provided information about the impact of the expert panel meeting. Results: The final ratings had face validity and minimal variability, indicating considerable consensus among experts. Comparisons of the pre-and post-meeting treatment utility ratings indicated substantial convergence as a result of the expert panel, and evidence exists for the assumption of scale equivalence across the two health conditions. Conclusions: Valid and comparable ratings of treatment utility for distinct health conditions can be obtained from experts using the expert panel methodology. Aims: Parkinson's disease (PD) is characterized not only by its motor impairments, but also by cognitive, mood, sleep, and autonomic dysfunction. Many of these features are related to feelings of control. This hampers the customary assessment of the impact of mastery and self-efficacy on quality of life. Aim of our study was to assess mastery (MA) and self-efficacy (SE) in PD patients, to assess the association of these with disease severity and with patients' valuation of their quality of life. Methods: Three hundred and seventy-six PD patients filled out an extensive questionnaire, including MA (6 items, Pearlin & Schooler 1978) and SE (2 items, adapted from Sullivan et al. 1998) . Patients were examined to assess their Hoehn and Yahr stage and interviewed to obtain a valuation of quality of life (Time TradeOff, TTO). Psychometrics were assessed for MA and SE (factor structure, alpha). Associations between TTO and MA/SE were assessed using correlations and ANOVA (controlling for stage). Results: Based on factor and reliability analysis, we grouped 4 MA items pertaining to the present (MA-present), and 2 MA plus 2 SE items, all pertaining to expectations of the future (MA-future). Stage was negatively correlated with MA-present ()0.36), less strongly with MAfuture ()0.24), supporting a more physical interpretation of control. An interaction was seen between stage and MA-present: patients with low mastery showed a strong decrease in TTO-scores by increasing stage, patients with high mastery showed higher TTO-scores that did not decrease with increasing stage. This supports a more psychological interpretation of control. Conclusions: Mastery is a complex construct in PD patients, with both psychological and physical connotations. More in-depth studies are needed to explain the associations found, and to disentangle physical from psychological aspects. intervention group-completing QOL questionnaires with feedback to clinicians, attention-control group-completing QOL questionnaires without feedback and control group-no QOL data. Content analysis was applied to three consecutive encounters, coding mentioning emotional issues and any actions taken in response. Data was summarised over three encounters per patient and analysed using descriptive statistics and regression, controlling for patient gender and age. Results: 215 patients (900 encounters) were analysed. Emotional issues were discussed more consistently when QOL information was given to oncologists (p = 0.019). Emotional issues were more often mentioned with female patients (p = 0.020). Approximately 50% of the discussions were initiated by physicians regardless of study group. Actions were recorded in 25% of intervention encounters vs. 15% control arm encounters. The difference was due to more frequent discussion of psychological distress. Referral rate to psychology services was low (12%) with no difference between the groups. In 6% of intervention encounters referral was offered but not taken up by the patient. Drugs were prescribed in 4% of intervention patients vs. 5% of controls. Conclusions: Informing oncologists of patient QOL leads to more consistent discussion of psychological problems and a trend to offering referrals to psychosocial services. No increase in referrals or use of psychotropic drugs was observed. Regular discussion of psychological distress by oncologists seems to result in a positive impact on patient emotional wellbeing. Aims: Extremity soft tissue sarcoma (STS) can often result in severe functional disabilities. Fatigue and sleep disturbance, common complaints of people treated for cancer, can affect these patients' quality of life. The purpose of this study was to determine the impact sleep disturbances and fatigue had on functional disabilities (impairments, activity limitations, and participation restrictions) and overall well-being (OWB) in extremity STS patients one year post treatment. Methods: We evaluated 100 extremity STS patients for impairment (MSTS, 1987) , activity limitations (TESS), participation restrictions (RNL), OWB (EQ5D-VAS), sleep disturbances (SCL-90) and fatigue (FACIT). All patients were over the age of sixteen and non-metastatic at diagnosis and follow-up. Pearson correlations, linear regressions, and stepwise multiple linear regressions adjusting for clinical and demographic variables were used. Results: Sleep disturbances explained 19% (p < 0.01) of the variation in participation restrictions and 28% (p < 0.01) of the variation in OWB. Sleep disturbances did not have a significant impact on impairment or activity limitations. Fatigue explained between 45% and 61% (p<0.01 for each) of the variation in functional disability and OWB, having the strongest impacts on participation restrictions and OWB. Conclusions: Fatigue had a stronger impact on the functional outcomes and OWB than sleep disturbances. Interventions that target fatigue may improve patient outcomes. Aims: The aims of this study were to compare complementary therapy users to non-users, to compare those who stopped using to continuous users, to describe the frequency and patterns of use, and the relationship of use to quality of life in persons with brain tumors over time. Methods: A secondary, data analysis of data collected for the Glioma Outcomes Project Database was performed. Enrollment included 787 adult patients with high-grade, primary brain tumors, undergoing surgery and treatment. Data was explored in 186 high-grade glioma participants with complete data at three study points.

Frequencies, percentages and chi-square statistics were used to examine data comparing the demographic, disease and treatment related factors, and QOL characteristics of users and nonusers of CAPPs, those who stop use and those who continuously used CAPPs over the illness course. T-tests, ANOVA and MANCOVA were calculated to explore the differences in age and QOL, between users and non-users of CAPPs, continuous users, stoppers, and never users of CAPPs. Results: Of the 186 participants with complete data on initial study enrollment, 32.2% had used at least one form of CAM. At 6 months, 40.9% used at least one form of CAM. Patterns of continuous use, and stopping and starting use emerged over the study course. CAM users were more likely to be younger (mean age of 49.6 vs. 53.41 years) and to have undergone chemotherapy than non-users at the time of study enrollment. CAPPs users at all three measurement points rated their QOL higher, although not significantly higher, than those who were non-users. Conclusions: This study, is important for population health because it provides new information relevant to improvement of care for persons with brain tumors. Innovative interventions that ameliorate undesirable symptoms in brain tumor patients and improve functional status and quality of life are needed for persons with brain tumors, particularly high-grade glioma. Aims: Many children with cancer experience HRQL-related problems, especially when treatment ends. However, these problems are not always known to and treated by their physician. Incorporating standardized HRQL assessments as a routine part of the outpatient care is a method to improve communication regarding HRQL. However, to be effective we believe this should be carefully guided. Methods: In two outpatient pediatric oncology clinics a study is conducted which evaluates the efficacy of incorporating standardized HRQL assessments as a routine part of the outpatient care of children with cancer in terms of three primary outcomes: (1) tailoring child management with regard to HRQL problems; (2) increasing child-parent satisfaction with care and; (3) increasing quality of life. To optimize the effectiveness of giving feedback of the HRQL profile to the pediatric oncologist, a new implementation method is developed. Results: AIMS of the new method are: (1) to educate the oncologist in interpreting the HRQoL profile; (2) to inform the oncologist about the actual map of psychosocial and educational institutions and criteria for referral and (3) to practice supportive talks with the child and the parents. In the training-sessions, the oncologists learn the meaning of high and low scores on different HRQoL domains. They are also trained in asking specific additional questions. Criteria for referral to other departments (e.g. psychology) are formulated. Pediatric oncologists are provided with an information booklet. Conclusions: A step has been made to incorparate HRQoL assessment in clinical practice. To optimize the effectiveness a bridge should be built between patients and physicians by providing guidance in interpreting and discussing these data. Aims: We compared diabetic men with non-diabetic men with prostate cancer to understand whether the co-occurrence of both conditions imposed a greater burden on health-related quality of life (HRQOL), adjusting for obesity. Methods: Data were abstracted from CaPSURE, a disease registry of 10,581 men with prostate cancer. 1612 men were included who were treated between 1989 and 2002, had BMI information available, and had completed both a pre-treatment and at least one posttreatment HRQOL questionnaire. A repeated measures model adjusted for baseline clinical and demographic variables was used to evaluate group differences. Results: Men were divided into three groups based on diabetes history, reporting of diabetes medication, and BMI: (1) diabetic and overweight/obese (N=155), (2) non-diabetic overweight/obese (N=1067), and (3) non-diabetic normal weight (N=390). Diabetic/overweight men had poorer sexual functioning at pre-treatment and over time than the either of the two non-diabetic groups. Diabetic men also reported greater sexual bother and worse bowel functioning over time. No group differences were found in urinary functioning or bother or bowel bother. Conclusions: While previous studies among men with prostate cancer have found differences in HRQOL by obesity level, our current results indicate that the combined presence or absence of diabetes mellitus and being overweight or obese may have a greater impact on HRQOL than obesity alone. Aims: The aim of the present study was to identify variables related to quality of life in Korean patients with cancer. From a literature review, associated variables were derived: host variables (age and gender), disease/treatment variables (stage, treatment status, symptom distress), psychological variable (mood disturbance), and performance status. Methods: Participants were recruited from a cancer center. The sample consisted of 364 patients with the most common types of cancer (stomach, lung, liver, colon, breast, and cervix) in Korea. Patients, who agreed to participate in this study, were asked to completed questionnaires, measuring symptom distress (SDS), mood disturbance (LASAS), ECOG performance status, and Korean cancer-specific quality of life (QOL-C). The obtained data were analyzed using descriptive statistics and a multiple regression. Results: The host, disease/treatment, psychological and performance variables accounted for 46% of the variance in quality of life (R = 0.68, F = 43.55, p < 0.001). All variables, except gender, significantly contributed into the explanation of the quality of life. Conclusions: To promote quality of life in patients with cancer, it needs an integrate care program reflecting host, disease/treatment, psychological and performance status related variables. Aims: Women at increased hereditary risk of ovarian cancer may opt for gynecological screening (GS) or prophylactic oophorectomy (PO) to manage their cancer risk. Hormone replacement therapy (HRT) is often prescribed to compensate for post-surgical endocrine deficiencies, and therefore to prevent premature menopause. The purpose of this study was to determine the impact of HRT use on the levels of menopausal symptoms, sexual functioning and QL in younger women undergoing PO. Comparisons have been made with similar women undergoing GS. Methods: Questionnaire data (FACT-ES, SAQ, SF-36) were obtained from 450 premenopausal, high-risk women (response 79%) who participated in this nationwide, cross-sectional, observational study. 36% of women had undergone PO and 64% had opted for GS. Results: Following PO, 47% of the women had used HRT. Women who had undergone PO and had used HRT reported significantly fewer hot flushes and night sweats than did non-users of HRT (all p values <0.05). However, compared to premenopausal women undergoing GS, oophorectomized users of HRT were more likely to suffer from vasomotor and sexual symptoms (all p values <0.01). Majority of women reported being sexually active. Users and non-users of HRT reported comparable levels of sexual functioning. Compared to premenopausal women in the GS group, women who had used HRT after PO reported significantly more sexual discomfort (p < 0.01). No statistically significant between-group differences were observed in SF-36 scale scores, with all groups scoring similarly to the general population. Conclusions: Although PO has no measurable negative impact on generic QL, it results in a number of endocrine symptoms and sexual problems due to the premature onset of menopause. HRT appears to have only limited beneficial effect on PO-induced symptoms. Aims: The goal of palliative care is to achieve the best quality of life (QOL) not only for patients but also for their families. The study was conducted to investigate the HRQOL of families of terminal cancer patients and families who lost their loved ones to cancer. Methods: A cross-sectional survey was carried out using the Japanese version of SF-36v2, and a questionnaire designed to assess the level of satisfaction towards palliative care. Questionnaires were sent out by post to 125 families of patients who had once been admitted to the department of internal medicine of the university hospital to receive palliative care and who had been dead for more than 10 months. Seventy-seven families provided valid responses. Responses were also obtained from 12 families of patients currently receiving palliative care at the university hospital, who completed the questionnaires within the hospital ward. A total of 89 respondents were included in the analysis. Results: Respondents had lower levels of QOL compared to the average Japanese citizens. The QOL scores were examined using one-way analysis of variance between five groups. The mental health summary scores differed across groups. A statistically significant difference (p<0.05) was observed in social functioning scores and mental health scores between the families of patients currently receiving palliative care and the families of patients who had been dead for more than one year. Conclusions: The families of patients receiving palliative care and the families of patients who have been dead for less than a year have lower levels of QOL and greater needs for appropriate care and support compared to the families of patients who have been dead for more than a year. Aims: There are a few reports of HRQOL in relation to type of surgery in breast cancer patinets who are disease free for a long time. This study was performed to clarify it. Methods: Between July 04 and March 05, 100 women who had received treatment for breast cancer in a single institution and had been alive without recurrence and serious co-morbidity for more than 5 years (median 84 months) were asked to answer the several patient-administered instruments to assess HRQOL (FACT-B), psychological distress (HADS) and disease/treatment-related symptoms. Ninety-three answered them. The demographic characteristics at the study and medical ones were used as independent variables. T-test, analysis of variance, and multiple regression analysis were used for statistical analyses. Results: The univariate analyses revealed that an older age at the study was associated with worse scores of FACT-G (p = 0.012), FACT-B (p=0.029), and social subscale of FACT (p=0.001), and a higher educational status was associated with a better score of social subscale of FACT (p<0.001). There was no difference between breast-conserving therapy (BCT) and mastectomy (M). On multivariate analyses including all variables examined on univariate analyses, BCT was significantly better in the score of breast cancer subscale of FACT-B than M (p=0.045). Additionally, an older age was associated independently with a worse score of the social subscale of FACT-B (p<0.001). Conclusions: After adjustment for clinical and social factors, BCT was statistically significantly better than M in breast cancer survivors with regard to breast cancer subscale of FACT-B. Aims: Patients undergoing limb preservation surgery for soft tissue sarcoma (STS) may be left with significant functional disability and disfigurement, potentially affecting their quality of life (QOL) and satisfaction with treatment. This study hypothesizes that these patients' satisfaction with treatment outcome is linked to the way in which they experience their bodies, specifically, their state of embodiment. In this study, embodiment is defined as a feeling of either unity or disunity between self and the affected body part. Methods: One hundred and twenty seven patients undergoing limb preservation surgery for STS were followed prospectively. Patients included were over the age of 16 at diagnosis and were non-metastatic at diagnosis and follow-up. Variables completed one-year post surgery were embodiment state, satisfaction with treatment, as well as the following QOL components: impairment (MSTS, 1987) , activity limitations (TESS), and participation restrictions (RNL). Pearson correlations, linear regression and stepwise multiple linear regression were used for statistical analysis. Results: Independently, embodiment explained 49% (p<0.01) of the variation in treatment satisfaction. After adjusting for impairment, activity limitations and participation restrictions individually, embodiment explained between 56% and 70% (p<0.01 for each) of the variation in treatment satisfaction. Using a stepwise procedure to adjust from the demographic, clinical, impairment, activity limitation and participation restriction variables, embodiment explained 79% (p=0.03) of the variation in treatment satisfaction. Conclusions: Embodiment is a significant factor in patients' treatment satisfaction. Patients who described feeling a unity with the affected body part tended to be more satisfied with treatment even after accounting for their demographic, clinical, and QOL variables. Aims: Breast cancer is a main health problem in Uruguay in spite of prevention efforts. In the framework of a longitudinal study, the purpose of present work is to assess the influence of coping strategies and emotional state over health related quality of life in women who underwent breast surgery. Methods: Women who consulted in the Mastology Unit (University Hospital),during 2004, with positive diagnosis of cancer and who have undergone breast cancer surgery, were included. Prior written informed consent was required. Five percent refused to participate. Patients were interviewed and a battery of scales was applied 1 week after the surgery. Battery included Hospital Anxiety and Depression Scale (HADS), POMS (Profile of mood state), Mental Adjustment to Cancer (MAC) and SF-36. Nonparametric correlations were used to analyse the data. Sample included 91 women, mean age 60±12. Results: SF 36 subscales showed no significant differences except for mental health (MH). MH was below 55 in 29% of the patients. These patients also showed significantly higher scores (p<0.05) using Mann-Whitney's test in anger-hostility, depression-dejection, tension-anxiety and lower fatalism scores. Fatalism, scoring high in women with better MH, would be representing a mechanism of realistic acceptance (''At the moment I take one day at a time'', ''Since my cancer diagnosis I now realize how precious life is and I'm making the most of it''), opposed to a higher degree of anxious preoccupation as coping mechanism in the other group (''I am apprehensive'', ''It is a devastating feeling'', ''I feel very angry about what has happened to me'') . Conclusions: These results give light into underlying determinants of HRQL results, supporting the importance of quality of life-enhancing psychosocial interventions centred on emotional regulation and coping styles. Future longitudinal data will confirm or refuse this hypothesis. With that variable, we conducted analysis of variance (ANOVA) for well-being and analysis of covariance (ANCOVA) for the self-rated health with age as a covariate. Results: In well-being, the significant difference was shown by ANOVA. 'I am a family caregiver now' was significant lower value compared to other two groups. According those means, ex-caregivers of the family could recovery the well-being. No significant difference could be seen in the self-rated health (F=0.430, p=0.650). The results showed that ex-caregivers of the family could recovery the well-being, however, might not recovery their subjective health. We need to conduct cohort studies to improve the evidence level. Conclusions: We tried to find whether excaregivers could recovery their well-being and health. We conducted cross-sectional analyses, using database of JGSS. We found that ex-caregivers could recovery the well-being, but some ex-caregivers could not recovery their health. We need to conduct cohort studies to improve the evidence level. Aims: To describe the sleep quality (SQ) and quality of life (QOL) of chronic pain patients. Methods: The sample consisted of 21 patients from a tertiary care outpatient pain center in Sa˜o Paulo. The WHOQOL-bref measured QOL. It is composed by 26 questions in 4 domains: physical(PYD), psychological (PSD), social relationship (SRD) and environment (ED). Its score range is 0-100, being 100 the best. The Pittsburgh Sleep Quality Index (PSQI) measured SQ, and higher scores indicate worse SQ. Results: The mean age was 56.52 years (range 39-83), the majority (81%) were male, 47.6% were married, 66.7% (14) had upto 8 years of education, 52.4% (11) were Catholics, 38.1 (8) Protestants. The sleep latency (SL) was 55.9 min(SD=30), its duration was 5.24 hours (SD=1.7), and efficiency (SE) 70.0% (SD=18.8). The patients with SL more than 60 min had lower means in the PYD (43.4, p<0.05) and SRD (51.4, p<0.05) than patients with SL equal or lower than 15 min (57.9 and 71.7, respectively). Those who referred very good SQ had higher mean (91.6) in the SRD than those with very bad (61.1, p<0.05). PSQI score was correlated with PSD (r=)0.46, p<0.05), and general QOL (r=)0.49, p<0.05). The mean SE was higher (82.1) for those who referred good QOL than for those with very bad (35.5, p<0.05). The sleep efficiency was correlated with general QOL (r=0.6, p<0.05); the SL with PSD (r=)0.5, p<0.05) and PYD (r=)0.5, p<0.05). The patients who referred that the pain severely impeded them to execute their daily activities had higher mean sleep efficiency (87.5) than that for those who referred nothing (67.2). The latency time was higher (62.5 min) for those who frequently had bad feelings (depression, bad humor and anxiety) than who had only some times (53.4). Conclusions: QOL might be modified by sleep disturbance in chronic pain patients. If SQ could be improved, the QOL will improve too, especially in the PYD, SRD and PSD. Aims: The purpose of this study is to support the elderly living in mountainous regions showing progression of depopulation. Methods: In October 2003, we performed a self-described questionnaire survey by mail in 2887 aged. The assessment tool described by Okamoto et al. was used. Eight items of healthy longevity were evaluated (health promotion, physical condition, activities of daily living, social exchange, environment, family, feelings, and use of services). Higher scores indicated higher degrees of healthy longevity. Concerning QOL, the subjects expressed the degree of satisfaction in terms of scores from 0 (very unsatisfied) to 10 (very satisfied). The mean score in the degree of healthy longevity was compared between 0 points and 10 points groups. In addition, multiple regression analysis was performed using QOL as the criterion variable and the healthy longevity items as explanatory variables. The statistical software was Excel 2003. Results: The questionnaire recovery rate was 94.1%. Responders were 1183 males and 1524 females, which accounted for 87.4% of the population aged >65 years in this areas. The QOL score was 0 point in 118 subjects (4.4%) and 10 points in 949 (35.4%). Multiple regression analysis showed a multiple correlation coefficient r of 0.436 and a coefficient of determination with corrected degrees of freedom of 0.188. Seven of the 8 items were variables significantly associated with QOL. The regression coefficient was particularly high for ''environment'' and ''use of services''. Conclusions: The QOL in the elderly in mountainous areas, which was evaluated in terms of the degree of satisfaction, was high, suggesting the influences of physical residential environments and psychological environments with life support on QOL. P-269/1496/SUBJECTIVE QUALITY OF LIFE AND HEALTH -CROATIAN NATIONAL HEALTH SURVEY Gorka Vuletic, Statistics, School of Public Health, University of Zagreb, Zagreb, Croatia

