| SERVICES RESEARCH (0) MONOGRAPH SERIES) Management Information Systems in the Drug Field © Management Aut Binformation Systems | in the Drug Field / Edited by George M. Beschner, M.S.W. Neil H. Sampson, M.P.A., M.P.H. National Institute on Drug Abuse and Christopher D’Amanda, M.D. Coordinating Office for Drug and Alcohol Abuse City of Philadelphia U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE Public Health Service Alcohol, Drug Abuse, and Mental Health Administration Services Research Branch Division of Resource Development National Institute on Drug Abuse ny 5600 Fishers Lane Rockville, Maryland 20857 A vg } Klee unl The Services Research Reports and Monograph Series are issued by the Services Research Branch, Division of Resource Development, National Institute on Drug Abuse. Their primary purpose is to provide reports to the drug abuse treatment community on the service delivery and policy-oriented findings from Branch sponsored studies. These will include state of the art studies, innovative service delivery models for different client populations, innovative treat- ment management and financing techniques, and treatment outcome studies. Chapter 2, State of the Art Review: Drug Abuse Management Information Systems in Single State Agencies was written under NIDA contract #271-76-5506. All of the other chapters in this report were prepared under NIDA grant #5 H81 DA 01729-03 to the Coordinating Office for Drugs and Alcohol Abuse, City of Philadelphia. The material contained herein does not necessarily reflect the opinions, official policy, or position of the National Insti- tute on Drug Abuse of the Alcohol, Drug Abuse, and Mental Health Administration, Public Health Service, U.S. Department of Health, Education, and Welfare. DHEW Publication No. (ADM) 79-836 Printed 1979 For sale by the Superintendent of Documents, U.S. Government Printing Oflice Washington, D.C. 20402 Stock Number 017-024-00938-5 ii Acknowledgments The completion of this publication is due largely to the work of Birch and Davis Associates, Inc., Silver Spring, Maryland. Through a subcontract with the city of Philadelphia, this firm assumed a management role in coordinating the Management Informa- tion System (MIS) demonstration project that led to this publication. The editors are pleased to acknowledge and extend their thanks to Ms. Elaine Palusci Bencivengo, Director of the Philadelphia Central Medical Intake, who provided ongoing consultation regard- ing treatment implications and issues. In addition, we are grateful for the critique by an external technical reviewer, Louis Wynne, Ph.D., of the Bernalillo County Mental Health/Mental Retardation Center, Albuquerque, New Mexico. iii CONTENTS ACKNOWLEDGMENTS INTRODUCTION George M. Beschner and Christopher D'Amanda 1. MIS IN DRUG ABUSE PROGRAMS: A REVIEW OF THE STATE-OF-THE-ART Thomas L. Foster STATE-OF-THE-ART REVIEW: DRUG ABUSE MANAGEMENT INFORMATION SYSTEMS IN SINGLE STATE AGENCIES Paddy Cook, Barry Rosenthal, and Cheryl Davis A CASE FOR MANAGEMENT INFORMA- TION SYSTEMS: HELPING THE MANAGER MAKE DECISIONS REGARDING DIFFICULT RESOURCE ALLOCATION PROBLEMS Edward Leibson PEOPLE AND DATA SYSTEMS: SOME ISSUES OF INTEGRATION George De Leon COMPUTER SOFTWARE: BRIDGE TO INFORMATION UTILIZATION Clyde B. McCoy, Anne C. McCoy, and James E. Rivers AUTOMATION ALTERNATIVES IN THE DRUG ABUSE TREATMENT SETTING Herbert M. Birch, Jr., and Kerry G. Treasure FUNDAMENTAL CONSIDERATIONS IN DEVELOPING AN MIS Willie Davis and Kerry G. Treasure page iii 34 79 107 121 139 172 Introduction George M. Beschner, M.S.W., and Christopher D’Amanda, M.D. During the past few years there has been a surge of interest in the acquisition and utilization of data in the drug abuse field. Managers at all levels have become increasingly involved in collect- ing, analyzing, and communicating data. In addition, drug abuse treatment programs have been inundated with data requests from program evaluators in monitoring and funding agencies, from public officials, and from the media. It has become more and more apparent that management information systems (MIS) are essential to the organizations that make up the drug abuse field. Yet little is known about the composition and utility of the systems that have been designed and installed. In order to understand the state of MIS in the drug field, one must first know something of the recent history of drug treatment programs and their impact on MIS development. One of the major factors influencing the development of information systems was the dramatic expansion in treatment services during the early seventies. In 1971 the National Institute of Mental Health, which was the foremost administrative drug rehabilitation agency, was supporting 23 community-based treatment programs with a total treatment capacity of some 13,000 patients. As a result of the Drug Abuse Office and Treatment Act of 1972 (Public Law 92-255), the number of federally supported drug treatment programs in operation by 1973 increased to 183. These provided services to approximately 75,000 drug abusers. The Act also established the National Institute on Drug Abuse (NIDA) as a separate institute to coordinate the new health delivery systems established by the Federal Government. Initially emphasis was placed on implementing treatment services in areas with high rates of heroin addiction. In the 1972-73 period, grants, the primary means of allocating funds to drug treatment programs, rapidly expanded the number of treatment programs. Immediately following this short period of unprecedented growth in the drug treatment field, Federal support became more unpredictable, and there was a mounting demand for fiscal and program accountability. A series of audits was initiated by NIDA to evaluate program management and service capabilities. New treatment guidelines, standards, and regulations were formulated and the quest for information continued to grow. In 1973 the Food and Drug Administration (FDA) specified that methadone maintenance programs must maintain clinical records on each patient treated to include urinalysis results, the type of rehabilita- tion services employed, accounts of the patient's progress, and other information on relevant aspects of the program. Treatment staff including physicians, nurses, and counselors were required to record services in the patient's clinical records and periodically to evaluate results. More recently (1978), NIDA produced a monograph entitled Clinical Record System for Drug Abuse Treat- ment Programs. This elaborate client record system was designed to help programs develop treatment plans and foster continuity of care. Funding policies and procedures also changed. NIDA, through Statewide Services Contracts, began transferring responsibility for administering federally funded treatment services to State agencies by the end of 1973. A program/budget matrix, which specified a unit cost for each treatment slot, was used as a funding guide by NIDA and the State agencies. Treatment pro- grams were encouraged to pursue other funding sources, and there was a concerted effort to help programs generate third-party payments. These changes put more pressure on programs to collect information, maintain records, and report data. CODAP (Client Oriented Data Acquisition Process), which was designed to provide comparable data for financial resource manage- ment at the national level, became a household word in the drug field. CODAP was used primarily as a management tool to deter- mine how treatment slots were actually being utilized. For many treatment programs and State agencies, it was their first informa- tion system and provided some structure for data acquisition and clinic recordkeeping. Additional evaluation activities evolved in response to the need for drug abuse treatment information. The Drug Abuse Reporting Program (DARP), an elaborate reporting system administered by Texas Christian University's Institute for Behavioral Research (IBR), was designed to evaluate treatment outcome. The first series of DARP research reports was produced by IBR in 1972 and 1973 and stimulated interest in the use of data for evaluation. Research and evaluation studies were also conducted by agencies interested in fostering new knowledge. Journals and conferences served as the principal communication sources for sharing and reviewing research findings and evaluation results. With their growing appreciation of the potential value of data for reporting, management, and evaluation purposes, State and county agencies began investing in management information systems. The Division of Scientific and Program Information (DSPI), NIDA, attempted to support these efforts by providing technical assistance and staff support. In addition, software packages were developed to help county and State agencies make use of CODAP data. Programers, computer specialists, and system designers were employed to help manage systems, and these specialists comprised a new source of knowledge and skill for the field. Many of the program managers who were able to implement information systems and derive benefits from having data readily available became MIS proponents. However there was little opportunity for people involved in MIS work to communicate with one another. Many of those who were involved in the developmental stages of designing and implementing systems expressed interest in learning what others were doing and in sharing information about their own work. In recognition of this, the Services Research Branch (SRB), NIDA, organized a planning conference to bring together those who had been successful in establishing information systems. The conference, held in Washington, D.C., on January 26-217, 1976, was structured to identify and review the different systems that had been developed by the participants, to determine how they functioned, and to discover the extent to which they had met various program and management information needs. The conference was followed by site visits to five exemplary information systems. Creative Socio-Medics, under contract to SRB, conducted the visits to study and compare the design characteristics, the format, and instruments used to collect data, staffing patterns, quality-control procedures, and outputs of these systems. Since the five systems were primarily at the county level, another series of site visits was conducted by the Berkeley Center for Drug Studies to individual drug treatment programs. These visits were structured more specifically to determine the extent to which the information being generated by the systems actually addressed decisionmaking needs at the treatment program level. This publication is an attempt to share what has been learned about MIS in the drug abuse field as a result of these exploratory efforts. We recognize the many limitations in the state-of-the-art review and that the programs studied were not representative of all programs in the field. We hope, however, that others will be encouraged to share their experiences in organizing and implement- ing MIS and to provide information about other systems that have been developed. There have been some attempts by the Services Research Branch to construct model systems that could be replicated, but this objective has not yet been achieved. It was felt that, at this early stage of MIS development, it would be more appropriate to share information about the process involved in developing informa- tion systems, the various problems and issues involved, and some of the software and hardware alternatives. Individuals who had practical knowledge and experience in drug field MIS development were contacted during the course of the state-of-the-art surveys and encouraged to write about their experiences. New systems currently are being developed and implemented in Single State Agencies, county coordinating agencies, cities, and programs. In addition, NIDA-supported MIS demonstration programs were implemented in Detroit and Miami in 1978. The Wayne County Department of Substance Abuse Services in Detroit is in the process of developing and testing an automated budget-modeling system to be used by local, county, and State organizations. The University of Miami is attempting to develop and test the usefulness of MIS in different types of drug programs. Both of these demonstration projects are funded by the Services Research Branch for a 3-year study/demonstration period. In chapter 1, Thomas L. Foster reports on the findings from an MIS conference and two series of site visits organized by the Services Research Branch to assess the state-of-the-art in the drug field. He reviews such features as: ° Design characteristics ° Sources of information ° Data collection procedures ° Availability and presentation of data ° Utilization of data ° Information gaps ° Staff attitudes and staff capability He assesses the extent to which MIS is supported by managers, system users, and individuals involved in data collection and recording. Foster recommends steps that can be taken to enhance the availability of information, reduce the MIS burden on staff, and increase support for MIS at all levels. Chapter 2 provides the results of another state-of-the-art review of management information systems. This study of MIS in the 50 Single State Agencies (SSA) and 5 county agencies was conducted during April and May 1977 by Richard Katon and Associates under contract to NIDA. It should be understood that their results would be different if the same survey were conducted today, since many SSAs were then actively engaged in developing MISs. The Katon survey showed the extent of MIS implementation as follows: ° Slightly over half of the Single State Agencies (56 percent) had a more or less stable system in place to meet at least the CODAP requirements. ° Thirteen States (27 percent) were in the early stages of MIS development, having only recently overcome such obstacles as: . Inadequate State support . Inappropriate or insufficient staffing . Lack of leadership at the SSA level . Organizational barriers . Financial constraints Ol WDD An additional 7 percent of the States (eight in number) were in the process of revising their MIS. In chapter 3 Edward Leibson discusses how MIS can assist managers in procuring, allocating, and redeploying limited resources. He points out that minimum data sets are required to manage a program including data about staff, clients, daily client counts, budgets, material resources, and client progress. Examples are provided on how the management-by-exception approach may be used to assess available variance figures and determine whether resources are being utilized appropriately. Leibson also discusses the barriers to MIS and strategies for overcoming them. George De Leon, who has had considerable experience in designing and implementing evaluative information systems in a large thera- peutic community, discusses some of the strategies required to elicit staff cooperation and support in chapter 4. He starts with the premise that an information system will not be successful if the participating staff are not actively involved in its development and do not obtain ongoing benefits from it. He discusses how to get staff involved--by reviewing the purpose of the information system, alleviating personal fears and difficulties, providing education and training, and helping staff to understand and experience the benefits of self-evaluation. The importance of improving communication between data people and clinicians is also stressed. In chapter 5, McCoy, McCoy, and Rivers focus on the development and utilization of computer software--the instructions and program- ing developed by people--to handle large arrays of data. They point out that, although the need for computerization is usually recognized, drug treatment programs have been slow to make use of computers, partly because of such factors as the scarcity of funds, inadequate knowledge, the lack of experienced people, and poor communication between clinicians and computer personnel. Although advances have been made in the Electronic Data Process- ing field and new software packages are now available to relatively unsophisticated users, there is still a need to modify and adapt systems to meet the special needs of particular users. Potential software users must be aware of the various technical requirements, including the formatting of the data, file structure, instructions, etc. McCoy, McCoy, and Rivers describe a software system called "The Quick Interactive System" developed in Dade County, Florida. The advantages of automation are becoming more apparent to drug programs, but because of the lack of knowledge about computers, most program managers do not know how to select a data process-— ing system to meet their needs. Birch and Treasure in chapter 6 describe some of the steps and functions involved in data process- ing, e.g., the collection, input, and manipulation, storage, and output of data. Such factors as availability, reliability, security, turnaround time, software suitability, programing consultation, flexibility, and ease of installation, implementation, and operation are also covered. There is also some attempt to describe the computer alternatives that are available including manual systems, small-business computers, and the services of computer service bureaus (CSB). The authors point out that CSBs offer two modes of operation: a terminal for online processing, and batch processing. In the final chapter, Davis and Treasure present some factors that should be considered by drug treatment managers in establish- ing an MIS. They discuss the process involved in (1) identifying information needs, (2) classifying information, (3) developing a format to capture and display the data, and (4) establishing procedures for collecting and processing data. The authors also demonstrate how the data needs of drug programs can be classified into a module structure. The modules--Planning, Client Manage- ment, Accounting and Financing, and Program Evaluation--corre- spond to the major categories in which data are needed to manage a drug program. AUTHORS GEORGE M. BESCHNER, M.S.W. Services Research Branch National Institute on Drug Abuse Rockville, Maryland CHRISTOPHER D'AMANDA, M.D. Chief Medical Officer Coordinating Office for Drug and Alcohol Abuse City of Philadelphia Philadelphia, Pennsylvania 1. MIS in Drug Abuse Programs A Review of the State-of-the-Art Thomas L. Foster It must be