Aims: This paper presents results from the 2003 Croatian national health survey, the normative data for subjective quality of life (SQOL) and self-perceived general health. Methods: Survey comprised the representative sample of Croatian population (n=9068)based on Croatian census 2001. Survey targeted persons aged 18 years or older living in private dwellings in the Republic of Croatia. In order to estimates be representative for the Croatian population, not just the sample itself, a survey weight is given to each person from sample. This weight corresponds to the number of persons represented by the respondent for the entire population. Results presented were based on weighted sample values, and represent 84% (n=3,478,892) of Croatian population. SQoL was measured with the question ''How do you satisfied with your life as a whole?'' on 11-point Likert scale. Scores were transformed in standardized form: percentage of scale maximum (%SM). General health was assessed on 5-point scale from excellent to poor (when value 1 is excellent and 5 is poor health). Results: Croatian average satisfaction with life was 60.11±25.12%SM. Significant decline in SQoL as health become worse. People who report excellent health (10.6% participants) report SQoL 74.50±22.57%SM; those who report health good (29.2%) have average SQoL 60.73±23.21%SM, and those with pore health (12.5%) also report very low SQoL (39.25±25.71%SM). Further analysis was made according to self-reported socio-economic status. Results show strong relationship between socioeconomic status, general health perception and subjective quality of life. Furthermore, statistically significant gender difference was found on a population level. Men report lower SQoL than women (59.86±24.92%SM and 60.32±25.30%SM, respectively). At the same time, men report to have better health than women do. Conclusions: The value and importance of this research is that for the first time in Croatia one national survey includes measure of subjective quality of life, and this results provide the normative value for Croatian population. The normative value for self assessed health status were also obtained and this will be the basis for the health policy and health prevention programs development. Aims: The PAC-SYM is a measure of the severity of symptoms associated with constipation. However, the severity and/or frequency of symptoms are not accurate proxy measures of symptom bothersomeness, defined here as the degree of discomfort reported by patients in relation to the symptoms they experienced. Similarly, using the PAC-QOL in addition to the PAC-SYM is not a suitable replacement for the systematic assessment of symptom bothersomeness. The inclusion of separate bothersomeness items would enhance understanding of the experience of opioid-induced constipation. Methods: The PAC-SYM was extended to include a measure of bothersomeness for each symptom. An additional question on overall bowel frequency satisfaction was also included. This amended PAC-SYM was discussed during three focus groups with participants who experienced constipation as a result of opioid treatment. The focus groups also discussed experiences of constipation in terms of the symptoms, treatments and impact on quality of life. Participants were asked to consider the adapted PAC-SYM and the severity and bothersomeness of their symptoms. Results: Focus group participants reported that severity and bothersomeness were separate concepts and both were necessary to gain an in-depth understanding of their experience of the symptoms associated with constipation. The adapted PAC-SYM, with the addition of a bothersomeness scale and an overall bowel frequency satisfaction question was acceptable to the participants. Conclusions: This qualitative study advocates the adaptation of the existing PAC-SYM to include a measure of the bothersomeness of each of the symptoms associated with constipation. The inclusion of an overall bowel satisfaction question is also recommended. The inclusion of these items would produce a more accurate picture of the experience of opioid-induced constipation. Aims: The measurement of pain, and the difficulties of capturing its impact, have a long history. However, neuropathic pain (NP) has received relatively little attention compared to nociceptive pain. In particular, patient reported outcome measures (PROMs) of NP have not been constructed using both a conceptual model of the impact of pain and a mathematical model aimed at securing scientifically sound, interval level measures. The aim of this study was to define a conceptual model for the health impact of NP, upon which to develop a new PROM. Methods: We conducted a qualitative study in three stages. First, we carried out an extensive review of the current NP literature. Second, we consulted relevant health care professionals to explore further the important clinical issues surrounding the impact and management of NP. Third, we conducted extensive in-depth, semi-structured qualitative interviews with thirty NP patients. Results: A framework was used to arrange the thematic content to allow grouping and comparison which resulted in a conceptual model for NP that includes physical and psychological perceptions of symptoms, environmental mediators, limitation in daily activities involving upper and lower limbs, sleep disturbance, avoidance behaviours, selfefficacy, psychological and social functioning. Conclusions: Our model highlights important areas which current NP PROMs do not address. This work will form a basis from which we can construct stable linear measures to assess the wider health impact of NP. A survey of 390 people with consultant diagnosed NP closes at the end of May and we expect the new measure to be developed by Summer 2005. Aims: To identify the advantages and disadvantages with respect to results and efficiency of two qualitative methods used according to two different approaches and to come forward with a proposal of the most appropriate method and approach to validate Core Sets of the International Classification of Functioning, Disability and Health (ICF) from the patient's perspective. Methods: Focus groups (FGs) and individual interviews (IIs), and of two different approaches, i.e. an open approach and an ICF based approach were used, respectively. The sampling of the patients with rheumatoid arthritis followed the maximum variation strategy. Sample size was determined by saturation. The FGs and IIs were digitally recorded and transcribed verbatim. The meaning condensation procedure was used for the data analysis. The results were linked to ICF categories according to established linking rules. Results: 49 patients participated in the FGs and 21 patients in the IIs. Saturation was reached after 5 FGs and 13 IIs in the open approach and after five FGs and 8 IIs in the ICF-based approach. Seventy four categories (FGs) and 65 categories (IIs) out of 76 ICF categories contained in the current version of the Comprehensive ICF Core Set for RA were reported by the patients based on the open approach and the ICF-based approach, respectively. The average time to perform the sessions in the open approach was in the FGs 1.02 hours (±.02) and in the IIs 0.25 hours (±.07). The average time to perform the sessions with the ICF-based approach was in the FGs 1.24 hours (±0.23) and in the IIs 0.34 hours (±0.03). Conclusions: FGs and IIs are highly useful qualitative methods to validate the Comprehensive ICF Core Set for RA from the patient's perspective. For further validation studies and ongoing development of Comprehensive ICF Core Sets in other countries and in other diseases the choice of the appropriate method and approach will depend on efficiency issues and costs and the availability of patients. Aims: Many patients with chronic disease do not take medications as prescribed. The purpose of this study was to establish the impact of diabetes drug treatment on the lives of patients with type 2 diabetes. Methods: The same professional facilitator conducted 18 focus groups (7 male, 11 female) with 138 socioeconomically diverse individuals with type 2 diabetes. The diabetes regimens for participants included oral and insulin mono-and combination therapy. Content analysis of focus group transcripts was used to establish themes. A frequency count of how often each theme was discussed across focus groups was calculated. Results: The most frequently mentioned dislikes of diabetes treatment were: 1. injections (100%), 2. fluctuating blood sugars (83%), and 3. ''everydayness'' (frequency and timing) of diabetes drug treatment (75%). The most frequently mentioned physical effects of treatment were sexual dysfunction (67%), hypoglycemia (50%), and gastrointestinal disturbances (50%) while anxiety and depression were considered emotional effects. Patients evaluated the convenience and flexibility of treatment on how the administration, timing, and frequency impacted their lives. Patients suggested that a patch or a once-a-day time-release pill would significantly increase treatment convenience and flexibility. Patients on insulin expressed trepidation about beginning insulin but acknowledged improved physical and emotional well-being after taking insulin. Assuming equal effectiveness of two drug treatments, patients would prefer the one that was most convenient/flexible (87%), non-injectable (73%), had least physical/emotional effects (60%), or cost less (60%). Conclusions: The impact of diabetes drug treatment on quality of life is tremendous. Patients with type 2 diabetes may take medications more appropriately if prescriptions for their treatment are based not only on clinical effectiveness but also take into consideration their individual needs and preferences. Aims: To gain more insight into the problems that families with a child suffering from a chronic renal condition experience, and into the solutions they desire, use or tried out for these problems. Methods: Families were recruited for interviews by five hospitals specialized in child nephrology and dialysis. Care was taken to vary treatment condition (haemodialysis, peritoneal dialysis and transplantation), age of the child, ethnicity, and family composition. Twenty-seven families took part in the study. From each family, a parent was interviewed. Children aged 8 or over were also interviewed. Semi-structured interviews were conducted in the families own homes, following a checklist of topics to ensure that the same topics were covered with each family. Topics included experienced problems, desired solutions and solutions the family used or had tried, experienced support and experienced control over family life. Key themes reflecting major problems and categories of solutions were defined on the basis of transcripts of the interviews. Results: The majority of the families indicated that the disease had a great impact, and that they had (far) less than desired control over family life. However, about one-third of the families indicated that they did not experience a great burden of the disease, and that all went fairly well. Major problem areas that emerged from the data were 'burden of the disease' (time as well as emotional), 'responsibility', 'lack of possibilities for relaxation', 'finances and regulations', 'social support/asking for and getting assistance', 'uncertainty' and 'child functioning'. Although problem areas were the same for many families, desired and used solutions varied considerable across families. Conclusions: This study adds to our knowledge about problems experienced by families with a child suffering from a chronic renal condition. Furthermore, it gives insight into solutions families desire, use or tried out. The results show that these solutions vary considerable across families and show the importance of tailoring interventions to families' needs. Aims: To measure the effect of different health care systems on the quality of life of physicians, this study compares the quality of life of working physicians in Germany and the USA in a representative cross-sectional study using comparable instruments under similar conditions. The focus was laid on factors promoting and impairing the physicians' well-being. Methods: We studied 342 German licensed physicians from three central German states (Hessen, Nordrhein-Westfalen, and Rheinland-Pfalz) and 303 working American physicians from upstate New York who were randomly assigned to the study. The selected regions are comparable regarding their socio-demographic and socio-economic characteristics. Physicians were working in hospital or in private practice, all positions and specialties being included. As main measuring instrument we used a specifically developed Questionnaire on the Quality of Life of Physicians (Jurkat & Reimer) supplemented by the cross-culturally validated SF-36 Health Survey. Response rate for the German sample was 39%, for the US sample 41%. For statistical analysis Mann-Whitney U tests and Spearman correlations were used. Results: Although the amount of working hours was nearly identical in the two samples, US-physicians declared significantly more often to be very satisfied with their working situation and with life in general. American physicians evaluated their lifestyle to be rather promoting than impairing their own health, which contrasts the majority of German physicians who felt impaired or strongly impaired. The SF-36 test results also indicate a higher health-related quality of life in US-physicians. Conclusions: Since the high workload affects the quality of life in both countries negatively, it seems necessary to reduce working hours. Similar factors seem to be responsible for a high well-being in both health care systems. Nevertheless, reforms are needed more in Germany due to the adverse effects of this countrys health care system. Aims: Study conducted to examine the coping strategies and its relationship with the quality of life and future concerns of HIV positive individuals at different stages of disease, attending HIV referral clinic. Methods: Cross-sectional qualitative study conducted on 97 individuals during Feb '02 Mar '03 using openended questions, in quality of life instrument.Univariate analysis of active coping with different demographic characteristics and QOL scores. Results: 56 participants above 30 yrs age, 64 men with equal number with secondary level education and living with spouses. As information about immediate reaction to diagnosis was elicited at later date, only 57 individuals could recall their reactions, which ranged from being suicidal (6), and other stressful emotional reactions. Wives of affected individuals had mentally prepared themselves for report. Passive coping like denial and worry (10). Active coping like taking care of health, seeking scientific information, starting medicines, following advice and developing positive attitudes (52). Counseling/information from doctors/counselors (24). Social support by friends/family members (11). Univariate analysis of the active coping was found be significant only for the domain of work and earnings (p=0.030). Educating children (34), family (11) and health concerns (8), fear of discrimination (6) and troubling others when ill was their main concerns. Securing future through saving money (25) and through insurance policy (15) was reported while others indicated their inability to save money. Conclusions: Providing scientific information, counselling and social support helped many persons to adjust to their disease. As their main concern was childrens' education and future of their families, efforts should be made for he rehabilitation of family members and sponsorship for children's education. Aims: The aim of this study was by means of Grounded Theory (GT) to obtain insight into what it is like to live with multiple sclerosis, in a Norwegian context. The idea was to focus on quality of life, and to gain understanding of aspects that enhance as well as inhibit quality of life from the perspective of the participants of the study. Methods: In accordance with the GT methodology the main purpose of the study was to develop a substantive theory on how people are able to live a good life in spite of an MS diagnosis. This was done by searching for the elements that the participants experience as a basic social-psychological problem and how they cope with this problem. Data collection was done mainly through qualitative interviews and participant observation Twentyone interviews with seventeen participants diagnosed with MS, recruited from different health care settings, were carried out. Results: The basic social problem experienced by participants was how to go on living a life as good as possible in spite of deterioration of health. The problem as it appears is managed through the process of redefining life course. Aims: Acquired hearing loss is one of the most common chronic health problems affecting approximately 28 million individuals in the US. Despite the fact that adverse affects on physical, emotional, behavioral, occupational, and social function, a paucity of measures currently exist to assess these effects or evaluate intervention outcomes. This paper will review the status of generic health-related quality of life measures used with hearing-impaired adults, followed by a description of an ongoing project to develop a quality of life instrument for individuals with adult-onset hearing loss. Methods: Selected studies utilizing generic quality of life measures with the hearing-impaired population will be briefly reviewed. Particular attention will be given to results of a study conducted by the presenting author in which 392 adults with hearing loss completed the WHOQOL-100, including 42 individuals with cochlear implants. Next, the methodology currently being used to develop a quality of life instrument sensitive to the consequences of hearing impairment will be presented, followed by preliminary research findings used to guide the development of this measure. These include results of focus groups, cognitive interviews, and the initial field testing of the pilot instrument which consists of the SF-36 and items generated by individuals with hearing loss, professional who serve them, and literature review. Results: Generic health-related quality of life measures used with individuals who have difficulty hearing lack sensitivity to the consequences of hearing impairment. These measures fail to adequately probe communication function or address the complex interactions that communication difficulties often cause in the social, emotional, and occupation domains. Conclusions: A need exists to develop a multidimensional health-related quality of life measure that is sensitive to the consequences of hearing impairment. The potential impact of the measure being developed in both rehabilitation research and service delivery should be widespread and important.