remembered that there is nothing more difficult to plan, more doubtful of success, nor more dangerous to manage than the creation of a new system. For the initiator has the enmity of all who would profit by the preservation of the old institution and merely lukewarm defenders in those who would gain by the new one. Machiavelli, The Prince, 1640 This chapter bears testimony to the observation made above. Results from efforts to assess the state-of-the-art and current strategies developed to implement and operate MIS in the drug field are reviewed. These efforts included a conference’ involving individuals who had been successful in developing MIS, and two series of site visits (see Introduction) to review systems that had been established. The chapter is organized into four parts, and is concerned respec-— tively with MIS subsystems, presentation of MIS information, factors affecting MIS implementation, and recommendations. The content represents no more than a summary of what has been learned and does not pretend to reveal a definitive or final state- ment on management information systems. Client Oriented Reporting Subsystems The primary external recipients of client oriented information are the Single State Agencies (SSAs) and ultimately the National "National Institute on Drug Abuse. "Technical Review of Manage- ment Issues and Systems in Drug Treatment." Unpublished conference transcript. Washington, D.C., January 1976. Subse- quent citations will read NIDA MIS Conference, 1976. Institute on Drug Abuse. The principal vehicle for the transfer of this information is the Client Oriented Data Acquisition Process (CODAP). CODAP was originally designed to provide cross-program comparable information for financial resource management at the national level. In recent years there have been efforts to adapt the system to meet management needs at the State, county, city, and program levels. The CODAP reporting system is composed of two major parts, an admission report and a discharge report, which provide information on the two crucial points in the treatment process. In addition, an activity report provides information on the treatment process to meet NIDA grant and contract specifications. Information from the CODAP system has served a variety of interesting and impor- tant decisionmaking needs in recent years. Basic demographic data collected at intake have been aggregated to obtain a more detailed picture of the drug-using population. In particular, data on client age, age at first use, and age at first admission have been combined into trend-analytic studies to shed light on the prevalence of drug abuse and, hence, upon the overall need for treatment services (Hunt 1974). Moreover, CODAP data have, through aggregation by program of the quantity of services delivered, served as a source of operations and fiscal information as well. There are still, however, considerable informational gaps. Present reporting requirements provide little information regarding treatment and ancillary service needs. As a consequence, admis- sion to a particular treatment modality or service is frequently assumed to be an indicator of need for that modality or service; and subsequent planning of service capacity (e.g., the ratio of methadone to residential slots) is then based upon assumption rather than upon actual needs for services in a particular area as determined by rigorous needs-assessment techniques, Similarly, there is little information available to Federal and State decisionmakers. Some States have attempted to meet this informa- tional need through extensive (and expensive) followup studies (Social Issues Research Associates 1977). But for the most part, relatively little is known at the State and Federal levels about the efficacy of drug abuse treatment. One piece of available information which is rarely utilized is the attrition profile of clients in a particular program or modality. If one assumes that longevity in treatment is at least partially related to the satisfaction of client needs, then comparisons of aggregated attrition data can become a useful evaluative tool. (See Harris and Moitra [1977] for a review of this approach in the evaluation of criminal justice programs.) Client Oriented Planning and Control Subsystems The client oriented information needs of program and county-level managers are in many ways similar to those at the State and Federal levels. There is a need for descriptive information about age, sex, ethnicity, etc., so that services can be tailored to the needs of special population groups. In this capacity, the CODAP system or county systems which satisfy CODAP requirements serve adequately. Similarly, county-level planners have need of information bearing upon the geographic distribution of client residence in order to properly allocate services to appropriate catchment areas. Again, where county systems have been imple- mented to satisfy Federal reporting requirements, geographic data are typically included in intake documents and are therefore available for county planning purposes. However, neither county-level planners nor local program managers typically have access to information relating to treatment and other service needs. The assumption is usually made that clients applying for or admitted to a particular service are in need of precisely that service rather than something quite different. The result is that the existing mix of services tends to be perpetuated rather than modified to conform to changing service needs. The deficiency of client oriented evaluative information is much less severe at the county and local program levels. In those counties where centralized information systems have been developed, there is the capability, in principle, of aggregating recidivism data on a program-by-program basis. This capability derives from the practice of uniquely identifying each client within the information system and of maintaining that same client identifi- cation through subsequent treatment episodes. Despite the fact that program-level administrators have consistently indicated that existing reporting systems do not adequately meet their evaluative needs (Creative Socio-Medics Corp. 1976), the sources of client oriented information available to them are substan- tially richer and more diverse than those otherwise available. The following recent in-house studies are examples of the kinds of information currently being developed and utilized at the program level: ° A residential program in California found that clients who completed the yearlong treatment regimen did not fare sig- nificantly better than those who had terminated against program advice after 4 to 6 months, despite the fact that both groups appeared to fare better than those who completed less than 4 months. This finding led to a restructuring of the program's approach to residential treatment. ° A program counselor in Delaware suggested that program "splits" were highly correlated with certain types of thera- peutic activities. These activities have been reduced in frequency and intensity, and program retention rates have increased significantly. ° A multimodality program in California utilized routinely collected personality data to determine that clients who appeared to be most responsible to treatment (reduced their drug use, obtained jobs, etc.) also exhibited greater signs of dependency. Special attention has since been directed toward reversing this unanticipated apparent program effect. Finally, directors of some exemplary programs have established extensive informal systems to obtain additional client oriented information including weekly clinical staff meetings and frequent meetings with clinical supervisors and individual counselors. Client Oriented Information Sharing Subsystems As treatment programs have grown in size and complexity, it has become apparent that clients are likely to receive services from several different program staff members. To the extent that this occurs, it is imperative that each such staff member have access to information about the client, about his or her treatment needs and goals, about treatment progress, and about what each of the other staff members is doing. The earliest attempts to meet this informational need relied upon extensive informal contact between counselors and upon the use of a general program log as a vehicle for intraprogram communication. When these efforts proved inadequate to the task of information sharing, some programs moved to a caseload system wherein each program counselor would function as the locus of information regarding a particular group of clients. More recently, as pro- grams continued to expand and to increase the diversity of avail- able services and the diversity of personnel, greater reliance has been placed upon the development of a detailed client file contain- ing records of each service contact. Yet even a cursory review of client files in practically any treat- ment program suggests that this approach is also inadequate. While files typically contain a great deal of data in the form of service documentation, they contain relatively little useful infor- mation. What appears to have gone awry is that the Federal (and State) monitoring emphasis upon documentation of services delivered through the use of client files has supplanted the normal and beneficial use of these files as a means of staff communication. Some programs have developed new and perhaps more constructive approaches to the need for client oriented information sharing. For example, several programs have established two parallel sets of files, one for service documentation to satisfy the external auditors, and a separate one for counselor treatment notes. While the necessity for this distinction is unfortunate, it is at least possible for each staff member to obtain the needed information about a particular client's progress and goals. A second approach has been the establishment by some programs of a series of clinical staff conferences. A typical format has been that each client is discussed at length at least once every 60 or 90 days, and that the progress of other clients is highlighted as necessary during the interim. Not only do these regular 10 conferences serve as a valuable means of information sharing, they also provide a useful vehicle for ongoing staff training. Operations Reporting Subsystems Extensive information is available to State and Federal officials regarding the operations of drug abuse treatment programs. Not only is information available through the CODAP documentation of services delivered, but it is systematically augmented through periodic site visits by both State and Federal program monitors. Moreover, several States have established additional channels for information regarding program operations through the use of periodic narrative reports prepared either by the programs them- selves (e.g., Michigan State Office of Substance Abuse Service 1976) or by outside evaluators working in conjunction with program personnel (e.g., Marlo et al. 1975). Operations Planning and Control Subsystems At ihe county level the potential for useful information regarding program operations is substantial. A recent innovation adopted by Wayne County, Michigan, focuses basic data collection on the program counselor rather than on the individual client. Thus, each counselor accounts for his or her time indicating which clients were served, what services were delivered, and for how long. These data are then restructured to comply with client oriented reporting requirements and aggregated by counselor and by program to provide a systematic accounting of program activities. In addition, in several metropolitan counties individual staff members are assigned to work with local programs. While a primary function of these contacts is to insure that program documentation is in compliance with Federal and State requirements, they also serve as a valuable source of information to county officials regarding program operations. Similarly, at the program level managers appear to have available a variety of informational sources regarding program operations. In most larger programs directors have moved to establish more clear-cut lines of administrative and clinical responsibility, thus creating an organizational hierarchy which facilitates the flow of information (rather than data). There is a correspondingly greater reliance upon clinical and other supervisory personnel as sources of information through routine administrative meetings and conferences as well as through ad hoc requests for special reports and briefings. Fiscal Planning and Control Subsystems Despite the fact that most larger programs retain the services of an accountant, or at least a professional bookkeeper, and that the fiscal data collected in almost all programs are adequate to insure 11 at least minimal standards of fiscal accountability, it is nonetheless true that relatively little useful fiscal information is available either at the county or treatment program levels. This is primarily because program accounting and fiscal management systems have been developed almost wholly independently from those MIS sub- systems which focus on clientele and/or operations. While this is less keenly felt at the county level, where each program can be viewed as a cost center in its own right, it is of particular concern at the program level, where a variety of serv- ices is delivered to each individual client. Program directors are as a result unable to allocate various program costs to particular services or to particular groups of clients. The significance of this inability lies in the growing need to seek fee-for-service payments from alternative third-party sources such as private insurers, title XX agencies, etc. In many cases these alternative sources require firm and justifiable cost estimates for particular services. Thus programs lacking a system of cost accounting will encounter difficulties in obtaining third-party funding. External Environment Planning and Control Subsystems While program directors are usually well enough informed about their own programs, they suffer from a paucity of information about the external environment. Most program directors are only vaguely aware of the variety of funding resources available through various State, Federal, and private agencies. While NIDA has made some attempts to disseminate this sort of information le.g:» Priesman 1976), there has been little followthrough by State and county officials. It would be helpful if county drug abuse coor- dinators would begin to view their role at least in part as a clearinghouse for information which is vitally needed at the pro- gram level. A similar informational gap exists regarding the activities of other programs. Many program directors are unaware of what is being done by other service providers--both substance abuse programs as well as other human service agencies--within their own commu- nities. In this context, the failure of county- and State-level officials to facilitate the flow of information is exacerbated by the continuing climate of suspicion and mistrust which pervades interprogram relations. Moreover, even at the county or umbrella agency level, planners and other officials are frequently not cognizant of the services available through providers outside their own sphere of control. In the past, the consequence of such informational gaps was thought to be an unnecessary duplication of services. In the present era of limited resources, it seems likely that the result will be the failure to make needed services available. The ultimate victim will be the client rather than the taxpayer. 12 External Environment Information Sharing Subsystems Just as it is important for program directors to be aware of those services that are available elsewhere within the community, it is important for individual counselors to have access to the same information. The result of a counselor's ignorance of available services is that the client will not have access to them. It is usually those counselors most in need of such information who are also the least informed counselors in central intake or central referral units. Not only do these counselors need to know about the availability of various services, they also must be aware of which programs are functioning well and which poorly, of which programs have open slots and which have waiting lists, of which prefer certain types of clients, and of which have eligibility requirements that might hinder prompt admission. Relatively little has been done to make this information available. Instead, intake and referral workers remain largely uninformed even about the programs to which they make frequent referrals. SUMMARY With the exception of subsystems designed to provide information regarding program operations, there are serious gaps in the information available to decisionmakers at all levels. Information regarding service needs and treatment results is typically unavail- able to managerial and external decisionmakers. Even at the counselor level, relatively little useful information is communicated through client files and other program records. Fiscal subsystems, while adequate to the task of financial accountability, offer little assistance to the program manager in relating costs to specific services or program activities. Finally, neither program directors nor key staff members have sufficient information available regard- ing the environment in which their programs function. THE PRESENTATION OF INFORMATION: FORM AND PROCESS An important design characteristic of information systems is the extent to which decisionmakers actually have access to information and the extent to which that information can be accessed in a form and format compatible with the decisionmakers' needs. Observations regarding this aspect of the MIS art follow two major themes: ° Much of the information which is made available to decision- makers is obscured by a substantially greater volume of irrevelant data. ® Much of the information which is in principle available to decisionmakers at the county and treatment program levels is in practice not available at all or, at best, not available in a timely fashion. 13 Access to Information In a recent article extolling the virtues of modern EDP techniques, Beehler (1976) notes that one of the traditional complaints of MIS users has been the inability of both manual and automated systems to respond quickly and easily to special requests--for information aggregated in new and perhaps unanticipated ways, and for information which spans multiple data files. His argument is that recently developed techniques of data base management allow for prompt and flexible user access across a broad range of data sets and have thereby resolved these complaints. Drawing upon these advances in computer sophistication, most of the newer automated information systems in the drug abuse field have made special provisions for ad hoc user access. Thus, the program director can elect to initiate series of special requests. Yet, as noted by Creative Socio-Medics Corp. (1976), most pro- gram personnel have quickly become disenchanted with this MIS feature, citing lengthy delays as the primary source of their disaffection. Such delays have arisen for a variety of reasons, each of which can be remedied. First, there has been the procedural hurdle: Special requests must usually be approved at one or more levels. Next, there is the programing problem: All too often special one- time programs need to be written. Finally, there is the scheduling obstacle: Few, if any, county substance abuse agencies have access to their own computer facilities; therefore, many special requests receive rather low priority in the county or private time-share system. Indeed, the impression is hard to avoid, after following a set of special requests from inception to response, that had the system designers deliberately wished to discourage system use, they could not have done a better job. Several automated systems contemplate the eventual implementation of online user access, the plan being to install cathode-ray tube terminals in each major treatment program. Yet the present state of the programing art, at least in the drug abuse field, argues against such aspirations. For example, one of the few online systems operating at the county level can handle user queries involving only a handful of variables. Unfortunately most of the truly interesting and important questions require substantially greater flexibility. Accuracy of Information Accuracy of information is a crucial consideration in any MIS. However, the accuracy of an information system must be evaluated in the context of particular decisions rather than in the abstract. Certain decisions are highly sensitive to small errors in data collection or manipulation, while other equally important decision situations require only the crudest of estimates. Increased accur- acy is never without cost. It must be paid for through either more staff time spent in data collection, more extensive staff 14 training, broader sampling of data or, alternatively, slower responses to informational requests, more expensive edit routines, etc. In effect, accuracy which substantially exceeds the organiza- tion's decisionmaking needs represents wasteful expenditure of resources. Two basic flaws in the design of data collection procedures in the drug abuse field are: ° Data-collection forms designed primarily to facilitate key- punching (or other data handling procedures) rather than to insure accurate completion by treatment staff. ° Staff ability to report information rather than data is often overestimated. One has only to examine the forms currently utilized in treatment programs to recognize this. The forms are reminiscent of finance company loan applications. Each page (together with its multiple- color, self-carbon copies) is replete with fine print, with large boxes (captioned, "For County Use Only"), and with small boxes (each accompanied by mysterious, italicized numbers). Somewhere, usually at the bottom of each page, are several lists of codes for individual data items (e.g., "Heroin=01, Amphetamines=02," . . .). In short, the entire set of forms often appears bewildering and confusing. It is little wonder that subsequent edit checks report a myriad of missing data items, failures to follow indicated skip patterns, out-of-range responses, etc. Even clinicians who can prepare accurate and insightful narrative reports based upon their intake interviews may not be able to combine their clinical skills with the process of codified data collection. What is needed is a method of collecting and codifying client oriented information that will take advantage of the diagnostic and clinical skills of program counselors rather than conflict with and obscure those skills. In the absence of innovation along these lines, it seems likely that much of the clinical information to be adduced from present systems will be built upon a limited or only skeletal base. Quality Control Procedures Most of the newer automated systems being developed in the drug abuse field include extensive computerized edit routines designed to insure adequate data quality. These edit procedures typically include: ° Document edits, i.e., checks for missing data, range and plausibility tests, performed on each document submitted. ° Merge edits, i.e., comparisons between newly submitted client records and existing files. 15 ° Cross-file validations, i.e., monthly comparisons between active client files and summary client flow statistics as re- ported by each program. In addition, periodic site visits by auditors are utilized to further validate reported data through comparisons with documents con- tained in client files. However, there are several problems inherent in this approach. First, while program directors are notified through a monthly printout of all edit exceptions, there is relatively little control over their eventual turnaround, so that the majority of errors go uncorrected. Second, the edit procedures themselves are not conducive to the eventual reduction of errors. The majority of all computer- detected errors are classified as "nonfatal," i.e., the input docu- ment is accepted into the data base despite certain minor errors or inconsistencies. Each of these data collection errors is subse- quently noted in the director's monthly edit report, a document several pages long. These edit reports are merely data, whereas useful information might include, for example, a cross-tabulation between error type and reporting counselor. With information, rather than data, in hand, program directors are better able to respond effectively, e.g., focus training for individual counselors to reduce error commission rates. Instead, what commonly happens is that lengthy lists of edit exceptions are treated as just so much wastepaper. Finally, to the extent that error correction and document resubmit- tal are required, the process tends to effect an overall bias in reporting practices. Counselors simply do not like additional paperwork, and as a consequence, they learn very quickly what kinds of discrepancies result in fatal errors; that is, errors that must be corrected. Then, rather than becoming more meticulous in their initial reporting, temptation may lead to the habit of underreporting through simplification. For example, if counselors are frequently confronted with fatal errors regarding prior treat- ment history, they will soon begin to report almost all admissions as having had no previous treatment experience. System Costs It is difficult to discuss either the developmental or the implementa- tion costs of existing information systems in the drug abuse field, because there is little available documentation of actual costs incurred. For example, officials in one metropolitan county estimated system design costs at approximately $50,000-$60,000 per year for the first few years (Creative Socio-Medics Corp. 1976). (A modular approach to system development was adopted by this county, making it difficult to specify the duration of the design process.) This estimate is based solely upon the salaries 16 of the system programers and the computer time actually utilized for system development. No estimates are available for such cost factors as preliminary, conceptual design; obtaining input from intended system users; training of system users; or training of data-collection personnel, e.g., counselors, Officials responsible for the drug abuse MIS in a second county estimated total develop- mental costs at over $500,000; however, there are no estimates available of individual cost factors. Implementation costs may also be estimated on a different basis: the anticipated volume of clients to be "managed" by the system. Figures range from a low of $30 per client per year to a high of $72 per client per year. Yet none of these estimates include such ongoing cost factors as printing and distribution of data-collection forms, program staff time involved in data collection, ongoing training of system users and program data-collection personnel, or administrative support. Some of the cost factors involved in systems development and implementation are likely to be misleading. For example, the costs of staff time for personnel not specifically designated to MIS operations must be evaluated in the context of alternative uses of their time. Perhaps the most conspicuous instance of this problem is the time spent by local program personnel: their input into system design, their time spent in data collection and processing, etc. Program personnel, especially counselors, are almost univer- sally critical of these impingements upon their time. Yet it must be noted that at least some county-level systems require less counselor time than would have been spent in complying with Federal reporting requirements in the absence of a local MIS. Nevertheless, substantial further reductions in the amount of counselor time presently expended in data collection might be effected by more careful planning at the program level. In short, we simply do not know how much systems actually cost. All we can say is that they seem likely to cost far more than is presently estimated. What is needed, of course, is an accurate and comprehensive cost-accounting system at the county level. IMPLEMENTATION CHARACTERISTICS OF EXISTING SYSTEMS Much of the recent MIS literature has called attention to the process of system implementation. Lucas (1975) has observed, "Concentration on the technical aspects of systems and a tendency to overlook organizational behavior problems and users are the reasons most information systems have failed." This section explores the extent to which drug abuse treatment MISs have achieved the following important objectives for success- ful implementation: management support; user support; data collection and recorder support; application to decisionmaking. 17 Support by Management To obtain effective and visible support of top management, Dickson and Simmons (1970) have noted that in many organizations outside the drug abuse field the task must begin with an effort to overcome a natural resistance of managers. This may result from feelings of insecurity, role ambiguity, anticipated increase in job complexity, and uncertainty or unfamiliarity. Fortunately, most countywide information systems in drug abuse treatment agencies have been initiated directly by county-level managers so that these and other similar sources of resistance have not proven to be major stumbling blocks. Even without resistance by top management, however, their active support has often been neither visible nor effective. Rather than directing their attention to the ways in which county-level subordinates and local program personnel might be convinced of the desirability and importance of the newly developed systems, many officials have relied largely upon "management by fiat" to obtain compliance with system input requirements. Thus, despite protestations to the contrary, the appearance is that systems are intended solely to channel infor- mation to top management rather than to meet the informational needs of the organization as a whole. For example, during interviews most county agency directors went to great lengths to describe the virtues and importance of their information systems. Yet local program personnel indicated that they had little understanding of information needs at the county level. Furthermore, many program directors indicated that efforts by county managers to encourage and stimulate MIS use at the treatment program level were at best perfunctory. Support by Users Lucas (1975) has argued that the fate of any information system depends upon obtaining the support of the intended system users. This task can best be examined in terms of three rather distinct objectives: ° Overcoming the natural resistance within any organization to the development and implementation of any new system; ° Cultivating generally positive attitudes and perceptions on the part of intended users toward the particular system to be implemented; and ° Training users in the uses of an MIS. In their examination of organizational resistance to an MIS, Dickson and Simmons (1970) suggested a variety of factors which appear to motivate the typically negative response of operating managers. These factors include: 18 ® Feelings of insecurity ° Role ambiguity ° Threats to status or power ° Threats to economic security ° Increased job complexity ° Increased uncertainty ° Changed work patterns ° Increased rigidity ° Added time pressure It is not difficult to appreciate the impact of these pressures upon treatment program directors and middle-management person- nel. A countywide MIS provides a direct channel for information to county-level officials; a channel which, in effect, bypasses the program manager and eliminates his or her opportunity to filter selectively the upward flow of information. Moreover, the informa- tion obtained at higher levels is very likely to be used to evaluate the program's performance, a threatening contingency to many managers. Finally, the implementation of any MIS involves added supervisory time and increased responsibility upon the program director for the performance of the system itself. One potentially effective response to these sources of resistance by intended users is to allay their concerns and suspicions by involving them in the system design process itself. Lucas {1975) places considerable emphasis on this strategy to develop user support. Even with the efforts that have been made in this direction in the drug abuse field, it was reported that those efforts have typically failed to be responsive to the concerns of local program personnel. During the development of most existing systems, program directors have been asked about what information they would like. Yet only rarely have they been asked about how and when they would like to receive that information. Perhaps more significantly, program directors are never asked, at the outset, whether they would like an MIS at all. In short, while gestures have been made toward user involvement there has been little attempt to operationalize that involvement. As a consequence many systems are viewed with anxiety and suspicion by the intended system users. Beyond placating such anxieties and suspicions, there remains the ongoing task of cultivating positive user attitudes toward the developed MIS. Lucas (1975) cites several factors which underlie user perceptions and attitudes, including the technical quality of the system; the policies and apparent attitudes of the MIS department and staff; 19 and the nature of the ongoing contacts with MIS personnel. Lucas also includes in this list the level of support by top manage- ment. Most program directors interviewed were skeptical about the technical quality of the information systems developed at the county level. This skepticism often focused on the accuracy of available information. For example, despite the ability in principle of several automated systems to track client reentries into the treatment network, this source of information is not utilized by program directors because they do not believe that the data are sufficiently accurate. Similarly, in those counties where centralized intake units complete initial client "workups," most program counselors report that they cannot rely upon the accuracy of data collected by these intake centers; instead, they typically repeat the intake process in order to develop more accurate internal records. Finally, in one county where the automated system appears to be well suited to responding to Federal reporting requirements, the monthly client flow summaries are completed manually, based upon telephone queries, in part because of a general feeling that the MIS data base is often inaccurate. Additional user skepticism can be traced to the format of the available output and to the problems and delays encountered in accessing the system. Many program directors have emphasized that existing systems do not meet their information needs. More specifically they cited: ° Bulky computer printouts that contain masses of data (but little information) and which are difficult to comprehend; ° Frequent and lengthy delays in obtaining special reports; ° Failure to meet schedules for routine output; and ° Inflexibility of routine reports, i.e., the inability of the MIS unit to incorporate the special requests of a particular program into the regular reports issued to that program. Despite several of the newer drug abuse information systems having the potential for producing special reports and even for online user access to the data base, many program directors pointed out that it is not an easy matter to obtain specific infor- mation on an ad hoc basis. Instead they impute a general feeling at the county level that if the potential system users were to become actual users, system operating and maintenance costs would substantially exceed projected budgets. Thus, program personnel typically must submit requests for information and await the results of lengthy administrative approval processes. Several directors expressed resentment over this, saying in effect, "They don't really want us to use the system; it's for them, not us." Finally, it should be noted that by far the most frequent circum- stance of contact between MIS personnel and program-level users is the correction of input error. As Lucas (1975) points out, "Contact is generally thought to improve the attitudes between two groups of individuals by increasing understanding. However, 20 when contact occurs under unfavorable conditions, worse attitudes may result." It is not enough that the attitudes and perceptions of the intended users be positive. Even if this were true, the MIS might still falter because of unfamiliarity by human services personnel with the effective use of information. Program managers often do not know what information they actually need. "Instead, expecting to use nothing, they ask for everything" (Foster and Evans 1977). This dilemma was also noted by Ackoff (1967), who made the observation that such requests ought to be viewed as rational, defensive behavior on the part of managers who do not entirely understand the phenomena they control. It appears that many program administrators are uninformed about the ways in which quantitative information can be utilized to enhance decisionmaking. Not only are they unfamiliar with the ways in which data can be brought to bear upon specific infor- mational questions, they are equally unprepared even to ask appropriate questions. County-level managers have made little effort to provide program personnel with the training necessary for competent and productive system use. In summary, the implementation of existing systems in the drug abuse field has suffered seriously because of the failure to enlist the support of intended system users. The need is to recognize and adequately cope with the natural sources of anxiety and to cultivate positive user perceptions of and attitudes toward the MIS system, its component units, and county administrations. The effort must include training the program managers in the use of the systems. Support at the Data-Collection Level A third group whose support is necessary to the successful implementation of any information system is the personnel respon- sible for data collection and recording. Pittel (1974) has commented at length on the many subtle and not so subtle ways in which systems are sabotaged at the data-collection level. These range from active and vocal resistance at the outset to: ° Failure to report accurately on complicated or sensitive cases. ® Frequently missing data items. ® Data-collection biases toward simplification and underreporting. ° Failure to submit data in a timely fashion. ° Failure to correct data-collection errors as noted by subse- quent edit procedures. 