Keiko Tsuji, Nursing, Nagasaki University, Nagasaki, Japan; Yuki Mase, Nursing, Kanagawa University of Human Services, Yokosuka, Kanagawa Pref., Japan; Akemi Terasaki, Nursing, Nagasaki University, Nagasaki, Japan Aims: A survey through a questionnaire was conducted on a self-help group of 710 laryngectomy patients to obtain data concerning the support they receive from within the group. The beginners' class members were not fully aware of any support from the group in this survey. Then, an interview-based survey was conducted on 40 beginner participants. Methods: Based on word-for-word records, survey findings were analyzed both qualitatively and inductively. Results: 238 codes were extracted and divided into 48 core categories. An association between self-efficiency in terms of esophageal speech and the self-help group was found in 14 core categories, including two discouraging ones. They participated in the group only in order to obtain the support they needed to help in their implementation of self-expression and in finding a role in society. Their behavioral patterns included actively attending the classes for exercise, and sharing their own experiences with those experiencing similar problems. In addition, the instructor's existence as a model of themselves, and support from peers were of great significance regarding their efforts to acquire esophageal speech skills. Accepting their degree of aphonia, receiving counseling, and assisting group solidarity by disclosing themselves demonstrated their adaptation to the new environment, and helped to lessen their personal grief response to aphonia. However, any reproachful attitude of the instructor were considered to have a negative effect on their efforts to acquire esophageal speech skills. Those who selected esophageal speech after first trying other means of communication and then joined the self-help group were engaged in daily training with a specific goal set at each session and phonation during exercise. Conclusions: These categories represented an essential process for laryngectomy patients to accept their identity as aphonics, and to continue esophageal speech training. Aims: To demonstrate the relationship between utilities and relative risk reduction of angina pectoris attacks in type 2 diabetes patients. Methods: A base risk for type 2 diabetes patients to develop an angina pectoris attack in dependence of systolic blood pressure (SBP) was derived from literature data. This base risk was adjusted with several hypothetical relative risk reductions following an assumed exponential distribution of UKPDS data. The published disutility value for cardiovascular disease in type 2 diabetes patients was then re-calculated for the different adjusted base risk levels. Three populations derived from Diabetes in America were created: 1. non-diabetic 2. medical history of NIDDM and 3. impaired glucose tolerance (IGT) group. Age range was 20-74 years in all groups. Groups 2 and 3 were compared to group 1 with respect to their disutilities corresponding to the different probabilities to develop an angina pectoris attack in dependence of SBP values. Results: Angina pectoris disutilities of type 2 diabetes patients were 44% higher compared to the normal population. Comparing only patients aged 20-64 resulted in 89% higher disutilities for the diabetes patients. Even for IGT patients the disutility was 21% higher when compared to the control group. For the age group 20-64 the respective disutility was increased by 45%. Conclusions: The high increase of disutilities for development of cardiovascular events implies that diabetes treatment should also comprise tight blood pressure control additionally to glycemic control. This consideration should also be implemented in the various treatment guidelines for diabetes.

Feng Xie, Pharmacy, National University of Singapore, Singapore, Singapore; Shu-Chuen Li, Pharmacy, NUS, Singapore, Singapore; Julian Thumboo, Rheumatology and Immunology, Singapore General Hospital, Singapore, Singapore Aims: To determine important health-related quality of life (HRQoL) domains and items within each domain affected by knee OA, and identify ethnic variations in the importance of these domains and items among three ethnic groups in a multiethnic urban Asian population in Singapore. Methods: Focus groups were conducted among subjects with knee OA categorised by gender, ethnicity, and language spoken. All focus groups were audio-taped and transcribed verbatim, with subsequent translation into English for groups conducted in other languages. Data analysis was performed by combining the key elements of grounded theory and content analysis with the assistance of the qualitative software ATLAS/ti 5.0. Results: Five domains (pain, physical disability, other symptoms of OA, mental health, and social health) were identified from the 74 items reported as important by at least one subject. These domains were important for subjects from all ethnic groups with the exception of social health, which was more often important for Malay subjects. Items more commonly reported as important in the pain, physical disability, and other symptoms of OA domains were generally similar across ethnic groups, In contrast, important items in the mental and social health domains differed among ethnic groups. Conclusions: The impact of knee OA on HRQoL is broadly similar in both Asian and Western socio-cultural contexts. Both similarities and differences in important domains and items were identified among subjects with knee OA from 3 major Asian ethnic groups. Aims: Aim of this study was to evaluate change in health-related quality of life (hrqol) and BMI in overweight adolescents who participated in a Dutch summer camp program (Victory Camp). Methods: Adolescents participated in a 2-week program focusing on physical activity, healthy nutrition, social and emotional functioning and pleasure. BMI and hrqol were measured on the day of arrival and after 6 months. Height and weight were measured by professionals. The BMI was standardized by age and gender using international guidelines. Hrqol was measured using KIDSCREEN-52 item version selfreport. Differences between the two measurements were calculated using paired t-tests. Results: At the day of arrival 46 girls and 26 boys were included, after 6 months 32 girls and 19 boys (71%) responded. Paired t-tests identified positive changes in three hrqol scales, namely in physical functioning, social acceptance, and social support and peers (p<0.001). In contrast, a significant lower hrqol score was found on parent relations and home life (p<0.001). BMI decreased significantly from 2.66(0.59) to 2.51(.61). Reasons for non-response were holiday, family activities or illness (n=10), unknown (n=10) and one adolescent followed a clinical program for overweight. Conclusions: Positive changes were found in hrqol scores on the physical and social domains after participation in summer camp (Victory Camp). In addition, a positive change in BMI was found. In contrast, a negative change was found in parent relations and home life. As the KIDSCREEN measure is relatively new, this is one of the first studies in which change in hrqol was studied using KIDSCREEN. The discussion will elaborate on change in BMI in relation to change in hrqol and methods used. P-289/1800/PRELIMINARY PSYCHOMETRIC TESTING OF THE FOX SIMPLE QUALITY OF LIFE SCALE Sherry Fox, Nursing, University of Virginia, Charlottesville, Virginia Aims: The aim of this study was to develop and test a QOL measure that would be easy to use clinically and capture both cognitive and affective components of quality of life. Methods: The FSQOLS Scale was tested in 177 persons with lung, ovarian and colon cancer, who were contacted via an internet support group. Of the 177 surveys mailed, 145 were returned and usable for a response rate of 84%. Ten percent of participants were selected to participate in qualitative interviews for the purposes of content analysis on scale questions. Standard procedures for the development of instruments were used including the testing of content, and construct validity, factor analsyis and reliability. Results: Factor analysis revealed an initial 6 factor solution explaining 69% variance. The Screen test revealed one very clear factor and a questionable second factor. Content validity was supported with qualitative data reflecting the impact of having cancer on their health, wellbeing, satisfaction with life and overall functional ability. Convergent validity was suported when four commonly used QOL scales correlated significantly with the FSQOLS. Cronbach's alpha was 0.93. Conclusions: Results of this study suggest population health may be more easily addressed by the use of a user friendly instrument that captures the cognitive and affective components of QOL. Additional studies will be needed to further validate the scale. It must be tested with additional populations to establish the normative characteristics of the scale. Future samples must include differing patient populations, larger numbers, more males, and varied levels of education, ethnicity and religion. Specifically, more validity assessments are needed. Quality of life over time must be evaluated, and the predictive value of the instrument in those who will and will not have QOL must also be addressed. Aims: Bipolar disorder, as well as some of the pharmacological agents used in its treatment often impact patients' ability to function on a day-to-day basis. The goal of the present study was to evaluate the psychometric properties of a new 33-item patient-reported instrument, the Bipolar Functional Status Questionnaire (BFSQ). Methods: A total of 596 patients completed the BFSQ, as well as a series of other patient-and clinician-reported assessments at 11 clinical sites. Study participants included 148 patients who were currently hypomanic or recently recovered from a manic episode, 215 patients who were currently depressed or dysthymic, and 233 patients whose mood had been stable for a minimum of 2 months. In addition, 187 stable patients completed the full assessment a second time (7-14 days later) for test-retest purposes. Analysis evaluated the factor structure, reliability, and validity of the instrument. Results: Exploratory and confirmatory factor analyses indicated that a one-factor structure best fit the data. Item-level descriptive statistics, Cronbach's alphas, and validity correlations all met standard criteria. Most notably, the BFSQ demonstrated superior ability to discriminate among the three patient subgroups. Conclusions: The BFSQ is a psychometrically sound measure of functional status among bipolar patients. Efforts are underway to gather evidence for the responsiveness of this instrument, as well as to establish a minimal clinically important difference (MCID) criterion. It is expected that use of this instrument will facilitate the identification of treatments which maximize the functional status of bipolar patients and ultimately improve their adherence to treatment. Aims: Fatigue is a common and distressing symptom in cancer patients and studying fatigue in cancer patients now became a global concern. This study aimed to translate and test the reliability and validity of the Iranian version of the CFS. The scale consists of 15 items and 3 subscales: physical, affective and cognitive fatigue. Each item is rated on a scale of 1 (not at all) to 5 (very much) and the possible scores range from 0 to 28 for physical, 0 to 16 for affective, and 0 to 16 for cognitive subscale. The scale provides a total score so called total fatigue ranging from 0 to 60. Methods: The English language version of the CFS was translated into Persian (Iranian language) and was used in this study. The questionnaire was administered to a consecutive sample 112 of breast cancer patients attending the Iranian Center for Breast Cancer either for their treatment or follow-up examination. Statistical analysis was performed to test the reliability and validity of the CFS. Results: The mean age of the respondents was 45.7 years (SD=11.0), mostly had stage II breast cancer (67%) and had completed their initial treatment (45%). The reliability of the scale was investigated and it was found to be satisfactory (Cronbach's a coefficients were 0.85, 0.82, 0.83, and 0.94 for physical, affective, cognitive and total fatigue respectively). Validity as performed using known groups comparison analysis showed satisfactory results. The questionnaire discriminated well between sub-groups of patients differing in clinical status as defined by disease stage. In addition convergent validity was assessed between the CFS and a single measure of fatigue and the results showed significant correlations (Pearson's correlation coefficients were 0.61, 0.58, 0.60 and 0.62 for physical, affective, cognitive and total fatigue respectively, all p-values<0.01). Conclusions: The Iranian version of the CFS proved to be a reliable and valid measure of fatigue in cancer patients. However, there is need to carry out a similar study with a heterogeneous sample of cancer patients. P-314/1599/CHANGES OF QUALITY OF LIFE FOR XIAO (ASTHMA) PATIENTS WHO WERE TREATED BY CHI-NESE MEDICINE Hsueh-Erh Liu, Nursing, Chang Gung University, Tao Yuan, Tao Yuan, Taiwan, ROC Aims: Xiao (as asthma in western medicine) is a popular illness in both adult and children worldwide. Chinese medicine is one of the popular treatments for this specific disease. However, none has been discussed its impact of quality of life.The purposes of this follow-up study are to identify the changes of quality of life when patients receive Chinese medicine and to identify the associated factors. Research questions include. (1) Will the quality of life of Xion patients change during the treatment of Chinese medicine?

(2) Are there any factors asssociated with the quality of life of Xion patients? Methods: This is a survey with 3 point follow-up design. Asthma Quality of Life Questionnaire (AQLQ), symptom distress scale (SSS), and Verran and Snyder-Halper Sleep Scale (VSHSS) were selected for data collection. Subjects are recruited from the Chinese medicine OPD in a medical center. Subjects have been diagnosed as asthma by physician and are receiving Chinese medicine as major treatment now. Subjects were contacted and exaplained. Obtaining their written permit, then the first assessment was conducted. Each subject is followed by their OPD schedule. A total of 106 subjects were recruited. Only 66 subjects completed the 3-point follow-up. The major reason of incompletion is the subjects change their appointment or their stop this treatment. All data were managed by SPSS-PC 10.0. descriptive and inferral statistics (i.e. t-test, ANOVA, Pearson correlation, repeated measures) were performed. Results: The characteristics of these 106 subjects were Taiwanese (68.4%), unmarried (63.2%), receiving financial support from their family (71.7%), and self-rated health as ordinary(64.2). In regard to quality of life and sleep, subjects reported significant changes during research period. Variables i.e. gender, age, marriage, employment status, quality of sleep,and symptom distress have different impact on each assessment. Aims: To assess the psychometric properties including minimal clinically important difference (MCID) of the OnyCOE-t, a modified questionnaire specifically designed to assess PROs of patients with toenail onychomycosis. Methods: About 504 patients with toenail onychomycosis were randomized to receive 12 weeks of terbinafine 250 mg/day with or without debridement in the IRON-CLADÒ trial. The OnyCOE-t was completed at baseline, weeks 6, 12, 24, and 48. Reliability, validity, and responsiveness were evaluated. To enhance interpretation and quantify a MCID, we examined the extent of change in scale score relative to percent clearing of mycotic involvement in the target toenail. Results: Most patients were Caucasian (63.5%), male (61.3%), with mean age of 48.5 years. Variable clustering confirmed 6 scales [toenail symptom frequency and bothersomeness; physical appearance-problems; physical activities-problems; overall problems (OP); stigma; and treatment satisfaction (TS)]. Internal consistency reliability was good (Cronbach's alpha coefficients: 0.84-0.94). Reproducibility, for patients reporting no clinical change from baseline to Week 6, was generally acceptable (ICCs: 0.29-0.73). Construct validity was demonstrated by inter-item and scale correlations. For known-groups validity, cured patients (negative mycology and 100% clearing) reported better scores than those not cured (p<0.0001, most scales). Scales discriminated well based on age and gender (p<0.0001, most scales). Responsiveness was good, with TS and OP most responsive (Guyatt's statistic of 1.72 and 1.13, respectively). MCID results were fairly consistent across scales and indicated that an 8-point change (on a 0-100 scale) was clinically meaningful. Conclusions: The OnyCOE-t questionnaire is a unique, toenail specific PRO questionnaire that can be used with confidence in future studies of toenail onychomycosis. Aims: Despite extensive literature on cross-cultural equivalence, there is no consensus on a practical procedure to assess the equivalence of PRO instruments in different languages. A ''universalist'' model has been proposed but remains theoretic. Its operationalization should naturally combine qualitative and quantitative methods. Our aim was to identify statistical questions implied by the ''universalist'' model and find quantitative methods to answer these questions. Methods: Each level of the model was studied from a statistical perspective. Results: The first level of the model, conceptual equivalence, is mostly involved with qualitative issues. Nevertheless, studying the item variance-covariance matrix structure, for example with confirmatory factor analysis, could constitute an a posteriori assessment of conceptual equivalence. Item equivalence and semantic equivalence focus on individual items and lead to the following statistical question: Do the answers to the item depend on the instrument version? Various quantitative methods exist to study this problem of differential item functioning, such as contingency table analysis and item response theory. The next level, operational equivalence, is mainly a qualitative issue, even if specific quantitative studies (e.g. about the response scale format) could be useful information sources. The quantitative methods for the measurement equivalence level have been extensively explored since this step has been described as the core of the statistical analysis. The study of psychometric properties and scale equating are relevant to this level. Conclusions: Various quantitative methods could be incorporated in the operationalization of the ''universalist'' model and could constitute a complementary information source to the qualitative methods in order to assess cross-cultural equivalence.

Aims: The aim of the study was to identify psychosocial predictors of quality of life and glycemic control over a 3 year follow-up period in adults with type 1 diabetes. Methods: The German Multicenter Diabetes Cohort-Study is a prospective longitudinal cohort study of adults with newly onset type 1 diabetes. About 313 adults (1-12 weeks after diagnosis)were included at 12 sites. Follow-up measurements were performed one, 2 and 3 years after diagnosis, including treatment and psychosocial variables, e.g. insulin units/kg, coping behavior, locus of control, social support, self esteem, or hopelessness. As outcome variables the SF-36 and the HbA1c was used to measure health related quality of life (HRQoL) and glycemic control. The sample was randomized before data analysis to develop different Structural Equation Models (SEM) in one half of the sample, and to investigate their fit in the other half (AMOS). Results: Complete data for all four measurements could be obtained from 85.3% of the patients. We observed a continuous deterioration of glycemic control, with 19.2% of the patients showing HbA1c values above 8% after 3 years of treatment. Different SEMs have been applied, showing a satisfactory to good fit (RMSEA 0.038-0.045, 90% CI 0.000-0.084; CFI 0.972-0.980). Patients who demonstrate a higher self confidence, and an active participation at the time of the diagnosis were more likely to show better glycemic control after three years (b=0.17) as well as a better HRQoL (b=0.19)0.34). Higher education served as an independent predictor for better glycemic control (b=0.13). About 12% of the variance of the HbA1c after 3 years could be explained by measuring psychosocial variables at onset of the disease. Conclusions: Evaluating the individual characteristics of a patient with diabetes mellitus after she/he gets her/his diagnosis helps to identify those patients who are at particular risk for lower therapy adherence and an unfavourable course of the disease. Aims: Although the QOL assessment is important in dentistry, QOL in orthodontic patients has not been fully investigated.

The AIMS of this study were to clarify their QOL and psychological status, and explore the relationship between severity of malocclusion and QOL. Methods: The subjects consist of 110 orthodontic patients and age-matched 27 persons with normal occlusion. The subjects were divided into following three groups based on their treatment plans; namely, jaw surgery group (SURG; n=48), non-surgery group (NONS; n=62), and control group with normal occlusion (CONT; n=27). Their dentofacial morphologies were assessed using specific severity scores. Generic QOL was assessed using the SF-36 and diseasespecific QOL was assessed using Subjective Oral Health Status Indicators (SOHSI), the Orthognathic QOL Questionaire (OQLQ), and recognition and satisfaction scores of surgical correction. Anxiety and depression were assessed using STAI and SRQ-D, respectively. QOL data were computed using the simple count and weighted score methods. Results: There was no significant difference in SF-36 among the three groups. In some scale scores about disease-specific QOL, i.e., 'ability to speak, other oral symptoms, communication/social relationship and worry/concern' (SOHSI), 'social aspects of dentofacial deformity, facial aesthetics and oral function' (OQLQ), 'recognition and satisfaction', SURG was significantly lower than NONS. In STAI and SRQ-D scores, there was no significant difference between SURG and NONS, but the patients with high score of STAI tended to belong to SURG. Conclusions:

The results of this study suggested that disease-specific QOL was significantly lower in patients required jaw surgery than in patients without need of jaw surgery. The QOL assessment of orthodontic patients is indispensable to establish individualized treatment goals in modern clinical orthodontics. Aims: Governments implicitly consider the enhancement of the quality of life as a touchstone of the effects of their policies. In our paper we focus on the subjective well-being. In this approach the emphasis lies on the evaluation of the individual's life, both in the sense of the evaluation of contentment (the cognitive aspect) and in the sense of affective reactions (state of mind and emotions). Besides the evaluation of life in general, we also analyse the satisfaction with more specific aspects of the daily life, such as health, the own financial situation and the neighbourhood. Based on the resource-theory on subjective well-being, we analysed the main determinants of people their evaluation of life as a whole and more specific aspects, like their own health. Methods: In order to study this, we have set up a face-to-face survey research design within a random sample of 4652 people entitled to vote, spread over the three biggest Flemish cities (Ghent, Bruges and Antwerp). After the first wave of the survey we have a response of about 40%. More specific, in our analysis we present the empirical data of 1873 adult citizens in Ghent, Bruges and Antwerp. The sample comprises a representative sample of citizens. In our search for the main determinants we took a wide range of variables into account. There relative importance was analyses using multivariate regression analysis. Results: In our paper we demonstrate that the determinants of the subjective well-being in general do not conform with the evaluation of all more specific aspects of the daily life. A distinction has to be made between the satisfaction with more private aspects (as the family life and health) on the one side and more public aspects (as the neighbourhood) on the other. The former correlates stronger with the general subjective well-being. Conclusions: Concerning the satisfaction with health, we found that that judgement is not only based one the appreciation of ones health as such. Other factors such as social support, social integration and self-esteem are important determinants of the satisfaction with health and subjective well-being more in general. Those factors had to be taken into account in the set-up and evaluation of the health care policy. P-322/1457/BEHAVIOR AND ANALYSIS OF SF-36 DATA IN SURGICAL QUALITY OF LIFE RESEARCH Velanovich Vic, Ilan Rubenfeld, Surgery, Henry Ford Hospital, Detroit, Michigan Aims: The purpose of this study was to determine the behavior of SF-36 data and to assess whether this data is best analyzed and presented as means (continuous), medians (ordinal), or topbox frequency (ordinal). Methods: About 400 preoperative and postoperative surgical patients with a variety of surgical disorders were asked to complete the SF-36 generic quality of life instrument. Data were analyzed for goodness of fit in a Gaussian distribution with the Wilk-Shapiro test, skewness, kurtosis, mean, median, mode, and top-box frequency. Results: See Table. Conclusions: In surgical quality of life research, SF-36 data does not follow a Gaussian distribution and should not be analyzed using parametrics statistical techniques. In fact, most responses for six of the eight domains are clustered in the one or two highest values. Therefore, top-box analysis may be a better alternative for data analysis. Aims: Importance weighting is a common idea in quality of life (QOL) measurement. The weighting procedure of multiplying item satisfaction by item importance was adopted in many QOL instruments. However, in Locke's [1,2] range-of-affect hypothesis, he indicated that given a mount of real-idea discrepancy, people would have stronger satisfaction/dissatisfaction on important items than on unimportant items, implying that item satisfaction has incorporated the judgment of item importance, and then, weighting item satisfaction score with item importance score is redundant. The purpose of this study was to address the issue of importance weighting by examining Locke's hypothesis in the context of QOL research. METHODS: Two studies were conducted to address this issue. The first study was conducted to examine the range-of-affect hypothesis by a survey to test whether the association between item real-ideal discrepancy and item satisfaction was stronger among people with higher item importance than people with lower item importance. Three hundred and thirty two undergraduate students at National Taiwan University (NTU) participated in the study voluntarily. In the second study, a within-subject experiment was conducted to show that given the amount of real-ideal discrepancy, people would have stronger satisfaction/dissatisfaction when they focused on a higher important aspect than a lesser important aspect. Forty undergraduate students at NTU voluntarily participated in the study. Results: Result of study 1 showed that item importance and perceived have-want discrepancy has significant interaction effect on item satisfaction. In addition, results of study 2 also revealed that the association between objectideal discrepancy and object satisfaction is stronger on higher important aspect than lesser important aspect. CONCLU-SIONS: The results of the two studies were generally consistent with Locke's range-of-affect hypothesis, revealing that item satisfaction has incorporated the judgment of item importance, and then, the procedure of importance weighting is unnecessary. Aims: Generalised Anxiety Disorder (GAD) is characterised by excessive uncontrollable worry or anxiety causing considerable stress and substantial social and functional impairment, suggesting that GAD has a negative impact on the patients' quality of life (QoL). The aim of this study was to assess the impact of escitalopram vs. placebo on the QoL of patients with GAD using the SF-36 scale. Data was collected along a clinical trial evaluating the effect of escitalopram (20 mg/day) vs. placebo on the time to GAD relapse in-patients who responded to escitalopram treatment. Methods: QoL data, assessed by the SF-36 scale, was collected along a multinational clinical study of out-patients (18-65 years) with a primary diagnosis of GAD (DSM-IV). After 12 weeks of open-label treatment, responders were randomised up to 76 weeks of double blind treatment with escitalopram (n=104) or placebo (n=91) to measure the time to relapse. QoL was assessed at randomisation, week 24, week 48 and week 76 post-randomisation or at early termination using the SF-36 QoL scale. Results: Escitalopram treated patients had significantly better SF-36 scores for the three mental health subscales; social functioning (p=0.005), mental health (p=0.002); and role emotional (p=0.004). Clinical results showed that significantly fewer escitalopram-treated patients relapsed, and the risk of relapse was 3.6 times higher in placebo treated patients. Conclusions: The SF-36 scale is a sensitive tool for measuring QoL in GAD patients. Escitalopram treatment resulted in better patient reported QoL assessed by the SF-36 scale compared to placebo, and escitalopram treatment lowers the risk of relapse significantly. This suggests that escitalopram treatment improves QoL of GAD patients and is more effective than placebo in the prevention of relapse in GAD. Aims: In order to reveal influences of different health-care systems on gender-specific differences in physicians' quality of life, a cross-cultural comparison was carried out. The objective was to determine if and to what degree working female physicians in the USA as well as in Germany differ from male physicians regarding quality of life and health-related behavior. Methods: We studied 342 German licensed physicians from three central German states (Hessen, Nordrhein-Westfalen, and Rheinland-Pfalz) and 303 working American physicians from upstate New York who were randomly assigned to the study. The selected regions are comparable regarding their socio-demographic and socio-economic characteristics. Physicians were working in hospital or in private practice, all positions and specialties being included. As main measuring instrument we used a specifically developed Questionnaire on the Quality of Life of Physicians (Jurkat & Reimer) supplemented by the cross-culturally validated SF-36 Health Survey. Response rate for the German sample was 39%, for the US sample 41%. For statistical analysis Mann-Whitney U-Tests and Spearman correlations were used. Results: German female physicians enjoy a slightly higher quality of life than their male colleagues; in the USA, however, the reverse is true. In the cross-cultural comparison, male physicians display highly significant differences regarding their quality of life in favor of the Americans. Between women, there are fewer significant differences. US female physicians are mentally more at risk than their male colleagues, whereas this result could not be verified in Germany. With regard to the reasons for studying medicine, men in both countries more often stated helping as a motive, whereas women most frequently had chosen their medical profession out of idealistic motives. Conclusions: Germany's job conditions, being obviously worse than those in the USA, significantly influence male physicians negatively in their quality of life. Assumed genderspecific and country-specific stereotypes/cliche´s cannot be confirmed. Generally, gender-specific research should increasingly take cross-cultural differences into consideration. Aims: Nursing home resident (NHR) self-report of QoL is hampered by cognitive deficits; yet, the US federal government holds NHs accountable for QoL outcomes. We sought to facilitate QoL assessment in NHRs by cognitively testing items written at a 4th-grade reading level. Methods: Subjects were recruited from a list of NHRs provided by NHs in two states. Following a standard protocol, trained interviewers asked NHRs 20 questions with probes to assess how well the NHR understood the item. QoL domains included autonomy, dignity, independence, meaningful activities, privacy, security, and spiritual well-being. Results: Just 25% (n=16) of NHRs approached completed the interview. Others refused (n=17), were unavailable (n=17), were unresponsive (n=9), or completed just a portion (n=5). Causes of partial interviews included fatigue and loss of interest. Easily understood questions addressed independence in reaching things, being treated with dignity, choice in activities, availability of activities, safety and security, and enjoyment of mealtimes. Autonomy in daily routine, spirituality, and global QoL were not readily understood. Most NHRs reacted with anxiety to a privacy item asking whether they could find a place to be alone if they wished, stating that they preferred not to be alone. Conclusions: It may be difficult to get a representative sample of NHRs to report on their QoL: only 25% of those approached completed an interview. NHRs who participated were able to comment on abstract concepts such as dignity and independence when these were grounded in specific and common events. With the exception of a general question on safety and security, clearly understood items were concrete; whereas poorly understood items were abstract. Results suggest that it is more productive to ask about religion than spirituality, and that respondents may be disturbed by questions that refer to being alone. Finally, a global question about QoL will be interpreted as referring to ones life before entering the NH unless it is limited to the proximal now. Aims: The purpose of this study was to identify central aspects of quality of life in relation to the experience of being a brother or sister of persons who suffer from schizophrenia. Methods: 16 people participated in the study.The interviews were recorded on audiotapes. The interpretation emerges from the relation between empirical substance and the theoretical context of theories concerning quality of life. This process could be understood as abduction. Abduction could be seen as a combination of induction and deduction, but it contributes with some new factors. In this process theories concerning interrupted feelings and health related rhythms created new hypothetical patterns for further interpretation and understanding of quality of life in this group. This spiral movementAIMS to adjust and purify the theoretical model. Results:

The key component of the siblings feelings grief, hope, anger, guilt and shame are interrupted by four interrelated factors; ambiguous loss, the fluctuation, an inner prohibition and invalidation. The ambiguity and the fluctuating nature of the illness is experienced by the subjects as having strong influence on their feelings toward the sick sibling and indirectly to their own quality of life. This led to new hypothetical patterns approaching in the tension between the empirical data and the assumption that health related rhythms are an essential factor in the siblings experiences. The characteristics of the health-rhythm were identified as: the rhythm in opening and closing, the rhythm in movement and the rhythm in exceeding. The essence of the health-rhythm is the rhythms intertwinement, the rhythms paradox and the rhythms possibility. Conclusions: Visibility and understanding of the character of interrupted feelings and health-rhythms may open for greater freedom of choice, which can guide to better health by transforming possibilities into action. These rhythms inn feelings are pointing at the fluctuation and ambiguity in the experience of quality of life. These assumptions should lead to further studies and to further investigation on the complex nature of experience of quality of life using different methodological traditions. Aims: To explore the relation between quality of life (QoL) and different healthcare systems as well as the relevant influent factors in China. A comparison was made between the middle-aged population who live in Guangzhou, a developed coastal city, and Guiyang, an underdeveloped inland city of China. Methods: A questionnaire designed on the basis of SF-36 was conducted. The collected data were stored in a database and analyzed by methods including descriptive analysis, t-test, and ordinal logistic regression by golden mean (ranking). Results: The mid-aged people who enjoy good healthcare measurement have better QoL than those have inefficient or without healthcare measurement. The results are all statistically significant (p>0.01). The analytical outcome also implies that factors like living region, age, teadrinking, self-health-assessment and chronic disease may have influence upon QoL. Conclusions: The QoL of the middleaged people who have good healthcare measurement is better than those have inefficient or without healthcare measurement. The QoL of the mid-aged people who live in the two Chinese cities is different due to the diversity of their healthcare payment patterns. Good living habits like tea drinking can keep people healthy and improve their QoL. Therefore, good living habits and dissemination of healthcare knowledge should be advocated among mid-aged people. Aims: Generic, preference-based measures of health-related quality of life (HRQoL) are useful tools for clinicians and policy makers wanting to quantify the potential HRQoL benefits of different treatment and intervention strategies. The Classification and Measurement System of Functional Health (CLAMES) is a new, multi-attribute health status instrument representing the broad range of health-related functioningphysical, mental, and social -from a ''within-the-skin'' perspective. CLAMES assesses functional capacity using 11 basic health status attributes: Pain or Discomfort, Physical Functioning, Emotional State, Fatigue, Memory and Thinking, Social Relationships, Anxiety, Speech, Hearing, Vision, and Use of Hands and Fingers. The objective of the presentation is to show how we constructed the scoring function for CLAMES. Methods: Preferences for a subset of 238 CLAMES health states were elicited from lay panels (n=146) in 9 Canadian communities, using the standard gamble (SG) technique. In order to develop a convenient scoring function for computing preferences for CLAMES health states not used in the survey, the obtained mean preference scores for the 238 states were modeled using both the statistical and decomposed approaches. Results: A log-linear model based on the statistical approach and with a scaling adjustment provided the optimum fit to the data; and is recommended as the preferred scoring function.This function and other functions tested will be presented with evaluation of their performence. Conclusions: Preferences for health states generated by CLAMES could be useful for the economic evaluation (i.e., cost-effectiveness and cost-utility analysis) of health care programs and interventions, as well as in the construction of summary measures of population health.