21 Based upon interviews with counselors, there appear to be three factors which underlie these problems. First, many counselors have yet to be convinced of the integrity of recently instituted confidentiality requirements. Moreover, their general lack of familiarity with the potential for safeguards on access to computer- ized records and their perspective of county-level officials as being "outside" the actual treatment system do little to allay these concerns over confidentiality. Finally, the occasional, well- publicized abuse of confidentiality in an MIS serves to exacerbate these concerns. Second, counselors frequently cited excessive and complicated paperwork as a primary reason for their misgivings about MIS implementation. The unnecessarily complicated appearance of many data-collection forms and an examination of the data items actually required by most county information systems suggest that counselor time spent completing forms need not be excessive. To account for this disparity between widespread counselor perceptions and apparently minimal time requirements, one must recognize that in most treatment programs compliance with county MIS mandates is only part of a much larger system of required paperwork. Programs are typically required to respond to the informational demands of a number of county, State, and even private agencies to support their funding. Examples include revenue-sharing authorities (often at both the city and county levels), title XX agencies, local foundations and philanthropic organizations, school districts, welfare departments, and criminal justice agencies. And each of these may require that program personnel complete yet another set of forms with slightly (often ever so slightly) different data elements. A third factor which seems to underlie counselor resistance to data collection is that from their perspective the effort is apparently to no avail. Just as program directors remain for the most part unfamiliar with the uses of information at the county and State levels, counselors are allowed only rarely to observe the use of an MIS in decisionmaking at the program level. In part this results from the little use that is actually made of existing systems. But it also appears that to the extent that these systems are used by program managers, little effort is made to demonstrate to counselors and other program personnel that the data which they have collected have any impact upon program operations. System Use Hirsch (1968) has pointed out that information, even the best of information, has value only to the extent that it can influence decisions. Conversely, information systems that are ignored by key decisionmakers are of little value to the organization. One NIDA MIS conference (1976) participant commented that, "It is not our experience that programs are using the information that they have." Interviews with program and county-level personnel across the country confirm this observation. This widespread failure to utilize existing systems can be traced to many of the deficiencies noted earlier in this chapter including: 22 ° The failure of current systems to provide information relevant to actual decision needs. ° The difficulties encountered by program personnel attempting to access specific informational items. ° The typical design of MIS output which seems calculated to discourage the use of information by deluging users with reams of unwanted data. ® Inadequate support for system use by county-level manage- ment. ° Skepticism by intended users regarding the relevance, quality, and utility of available information. ° Inadequate training of intended users. Yet failures in the design and implementation of existing systems do not tell the entire story. Even if each of these deficiencies were to be corrected, it might still be that the use of MISs as presently conceived would not be significantly increased. For what has been largely ignored thus far is the organizational context in which treatment programs and county agencies alike must operate. Wildavsky (1973) has said that planning and decisionmaking are meaningful activities only to the extent that they are accompanied by the power to implement. Yet administrators at all levels of the drug abuse treatment system are subject to a significant array of organizational and political constraints which substantially diminish the effective range of their decisionmaking ability. For example, while many county-level planners might wish to base their decisions about program funding upon a host of MIS-generated information, e.g., needs assessment, program performance, costs, they rarely are afforded such latitude. Instead their decisions are constrained by political pressures from neighborhood groups, ethnic organiza- tions, and county boards of supervisors (Foster 1976). Similarly, while program-level administrators might wish to encourage counselor performance by instituting a system of bonuses tied to performance (as documented by an MIS), they find that such a strategy is proscribed by current funding regulations. In short, many of the decisions which are potentially available to managers in the drug abuse field are not, in practice, currently realistic alternatives. Many of the truly important decision situations which do arise are perforce settled by political and bureaucratic considerations which have little to do with existing information systems. Thus, even at the county level the most prevalent uses of current MISs are: ° To document and justify decisions already made for quite different reasons. 23 ° To convince State and Federal funding authorities that the county agency is doing its job adequately, as witnessed by the mere existence of its technologically advanced MIS. ° To provide information to State legislative bodies in an effort to increase the flow of funds into the county. While these are perhaps laudable and certainly understandable objectives in their own right, it seems unlikely either that they reflect the original intentions of the county-level managers or that their attainment is ample justification for the substantial costs incurred. Designers of information systems would do well to examine carefully the nature and extent of actual decision alternatives available to managers (as well as to subordinates) and to design systems which respond to those perhaps more limited decisionmaking needs, rather than to the hypothetical needs of ideal managers in rather different kinds of worlds. That existing systems appear to be technologically sophisticated should not be allowed to obscure their meeting decisionmaking needs which are at best merely hypothetical, while failing with equal frequency to meet the actual needs of real-world managers. RECOMMENDATIONS The development of information systems in the drug abuse field is a relatively recent endeavor and the lessons and examples, both good and bad, which could have been drawn from other organiza- tional contexts have been largely ignored. The recommendations which follow are an attempt to apply some of those lessons. Three criteria have guided their selection. First, to effect a significant and lasting impact upon MIS performance and hence upon the programmatic performance of treatment organizations. Second, to appear feasible in light of the existing bureaucratic and political environment in which such agencies must operate. Finally, to promote a strategy which builds upon the nature of existing systems rather than advocating a wholly new beginning. Four broad categories of intent are discussed: ° Recommendations designed to increase the congruence between available information and actual decisionmaking needs. ° Recommendations designed to enhance the availability of information to decisionmakers. ° Recommendations designed to reduce the burden of information systems on treatment programs. ° Recommendations designed to increase the level of support for existing systems throughout the treatment network. 24 Meeting Real Decision Needs The stronger single theme throughout this chapter has been the importance of matching the generation of information with actual decisionmaking needs. Paretta (1975) has commented that this matching of information flow to the requirements of system users is "critical to the success of all information systems work." Ackoff's (1967) emphasis on gaining an understanding of the nature and locus of actual and potential decisionmaking goes directly to the crux of the problem: The function of any MIS is to supply information which bears upon particular decisions. Yet to organize the design of information systems around particular needs encountered in an organization requires an adequate under- standing of those needs. There is no such understanding, and instead the design of information systems has typically followed two rather different strategies. First, it has been assumed that the intended system users know best what information they need. The design of practically every operating MIS in the drug abuse field has included as least some gesture to this maxim either through meetings with treatment program managers or through formal or informal surveys. This approach is unlikely to be successful. Ackoff (1967) contends that one of the major fallacies of MIS design is the assumption that the manager knows what information he needs. For a manager to know what information he needs he must be aware of each type of decision he should (as well as does) make and he must have an adequate model of each. These conditions are seldom satisfied. Most managers have some conception of at least some of the types of decisions they must make. Their conceptions, however, are likely to be deficient in a very critical way, a way that follows from an important principle of scientific economy: the less we understand a phenomenon, the more variables we require to explain it. Hence, the manager who does not understand the phenomenon he controls plays it "safe" and, with respect to information, wants "everything". Second, regardless of the extent to which suggestions have been solicited, designers typically proceed to develop systems to sup- port decisions which they believe ought to be important. In effect, it is assumed that the planned information system will function as somewhat of a change agent within the organization, focusing attention upon those issues which appear to be important. A value judgment is made, or inferred, instead of exercising an objective, analytic, and facilitative stance responsive to real, idiosyncratic system needs. Therefore, rather than leading managers to new and presumably more sophisticated decision- making, such systems are often soon ignored by program directors and counselors alike. As an alternative, it is recommended that those who design information systems become more intimate with the decisions which 25 are actually made at various levels within each local program, not merely by asking but by becoming involved, at least as a partici- pant observer, in the program's decisionmaking processes. The spirit of such a recommendation runs counter to a major trend in MIS development in the drug abuse field: the design of model systems. Instead, the implication is that systems ought to be tailored to meet the unique decisionmaking needs of each organi- zation. As Paretta (1975) has insisted, "No one model exists which has universal applicability. Each model must be matched against [an organization's] changing needs dictated by its state of development and available resources." Several recommendations relating to each kind of informational gap appear in table 1. Enhancing the Availability of Information One of the primary tasks of MIS design is the specification of the ways in which information is moved to the decisionmaking point. Among the most conspicuous failings of existing information systems is that much of the needed information which is available within the system is in practice difficult for decisionmakers to obtain. Sprague and Watson (1975) have emphasized the importance of an inquiry system for special informational needs not covered by routine reports. Most existing information systems do include such a plan in their design. Yet in many cases the process of accessing needed information through such inquiry systems is fraught with lengthy delays. As a consequence, the consensus of most program directors is that the county information systems which ostensibly serve them were, in fact, never intended for their use but were instead designed to meet the needs of county and State-level personnel. One of the most significant and far- reaching changes which might be made in existing information systems would be the restructuring of these inquiry and special report procedures to facilitate ad hoc access by program personnel. Of equal importance to MIS performance is the form in which information is provided to decisionmakers. Consistent with the observations of Dickson (1968) regarding MIS design in other fields, managers and other decisionmakers in the drug abuse treatment field are typically deluged by irrelevant data. The effect of this deluge is to obscure the relatively few items of truly useful information and hence to minimize the likelihood of effective system use. Ackoff (1967) has suggested that the emphasis in systems design should be upon the processes of "filtration and condensation." Two specific contexts in which this recommendation has gone unheeded are: ° The routine, computer-generated reports which are provided to program directors ° The potential use of client case files as a source of clinical information 26 TABLE 1.—Suggested strategies for reducing informational gaps Le Informational gap Typeaf Licks if Recommendations Where action Information regarding | Treatment Counselor Training in techniques of treatment and reentry planning. | Programs client needs for and service various treatment referral. and ancillary services. Planning County Further research regarding alternative approaches to dif- NIDA ferential diagnosis and the specification of service needs. For the most part, the requisite data are already being collected. What is needed is an approach to syn- thesizing these data and/or aggregating needs across client populations. Information regarding | Treatment Counselor Provide intake and referral workers with time specifically | Programs the external environ- and service allocated to visiting programs and other agencies. ment. referral. Funding Programs County substance abuse agencies should serve as a clear- County inghouse for information regarding potential funding sources. Information regarding | Budgeting Programs Integration of fiscal subsystems into operational and/or County, State the relationship of costs to specific program activities and services. client-oriented subsystems. Standardization of units of service. NIDA 8¢ TABLE 1.