277/1718/DETERMINING THE RELATION BETWEEN INCONTINENCE AND QUALITY OF LIFE FOR PER-SONS WITH SPINA

Viroj Tangcharoensathien, International Health Policy Program, MOPH, Nonthaburi, Thailand Aims: A new health status measure for the Thai general population called 9-THAI was developed for using in the national household survey. The objective was to assess construct validity of the 9-THAI using confirmatory factor analysis (CFA)

Because of the very skewed distribution of the large sample data, a weighted least square estimation method was applied through the LISREL program. Results: Multiple goodness of fit indices indicated that the 2F model was the best fitting model based on standard criteria used

Group P-290/1618/VALIDITY ASSESSMENT OF TWO MEA-SURES OF ONES READINESS FOR CHANGE: THE SINGLE ITEM STAGING SCALE (SIS) AND THE UNI-VERSITY OF RHODE ISLAND CHANGE ASSESSMENT (URICA)

Cluster 2 had higher pain and arthritis severity, lower SE-C and CE, and scored higher for arthritis management, readjustment and redefinition than Cluster 1. In the SIS, the PC group had the highest pain and arthritis severity and lowest SE-P and SE-C followed by the C, PA, M and A groups. PC and C groups had lower CE than the PA, A and M groups. PA group had higher scores for arthritis management, readjustment and redefinition than the C group. The 5 cluster solution based on URICA subscales was not reproduced

Rotterdam, the Netherlands P-292/1574/ANOREXIA/CACHEXIA RELATED QUALITY OF LIFE FOR CHILDREN WITH CANCER: TESTING THE PSYCHOMETRIC PROPERTIES OF THE PEDIATRIC FUNCTIONAL

Methods: 96 cancer patients (51 children aged 7-11 and 45 teenagers aged 12-17) from Chicago Metropolitian area were recruited. The psychometric properties of the peds-FAACT were examined using item-total correlations, alpha coefficients, confirmatory factor analysis, and Rasch model (mean square fit statistics, MnSq) to examine the unidimensionality of the scale and the quality of items. Logistic regression was used to examine differential item functioning (DIF) between genders, ages, cancer types and races. Results: A 6-item core form with acceptable itemtotal correlation (>0.3) and alpha (>0.7) was validated for all samples (ages 7-17). CFA confirmed its unidimensionality with the acceptable fit indexes (GFI=1, AGFI=1, non-normed fit index=0.99, normed fit index=0.90) and Rasch analysis showed acceptable fit (MnSq<1.4) values. None of the items demonstrated DIF on gender, race, cancer type or race suggesting these items are stable across these conditions. Four additional items were identified as the age-appropriate peripheral items for patients with ages ‡ 10. Conclusions: The core peds-FAACT demonstrated good psychometric properties using both classical test and

South Korea P-298/1786/DEVELOPMENT OF A DIALYSIS PATIENT SATISFACTION QUESTIONNAIRE Phi Linh Nguyen Thi & Serge Briancon

Intensive Care

/1225/NEW QUALITY OF LIFE (QOL) QUESTION-NAIRE FOR CHILD CARING FATHER Yukari Yoshida & Michiko Kobayashi

Japan P-306/1022/A BOOTSTRAP MODEL-AVERAGING TECHNIQUE TO INVESTIGATE THE PROGNOSTIC VALUE FOR SURVIVAL OF QUALITY OF LIFE (QOL) INFORMATION A

Methods: About 573 glioblastoma patients were analyzed from a RCT which found a significant improvement in survival for patients undergoing radiotherapy with concomitant/adjuvant temozolomide (p=0.005) without any detrimental effect on QOL. QOL baseline data were collected with the EORTC QLQ-C30/BN20. The Cox model was used for the multivariate analyses of survival. The stability/magnitude of the QOL factors was examined using a model-averaging technique, based on repeated forward Cox PH model-building on 1000 bootstrap generated data-sets. Results: The final multivariate model identified 3 baseline biomedical parameters as independent prognostic factors for survival: performance status (p<0.001), Age (p<0.001) and neurosurgery (p<0.001). QOL baseline scores for physical, cognitive and social functioning (p<0.001), and dyspnea (p<0.001), independently predicted survival. The inclusion frequencies for these QOL variables were respectively: 44, 97, 93 and 65%. The final model with the above 7 variables was the one most frequently selected (13.2%), out of the 1000 bootstrap generated multivariate models

Germany P-311/1461/JUDGING QUALITY OF LIFE: THE INFLU-ENCE OF COGNITIVE PROCESSES Stefan Ho¨fer

Psychology, Royal College of Surgeons in Ireland

PATIENTS WITH MALIGNANT PLEURAL MESOTHELIOMA (MPM) TREATED WITH CHEMOTHERAPY J. Van Meerbeeck, Thoracic Oncology

The Hospital for Sick Children

WHOQOL group P-345/1732/IMPROVING PEDIATRIC FUNCTIONAL MEASUREMENT: AN ITEM RESPONSE THEORY ANALYSIS OF THREE CHILD HEALTH SURVEYS

Methods: Patients who had surgery but no CTX for NSCLC participated in a structured interview, acting as surrogate decision-makers. The interviewer provided the decision context (choice: CTX or not), toxicity, and survival outcomes. Participants rated their preference for CTX (0-10; 10 most preferred) and specified the additional survival threshold (SAT) required for accepting CTX. They then saw HRQL information, randomized to either small-magnitude or large-magnitude HRQL effects of CTX. We assessed the impact of HRQL on preference ratings and on SATs; patients also rated the usefulness of the information (0-10; 10 most useful). Results: 50 patients participated (52% female; mean age 68 years; 18% had post-secondary education). Mean baseline SAT for accepting CTX was 10.5% (absolute increase in survival). On average, patients presented with smallmagnitude HRQL effect decreased their SAT for CTX (mean SAT 8.4%) while those seeing large HRQL effects increased their SATs (mean 16.2%) [p=0.04]; similarly, small HRQL effect increased preference scores for CTX while large HRQL effect reduced average preference scores. The usefulness ratings of toxicity information and of each domain of HRQL data were high (range 7.5-8.2) and did not significantly differ from each other

Background: Items suitable for use in a disease-specific utility instrument for COPD were identified from the St George's Respiratory Questionnaire (SGRQ) by applying classical test methods and Rasch analysis to an existing large data set. Nine items were selected, three from each SGRQ domain: Symptoms, Activity and Impacts. The items for each domain were chosen on the basis of their location along a Rasch item map to reflect mild, moderate and severe disease (one of each severity per domain). An item could also be absent from the domain, thus making four severity levels for each domain. A total of 64 (4Â4Â4) health states (HS) were created to produce all possible combinations, but we used the most common 27 HS, selected on the basis of combinations found in existing data. Using a simple scale for item severity (absent=0, mild=1, moderate=2, severe=3), each HS corresponded to 1 of 10 severity levels. Methods: 50 COPD out-patients were studied (age 71 years SD 10, 50% male, FEV1 1.0 l, SGRQ score 55 SD 16 units). Each rated their current health and 9 HS using TTO and a Visual Analogue Scale (VAS)

The VAS score had a non-linear association with HS severity (r=0.96, 2nd order component p<0.0001) due to a ceiling effect that produced similar scores for HS in the highest severity tertile. Conclusions: Older patients with COPD used TTO to rate their preferences for health states in a reliable manner. VAS scores did not distinguish between the most severe health states

No significant differences in baseline characteristics were observed in either treatment group. Of the 165 patients, 145 completed the PAID and 160 completed the Fear of Self-Injection questionnaire in both treatment periods. After using InnoLet, patients scored lower on 17/20 PAID items, indicating a lower degree of feeling discouraged, scared, deprived, depressed, overwhelmed, angry, and unsupported (Wilcoxon, p<0.05). Patients reported significantly lower fear of self-injection after using InnoLet vs

/1605

/1389/THE UTILITY OF INTIMATE PARTNER VIOLENCE Eve Wittenberg

Aims: We compared the impact of tipranavir vs. comparator regimens on HRQL in two randomized clinical trials for ARVexperienced patients failing their current PI-based regimen. Methods: Patients were enrolled in 2 multinational phase 3 randomized clinical trials in North America and Australia

Analyses included all randomized patients who received at least 1 dose of study drug and completed at least 1 survey. The primary outcome was change from baseline to 24 weeks in Physical Health (PHS) and Mental Health Summary (MHS) scores. Multivariable longitudinal analyses used GEE modeling to explore treatment effects across all time points. Results: A total of 620 and 539 patients were included in RESIST-1 and 2. Patients on CPI/r were more likely to have missing data at all time points after baseline (at 24 weeks: RESIST-1: 46% vs. 15%; RESIST-2: 66% vs. 43%). Larger differences between groups were observed in RESIST-2 compared to RESIST-1. In RESIST-1, non-significant trends generally favored TPV/r. In RESIST-2, at Week 8 patients on TPV/r reported improvements in all MOS-HIV dimensions, with differences in change scores of 7.6 points for pain, 4.1 for cognitive functioning and 1.9 for PHS (p's <0.05)

Department of Health Policy and Administration

All analyses were nationally representative. Results: 26,558 individuals were included in the study sample, with 368 (national prevalence: 1.4%) underweight, 9754 (38.1%) normal weight, 9500 (36.0%) overweight, 5990 (21.4%) obese, and 955 (3.2%) extreme obese. Descriptive results showed that underweight, obese, and extreme obese patients had significantly lower SF-12 physical and mental scores and EQ-5D index scores, while overweight patients had slightly lower SF-12 physical scores and EQ-5D index scores (all p<0.01). The results were further supported by the multivariate analyses

Aims: To assess quality of life (QoL), depression, patient selfperceived burden, and caregiver burden in a cross-sectional investigation performed on amyotrophic lateral sclerosis (ALS) and multiple sclerosis ( 5%] were over the cut-off score for depression, vs. only 11 and 5 ALS patients and caregivers, respectively 27.5 and 12.5%. Conclusions: The two series showed remarkable differences in QoL, depression and burden, which are much better in ALS than in MS, both in patients and in caregivers. Differences in age, gender and premorbid personality can also play a role. It is to be noted that in our department an interdisciplinary approach focused also on psychological issues has been used for years in ALS, while it is still being implemented in MS care.