—Suggested strategies for reducing informational gaps—Continued . Type of Locus of i Where action Informational gap as oF decision Recommendations is neaded Information regarding | Service Counselor Greater reliance upon structured clinical staff meetings. Programs treatment progress planning. and outcomes. Training in keeping clinical case records (as distinguished | Programs from service documentation for audits). Program Programs Analysis of attrition data to determine which kinds of Programs, county planning clients terminate treatment and at what point in their Staff treatment plan. Also, correlation of attrition with evaluation specific services provided immediately prior to termination. Technical County Reinstitution of performance monitoring on a status at County assistance least equivalent to contract compliance monitoring. planning. Available data include treatment longevity (i.e., attri- tion rates) and—in some counties—length of time after Funding treatment completion until reentry into the treatment system. The observation by Dew and Gee (1973) that more than half of all data items contained in routine computer printouts are ignored by system users may substantially understate the case in the field of drug abuse treatment. Most program directors indicate that they do not make use of monthly MIS output made available by county systems. The implication is that routine statistical output ought to be drastically reduced. Moreover, the format of that output ought to be simplified so that key items of information, e.g., trends, comparisons, are readily recognizable. The importance of client files as a means of conveying information has been largely ignored. Instead client files are used primarily as a source for cross-documentation during periodic program audits. As a consequence they typically contain much data which are of little use to treatment personnel. In order to enhance the availability of clinical information to counselors (and to supervisors), it would seem appropriate to adopt the following maxim: "Treatment files are for counselors." To the extent that the documentation of service delivery is deemed essential for accountability, such documentation should be entered into a separate set of files designed specifically for that purpose. Reducing the MIS Burden One of the most frequently voiced complaints by treatment program staff at all levels has been that compliance with existing MIS and other reporting requirements substantially reduces the time avail- able for the delivery of services. In contrast with this objection to the paperwork burden, the attitudes of county, State, and Federal administrators have inclined increasingly toward stricter standards of programmatic and fiscal accountability, the develop- ment of countywide information systems, and a recognition of the need for sound management in the face of limited available re- sources. The logical synthesis of these apparently conflicting priorities ought to have been the design of information systems which would reduce rather than exacerbate the burden felt at the program level. While this objective was certainly considered in the design of existing systems, the accelerating trend toward multiple-source funding has worked in precisely the opposite direction. Program directors are obliged to respond not only to the MIS requirements of county substance abuse agencies, but also to the reporting mandates of title XX agencies, LEAA, CETA programs, courts, probation departments, school districts, and city and county revenue-sharing authorities. For the most part such requirements could be consolidated into a single system. County substance abuse agencies must recognize that the integration of the drug abuse treatment system into a broader network of human services delivery agencies necessitates the negotiation of interagency agreements at the county level to facilitate such a consolidation. At the same time, Federal and State funds ought to be made available for MIS design and 29 implementation only to the extent that such consolidations have been effected and multiple reporting requirements have been eliminated. At the program level as well, more careful planning of the data- collection process would contribute significantly to the reduction of burden. For example, much of that burden can and should be shifted to clerical personnel whose primary function is the comple- tion of required paperwork. Thus, rather than make documenting entries (as distinguished from clinical observations) in the treat- ment records of each participant at a group therapy session, counselors should be required merely to submit a list of those participants to a designated records clerk. Similarly, much of the paperwork required at intake and termination might easily be completed by competent clerical personnel. In short, a substantial increase in counselor time available for service delivery can be brought about by the creation of clerical positions assigned spe- cifically to the tasks of data collection and/or recording. Finally, additional MIS burden can be eliminated by a revision of the cumbersome and ineffective quality control procedures currently being utilized. Nonfatal errors need not be itemized and returned to the programs for input revision; this practice has typically failed to bring about the desired correction of errors. Instead, they should be summarized by reporting counselor and by error type so that subsequent training can be directed toward error reduction. Increasing the Level of Organizational Support There is a need to develop support for MIS implementation at all levels within the drug abuse treatment system from county adminis- trators to program directors to counselors, and even to clerical personnel. While it is true that almost all of the preceding sugges-— tions are calculated, either directly or indirectly, to enhance the level of organizational support, they accomplish that task largely by eliminating or reducing the sources of user disaffection. In contrast, the strategies recommended in this final section focus directly upon building positive attitudes toward MIS implementation. Even if the burden of a large-scale information system were minimal, program personnel would hardly become enthusiastic MIS advocates simply because there are rarely any visible payoffs either for themselves or for their clients. Program directors and counselors alike need to know about and see the kinds of decisions which an MIS supports at the county level, Similarly, counselors and other program staff members need to understand the ways in which information is used by their directors and supervisors. The objective would be to give personnel at all levels a greater appreciation of the importance of the system and of the benefits which accrue ultimately to their clients. 30 Second, Hanold (1968) has aptly remarked that in addition to management-oriented information systems, there must be information- oriented managers. Yet, as noted by Touche-Ross and Co. (1976), many program-level managers, while sincere and well intentioned, are neither adequately prepared as managers nor well schooled in the use of information for management decisionmaking. A number of useful training courses have been developed by NIDA in an effort to assist program managers in upgrading their skills. However, these training packages are directed to a diverse audience while the problems faced by individual managers are often particular to their own agency's needs. A strategy which appears to offer more direct benefits is to institute a program of continuing onsite technical assistance. Field personnel, thoroughly familiar with the possibilities inherent in a particular countywide system, could be assigned to work within a program to identify and demonstrate ways in which the available information can be brought to bear upon management decisions. Emphasis can be placed on decision situations which actually arise in each program rather than merely upon a demon- stration of the system's capability. While this process is likely to take some time, it offers the potential benefit of developing pro- gram directors who are committed to the use of information in their role as managers. System users should become more involved at the program level. Opportunities for user involvement in the drug abuse treatment field may include: ° User-controlled steering committees to plan necessary system revisions. ° The development by users of criteria for ongoing system evaluation (Lucas 1974). ° User task forces to investigate alternative resolutions to identified systems deficiencies. REFERENCES Ackoff, R.L. Management misinformation systems. Management Science, Dec. 1967. Beehler, P.J. Integrated MIS: A data base reality. Journal of Systems Management, Vol. 27, 1976. Creative Socio-Medics Corporation. "An Assessment of Data Systems in the Drug Abuse Field." Unpublished technical report, 1976. 31 Dew, R.B., and Gee, K.P. Preliminary findings. In: Management Control and Information: Studies in the Use of Control Infor- mation by Middle Managers in Manufacturing Companies. New York: John Wiley & Sons, 1973. Dickson, G.W. Management information--decision systems. Business Horizons, Vol. 11, Dec. 1968. Dickson, G.W., and Simmons, J.K. The behavioral side of MIS. Business Horizons, Aug. 1970. Foster, T.L. "Another MIS Story: A Review of the Design and Implementation of the Drug Abuse Treatment Monitoring System in Santa Clara County." Unpublished manuscript, 1976. Foster, T.L., and Evans, T.A. "Problems and Prospects for MIS in Human Service Delivery." Paper presented at the Joint National Meeting of the Operations Research Society of America and the Institute for Management Science, May 1977. Hanold, T. A president's view of MIS. Datamation, Nov. 1968. Harris, C.M., and Moitra, S.D. "On the Transfer of Some OR/MS Technology to Criminal Justice." Unpublished manuscript, 1977. Hirsch, R.E. The value of information. Journal of Accountancy, June 1968. Hunt, L.G. Drug Incidence Analysis. Series A, 3. Washington, D.C.: Special Action Office for Drug Abuse Prevention, 1974. Lucas, H.C. Why Information Systems Fail. New York: Columbia University Press, 1975. Marlow, R., et al. Development of a Process Evaluation Methodology for Drug Abuse Treatment Programs: Summary Report. Sacramento: California State Department of Health, 1975. Michigan State Office of Substance Abuse Services. Project ISAMIS. Lansing, Mich., 1976. National Institute on Drug Abuse. "Technical Review of Management Issues and Systems in Drug Treatment." Unpublished confer- ence transcript, Washington, D.C., 1976. Paretta, R.L. Designing MISs: An overview. Journal of Account- ancy, April 1975, Pittel, S.M. "Snakes in Iceland: Issues in the Evaluation of Substance Abuse Treatment Programs." Unpublished manu- script, 1974. 32 Priesman, I. Third-Party Payments: Alternative Funding Sources for Drug Abuse Treatment Programs. Washington, D.C.: National Institute on Drug Abuse, 1976. Project ISAMIS: Introduction. Michigan State Office of Substance Abuse Services, 1976. Social Issues Research Associates. "Drug Abuse Treatment Out- come Study: Final Report." Unpublished manuscript, 1977. Sprague, R.H., and Watson, H.J. MIS concepts. Part 1. Journal of Systems Management, Jan. 1975. Touche-Ross and Co. "Program Management Review Project: Final Report." Unpublished manuscript, 1976. Wildavsky, A. If planning is everything, maybe it's nothing. Policy Sciences, Vol. 4, 1973. AUTHOR THOMAS L. FOSTER Pacific Institute for Research and Evaluation Berkeley Center for Drug Studies Berkeley, California 33 2. State-of-the-Art Review Drug Abuse Management Information Systems in Single State Agencies Paddy Cook, Barry Rosenthal, M.S., and Cheryl Davis, M.A. Information systems within State governments specifically designed for the management of drug abuse treatment and prevention services are a recent phenomenon, not more than 5 years old, and currently entering a critical stage. Temporary funding incentives which were originally provided to the States by the National Institute on Drug Abuse, Division of Scientific and Program Information, (NIDA/DSPI) during 1974 to install a stand- ardized Federal system have expired. The Single State Agencies (SSAs) for Drug Abuse have meanwhile been delegated more authority for management in keeping with the decentralization goals of revenue-sharing guidelines and have assimilated Federal data requirements into their own organizational and information management environments. A variety of drug abuse management information systems now exist in the States with different levels of technical complexity, selections of data elements, and patterns of report use. Continued development and utilization of these systems will involve new relationships between NIDA and the States. Within the National Institute on Drug Abuse, the Division of Resource Development has also been involved in the development of information systems at the local program level. The Services Research Branch of that Division joined with the Division of Scientific and Program Information in February 1977 to sponsor a state-of-the-art review of drug abuse management information systems within the SSAs, the point at which Federal and local program interests converge. The purposes of this 5-month project were twofold: ° To survey the nature and extent of drug abuse information systems within the SSAs. ° To determine the areas of primary concern for future Federal technical assistance or collaboration among the States relative to MIS design, development, or revision. 34 HISTORICAL BACKGROUND To appreciate the significance of the survey findings and the recommendations based on them, one must be aware of the brief but complicated history of drug abuse management information systems. Many aspects of the current national framework for the delivery of drug abuse treatment services were established by the Drug Abuse Office and Treatment Act of 1972. This bill estab- lished a Special Action Office for Drug Abuse Prevention (SAODAP) which was charged with the responsibility for determining the extent of the drug abuse problem, identifying treatment resources and methods, and developing a strategy for drug abuse prevention and treatment. The law also mandated the creation of Single State Agencies in the States to carry out the planning functions, manpower development, and implementation of appropriate preven- tion and treatment activities. The organizational framework in which these Single State Agencies were to be located was not specified and the subsequent variations in implementation affected the priorities afforded to drug abuse problems and the later development of their information systems. SAODAP implemented the first drug abuse data-retrieval system called the Client Oriented Data Acquisition Process (CODAP) with the approval of six Federal agencies concerned with treatment and rehabilitation services. The original version of CODAP, which became a reporting requirement for all federally funded drug abuse treatment programs in 1973, provided demographic data and information on the drug-abusing behavior of clients admitted for treatment. Data were collected on individual admission forms, case sample progress reports, client census summaries, and funding information forms for all programs. In 1974, CODAP underwent a major revision. CODAP II added a client discharge form and an expanded admissions form which were designed to enable measurement of changes in client behavior during treatment. During the same year the functions of SAODAP, which had been formed as a temporary organization, were merged into the newly formed National Institute on Drug Abuse. The Division of Scien- tific and Program Information (DSPI) within NIDA was directed to implement an integrated and comprehensive drug abuse management information system that would satisfy the management needs of both the Federal and State governments. This new system was called the Integrated Drug Abuse Management Information System (IDAMIS) and was composed of three separate but coordinated systems: ° CODAP II--information on clients in treatment. ° Drug Abuse Program Reporting Unit (DAPRU) later known as the National Drug Abuse Treatment Utilization Survey (NDATUS)--data on all units providing drug abuse prevention and treatment services. 35 ° Financial Management Information System (FMIS)--information for budgeting, cost finding, program monitoring, and financial control. Original plans called for these subsystems to be transplanted intact (or with only slight modification) to the SSAs complete with fully workable software. Encouragement for State implementation of these modules was provided by DSPI in the form of 2-year contracts for establishing an Integrated Drug Abuse Reporting System (IDARP) in each Single State Agency. Sufficient resources for system installation and maintenance were offered in the form of (1) funds for staff, manpower training, and computer hardware (EDP equipment), and (2) technical assistance. The first IDARP contracts were negotiated in 1974 and all 50 States and 5 trusts and territories responded in less than a year. The Single State Agencies have focused primarily on two of the three subsystems: CODAP and NDATUS. FMIS remains an optional component but CODAP data are required from all federally funded treatment programs and the annual facility resources survey (NDATUS) is conducted through the IDARP managers. An intensive effort has been required to train staff and establish procedures for data collection and quality control. As these tasks have been completed, the emphasis of the SSAs has shifted from data collection to data utilization. Although the IDARP contracts expired in June 1976, the States have continued to support the Federal drug abuse MIS with extensions of unexpended IDARP funds and/or State moneys. The decisionmaking and reporting responsibilities assigned to the SSAs by NIDA (e.g., monitoring of statewide services, contracts, production of State plans, program management reviews, and mandatory reporting of CODAP and NDATUS) require a continued State involvement with federally defined information systems. New arrangements for NIDA incentives for continued SSA participa- tion may, therefore, be pending. In January of 1977, CODAP was revised again. The changes in the data elements were primarily motivated by a desire for more precise measures of change in client behavior from time of admission to time of discharge. Several modifications in the original plan for implementation of IDAMIS have discouraged the development of uniform management information systems within the States: ° The activity reports which were a part of the design of CODAP II provided aggregate data on clinic activities, client services, and clients' progress in treatment. These activity reports were soon dropped as requirements by NIDA, and States were then left with no uniform procedures for reporting client services or staff activities. 