Aims: The majority of the world's elderly population lives in low-income countries. In addition to poverty, disability and poor health prevents the group of elderly people to meet their basic needs. Social capital has been identified as a critical asset for creating opportunities that enhance well-being and for achieving greater security and reduced vulnerability. This studyAIMS to identify determinants of quality of life (QoL) and investigate its association with individual-and communitylevel social capital among older people in rural Bangladesh. Methods: A cross-sectional study of 1198 elderly persons (aged 60 years and older) was conducted in a rural district in Bangladesh. Independent variables include age, sex, education, marital status and economic status of the elderly person as well as individual-and community-level social capital. Self-rated QoL was the dependent variable. Descriptive analyses were done to show group differences in QoL and logistic regression analyses to identify determinants of QoL. Results: Advanced age, lack of education and poor economic status were significant determinants of poor QoL. Low social capital at individual-and community-levels was associated with poor QoL among the elderly. The odds of reporting poor QoL was significantly higher (OR: 1.7; 95% C.I: 1.2-2.4) among elderly people with low individual-level social capital (SCI), compared to those who had high SCI. Among elderly people with low community-level social capital (SCC), the odds ratio for reporting poor QoL was similarly significant higher (OR: 1.9; 95% C.I: 1.1-3.3) than for those with high SCC. Conclusions: This population-based study provides empirical evidence that social capital both at individual-and community-levels is directly associated with QoL of elderly people in rural Bangladesh.Aims: Quality of life (QOL) of older adults is particularly important given our aging populations. There have been few cross-cultural studies comparing QOL of older adults. In this study, we examined the differences and similarities in QOL and its predictors for older adults in selected regions of Canada and Brazil. Methods: The WHOQOL-BREF and a demographic data sheet were administered to random samples of 202 older adults from Canada and 288 from Brazil. In Canada, a mail survey was used and in Brazil, a household survey was conducted. Forward entry least squares multiple regression analysis was conducted to examine predictors of QOL of both samples. Results: The mean age was 72.9 years (SD=8.52) in Canada and 71.26 (SD=7.49) in Brazil. There were more female than male respondents in both samples. Overall QOL and scores on the physical, psychological and environmental domains of the WHOQOL-BREF were higher in the Canadian sample. However, scores on the social domain were higher in the Brazilian sample. The regression model explaining QOL of the Canadian sample was highly significant (F=27.29, p<0.001) and explained 57% of the variance in overall QOL. The strongest predictor of overall QOL was health satisfaction followed by meaning in life, enough money, and opportunities for leisure activities. In the Brazilian sample,the same variables explained 41% of the variance (F=19.44, p<0.001). Similar patterns of predictors were found to be significant,with health satisfaction, followed by enough money, meaning in life, opportunities for leisure activities, and healthy physical environments as significant independent variables. Conclusions: While there are differences in overall QOL of older adults in Brazil and Canada, many of the same factors predicted QOL of these samples. Further multi-national cross-cultural research regarding the challenges associated with an aging population would be valuable. Health interventions to promote QOL of older adults in both countries certainly should focus on maintaining health (and health satisfaction), facilitating meaningful activities, and supporting policy initiatives to ensure that older adults have adequate financial resources. Aims: Prior clinical trials have observed significant differences in patient HRQoL across different classes of antihypertensive medications and between drugs within a single drug class. However, despite a wealth of research in relating antihypertensive treatments to HRQoL, the bulk of these studies has largely focused on efficacy research, i.e., addressing the question: ''Does the pharmacotherapy affect HRQoL in the ideal setting of a controlled clinical trial?'' Thus, the primary objective of the study is to assess the effects of pharmacologic treatments for hypertension on patients' HRQoL (as measured by the Veterans SF-36, a modified version of the widely used Medical Outcomes Study SF-36) in actual clinical practice. Methods: In this study, we will take advantage of several unique databases from the Veterans Health Administration (VA), including the 1999 National Health Survey of VA Enrollees on HRQoL (n=887,775), pharmacy records from the VA National Pharmacy Benefits Management, and VA national administrative data (which include ICD-9-CM diagnosis codes), to conduct an effectiveness study among 440,218 hypertensive patients. Using t-tests of means and general linear modeling (GLM), we compared the SF-36 scale scores among patients who were on different single antihypertensive drug therapies. Results: Consistent with clinical trials, significant differences were found in patients' SF-36 scale scores across different classes of antihypertensive medications as well as between drugs within a single drug class. However, the study also revealed findings that have not been reported in clinical trials. Conclusions: Unlike clinical trials which include a selected number of antihypertensive medications, our study included almost all antihypertensive medications. Thus, the results of the study provide a more comprehensive understanding about the relationship between antihypertensive treatment and patients' HRQoL. Further research is, however, needed to examine the relationship between two antihypertensive medications treatment and patients' HRQoL.Aims: Spina bifida is the most frequent congenital malformation responsible of neurogenic deficiencies. Urinary and fecal incontinence are frequently encountered. Management of sphincter disorders represents a challenging and controversial goal. The aim of the study was to determine relationships among incontinence and quality of life (QoL) in patients with spina bifida. Methods: In a multicentric cross-sectional exploratory study between 2003 and 2004, 373 patients (178 females and 195 males, age 23.1±8.2 yrs) with spina bifida have fulfilled continence rating and QoL assessment using the SF-36 questionnaire in adults and the CHQ (Children Health Questionnaire-Children Form and Parent Form) in adolescents [10-17 yrs] . Relationships between incontinence and QoL were studied in a univariate and multivariate analysis. Results: Urinary and fecal continence was considered as satisfactory respectively in 68.7 and 89.3% of adults and 53.3 and 81.6% of adolescents. In adults, gender was the only predictor of QoL with significantly lower scores for women in physical functioning, vitality, social functioning, mental health, emotional role and bodily pain domains. Incontinence was not a significant factor associated with QoL. In adolescent self reported questionnaires, analysis showed lower scores for girls in all domains and a better self-esteem score associated with urinary continence. On the other hand, parents' form of CHQ showed that fecal incontinence was associated with significantly lower scores in physical role, behaviour, role emotional and behaviour, general health and bodily pain. Conclusions: From individuals' point of view, continence did not appear as a significant factor of QoL both in adults and adolescents contrary to closed relatives' point of view. These findings may influence incontinence management in patients with spina bifida. Aims: To evaluate the impact of visual impairment from diabetic eye disease on the function and well being of patients with diabetic retinopathy. Methods: Qualitative interviews were implemented at four eye clinics providing laser treatment in the UK. one hundred and twenty eight diabetic patients with proliferate diabetic retinopathy (PDR) and 118 diabetic patients with macular oedema (MO) were recruited and interviewed in clinic prior to their laser treatment. Of the 73 patients who participated in follow-up telephone interviews 2 weeks after their laser treatment, 34 had received their first time laser treatment and 39 were at later stages of disease progression and had received multiple treatments. Qualitative interview transcripts were summarized by type of response and results were grouped by question area. Results: While patients undergoing first time treatment reported greater levels of anxiety prior to their treatment and experienced more pain during their treatment, they also reported greater improvements in their visual impairments at two week follow up than did the multi-treatment patients. Multi-treatment patients reported lesser improvement of their visual symptoms following treatment, but more role limitations and activity restrictions due to greater amounts of visual impairment. Multi-treatment patients were also twice as likely to report no change in their quality of life at the 2-week-follow-up interview. Conclusions: The ability of patients with diabetic eye disease to perceive improvements in their condition following laser treatment is substantially influenced by the degenerative nature of their condition and complicated by the retinal damage from multiple laser treatments. Aims: Factors e.g. health service system, family budget, chronic disease and living habits are potential impacts on the QOL of meddle-aged population. By comparing QOL and other influential factors of the mid-aged people who are living in a coastal developed city (Guangzhou) and a inland developing city (Guiyang) of China, the study focused on QOL and other relevant factors affecting the middle-aged QOL in these two regions so as to set a preceding theoretical base for further research to enhance the QOL of middle-aged population. Methods: The sf-36 questionnaire and a self-developed questionnaire were applied with non-random sampling. Questionnaires were distributed to the middle-aged people aging from 40 to 59 who had lived in Guangzhou or Guiyang for/over 3 years by community committee members and graduate students, filled by the interviewee, and then collected by our staff. The data processing was performed by SPSS 11.0. The analytic methods included: descriptive analysis, reasoning analysis and ordinal regression. Totally 1200 questionnaires were distributed in which 1082 were recollected, and 1077 were considered as valid feedback (89.75%). Results: The QOL of the Guiyang group have higher grades than the Guangzhou group in four dimensions (RP, VT, RE, MH). The result is statistically significant (p<0.05). The Result indicates the factors including living location, age, family budget, smoking, tea drinking, breakfast, physical exercise, self-evaluation of health, chronic disease 2-week morbidity and health service system are relevant impacts on QOL. Conclusions: Middle-aged people living in different regions hold difference in QOL. Bad living habits may lower QOL; and good habits like physical exercises and tea drinking can help keep people healthy and improve QOL. Therefore, health education for middle-aged people to form healthy living and eating habits plays a significant role in protecting their health.Aims: The Longitudinal Investigation of Depression Outcomes (LIDO)was a study including 6 centers worldwide (Brazil, Australia, Russia, Spain, USA and Israel) . Several measures were made during this study. One of them was quality of life using WHOQOL-bref. The main aim of this present research is to evaluate cross-cultural validity of this measure between Brazil and other countries involved in LIDO Study. Methods: To range the purpose of the present study we use a differential item functioning (DIF) analysis from Rasch model. We investigated data from six countries, assessing quality of life and economic aspects of undiagnosed depression among primary care patients. Results: We found that Brazil as a factor shown DIF in 13 of 26 items of WHOQOL-bref. The most affected domains were Psychological with DIF in 4 of 6 items, followed by Environment which shown DIF in 4 of 8, whereas, Physical domain shown DIF only in 2 of 7 items and Social domain 1 of 3 items. Conclusions: These finding may be accounted by some cultural differences related to social economic resources (environment domain) and to the meaning of psychological suffering (psychological domain), which may differ among these cultures studied. Although WHOQOL-bref was constructed by some different cultures in the same time to ensure cross-cultural validity, it can not be demonstrated by this sophisticated statistical method. The quality of life may vary more strongly within some cultures than others. In this respect, Brazil is a country with a broad diversity of cultures, which can have a strong influence on its concept and understanding of quality of life. Aims: To assess which health beliefs predict and explain satisfaction with the facial health state of patients undergoing surgery for basal cell carcinoma(BCC). Methods: A disease-specific questionnaire was developed. Questions and items related to health perceptions like susceptibility for BCC and expectations were devised. The questionnaire was administered parallel to a clinical trial in which the effectiveness of two surgical treatments for BCC was compared. Patients were interviewed pre-operatively and 6 months after surgery. Data of the questionnaire were analysed by principal component and factor analysis. The internal consistency of the scales was tested using Cronbach's a. Regression analysis was used in which a health perception as an indicator for satisfaction at 6 months was regarded as dependent variable. Results: At baseline, 222 patients were interviewed. At 6 months after surgery, 184 patients completed the questionnaire. Internal consistency of the scales was between satisfactory and very good (alpha between 0.76 and 0.90). Results of the regression analysis show that satisfaction as measured by postoperative worrying, susceptibility and fear of developing a new BCC at other facial sites can be predicted by pre-operative health beliefs. In addition, some patients experience an increase in perceived susceptibility, fear of recurrence of BCC on the same site or fear of developing a new BCC at other facial sites. Conclusions: Both pre-and post-operative perceptions predict and explain for a substantial part the extent to which patients are satisfied with their facial health state 6 months after surgery. Administering a short questionnaire at the start of the treatment period will give physicians a better understanding of how patients experience this skin disease. It will also help them to adjust information about BCC and its consequences to the needs of the patient.Aims: Especially since the introduction of PSA-screening, a much higher proportion of prostate cancers has been diagnosed in early stages than before. We aimed at assessing the currently unknown impact of receiving a diagnosis of prostate cancer through PSA-screening on patients and on their evaluation of dysfunctions that might be experienced after treatment. METHODS: Participants of a prostate cancer screening trial (ERSPC) completed a questionnaire on health (SF-36 Mental Health and Vitality, EQ-5D VAS for self-rated health) before screening, and, if diagnosed with prostate cancer, again before treatment and 6 months afterwards. Results: Before screening 3780 screen participants (response 90%) completed the questionnaire on health. Of 82 men who received a prostate cancer diagnosis, 52 (response 63%) completed 2 additional assessments. Compared to pre-diagnosis, mental and self-rated health worsened significantly after the prostate cancer diagnosis. After initiation of treatment, mental and self-rated health improved, albeit not to the original level. One month post-diagnosis, men's view on dysfunctions typically occurring after prostate cancer treatment became less negative. Non-response does not seem to have affected these results. Conclusions: We found a significant negative impact of receiving a prostate cancer diagnosis through PSA-screening on patients' mental and self-rated health. Clinicians are aware of the favorable prognosis of early detected prostate cancer but may not always realize the impact that the diagnosis nonetheless has. We recommend clinicians to share their knowledge on the prognosis with patients. Additionally, we showed that a first assessment of health-related quality of life after prostate cancer diagnosis cannot be considered to represent a situation without cancer. This finding may have implications for research.Aims: This study evaluated the multidimensional constructs of the EORTC Quality of Life Questionnaire (QLQ-C30) in patients with cancer, employing not only commonly used multitrait scaling analysis and interscale correlations, but also factorial and multidimensional scaling (MDS) analyses. Methods: A total of 334 Korean patients with stomach, lung, liver, breast, or cervix cancer participated in a cross-sectional study. All patients completed the QLQ-C30. Results: On the multitrait scaling analysis, the cognitive functioning scale did not meet item convergent and divergent validity. On the interscale correlations, physical and role functioning scales were highly correlated this was also evident in the factorial analysis. The MDS showed that items in each of the social, emotional, global health status/quality of life, and nausea/vomiting scales were clustered close together and far from the other items. Conclusions: The authors conclude that evaluation of the QLQ-C30 in four ways produced results that supported the original hypothesized constructs. However, the physical functioning and role functioning scales were not distinctive, and the cognitive functioning scale was somewhat problematic in the Korean population with cancer. Aims: The Quick Inventory of Depressive Symptoms (QIDS-SR16) is increasingly used in depression studies. We sought to determine the reliability, validity, and responsiveness of a daily version of the QIDS-SR16. Methods: Outpatients with non-psychotic MDD (n=78) received fluoxetine, were randomized to two assessment groups and followed for 28 days. The standard arm (SA) completed a battery of self-report (e.g., weekly QIDS-SR16) and psychiatric assessments (e.g., HAM-D-17) at Baseline, Days 7, 14, and 28. The daily arm (DA) did the same weekly evaluations as the standard arm, plus a daily battery that included the daily QIDS-SR16 among other measures. We modified the wording of the QIDS-SR16 including instructions, questions, and response options so that all the items would be relevant for daily administration, instead of weekly. Results: The intra-class correlation coefficient was 0.64 for the daily version of the QIDS-SR16, which was comparable to the ICC for the QIDS-SR16 weekly version when completed in the clinic. Cronbach's a was 0.81 for the daily version on Diary Day 1. The reliability of the average of 7 daily measurements was 0.93. The mean score on Diary Day 1 was 13.54 (SD=5.18), which corresponds to a score of 17 on the HAM-D-17. The daily version of the QIDS-SR16 was able to detect both onset and response significantly faster than the self-report QIDS-SR16 administered in the clinic. Growth curve analysis indicated that the slope of change of the daily QIDS-SR16 during the first 7 days of treatment was the most rapid of the measures used in the diary. Conclusions: The daily version of the QIDS-SR16 has comparable reliability and validity to the weekly version. The QIDS-SR16 may be a particularly useful instrument in daily studies to measure early onset antidepressant effects.Aims: To develop and evaluate the psychometric properties of a dialysis satisfaction questionnaire in French language. Methods: The satisfaction questionnaire wasd built using items banking, based on the CHOICE Satisfaction questionnaire, and Satisfaction questionnaire and the Dialysis Patient Questionnaire. This work has been implemented in 2 steps: A. Translation and adaptation to French context. Items were translated by 2 independent translators. They were adapted to French setting; content and face validity was appreciated in cooperation with dialysis patients and nephrologists. The tool was adapted to the French setting as follows: 1. 38 items dealt with aspects of the dialysis patient experience; 2. 3 overall satisfaction items; 3. 1 open-ended question. Satisfaction scores are standardized from 0 (poor) to 100 (excellent). B. Estimation of the questionnaire psychometric properties. The sample consisted of all 969 adult prevalent patients treated in the 12 dialysis structures of Lorraine region on February 1st 2004. All of them were mailed a questionnaire. Results: Response rate was 44.3%. Mean age of patients was 65.9 years; 67% were men and 61% retired. Mean length of dialysis was 4.5 years. Principal components analysis evidenced 7 dimensions: organization of care, medical problems management, information given by nephrologists, relation between nephrologist and general physician, hospital environment, accessibility and overall satisfaction. Cronbach's a coefficients were greater than 0.91 for all of the dimensions. Correlation coefficients between the 7 specific dimensions ranged from 0.51 to 0.75 , while those between the 6 specific dimensions and overall satisfaction dimension ranged from 0.43 to 0.62. Mean scores of satisfaction varied from 53.9 (relation between nephrologists and GP) to 74.1 (overall satisfaction). Older, peritoneal dialysis patients tended to be more satisfied than younger, hemodialysis patients. A significant variability from one care team to another was evidenced. Conclusions: Dialysis patient satisfaction questionnaire in the French version exhibits satisfying psychometric properties. Aims: Few measures assessing clinician satisfaction are currently available. This study evaluated the psychometric properties of 2 new questionnaires assessing both clinician and patient satisfaction with treatment for non-malignant pain. METHODS: Patients enrolled in a randomized, 14-day, double blind, placebo controlled, multi-center, 4 arm (hydromorphone HCl extended release capsules (HHER), fentanyl transdermal system, respective placebo groups) trial completed the 2-item Subjective Ease of Medication (SEM) questionnaire, and physicians completed the 6-item Clinician Satisfaction (CS) questionnaire at Day 14. The BPI, EQ-5D, VSH sleep questionnaire, and other PRO measures were also administered. We evaluated internal consistency, reliability, construct validity, and known groups validity for these 2 new questionnaires. Results: Data are available for 363 patients. In general, patients were 50 years old, Caucasian, female, reported moderate or severe pain, and had osteoarthritis or low back pain. Variable clustering resulted in a single SEM scale of ease of use and convenience, and 2 sub-scales (Overall Satisfaction, Dosing/Titration) and an overall score for the CS. Internal consistency was acceptable (Cronbach's alpha coefficients ranged from 0.82 to 0.92). Construct validity was demonstrated on the basis of correlations between related items and scales. Known groups validity was upheld, as patients who withdrew from the study reported lower mean scores (lower satisfaction) on the SEM and physicians reported lower CS scale scores than those who completed the study (p<0.05). Summary measures, such as EuroQol (EQ-5D), have been used as standard population measures of health status, and as utility measures in cost-utility analyses. Performance of empirically derived value weights to predict health status and utilization remains problematic. Development of US EQ-5D1 weights allows evaluation of each summary measure compared with disease specific measures. Methods: In a study of hospitalization for acute coronary syndrome (ACS), we compared domain and disease specific measures to EQ-5D to predict mortality and rehospitalization in 5 mid-Michigan hospitals. Consenting patients were interviewed by telephone at 1 month baseline (n=525), 3 months (n=440), and 8 months (n=388) post-discharge. We compared EQ5D vs. depression (CESD), activity status (ASI) and ACS symptoms (QLMI-S symptoms subscale), to predict 2-yr mortality, and re-hospitalization at follow-ups, in logistic regression models adjusted for age, gender, race, comorbidity count, ejection fraction and use of aspirin and/or beta-receptor blocking medicines post discharge. Results: Results show that baseline ASI, QLMI-S and CESD scores did not add information. Low baseline EQ5 was a strong predictor of death (p=0.03) and showed moderate association with hospital readmissions at 3 months (p=0.02). EQ5D had a very strong association with readmission from 3 to 8 months (p<0.001); QLMI-S (p=0.006) and ASI (p=0.003) were significantly but less strongly associated with readmission. Analysis using US EQ5D1 weights is in process. Conclusions: EQ5D is a good predictor of mortality and hospital re-admission in ACS patients and performs as well as domain-specific measures. EQ5D and EQ5D1 weights will be used to model mortality and hospital re-admissions and to describe cost-utility of consistent medication use in ACS.