36 NDATUS, which was originally designed to be collected on all drug abuse services on a semiannual basis, was restricted to an annual effort covering treatment units only. Therefore its reports on resource availability have been less timely and comprehensive than originally anticipated. FMIS is an optional subsystem which has been implemented in only a few States. Some modifications will need to be made in the software before it can be used to calculate unit costs, and it is not a billing system. Software for generation of routine reports using CODAP data was never provided to the States. Instead, IDARP personnel were trained in the use of commercial statistical software packages. These software packages have provided greater flexibility in data analysis, allowing MIS personnel to design reports around State informational needs. Factors within the States have also contributed to the diversifica- tion of SSA-level MISs: Although NIDA produces reports based on CODAP data, the turnaround time is lengthy. Many Single State Agencies wanted to return timely clinical information to individual programs and, when there was a delay in the provision of CODAP software to the States, they either purchased soft- ware packages or developed proprietary software. In an increasing number of States, responsibilities for drug abuse and alcohol services have been merged into substance abuse agencies. These new organizations must meet the reporting requirements of both NIDA and the National Institute on Alcohol Abuse and Alcoholism (NIAAA). Attempts to handle these overlapping reporting requirements efficiently have resulted in many design innovations. The MISs in some SSAs have been affected by the larger organization of which they are a part. This is especially true of the SSAs which are located within Mental Health Departments since the National Institute of Mental Health (NIMH) was several years ahead of NIDA and NIAAA in the encouragement of MISs within the States. Again, design innovations have been required in order to avoid excessive paperwork. SURVEY METHODOLOGY In developing the methodology for the survey, one objective was to identify characteristics of State MISs that might be useful for further technical assistance efforts, or guidelines and manuals directed to the particular needs of groups of similar States. It was necessary, therefore, to discover general classification schemes applicable to drug abuse management information systems. It was 37 also necessary to structure the data-collection effort itself around defined content areas so that compatible descriptive information could be compiled and analyzed. In order to find appropriate classification schemes and link the survey with other work in the MIS field, a brief review of the literature was conducted. Four potentially useful classification schemes emerged: ° Organizational location and purposes of the MIS (environment) ° Information contained in the system (design) ° Processing technology (mechanics) ° Management utilization of the information (decision-assistance) One common classification of data banks divides them into statistical and intelligence systems depending on the regulatory purposes for which they are used (Westin 1971). Statistical systems aggregate data to study variations in group characteristics for planning and policy-setting purposes. Intelligence systems, by contrast, provide case dossiers on individuals for treatment, administrative, or punitive purposes. State drug abuse information systems and the CODAP files, operating under confidentiality regulations, restrict reports on clients to statistical aggregates; some States also produce internal clinically oriented individual client records of urine results, dosage schedules, or counseling sessions by client identification number. These individual case reports, however, are not interfaced with other systems such as health records, criminal activities, or welfare services. This classification scheme is further amplified (Westin 1971) by grouping governmental information systems according to the organizational situations in which they are found as well as the purposes they serve. Although these particular classification categories have limited direct application to drug abuse infor- mation systems, the different organizational arrangements of the SSAs are of potential descriptive value. Questions focused on these organizational or environmental differences were therefore included in the survey. A second potential classification scheme was suggested by differ- ences in system design or data components used as input documents. Descriptions of information systems for the management of mental health clients, which are analogous in many ways to those concerned with drug abuse patients, frequently delineate subsystem components that can be combined in several ways to generate a variety of output reports (Elpers and Chapman 1973; Weinstein 1976). Some systems only contain data on services rendered to clients; others include measures of client improvement or unit costs, available resources, revenue and expenditures, etc. The number of instru- ments and the types of variables with common definitions that are included are a crude measure of system design complexity. 38 Another method of categorizing MISs was based on technological complexity or the historical availability of data-processing equip- ment (Withington 1974). The tasks actually performed by the processing equipment can be used to describe important differences among systems: transactions such as time posting and printing paychecks can be performed by simple data-processing systems; automatic subfile updates in systems with common data definitions can be made by integrated information systems; selection of specific data elements is possible in information retrieval systems with random access capabilities; and the meaning of data can be interpreted in management information systems which interface elements from different subsystems. These "real" information systems are designed to match scarce resources with areas of critical need (Davis and Freeman 1976). A final classification scheme was derived from a systems analysis approach that differentiated types of management responsibilities and graded information systems by the management functions they support (Anthony 1965). Systems designed around operational control functions usually consist of simple accounting transactions; more complex systems for management control monitor variances from established goals or standards; and systems for strategic planning assist with the deployment of resources and definition of new objectives. This approach can be used (Kennevan 1973) to differentiate automatic data-processing systems from more complex management information systems that supply information condensations and analyses useful for setting policies and standards, forecasting resource require- ments, and controlling day-to-day operations. The decisions made by managers can be classified according to a similar scheme as structured, in operations such as client billing; semistructured, in managerial situations such as evaluating staff performance levels; and unstructured or ad hoc, in strategic planning functions such as agency reorganizations or predicting new high-risk target groups. Different information is needed to make structured decisions than to suggest acceptable alternatives when decisions are less predictable (Gorry and Morton 1971). Fairly sophisticated computer applications such as simulation models or interfaces of data from a variety of sources in online manipulations are required for meaningful analyses of social data where the variables describing behavior are numerous, inter- active, and seldom organized around a theoretical perspective (Pool, McIntosh, and Griffel 1968). Another decision-theory model cautions management against the organizational strains created by an overabundance of unorganized data (Ackoff 1967) and proposes distinctions between operations in which optimal solutions can be routinely identified and those management func- tions that require choices based on estimates and predictions. One clear message which emerged from the literature was that the true management information systems were those used to support organizational decisionmaking. Therefore the original survey design included an attempt to differentiate data-retrieval systems from management information systems. 39 The state-of-the-art review of MISs existing in SSAs was conducted February 1 to May 31, 1977. All 50 States and 5 additional counties were contacted and either visited or interviewed by telephone. Thirty field sites were selected based on such factors as systems complexity or the presence of innovative design features. An effort was made to visit States which used a variety of approaches in terms of data collection, utilization of CODAP, and scope of drug treatment and prevention services. In States which were not visited, documentation (input forms, manuals, or reports) was requested to supplement the information obtained by telephone interviews. The data-collection phase of the project was organized in less than a month; the site visits and telephone contacts were completed during the following 5 weeks. The Single State Agencies were generally quite cooperative in scheduling visits and providing the information requested. An Interviewer's Guide was the primary tool used for both the field visits and the telephone contacts. This handbook estab- lished the framework for the project, outlined procedures for arranging and conducting the site visits, and presented the general content areas for information collection. There was not enough time to devise and pilot test a standardized questionnaire nor to seek the necessary governmental clearance for such an instrument. Since it was not possible at that time to predict what information would be available in each SSA, how detailed that information would be, or which areas would be most fruitful for analysis, the format of the guide allowed for open-ended responses. States and counties were then described in a narrative report which reflected differing emphases and mixes of detail. This approach provided the flexibility needed to document widely varying organizational and technical developments at the State level but established a framework from which categories and classifications could emerge. The four general classification schemes discovered in the literature were repeated in the Interviewer's Guide which outlined the basic procedures for observing and describing MISs in the States and counties. These general areas were: ° The impact of the SSA/county organization on the design and installation of a MIS. ° The instruments and procedures used for input into the system. ° The processing modes, technology, and personnel used in data transformations. ° The output records generated and distributed for management information and utilization. 40 Whenever possible two consultants visited each site and shared responsibilities for interviews. An organizational consultant described the functions and structure of the SSA, the scope of drug abuse services, the expectations and attitudes toward quan- tified information, other constraints or supports for an MIS, and the actual generation and utilization of reports from the system. The technical consultant analyzed the types of data collected, the quality control procedures in effect, the hardware and software used, and the flow of data through the system. The major methodological problems encountered were related to the uneven quality of the information obtained. The data gathered reflected only questions that were probed; consultants differed in their perspectives and States did not always have sufficient information readily available. These difficulties were not unantici- pated and were considered normal for a study of this nature attempting to obtain baseline data. After all the narrative reports had been completed, a coding sheet was developed and the data contained in these reports transferred to the coded format for computer analysis. The code sheet contained the same four general content areas that had guided the interviews and narrative descriptions. However, the objective of differentiating data retrieval systems from management information systems could not be accomplished. Limitations on time and the availability of information made it impossible to make valid assessments of the manner and degree to which different SSAs were using their information systems. The fact that many systems were in a state of flux with many SSAs in the planning or implementation stage of a new system meant that there was not proof yet of what the output of the new systems would be or how that output would be used. Where reports were being produced by established systems, the format and distribution of the reports could be described but frequently the degree to which the information contained in the reports was actually inte- grated into the decisionmaking process could not be determined. FINDINGS This section reports the major findings of the MIS survey in six areas: ° Level of MIS Implementation ° Decisionmaking Assistance Expected From MISs ° System Mechanics and Technologies ° MIS Typologies 41 ° Most Frequent Problems o State Summaries LEVEL OF MIS IMPLEMENTATION One major finding was that many drug abuse management informa- tion systems were in a state of flux with no consistent develop- mental direction. Some States were adding new data components and automating subsystems while other were cutting back to minimal requirements and decreasing budgets for processing. This made it difficult to describe and classify the systems (table 1). TABLE 1. Level of MIS implementation Number Percent Phase! of States of States N=48 Minimal CODAP 14 29 Established systems 13 27 Developmental phase of new MIS 6 12 Piloting a new MIS 7 15 Major revisions in earlier MIS 8 17 "Missing States = 2. Fourteen states (29 percent) utilized only CODAP and NDATUS information as the basis of their MIS and had not moved beyond the NIDA instruments to develop additional components or different systems. Thirteen States had established systems beyond CODAP and were not currently undergoing major changes. Overall, 56 percent of the States had apparently reached a momentary equi- librium in MIS development, although information concerning either their immediate or long-range plans was not always obtained. The remainder of the States were either planning for change, piloting new systems, or undergoing major system revisions. The 13 States which were in a planning or piloting phase had not placed formal priorities on MIS development. The reasons for delayed development were not coded but included such barriers as lack of State support, staff limitations, or organizational changes. One State acquired a new SSA director who resolved the stalemate in information system development; another director reshuffled 42 internal MIS staff responsibilities; and still another had to await a bureaucratic reorganization before MIS plans could be approved. The grants awarded through the Council of State and Territorial Alcoholism Authorities (CSTAA) for MIS development were a decided stimulus to many of the States including several which were launching new systems at the time of the survey. In the eight States undergoing major revision, anecdotal informa- tion indicated that three factors were apparently at work: ° Recent CODAP revisions and similar modification in the NIAAA reporting system. ® Mandates from State legislatures. ° Pressures from organized community interest groups and task forces. Such changes in reporting requirements were often sudden, unpredictable, and extrinsic to the orderly and controlled develop- ment of the MIS as planned by the State Agency staff. In some cases the perspective for data-collection efforts changed completely. In one State, which had developed prompt evaluation methodologies based on treatment outcomes, a new policy on confidentiality curtailed the submission of any data on individual clients; the revised MIS concentrated on process measures instead. By con- trast, another State which had recently installed a system to measure program efficiencies and report variances from targeted goals had just been required by its legislature to develop new measires of program effectiveness and to implement followup studies of clients after treatment. The survey did not include enough questions about the maturity of systems to draw conclusions about developmental stages or cycles. However, two patterns of MIS evolution were observed. Some States moved slowly but directly from simple to more complex arrangements adding new information components gradually and incorporating them into the ongoing system. Other States began with grandiose schemes calling for "total" systems designed to provide all the information that might ever be wanted. In trying simultaneously to meet the information needs of program administrators, clinicians, researchers, elected officials, and the general public, they swamped managers with what Russell Ackoff, in his classic "Management Misinformation Systems," described as "an overabundance of irrelevant information." Experi- ence taught them that "facts" even when available were often less influential than other factors (such as political considerations) in real-life decisionmaking, and that computerized information sup- ported some decisionmaking tasks better than others. As a result some States were cutting back on the volume of data collected and focusing on using information for management pur- poses rather than simply acquiring and storing it. In one State whose MIS was undergoing modifications, the revised goal was to 43 produce the "minimum number of justifiable output reports." Several States were reducing the amount of extraneous research information collected routinely on clients but offered coded spaces on the data-collection instruments which could be used as needed for special studies. Still others were moving to adopt CODAP as their central module for client-related information in order to avoid unnecessary duplication of reporting, while adding other information to build a more comprehensive MIS. In general there was a growing awareness of the need for economy in information collection systems. Decisionmaking Assistance Expected From MISs One set of questions directed to SSA directors and their staff concerned expectations from an MIS: What types of information were needed to fulfill managerial responsibilities? Responses were categorized in the nine general areas depicted in table 2. TABLE 2. —Decisionmaking assistance expected from an MIS SSAs indicating expectation Area Number Percent TT TN=50 Program monitoring and evaluation: Outcome/impact/effectiveness 31 62 Process/utilization/efficiency 24 48 Program planning: Needs assessments 26 52 Resource inventories 14 28 Budgeting by unit costs 21 42 Meeting external reporting requirements/requests 11 22 Research 10 20 Clinical treatment needs 7 14 Standards development 6 12 By far the greatest emphasis was placed on assistance with program monitoring and evaluation. Sixty-two percent of the States ex- pected to measure program effectiveness, and 48 percent expressed an interest in process or efficiency measures. This concern with monitoring and evaluating treatment programs was a natural correlate of SSAs being required to allocate and account for funds and services--usually through subcontracting mechanisms rather than direct administrative control. 