To critically review measures of patient-related activity used to assess outcomes in critical (intensive) care. Specifically, to investigate the conceptual basis and psychometric properties of existing measures of patient dependency/severity and nursing dependency/intensity/workload. Methods: Critical review of the literature to identify measures, systematic evaluation of the psychometric properties of existing measures and critical analysis of the conceptual basis of measures in this area. Results: We identified several measures that are commonly and widely used in critical care settings to evaluate outcome. However, (i) most have minimal psychometric evidence, i.e. the majority have not been evaluated for inter-rater reliability, internal consistency, criterion-related validity or responsiveness; and (ii) there is a distinct lack of clarity about the concepts that these instruments purport to measure, i.e. although patient dependency, nursing dependency, patient severity, workload, patient need and nursing intensity are defined as distinct constructs, they are often used interchangeably with little analysis of the similarities, differences and interrelationships between these concepts. Conclusions: There is a need for a more scientifically robust approach to evaluating patient-based outcomes in critical care. Although several measures have been used to evaluate the impact of patient-related activities, few have been rigorously developed or evaluated using well-established psychometric methods and standards. Moreover, there is an urgent need for a clearer definition of the concept of patient-related activity in critical care to guide the development, evaluation and interpretation of measures in this area. Aims: To assess inter-rater agreement of HUI2 measurements in survivors of cancer during childhood in Honduras, El Salvador, Nicaragua and Panama. Methods: Patients and their parents completed Spanish-language interviewer-administered HUI questionnaires (43-and 46-items, respectively). Physicians answered the Spanish-language 15-item self-complete HUI questionnaire. Agreement between patients and parents and physicians, for HUI2 single-attribute and health-related quality of life (HRQL) utility scores, was assessed by t-test of differences in mean scores and single-measure one-way intra-class correlation coefficient (ICC). Differences in single-attribute scores >0.049 and HRQL scores >0.029 are clinically important. Results: Of 211 patients surveyed (aged 3.4-25.8 years at time of study, 56% males), there were complete HUI assessments for 191 patient/ parent pairs and 192 patient/physician pairs. Patient mean scores were lower than parent (diff=0.116, p<0.001) and than physician (diff=0.114, p<0.001) for emotion, and than physician (diff=0.065, p<0.001) for HRQL. Mean HRQL was lower for parent than physician (diff=0.042, p=0.005). HRQL agreement was less than moderate (ICC<0.41) between patient/parent and patient/physician pairs. There was moderate or better agreement (ICC>0.40, p<0.001) between both pairs for sensation, mobility and self-care. Agreement between both pairs was less than moderate for emotion, cognition and pain. Conclusions: Patient and parent assessed HRQL was clinically importantly lower than physician assessment. Patient assessed emotion was clinically importantly lower than parent and physician assessments. Parent and physician HUI2 reports should not be considered interchangeable with patient reports, especially for aspects of health not objectively assessable. The (2002). An evaluative instrument, the ePRO, was developed. It contains 39 items covering 8 dimensions: conceptual and measurement model (7 items); cultural and language adaptation (2); reliability (8); validity (6); responsiveness (3); interpretability (3); respondent and administrative burden (7) and alternate forms (2). The respondent expresses his/her agreement with each item's statement in a 4 options Likert scale ranging from 'Strongly disagree' to 'Strongly agree'. In a pilot test, groups of 4 evaluators independently used it in the assessment of the Spanish versions of 5 generic PRO questionnaires (MOS SF-36, Nottingham Health Profile, COOP-WONCA charts, EuroQol-5D and EORTC QLQ-C30). Results: The ePRO instrument showed high internal consistency (Cronbach's a=0.96). Inter-rater agreement (Intraclass Correlation) varied between groups (0.51-0.89). Global ratings ranged from 'Recommended' to 'Strongly recommended', and they were associated, though not significantly, with time since questionnaire development (r=0.79) and number of publications retrieved (r=0.56). Conclusions: The ePRO instrument is a new standardized tool for the evaluation of PRO measures, that can assist the choice of the most appropriate questionnaire for a given purpose. Funded by: Instituto de Salud Carlos III (Red IRYSS G03/202) and Generalitat de Catalunya 2001 (SGR0405) Aims: The purpose of this study was to develop the new QOL questionnaire for child caring father. Methods: 80 child caring fathers participated in this stydy. Our new original selfadministered questionnaire consistd of 53 questions divided into 9 categories. Results: Cronbach's alpha coefficients of our new questionnaire were high enough to accept for clinical use: 0.80 in connection with children, 0.79 in psychological problems, 0.78 in physical function, 0.73 in marital life, 0.72 in social participation, and 0.71 in living condition etc., respectively. Our questionnaire contained 16 main factors and cumulative contribution was 0.84. There were significantly positive correlations between total QOL and child caring r=0.72 (p<0.01), between total QOL and love for children r=0.54 (p<0.01), and between total QOL and connection with children r=0.50 (p<0.01), respectively. The fathers whose child caring scores were excellent showed the significantly prominent total QOLs (p<0.05).There was significantly positive correlation between connection with children and love for children r=0.34 (p<0.05). The more their self awareness as father elevated, the more their QOLs improved. Conclusions: These findings indicated that our new QOL questionnaire for child caring father was high enough to accept for clinical use. Although we have the tendency to understand that mothers should mainly play the important roles for child caring, these our findings indicated that fathers who play the more active role for the connection with children can possess the significantly excellent self awareness as father. We should pay more attention to the father's roles on child caring in future by using our new QOL questionnaire for child caring father. Aims: EGFRI block tumor growth by inhibiting cell growth pathways and commonly cause an acne rash on face and trunk. Clinicians may consider this minor, but patient (pt) perception of the rash is unknown. Intriguing relationships have been found between rash and survival; however QoL impact of the rash is unknown and so evaluated here. Methods: NCIC-CTG PA.3 is a RCT comparing gemcitabine (G) to G+EGFRI erlotinib (E) in pts with advanced pancreas adenocarcinoma. About 285 pts were randomized to E+G. QoL was assessed using EORTC QLQ-C30 and NCIC-CTG guideline QoL response analysis: a change score ‡10 points from baseline defined as clinically relevant. Mantel-Haenszel test for trend enabled exploratory comparisons of proportion of pts with improved/ stable/worse QoL depending on rash development and severity. Results: More pts who had a rash on E had improved Global QoL. Those with a more severe rash (grade ‡2 vs. 0-1) showed a trend to better Global QoL response: improved/stable/worse 57/14/29% vs. 44/14/42% (p=0.05). Rash development and severity were not associated with deterioration in QoL response in Social, Emotional, or any other domain/item. Conclusions: Self-assessment suggests that the acne rash associated with E may not impair QoL in pancreas cancer pts. In fact improvement in Global QoL response was seen in pts who develop a rash, with a trend to higher proportion of improvement with a more severe rash. These outcomes may guide management of cancer pts with EGFRI therapy. Aims: The objective of this work is to develop an evidencebased, generalized, a priori method to estimate the utility of multiple concurrent conditions. Methods: We evaluated a function that combines the individual utilities u1 and u2 of two conditions to derive a single estimate of the utility of the comorbidity u1,2. u1,2=(1)k)*minimum (u1,u2)+k* (u1*u2) This function is built from two possible estimates of comorbidity: the minimum of the two utilities, which assumes no cumulative effect of having both conditions and that the worse one prevails; vs. the product of the two utilities, which assumes a cumulative effect, proportional to the magnitude of the two utilities. To evaluate this approach, and estimate the value of (k), we used the Health Utility Index (HUI3) as a measure of utility as reported by 131,535 respondents to the Canadian Community Health Survey (CCHS) for 2000/2001. We estimated the optimal value of k with the method of least square error across all reported combination of two co morbid conditions. The possibility that the cumulative impact of conditions varies according to the severity was studied, by dividing the conditions into terciles according to their severity. Results: There were 281 combinations of two conditions reported on the CCHS. A comorbidity coefficient (k) of 1 offered the best fit between the observed and derived HUI3 value. When the co morbid pairs were divided into terciles according to level of severity, estimates of (k) were similar for the low-, mid-and high-severity terciles. Conclusions: This analysis of the HUI3 on the CCHS suggests that the utilities of separate conditions can be combined, a priori, to estimate the utility of their comorbidity. The level of severity of the conditions does not appear to play a role in the optimal value of (k). The simple rule can be applied iteratively to calculate utility associated with more than two conditions. This method can be applied in quality of life studies.Aims: Visual impairment and associated quality of life is more pronounced in developing nations. This study was to determine the difference in the quality of life between rural and urban dwellers who had their vision corrected with spectacle lenses for 3 months. Methods: Four hundred and ninety subjects drawn from urban and rural settings in Nigeria were enrolled in the study. The urban population included 225 subjects who had a comprehensive eye examination before their vision correction with spectacle lenses while the rural population included 265 subjects who obtained their spectacle lenses from untrained eye practitioners or from the open market. All subjects had worn their lenses for 3 months and were free of eye diseases. The NEI VFQ-25 was administered by trained interviewers to all subjects at their places of work. The principal outcome examined was quality of life score. Results were presented in means and percentile values, and an unpaired student t-test was used to determine statistical significance of differences between the two populations. Results: On a 0-100 scale, the mean QOL scores for the urban population was 75.8 while that for the rural was 57.5. The difference between these scores was statistically significant (p<0.05). Seventy nine percent of the urban population and 44.9% of the rural population scored 80 and above. General Health subscale showed the biggest difference in scores,followed by general vision, Near activities, Mental health in Vision specific and Ocular pain (p<0.05). There was no significant difference in scores in Driving, Color vision and Peripheral vision subscales (p<0.05). Conclusions: There is a lower quality of life among rural dwellers who could not afford an adequate vision correction. If adequate vision correction is made available to rural dwellers in developing nations, their quality of life will definitely be improved.Aims: MPM patients present with a poor prognosis, hence maintaining QOL is important. Little is known about QOL during initial treatment of MPM patients. We report the impact of treatment with raltitrexed and cisplatin (arm A) on QOL when compared to cisplatin (arm B). Methods: Patients with histologically proven unresectable MPM, not pretreated with chemotherapy, WHO PS O-2, and adequate hematological, renal and hepatic function were randomly assigned to receive cisplatin 80 mg/m 2 iv on day 1, without or with preceding infusion of raltitrexed 3 mg/m 2 . QOL was assessed with the EORTC QLQ-C30 and EORTC LC13 tools. Assessments were conducted at baseline and just before every other treatment.Results: About 250 patients were randomized; 80% male, median age 58 years, WHO performance status 0, 1, 2 in 25, 62, and 13% of cases. Clinical results found arm A superior to arm B with regard to overall survival (p=0.048). QOL questionnaire compliance was 91% at baseline, exceeding 85% throughout the treatment. Both arms showed substantial impairment in baseline scores for all QOL scales. The global QOL scale was comparable at baseline on both treatment arms (p=0.848) and at no point was any significant difference seen on this endpoint. Patients reported high levels of fatigue, and major limitations in physical, role and social functioning, although these were stable. None of the scales of the QLQ-C30 or LC13 showed any statistically and clinically significant differences over treatment time. Conclusions: This study provides key information about the QOL of MPM patients treated with chemotherapy. In many respects, QOL was positively maintained over the treatment.The clinical result of improved survival outcome with cisplatinraltitrexed is supported by a lack of detrimental effect of the combination on QOL.Aims: To assess the value of using 3 distinctly different methodologies, within one quality of life (QoL) study. Methods: We conducted a cross-sectional study to compare the QoL of youth (13-18 yrs) and young adults (23-33 yrs), with cerebral palsy, spina bifida and acquired brain injuries from 6 rehabilitation centers. We analyzed health services utilization data (n=1,606), survey data from a sub-set (n=290), and semi-structured qualitative interviews from a further sub-set (n=30+30 parents). In the health services data, the proportion admitted to hospital over 4 yrs was used as a marker of QoL, with higher proportions representing poorer QoL. A chi square test was used to compare the proportion of youth vs. adults hospitalized. The survey used the Assessment of Quality of Life (AQoL) and the Health Utilities Index (HUI) to measure QoL. High scores indicate best QoL. An un-paired t-test was used to compare youth vs. adults. The interviews were analyzed to identify themes and explore differences between youth and adults. Better health care, social and living situations were considered indicative of better QoL. A social interactionism analysis was used. Results: The health services utilization data identified more admissions among the youth (p<0.0001) suggestin the youth had worse QoL. The survey data showed similar AQoL and HUI scores for youth and adults (p=0.90). The qualitative analysis identified worse QoL among the adults. Thus, each method produced radically different conclusions. Because of the richness of the qualitative data, we are most convinced that its conclusions are accurate. Conclusions: The integration of results from 3 diverse methods generated a detailed picture that greatly improved the accuracy of the conclusions. Not only did this bridging of methods identify lower QoL among the adults, but it also provided sufficient information to suggest solutions. This approach is recommended in future examinations of QoL, particularly in small and or unique populations. . Patients with early stage disease and those who received single modality of therapy scored better than advanced stage disease and patients who received combined modality. Although patients did not articulate any problems to the treating surgeon, a significant number of patients scored poorly on various domains in the QOL questionnaire. Apart from treatment related physical problems, emotional disturbances and social problems were also reported. Conclusions: Quality of life questionnaires can provide vital information helping treating surgeons to focus their rehabilitative efforts. Even after completion of therapy, patients have various physical and emotional problems which need to be detected and specialized intervention may be required. Aims: To determine the correspondence in direction of change in 2 ways of assessing QOL: changes in McGill Quality of Life Questionnaire (MQOL) scores (Global; Overall; Physical; Psychological; Existential) and those reported in a semi-structured interview. Correspondence was determined for the overall group and also for each participant. Methods: MQOL was completed within 48 hours of admission to one of 3 palliative care units and again 7-8 after admission. After this last completion, those who had energy and were willing (n=27) were asked to tell the interviewer what had changed since admission to the unit. Although many patients on the units were physically or cognitively incapable of participating, the demographic information for those who completed the interview and those who completed only the two questionnaires was not significantly different. Interviews were audiotaped and transcribed verbatim. MQOL scores were considered to have changed if there was a change of 0.5 on the scores which all had a possible range of 0-10. This was chosen as it represents 5% of the scale range, which has been reported to often correspond to a minimal clinically important difference, and is greater than the standard error for MQOL scales in this study (0.17-0.33). Blind to the MQOL scores, interviews were rated, for each of the MQOL domains, as indicating (1) improvement; (2) deterioration; (3) no change; (4) rater not sure; or (5) domain not mentioned. For each MQOL score, we determined the percent of the group who showed improvement, no change, or deterioration in MQOL scores and in interview ratings. To determine the correspondence at an individual level, the percent of people whose MQOL scores changed in same direction as rated in the interview was determined. Results: On a group level, MQOL scores and interview ratings both indicated improvement, no change, or deterioration 80-83% of the time for the different domains. At the individual level, the correspondence was less, ranging from 53% for physical symptoms to 70% for the MQOL Total Score. Conclusions: This provides some new evidence that MQOL scores are responsive to changes in the QOL of groups. The results support the use of MQOL on an individual basis only with caution. Aims: There is increasing attention being paid to the qualitative research process by the FDA and the pharmaceutical industry. Rigorous qualitative methods are essential components when designing valid and reliable patient-reported instruments. The objective of this research was to use qualitative methods to initiate the development of a patient-reported measure of treatment satisfaction with medication (TS-M) for hyperpigmentation (HP), specifically for solar lentigines (age spots) and melasma (pregnancy mask). Methods: A conceptual model, based on a literature review, formed the basis of a structured focus group discussion guide. To explore and elaborate on key content, seven focus groups, an expert panel of treating physicians, and in-depth interviews were conducted. The expert panel (n=10) was convened to validate the findings from the focus group members (n=56). In-depth interviews (n=19) further refined the item pool. Content analysis was conducted on verbatim transcripts of the focus groups and the expert panel.Results: Qualitative analysis of the focus group transcripts was used to revise the conceptual model, describing the domains of Effectiveness, Convenience, Physical Attributes, Psychosocial Issues, Side Effects, and Overall Satisfaction. As expected, Effectiveness and Side Effects were the most influential domains for TS-M. Participants also identified Physical Attributes of Treatment (e.g., how the treatment feels on the skin) and Psychosocial Issues (e.g., self-consciousness) as domains that are specific to the measurement of HP TS-M. The expert panel supported and refined the qualitative information from the focus groups. Finally, cognitive debriefing methods were used to reduce the item pool from 100 to 38 items and prepare the HP specific TS-M measure for psychometric testing. Conclusions: In this study, qualitative data collection and analysis refined the conceptual model for HP TS-M and resulted in a rich item pool reflecting content of particular interest to people with HP.Aims: As erectile dysfunction (ED) impacts physical and psychological factors, it is important to assess both aspects in ED patients seeking to improve their sexual experience. Our objective was to develop a self-administered instrument to encompass the functional and emotional components of a quality sexual experience in men. Methods: Literature review, advisory boards, focus groups, and in-depth interviews in four countries (Germany, Spain, United Kingdom, United States) were used to identify concepts that defined a quality sexual experience. These potential concepts were explored using input from three groups: (1) males prescribed a PDE 5 inhibitor, (2) males with low to moderate confidence in sexual function, and (3) males with high to very high confidence in sexual function. As the erectile function (EF) domain of the International Index of Erectile Function instrument clearly assesses EF, these items were modified based on patient feedback; items for the remaining domains were generated from primary research. Physicians from the four countries provided input to refine the item pool that was finalized on cognitive debriefing. Results: Three concepts that defined a quality sexual experience were identified: improvement in EF, sexual relationships and satisfaction; 25 items were subsequently generated. Differences in these concepts across countries were not observed. Physicians confirmed appropriateness of the concepts, item content, 4week recall period, and 5-point Likert scales. Cognitive debriefing led to item reduction, item wording and refinement of response options. Conclusions: The 15-item draft SEX-Q is a self-administered instrument addressing functional and emotional aspects of a quality sexual experience. On-going psychometric evaluation of the SEX-Q may lead to further item reduction and may prove the SEX-Q to be valid in assessing improvement in EF, sexual relationships and satisfaction. Aims: We assessed the feasibility of including HRQL assessments in the routine care of pre-and post-heart transplant patients. Methods: The pilot study was undertaken in anticipation of a randomized controlled trial to be conducted at the cardiology division at the University of Alberta Hospital, Edmonton, Alberta. The pilot sample includes pre-and post-heart transplant patients. At baseline, before the visit, patients completed: Health Utility Index Mark 2 (HUI2) and Mark 3 (HUI3), Hospital Anxiety and Depression Survey (HADS), and adherence measures to medication and exercise. After the visit, a satisfaction questionnaire was completed. Patients were randomly assigned to intervention group (completion of touch screen HUI2 and HUI3, and feedback to clinicians) and control (completion of touch screen HUI2 and HUI3 but no feedback to clinicians). Feedback involved a graphic representation, HUI score card. As part of the intervention, training sessions were organized with the clinicians. The interviews were tape recorded. Results: To date, 19 patients were asked to participate, 14 agreed, 2 were reluctant to tape recording, 3 refused to participate. Patients did not need to come to the clinic earlier.The average time for completion was 15 min. After the first visit patients became comfortable with the tape recording. Weekly training sessions were held for members of the team but these sessions were difficult to organize with the cardiologists. Five of seven cardiologists and most patients were willing to participate. Results displayed on HUI score card were easy to interpret. Conclusions: Routine use of HRQL has been shown to benefit patients in several studies in oncology. We confirmed the feasibility in routine cardiac care. The inclusion of HRQL measures in routine care may improve clinical practice.

Kiyako Takai, Nursing, Nagasaki University, Nagasaki, Japan; Yuki Mase, Nursing, Kanagawa University of Human Services, Yokosuka, Kanagawa Pref., Japan; Akemi Terasaki, Nursing, Nagasaki University, Nagasaki, Japan Aims: In our previous report we investigated the process of empowerment for aphasics belonging to a self-help group, and their efforts to achieve better lives with aphasia based on subjective experience, from onset to current status. The present study was conducted to identify the factors influencing this process. Methods: (1) Subjects: The study population comprised 30 aphasics (16 males, 14 females), from a self-help group for aphasics,who had been living with their condition for 5 years or longer, and who cooperated voluntarily in the study. (2) Method: Each subject was interviewed as per a semi-constitutional interview guide we prepared independently based on the concept of the trajectory of illness proposed by Strauss (1984) .(3) Survey period: February 2002-March 2003. (4) Analytical method: Based on word-for-word records, survey findings were analyzed both qualitatively and inductively. (5) Ethical considerations: The established ethical procedures of the study institutions were followed. Written informed consent was obtained from the representative of the self-help group and from each interviewee. Results: Regarding prospective factors that promote the acceptance of aphasia and support the empowerment of aphasics, and were identified for all time frames, i.e., during hospital stay, after discharge, and after joining the selfhelp group. For the period after discharge, and were identified as priorities. For the period after joining the self-help group, was identified. As for interference factors was identified for all the time frames, and was identified for the periods after discharge and after joining the self-help group. Conclusions: In summary, the importance of the influence of the environment in which the aphasic is living was revealed as highly significant.Aims: This qualitative study assessed hepatitis C (HCV)-infected methadone maintenance treatment (MMT) patients' comprehension and execution of Standard Gamble (SG) and Rating Scale (RS) utility assessments. Recent changes in treatment guidelines have led clinicians to call for individualized risk-benefit assessments by MMT patients and physicians prior to commencing treatment for HCV. Studies have found that 65-90% of MMT patients are HCV-infected and HCV incidence rates are 10-20% per year. Methods: MMT patients (n=23) were asked to evaluate hypothetical health states associated with mild and moderate HCV illness and moderate and severe side effects due to HCV treatment. Other measures included the SF-12 and CES-D 10. Results: The quantitative and qualitative responses during the interviews indicated that most patients were willing to ''play the game'' of the SG. The poor physical and mental health (mean PCS=39.44; MCS=31.16) and high prevalence of depressive symptoms (mean CES-D 10 score=18; ‡10 =83%) in this population made it challenging for patients to distinguish their current health from the different hypothetical health states. Interviewer prompting to compare the hypothetical health states to current health was helpful. Several patients recalibrated previous responses to adjust for more or less severe health states as they were presented, demonstrating that they understood the questions and were able to perform the tasks in a thoughtful manner. Utility assessments are uncommon in this population and some authors doubt whether MMT patients can provide health assessments due to the need for abstract thought (hypothetical health states) and numeracy. We found that utility assessments can be completed by MMT patients. Conclusions: Using a guided and flexible administration approach provided greater insight into preferences. Further research is warranted to evaluate the use of this approach in other disadvantaged populations. P-340/1809/AN ITEM RESPONSE THEORY MODEL FOR VISUAL ANALOGUE SCALES Karl B. Christensen, Copenhagen, Denmark; Mark Kosinski, QualityMetric Inc., Lincoln, RI; Jakob B. Bjorner, Quality-Metric Inc, Lincoln, RI Aims: To develop an item response theory (IRT) model for visual analogue scale (VAS) items. Such a model will allow VAS and multi-item scales to be calibrated to the same metric and permit a better description of the measurement properties of VAS items. Methods: We treat the VAS items as categorical items with numerous (e.g. 101) categories. A model for such items can be achieved by extending an IRT model proposed by Andrich and Luo (2003) . Instead of estimating the threshold parameters directly, the principal components of the thresholds are estimated, from which threshold estimates are then recovered. This model can be estimated even when the number of thresholds is large and many categories are unobserved. The arbitrary choice of number of categories can be handled by including a discrimination parameter in the model. We have developed a marginal maximum likelihood estimation procedure for the VAS IRT model and a preliminary test of fit. We used the model in the reanalysis of a clinical trial data set and a general population study. Results: A model with two threshold components (mean threshold and distance between thresholds) showed reasonable fit. The model revealed that the VAS scale provided more information than a single item with five categories, but less information than e.g. a three-item scale. In line with these results, a three-item pain scale showed better responsiveness in the clinical trial. Based on the IRT model we developed a cross-walk table which successfully predicted group mean VAS scores from the categorical items. Aims: To determine whether the WHODAS-2 can be considered sufficiently unidimensional for applications; and to determine its measurement precision across the spectrum of disability. Methods: The NCS-R was a nationally representative household survey with a 71% response rate conducted between 2/2001 and 12/2002 of 9282 adults who were administered a psychiatric diagnostic interview; 5692 also received a disability assessment. We mapped items from the NCS-R to their WHODAS-2 progenitors. We used a bi-factor model to determine whether standardized loadings on a general factor were salient when the six sub-domains of the WHODAS were included using MPLUS, incorporating WLSMV estimation of the polychoric correlation matrix. We then used the Graded Response Model using PARSCALE. Results: These data were consistent with a sufficiently unidimensional structure of the WHODAS-2, as almost all items had large loadings on the general factor. A few items addressing employment performance had negligible information curves but were retained in the final metric to avoid a floor effect in the NCS-R sample. The WHODAS-2 had poor measurement precision at the extremes of severe disability and negligible disability, with good measurement precision for modest amounts of disability. Conclusions: Item response theory provides a valid scoring approach to the WHODAS-2. Measurement precision varied widely across the range of disability measured by the instrument, implying that error terms should be considered in models using this score as either a covariate or an outcome variable. The WHODAS-2 may prove useful for assessing disability levels in epidemiological studies.Aims: The WHOQOL-Bref (Bref) is a generic quality of life measure, which has been developed simultaneously in many cultures and languages. With the data from this study, it is our goal to further compare the Bref (items and subscales) across 23 cultures worldwide.Using the RUMM 2020 software, we conducted a Rasch analysis to explore the domain structure of the six countries with regard to one-dimensionality. Methods: The sample consisted of 11,239 subjects from an obtained survey of adults. Sick and well respondents were sampled from the general population, as well as from hospital, rehabilitation and primary care settings and with respect to quotas of important socio-demographic variables. Results: We found a separation index of 0.89 from the total scale and it didn't fit the unrestricted model (v 2 =36.56; p<0.0001). As we know that quality of life is not a one-dimensional construct, we re-analysed each domain as an independent scale, so we found a separation index of 0.53; 0.61; 0.81; 0.68 for Physical, Psychological, Environmental and Social domains, respectively. Of the Bref's 26 items, only 6 (pain, medication, environment, sleep, sex and negative feelings items) needed to be rescored because they had disordered thresholds. Many items showed fit residuals greater than 2, which indicate that they didn't fit the Rasch Model. The items locations in Physical, Psychological, Social and Environment subscales ranged from )0.37 (mobility) to 0.74 (medication); )0.34 (body) to 0.83 (negative feelings); )0.38 (support) to 0.81 (sex); )0.36 (home) to 0.52 (environment) respectively. Conclusions: We conclude that as Rasch analysis uses only one-parameter IRT model, this data set can be better understood by using a more complex model taking into consideration multiple parameters such as discrimination and guessing, mainly because we are studying a very diverse sample. Aims: Rating scales are increasingly the primary outcome measures for clinical trials. As such, they are the central dependent variables on which treatment decisions are based, and every effort should be made to maximize their scientific quality. The Multiple Sclerosis Impact Scale (MSIS-29) measures the impact of MS on physical and psychological functioning. It was developed using rigorously applied traditional psychometric methods and multiple studies demonstrate it satisfies standard criteria for reliable and valid measurement. More recently, there has been increasing interest in new psychometric methods, such as Rasch analysis, which enable more sophisticated and comprehensive evaluations of rating scales. Our aim was to examine the MSIS-29 using Rasch analysis and determine any implications for scale development. Methods: MSIS-29 data from n=2526 people with MS were analysed to determine the reliability and validity of the physical and psychological scales according to Rasch's criteria. Results: Rasch analysis largely supported the scientific soundness of the MSIS-29. However, the analyses highlighted a number of areas for improvement. Specifically more reliable measurement is achieved if the item response options are reduced from 5 to 4. More valid measurement would have been achieved if the scales had been constructed on the basis of strong conceptual definitions rather than the factor analysis of an item pool; more responsive measurement would have been achieved if the items had been better spread across the disability continuums. Conclusions: These results have implications for scale development. They indicate the need for stronger conceptual definitions, empirical testing of assumptions underpinning item response options, and the value of using the distribution of item calibrations to guide ongoing development. Aims: Items measuring Physical, Social, and Behaviorial Functioning in three widely used pediatric surveys were analyzed using IRT. Model fit, range measured, and opportunities for scale improvement and item banking were assessed. Methods: Instruments: (1) Child Health Questionnaire (CHQ): Data from a representative sample of the U.S. general population who completed the CHQ-PF in 1994 was available for 391 children ages 5-18; 55% had one or more physical or psychosocial conditions. (2) Child Health Survey of the Health Insurance Experiment (CH-HIE): The HIE database includes parent-completed health questionnaires for 1844 children aged 5-13 years, sampled from six areas across the U.S. (3) Pediatric Evaluation of Disability Inventory (PEDI): Clinician-completed data for 412 healthy children ages 6 months to 7.5 years, and 444 children age 6-17 years receiving rehabilitation services were available. Analysis: Within each dataset, a series of categorical confirmatory factor analyses were completed to assess unidimensionality and local independence of items hypothesized to measure the same construct. For items meeting model requirements, IRT calibrations were developed using the generalized partial credit model (GPCM). Item characteristics, IRT model fit, and differential item functioning (dif) for age, gender, and ethnicity were examined for each survey. Results: The hypothesized structure of domains was generally confirmed, and many items fit the IRT model. Range measured by the three surveys overlapped to some degree, and surveys differed in extent of floor/ceiling effects. Conclusions: Item response theory (IRT) holds much promise in improving health outcome measurement for children. Combining items from different surveys that measure the same construct into an item bank can extend the range of measurement possible and allow scores to be reported in original scale metrics. Aims: The Neck Disability Index (NDI) is the most widely used condition-specific instrument for evaluating change in neck painrelated disability in published trials. It consists of 10 items, each with 6 response options. It has not been assessed for interval level scaling. Item-response theory (IRT) methods were used to determine whether interval level scaling could be achieved. Methods: Baseline data on subjects from two randomized trials comparing chronic neck pain treatments were pooled (n=523). The generalized partial credit model was fit to data using PAR-SCALE (SSI, 2003) . Results: The NDI did not achieve interval level scaling. After collapsing response categories with no responses or responses with<1% of the total sample, item information curves were examined. All items exhibited poor discrimination (slope parameters <1.000) and response curves did not distinguish across levels of disability. The exception was the item 'Recreation' (slope parameter=1.052). The range of location parameter estimates was narrow and skewed to higher disability levels ()0.051 to 1.729). Interval level scaling was achieved by modifying the NDI. Response options were reduced until slope parameters were approximately 1.000, and poorly discriminating and redundant items were excluded, producing a four-item instrument with dichotomous response options. Somewhat improved sensitivity to lower levels of neck pain-related disability was achieved with these modifications. The NDI's inability to assess lower neck pain-related disability may have been the result of sample characteristics or a lack of items that capture lower levels of disability. Conclusions: These results suggest that modifications may improve the NDI's ability to detect lower neck pain-related disability levels required to measure positive treatment responses. Further analyses are needed to assess the NDI's responsiveness to change over time and assumptions of measurement equivalence across subpopulations. Aims: To assess the generalizability of the HUI3 MAUF by estimating and comparing two HUI3 MAUFs: Group A (those who rated ''all-worst'' < dead (n=193)) and Group B (dead < ''all-worst'' (n=22)). Two questions were investigated: (1) is there any difference in estimated MAUFs, (2) do any observed parameter differences in the two MAUFs affect the health state valuations? Methods: A decomposed approach was used to estimate the two MAUFs. Original HUI3 MAUF estimation project data were used. Parameters in two MAUFs were compared. Agreement of predicted HUI3 scores from the two MAUFs was examined by obtaining mean difference (MD), mean-absolute difference (MAD),standard deviation of the difference (SDD) and intra-class correlation (ICC) using 217 unique health states from Statistics Canada 1996/7 National Population Health Survey. Results: Additive utility form was strongly rejected in favor of a multiplicative form for both MAUFs. Both groups demonstrated preference complementarity. The importance of hearing, speech, ambulation and emotion differ between the two MAUFs. Agreement in predicted overall HUI3 scores from the two MAUFs was poor(MD=MAD=0.5, SDD=0.3, ICC=0.35). Stratified analyses showed moderate agreement for health states with mild burdens(HUI3>0.88; ICC=0.43) and poor agreement for health states with moderate/severe burdens (HUI3<0.89; ICC=0.32). The weighted MD (217 health states, each weighted by its prevalence) was 0.01, less than the clinically important difference of 0.03. Conclusions: Estimated scoring functions and valuations for health states importantly differ between the two groups. An implication is that in future preference elicitation surveys, it may be important to over-sample Group B to gain precision. However, when the 217 health states were adjusted for their prevalence, there was no important mean difference in utility scores between the 2 groups.Aims: As many as 50% of patients with diabetes who use insulin have poor glycemic control. Providing patients with alternative insulin delivery systems may lead to better outcomes. The study objective was to establish factors that contribute to preference for an insulin delivery system. Methods: Study participants were 119 patients with type 1 diabetes who completed a randomized, crossover trial comparing glycemic control of meal time injectable (IJ) vs. inhaled insulin (IH). Patient-reported outcome measures (PROs)-SF-36 Vitality Scale, subscales of the Diabetes Symptom Checklist-Revised, Perceived Hyperglycemia and Hypoglycemia items of the Diabetes Treatment Satisfaction Questionnaire, and Insulin Delivery System Questionnaire (ease of dosing, ease of blood glucose control, lifestyle impact) were administered at baseline, crossover, and endpoint. Patients were asked their preference for insulin delivery system at endpoint. Differences in PRO scores and glycemic control after 12 weeks of IH were evaluated using an ANCOVA model with baseline, after IJ scores, and sequence as covariates. All variables significantly discriminating between preference groups (IH vs. IJ) were entered into a stepwise linear regression using preference as the dependent variable. Results: IH preference was significantly (p< 0.5) associated with meeting glycemic goals; greater vitality; less perceived hyperglycemia; diabetes symptom reduction; and stronger agreement that IH could be integrated into patients' lifestyles, was easy to use, and easy to control blood glucose. The most significant predictor of IH was lifestyle impact (R 2 =0.24), followed by cognitive distress (R 2 change=0.06) and meeting glycemic goal (R 2 change=0.03). Conclusions: Lifestyle impact, symptom reduction, and clinical effectiveness drive patient preference for an insulin delivery system. Providing patients with their preference for insulin delivery system may lessen the burden of diabetes treatment and improve outcomes.Aims: New HIV treatment strategies have transformed the disease into a chronic state with an increased risk of treatment's toxicity. Hence, it is recognized that improving patient's quality of life (QOL) should be part of the HIV infection treatment goal. In the absence of suitable instrument, this study aimed to assess the cross cultural validity and reliability of Health Utility Index (HUI3) in UK HIV population. Methods: The study was approved by local research ethics committee. All adult patients receiving HIV care in Cardiff were approached in the out-patient clinic and recruited into the study after giving written informed consent. Participants were required to complete the HUI3. Variables analyzed include QOL score, CD4+ category, HIV stage, and viral load using Spearman's rank test, Kruskal-Wallis and Mann-Whitney U test. Results: In total, 103 (98%) of participants completed the questionnaire. The average age of the participant was 40.8 years (±10.7SD) and 81 were male. HUI3's was found to be reliable in most attributes (Cronbach's a 0.68), except in vision, hearing and ambulation. Four attributes (ambulation, emotion, cognition and pain) correlated significantly with QOL score (p<0.002) after controlling for antiretroviral use, clinical, and CD4+ categories. Findings also revealed no significant difference and correlation between QOL score and CD4+ count, viral load count and HIV clinical categories. There appears to be a stronger correlation (r=)0.193) and mean difference (p=0.083) between QOL score and antiretroviral use but this did not reach statistical significance. An interesting pattern was observed whereby asymptomatic patients (CDC A) and those not using antiretroviral (ARV) therapy had a lower QOL score than AIDS patients (CDC C). Conclusions: The findings of this study support validity and reliability of HUI3 in UK HIV population and could therefore be used with confidence in comparative study of HIV treatment. The results also suggest benefit of ARV use in improving patient QOL; however this requires further investigation in a controlled study.Aims: To estimate the utility of health states associated with intimate partner violence (IPV) and the independent effect of IPV on health related quality of life. Methods: A cross-sectional interview survey of a convenience sample of 231 abused and non-abused women recruited from hospital out-patient department waiting rooms in greater Boston was conducted. Abuse was measured with the Conflict Tactics Scale; health status data from the SF-12 were transformed to utilities. Linear regression modeling was used to estimate the association between IPV and utility while controlling for other health and demographic factors. Results: Utilities estimated from 93 women experiencing varying degrees of IPV showed a median value of 0.58 on a scale of 0 (dead) to 1.0 (perfect health), ranging from 0.64 for less-severe IPV (lower scores on the conflict tactics scale) to 0.53 for more-severe IPV (higher scores). Current IPV was independently associated with lower health related quality of life (HRQOL), while controlling for depression, comorbid conditions, past abuse, level of education and employment status (n=217; regression coefficient for IPV=)0.14, p=0.003; model adjusted r 2 =0.26). Of these other factors, depression was the largest contributor to lower HRQOL (regression coefficient=)0.12, p=0.001). Conclusions: Women experiencing IPV have significantly diminished HRQOL compared with unaffected women and people with other health conditions. IPV alone has a separate negative association with quality of life that remains after controlling for other effects that are known to diminish quality of life. The estimated utility values for IPV-related health states can be used in economic analyses of IPV prevention and intervention efforts to determine their cost effectiveness relative to other public health and medical programs. These results inform priorities and resource allocation decisions for efforts to improve health related quality of life.