44 Interest in measuring client "success" and relative program effec- tiveness was expressed more often than a desire for measurement of program efficiency. There was a general demand for outcome and impact studies to assess reductions in client drug problems, increases in client productivity, and social and financial benefits accruing to communities receiving treatment and prevention serv- ices. There was a desire to rank and compare programs so that allocations of funds, staff training, and technical assistance could be made on a rational basis. However, at the time of the study only a few States had actually implemented policies that tied review results directly to funding approvals. The expressed demand for effectiveness measures rather than efficiency rates may have been related to the type of data already in the system since CODAP-based systems already require client-oriented information while process measures such as staff activities or units of services must be added. Second only to the interest in program monitoring and evaluation was an interest in using data for program planning and needs assessment. The desired information included external "indicators" from the community of drug use and abuse rates and incidence and prevalence studies with which to make projections of demand for treatment and prevention services. Fifty-two percent of SSAs expressed a desire to use MISs as support for needs assessments but only 26 percent were regularly collecting such indicator data. A few States had developed formulas for converting needs assess- ments to weighted indices by geographic area to be used as input in determining funding awards and facility locations. This repre- sented a trend toward the utilization of available data for planning decisions. Another type of information required for planning is an up-to-date inventory of resources in order to identify gaps in services. Only 28 percent of SSAs expressed this need--probably due to the availability of the NDATUS data base. Another category of support desired from an MIS was unit-cost reporting to be used for budgeting, third-party reimbursements or rate negotiations, and measures of client costs. More than half the States expressed this need but only a few had made progress toward developing such a capability. Many States were planning to tie expenditure information to client service or modality of treatment in the very near future, or were piloting such a system. The other four categories of anticipated MIS support received much less attention from the SSAs. Assistance in meeting outside reporting requirements or requests was mentioned by less than a quarter of the States. The lack of stress on this area may have been related to experiences with political realities. At least two States felt helpless in the face of community power groups using emotional arguments; however, several others cited cases in which studies and special reports based on MIS data had been influential in obtaining funding or reversing adverse policies. 45 Research needs were considered important for MIS assistance by 20 percent of the survey participants, and clinical treatment supports by 14 percent. It could be argued that these two categories are not usually the primary functions of SSAs and are, therefore, less frequently mentioned as areas requiring support from State-level management information systems. The aggregate information needs at the SSA level would generally not be the same as the individual client status reports desired by clinics for treatment planning and intervention. As mentioned previously, many States were interested in reducing the amount of information in their systems that was geared to research rather than manage- ment needs. Only 12 percent of the States specifically mentioned a desire to use the MIS for developing standards or supporting policy decisions, a function strongly advocated by most designers of information systems who think of strategic intelligence as the primary role in which management should receive decision assistance from carefully selected information. Overall the types of support expected from information systems in the States indicated a movement away from reports related to individual clients and an increased emphasis on those useful for management control and resource deployment. At this point, however, most States had not yet conceived of an MIS in strategic terms as useful for setting objectives, determining policies, or acquiring appropriate resources. There was some evidence, based on the interest in development of formulas for resource allocations, that progress was being made in this direction. System Mechanics and Technologies The survey included a number of questions about the mechanics of MIS operations which are reported in the four tables that follow: ° Instruments for data collection ° Control of data quality ° Technology for operations ° Reports for management decisions The number of data-collection instruments in use among the systems varied widely between one form and an incredible 15 instruments (table 3). States using only the 3 basic CODAP reports had the highest single representation (15, or 32 percent). Two-thirds of the States used 5 instruments or less and a quarter of them reported between 6 and 10 forms, not all necessarily in the same programs and not all a part of the same system. These higher figures usually reflected the overlap of several systems such as CODAP, NIAAA, and a State MIS. 46 TABLE 3. —Instruments for data collection Number Percent Instruments used of States of States Number: 0-4 21 45 5 10 21 6-10 12 25 > 10 4 8 Type: ? CODAP 42 84 Non-CODAP admission 16 32 Client services 11 22 External indicators 13 26 Non-CODAP discharge 11 22 Expenditures 10 20 Change in status 8 16 Aggregate client activity summary 9 18 NIAAA 10 20 Staff activities 7 14 Prevention/education services 8 16 Preintake contact 6 12 Aggregate client census summary 8 16 Addendum 6 12 Followup 7 14 'Unknown=3; N=47. 2N=50. When the instruments were classified by type, CODAP forms were used in 42 (84 percent) of the States with 16 (32 percent) of the States preferring to use their own admission forms and 11 (22 percent) also using their own non-CODAP discharge questionnaire. The overlap can be accounted for by the duplicate reporting requirements in 70 percent of the States. A client services form (22 percent) and routine collection of drug-related "indicator" data (26 percent) were next in popularity. Only six States (12 percent) mentioned trailer sheets or addendum items of client- related data to supplement CODAP. Seven were undertaking client followup studies using instruments specifically designed for that purpose. Twenty percent of the SSAs required routine expenditure information from clinics and programs. Eight States (16 percent) had implemented at least a cursory attempt to aggre- gate information related to prevention and education services. Numerous manuals were collected from States using their own instruments in addition to CODAP and the quality of these manuals varied widely. The best provided a full overview of the MIS in a 47 series of volumes including explanations of the data flow and examples, with interpretations, of output reports. The best manuals had clear definitions of all data elements, consistent instructions, and illustrations of all the forms. However some were so poorly printed that the words were illegible or so poorly written that the instructions were unclear. The most appealing manuals contained some humor to alleviate the tedium inherent in such documents. In a few States new forms had been designed and implemented without an updated manual of instructions; this practice was not conducive to the maintenance of a high-quality data base. The data-collection instruments also showed variations in quality and design. One of the most imaginative forms combined admission, discharge, followup, and client-status-change reports on one-half sheet of paper. Separate coding instructions were available, and color codes on the instrument depicted the boxes to be completed at different points in the treatment process. Users were apparently quite satisfied with the form, which was not true of staff burdened with a six-page intake form in another State. In some States the instruments had coding instructions printed on the back but no separate instruction manuals. The effectiveness of this practice was not evaluated. The procedures for control of data quality were usually critiqued in the narrative descriptions of the States (table 4). Forty percent of the States rated themselves as having good practices, 38 percent had adequate quality control, and 22 percent were described as having limited control procedures. The need for continued improvement in data quality in over half the States was borne out by frequent complaints from IDARP managers about the time required to edit incoming forms, resolve errors, and provide technical assistance to the worst offenders at the clinic levels. Generally the States were aware of the great effort required to maintain quality control in a system even after completion of the initial training and implementation phase. This emphasis on the requirements for quality control may be at least partially attributed to the establishment of the IDARP function as a central control point in the SSA for CODAP monitoring. Several of the States had further decentralized this process to regional offices that were assigned the major responsibility for logging forms, batching the inputs, and resolving error reports. When broken out separately, the specific procedures for quality control that reportedly received the most consistent attention were manual edits (91 percent of the States), staff training (79 percent), and the use of instruction manuals (65 percent). Machine edits, producing either error reports based on consistency checks or turnaround documents, were used in 59 percent of the States, frequently in tandem with another manual edit routine. When questioned about the frequency of data submissions, most States (88 percent) reported a monthly schedule. Only 12 percent required weekly submissions; three States mentioned quarterly 48 TABLE 4.—Control of data quality Number Percent Quality control of States of States Degree of quality control exercised: Limited 8 22 Adequate 14 38 Good 15 40 Specific procedures for quality control: ? Manuals 22 65 Training 27 79 Manual edits 31 91 Machine edits 20 59 Reconciliations 21 62 Logging controls 19 56 Frequency of instrument submissions: 3 Weekly 5 12 Monthly 36 88 Quarterly 3 7 Annually 3 7 ' Not reported=13; N=37. 2Undefined=16; N=34, 3Unknown=9; N=41. collections. In States with varying submission dates for different forms, the most frequent submission requirement was coded. Seventy-three percent of the States were operating an automated information system with another 17 percent pending automation (table 5). This represented a significant increase in automation; in the previous year only 54 percent of the States were automated and 13 percent pending automation. However, almost half of the States still performed some manual data processing, e.g., tabula- tions of figures from the client-flow summary. No clear relationship could be established between the level of sophistication of the data-processing technology used in each State and system output in terms of quantity or quality of reports. Analysis of the relation- ship between level of technology and level of output was made more difficult by the fact that some States operated concurrent manual and automated systems. Most of the SSAs (77 percent) used State-operated computer facilities, usually available through a central department of admin- istration in the State's capital city. Another 15 percent found 49 TABLE 5. Technology for operations Number Percent Technology of States of States Processing mode: Automated 35 73 Pending automation 8 17 Manual 22 46 Computer facility available: ? State-owned 33 77 Private vendor 3 7 University 7 16 Special technical aids: ? Remote job entry 7 14 Interactive terminals 5 10 Automated system in use:* File maintenance only 4 11 File maintenance and report generators 3 8 Commercial software packages/ NIDA tapes 17 46 State tapes and commercial packages 13 35 Number of MIS staff:® 0-2 13 32 3-5 13 32 6 or more 14 35 Organization of MIS staff:® Within SSA 41 87 Outside SSA 6 13 'Unknown=2; N=48, ‘Unknown /none=13; N-37. 2Unknown/none=7; N=43. *Unknown=10; N=40, 3Unknown=1; N=49, *Unknown=3; N=47. 50 TABLE 5. —Technology for operations—Continued Number Percent Technology of States of States Programer available on staff: Yes 19 50 Use of CODAP tapes:® Received from NIDA and used 17 35 Produced for NIDA 3 6 Produces own tape for State use only 12 25 Not used/not received 16 33 § "Unknown=12; N=38. ®Unknown=2; N=48, universities to be cooperative, sometimes providing faster turn- around and more user assistance than the State facilities. Five States had developed an online capability using interactive terminals with display screens installed at the central MIS office or in other parts of the State. These offered immediate access to the data base and the ability to select individual records by client number or to provide cross-tabulations for different groups on a series of prepackaged table formats. It was not possible to determine whether sophisticated machinery served as an impressive gimmick or a useful management tool. Such machinery certainly increased the flexibility of the reports produced and eliminated problems associated with mountains of unused paper printouts. Seventeen of the States were using commercial software packages and the tapes received from NIDA as their basic system; 3 States produced CODAP tapes for NIDA from their own instruments; and 13 States made their own tapes for use with commercial packages for analytical purposes. States generally had small staffs to work in the management information system: almost two-thirds reported five or fewer available personnel. States with MIS staff located outside the immediate SSA (13 percent), often in a department of research and evaluation, usually had a strong commitment to the develop- ment of proprietary State systems beyond CODAP. Half the States had a programer available to them on their own MIS staff, most of these apparently being employed to write software for file building and maintenance rather than report generation. States with their own software for analyzing the CODAP II data were acutely aware of the expense and effort necessary to modify it and were amenable to using commercial software for routine anal- yses in the future. 51 The production of routine reports has begun only recently in many of the States that focused earlier efforts on data-collection procedures. At the time of the survey 67 percent of the States were producing regular reports; 53 percent of this group produced reports at least monthly (table 6). Almost half of the routine reports were reputedly available within 2 weeks after source documents were received. This figure was not validated but it did include manual as well as automated reports. According to the SSAs, 64 percent of clinics and programs did receive reports from the States, but some were cursory and consisted only of annual or quarterly analyses. One of the chief complaints of MIS staffs was that the program- and clinic-level personnel were not skilled in interpreting output reports. Many IDARP managers reported that printouts were discarded by program personnel who provided no feedback on the information contained in the reports. The sample State MIS output reports collected during the survey were analyzed for format and 58 percent contained only printouts. The greatest number (79 percent), however, compiled figures onto tables with clear labels which were readable and not difficult to interpret. Sixty-four percent of the reports used narrative summaries and 33 percent presented graphics to assist interpretation indicating a desire on the part of MIS staff to make quantitative analysis understandable to the users of the reports. Many MIS staff members stated that they planned to train program and clinic staff in data interpretation in order to increase understanding and utilization of the MIS reports. Output reports were also analyzed in terms of the categories or aggregations of the data. Most commonly cuts were made on a clinic or program basis (82 percent). Other usual aggregations were by planning area or some other geographic grouping (36 percent); by client demographics; by drug problem categories (36 percent); and by modality /environment of treatment (32 percent). The most frequent types of reports found in our samples were descriptive client profiles (78 percent), census or utilization studies (60 percent), drug problems or trends (40 percent), and outcome comparisons (40 percent). Twenty-four percent of the States produced either client registers to assist the unique identifi- cation and tracking of client treatment status or lists of clients active in treatment for verification at the clinic level. These report categories again reflected the primary concerns of SSAs with outcome evaluations, efficiency measures, and needs assess- ment. MIS Typologies The analysis of the information collected during this survey of drug abuse management information systems was directed toward answering the following questions: 52 TABLE 6. —Reports for management decisions Number Percent Characteristic of States of States Production mode:' Ad hoc only 16 33 Regular reports 32 67 Frequency of reports:? Weekly 2 6 Monthly 17 47 Quarterly 18 50 Annually 18 50 Timeliness of reports (after instrument submissions) :? 7-10 days 1 3 10 days-2 weeks 14 48 2 weeks-3 months 5 17 > 3 months 9 31 Use of CODAP data:* Yes 35 70 Format of reports:® Printouts 19 58 Graphics 11 33 New tables 26 79 Narrative summaries 21 64 Cuts of the data:® Clinic/ program 41 82 Planning area/geographic unit 18 36 Client group 18 36 Modality /environment 16 32 Categorical program 6 12 Funding source 3 6 'Unknown=2; N=48. *N=50. 2Unknown=14; N=36, *Unknown=17; N=33. 3Unknown=21; N=29. °N=50. 53 TABLE 6. Reports for management decisions—Continued Number Percent Characteristic of States of States Report types:’ Descriptive client profiles 39 78 Census/utilization 30 60 Drug types/trends 20 40 Outcome studies 20 40 Special research studies 15 30 Client lists/registers 12 24 Services 11 22 Unit costs 8 16 Individual clients 4 8 Report distribution/use:® SSA staff 46 92 Clinics/programs 32 64 External requesters 27 54 "N=50. *N=50. ° Are there distinct levels or types of management information systems which are identifiable and discrete? ° If so, what are their characteristics? ° What factors relate to the development of these characteris- tics? ° What common problems are experienced by MISs within each category? The effort to develop a typology was motivated not only by a desire to provide a framework for the analysis of the survey data but also as a possible basis for examining the feasibility of develop- ing a model MIS based upon the identified typologies. As discussed earlier, the data relating to the type of MIS were not forced into discrete categories but were descriptive and open ended. It was anticipated that if descriptive categories did exist, they would emerge following data reduction and analysis. Interest- ingly the first descriptive category to emerge was not based upon technological sophistication or data utilization for decisionmaking; rather, the overall design complexity based upon the scope and characteristics of the data collected appeared to provide the most cohesive descriptive categories. 54 The categories which emerged from this preliminary analysis were provided with labels which although somewhat descriptive were neither precise nor mutually exclusive. These typological cate- gories were: ° CODAP/IDARP: These systems processed only CODAP information. The degree of usage ranged from receipt of NIDA monthly tapes and use of commercial software packages for analysis to the development of State software which builds data files and produces output reports. The major distinguishing factor was clearly not technology but the scope of data-collection and analysis operations. Fourteen States were within this category. e CODAP and Addendum: Four States collected and processed additional client-oriented information as an addendum to the CODAP report forms. The primary focus of these systems was CODAP, with addendum sheets including data such as number of arrests, income, occupation, or other SES and drug-related elements. One State included on their addendum sheet clinic-level data which consisted of an expanded activity report. ° State Drug Systems: These States had designed and devel- oped their "own" integrated MIS using in-house data-collection forms and operational procedures. These comprehensive drug MISs went beyond CODAP in the range and scope of data collected and often included software capabilities such as file building, error recognition, automated editing, and report generation. Eleven States were within this category. ® Combined systems: Thirteen SSAs operated MISs which formed an integrated component of a larger State system combining information on drugs, alcohol, and often mental health. In these systems, the primary emphasis was not upon the collection and analysis of drug information. The drug-related data served as one component of the overall system. ® None: Eight States exhibited no MIS development and partici- pated in the CODAP system only to the extent of submitting hard copy to NIDA. These States were excluded from subse- quent analyses. The States classified as "None" differ from the CODAP/IDARP category in the extent of utilization of the information; the latter used the data to satisfy given informa- tion requirements. The above categories (excluding "None") reflected a continuum of design complexity. Based on the design characteristics it was suspected that there would be a related functional complexity in terms of software development and automation, with the States having more sophis- ticated design features also developing more sophisticated software and more often employing automated processing procedures. The 55 initial analysis did not show such a relationship. In fact the inverse was true: 79 percent of the CODAP/IDARP and 100 percent of the CODAP and Addendum employed automated proce- dures. This compared to 64 percent and 77 percent of the States operating a State Drug System or a Combined System using ADP procedures. Since the operation of many State systems required both automated and manual processing procedures, another coded question concerning the extent of manual processing indicated a greater rate of response among the latter two MIS types. Although the systems with the less sophisticated design complexity were more frequently utilizing automated procedures, the degree of expertise required to develop and maintain the systems was not measured. This may change the interpretation based on the knowledge that many of the CODAP/IDARP and CODAP and Adden- dum systems employ statistical software packages (e.g., SPSS) for most of their processing and analysis while those States using ADP procedures employ State-developed proprietary software. Another confounding variable is the level of implementation. Since the State MISs which fall into the latter two categories are in a more rapid state of change, the present use of manual process- ing procedures may not accurately reflect the intended functional design. A further refinement of our understanding of functional complexity was indicated by the number and type of available MIS staff. More than 50 percent of both the State Drug and Combined Systems groups, compared to only one State operating a CODAP and Addendum system and none of the States operating a CODAP/ IDARP system, had more than six staff members. In fact, 55 percent of the CODAP/IDARP States had only one MIS staff person. Recognizing that there are differences in the quality and responsibilities within various States, it is nevertheless apparent that the development of a more complex systems design was related to the availability of a larger MIS staff. The direction of the relationship is not known but it is reasonable to assume that more staff are required to design and operate a more complex system. The type of staff available is also crucial. Only 22 percent of the first two groups had a programer on staff compared to 55 percent of the States operating a State Drug System and 73 percent of the States with Combined Systems. The above indicators, although descriptive of staffing and automa- tion patterns, are vague in that they do not provide any informa- tion about what each system actually does for the SSA in terms of satisfying various information requirements. The most technically sophisticated and glamorous MIS is of negligible value if the information is not relevant to the SSAs decisionmaking needs or useful to the spectrum of potential recipients at various levels within the State administrative and treatment network. Two data elements which were helpful in examining the system's utility were SSA emphasis (perceived information needs) and SSA reporting (actual systems output/need satisfaction). 56 Table 7 illustrates the perceived State informational needs for each MIS type. These patterns seemed surprisingly similar despite varying levels of design complexity. The information needs reported most frequently from all groups were: (1) outcome evaluation, (2) planning and needs assessment, (3) process monitoring, and (4) unit cost/budget information. There were some minor differences between groups, most of which were expected. For example those SSAs operating State Drug Systems were most interested in process monitoring and clinical activities, and those operating CODAP and Addendum systems were most interested in outcome evaluation and needs assessment--two areas which required additional data to complement that collected by CODAP. We examined the extent to which the perceived informational needs were satisfied (actual data produced) in order to determine the relationship between design complexity and report production. It was assumed that the production of output reports is a necessary precursor to an indirect measure of data utilization. Table 8 provides the percent of States within each category and indicates the extent of positive responses to the production of 11 categories of reports. In general it appeared that States operating more sophisticated MISs made greater use of the data than those States operating less complex systems. A comparison of the States' reporting capabilities was based upon the number of reports produced by more than 50 percent of the States in a given MIS category. Only one report, client profile, was pro- duced by more than 50 percent of the CODAP/IDAPR States; three reports (client profiles, census utilization, and outcome studies) were produced by more than half of the CODAP and Addendum States; more than 50 percent of the States operating a State Drug System produced five reports (census utilization, client profiles, special research, drug types/trends, and "other"), while only three reports were produced by more than 50 percent of the combined Drug/Alcohol States. On the average (table 9), States operating a CODAP/IDARP system produced slightly less than three reports per SSA; the CODAP and Addendum States, about five reports per SSA; the State Drug Systems, almost five per SSA; and the combined systems slightly more than four per State. The largest differences were between those States operating a CODAP/IDARP system and those operating either a State system or an Addendum system. The reports most frequently produced by CODAP and CODAP and Addendum States were descriptive client profiles and outcome studies, while the reports most fre- quently produced by State drug systems and combined systems were client profiles and census utilization reports. There also was a greater emphasis upon activity and service reports in the comprehensive and combined systems--not an altogether profound finding since CODAP does not collect this information. One possible explanation for differences seen in the reporting and utilization of different systems categories might be the level of system implementation. For example, if a given system were in a piloting stage or undergoing major revisions, the reports actually produced might not be indicative of the system's potential for satisfying management information requirements. As mentioned 57 89 TABLE 7. —Percent of States, by MIS type and Single State Agency (SSA) emphasis MIS type I II II Iv CODAP/ CODAP and State drug Combined SSA emphasis IDARP addendum system drug/alcohol Outcome monitoring/evaluation 71 100 73 69 Process monitoring/evaluation 43 75 73 46 Planning/needs assessment 50 100 54 38 Resource inventory 28 25 27 46 Unit cost/budget 50 75 45 46 Standards development 7 25 27 7 Clinical activities 0 25 27 23 Research 14 25 27 31 External requests 21 25 9 39 Other 36 0 45 38 Total N 14 4 11 13 66 TABLE 8.—Percent of States, by MIS type and Single State Agency (SSA) standard reporting MIS type 1 II III Iv CODAP/ CODAP and State drug Combined Standard reports produced IDARP addendum system drug/alcohol Census utilization 36 75 73 85 Client profiles 85 100 82 92 Individual client listings 29 25 27 23 Individual client reports 7 25 9 7 Special research 14 50 55 38 Service reports 14 50 18 38 Unit costs 7 25 36 15 Drug types/trends 29 50 55 54 Staff activities 0 25 9 23 Outcome studies 43 100 45 38 Other 21 25 63 23 Total N 14 4 11 13 09 TABLE 9.—Number of Single State Agency (SSA) standard reports produced, by MIS type MIS type I II III Iv CODAP/ CODAP and State drug Combined Standard reports produced Total IDARP addendum system drug/alcohol Census utilization 27 5 3 8 11 Client profiles 31 12 4 3 12 Individual client listings 11 4 1 3 3 Individual client reports 4 1 1 1 1 Special research 15 2 2 6 5 Unit costs 8 1 1 4 2 Service reports 11 2 2 2 5 Drug types/trends 19 4 2 6 7 Staff activities 5 0 1 1 3 Outcome studies 20 6 4 5 5 Other _14 _3 _1 _1 3 Total reports 165 40 22 46 57 Total N 42 14 4 Hi 13 Mean (x) 3.9 2.8 5,25 4.2 4.4 previously, when examined by level of implementation (beyond simple CODAP reporting), it was found that both the State Drug Systems and the Combined Systems were in a more dynamic state of change. Hence the differences between their report generating capacities and those of the CODAP/IDARP and CODAP and Adden- dum States may have been even greater than that actually observed. In fact 69 percent of the combined systems and 53 percent of the State drug systems were either in a state of early development, pilot testing, or undergoing major revision. Eighty percent of the CODAP/IDARP and CODAP and Addendum States indicated that their systems were established and not undergoing any major changes. In analyzing the differences between various types of drug abuse MISs, the effect of each of the following factors was examined: (1) The SSA functional responsibilities (i.e., drugs only or drug/alcohol), (2) client static capacity, (3) total State drug budget, and (4) the existence of organizational or operational obstacles. In terms of SSA responsibility, 18 SSAs were responsible for only drugs, 31 for drugs and alcohol, and 1 for drugs/alcohol and mental health. There seemed to be only minimal conflict between MIS system responsibilities and administrative responsibilities (i.e., one State with drugs-only responsibility operated a combined system, and only three States with combined responsibilities operated a State drug system). The CODAP and CODAP and Addendum States showed no pattern based on administrative responsibilities, while the States operating more complex systems seemed to correlate more closely with the administrative respon- sibilities. However no causal relationship was established. For example, in more than one State an SSA with combined drug/ alcohol responsibility operated a CODAP system simply because of the small number of CODAP clients in the State. The supposition that a large client static capacity would dictate a more elaborate and complex management information system is partially borne out by the data. Table 10 shows a moderate relationship between type of MIS and the size of the client static capacity. Although a clear linear relationship is not present, there appears to be a greater probability that a State with a large treatment population will develop its own MIS. The obvious exception appears to be the combined drug/alcohol systems, where the size of the drug population appears to be considerably less than that in other States. Hence, the impetus for system development is probably based on some other factor, i.e., size of alcohol or mental health population, State priorities, etc. Another variable which might influence the type of MIS development is the size of the State drug abuse budget. Although data relating to the actual dollar expenditures for MIS development were not available, it was found that most States spend between 1 percent and 3 percent of the total State drug budget on MIS development and/or operation. 61 TABLE 10. Percent of States, by client static capacity and MIS type Client static capacity MIS type Less than 5,000 More than 5,000 CODAP/IDARP 79 21 CODAP and addendum 79 21 State drug system 45 55 Combined drug/alcohol 100 0 Table 11 compares the size of the total State drug budget for each category of MIS. The supposition that States with larger budgets would be more likely to develop a complex information system was confirmed by the data with the exception of the combined drug/alcohol systems. TABLE 11. Percent of States, by State drug budget and MIS type State drug budget Less than More than MIS type $5,000,000 $5,000,000 CODAP/IDARP 65 35 CODAP and addendum 50 50 State drug system 36 64 Combined drug/alcohol 60 31 Whereas 35 percent of the CODAP/IDARP States and 50 percent of the CODAP and Addendum States had budgets in excess of $5 million, 64 percent of the States operating their own system were included in this category. In fact four States in this latter category had total drug budgets in excess of $20 million a year. Although funding and expenditure policies often differ based on State priorities, there appeared to be a fairly strong relationship between the amount of funding available and the development of a State system. One additional explanatory variable examined was the percentage of total State drug abuse clients included in the CODAP system. Where CODAP is widely reported throughout the State and is "representative" of State treatment activities and clients, it was expected that there would be a greater reliance upon CODAP and less desire to operate an expanded system. 62 The data presented in table 12 indicate that the lower the amount of coverage in a State, the greater the likelihood that the State will implement a system which goes beyond the collection and analysis of CODAP data. The only exception to this finding was the group of States which collected addendum information. These data suggest the possibility that some State-level complaints about CODAP might be due less to the nature of CODAP data than to the percentage of clients on whom it is collected. TABLE 12.— Percent of States, by CODAP coverage in State and MIS type Percent of CODAP coverage MIS type Less than 75 75 or more CODAP/IDARP 27 73 CODAP and addendum 25 75 State drug system 45 55 Combined drug/alcohol 58 42 In summary, it appears that States with limited CODAP coverage, large client capacities, and large drug budgets are more likely to develop a management information system which goes beyond the scope and characteristics of the CODAP system. While the data presented indicate that there were some discernible patterns within the identified categories based on the perceived system complexity, these differences, in terms of the variables considered, seem more a matter of degree rather than a clear-cut distinction. Additionally, a variety of important environmental and operational factors were either not included in the data collection or not apparent from preliminary analyses of the data gathered. For example, one variable not analyzed was the effect of the visibility of the SSA within the State. A second factor which was clearly important but that could not be examined quantitatively was the effect of the personalities of individual SSA staff members. The limited scope of this survey permitted only the impression that personal differences, in terms of professional competence and ability to generate State support for MIS goals, had a significant effect on MIS development in several States. Analysis of MIS Problems Perceived by SSA Staff It was not surprising to learn that SSAs were experiencing a number of problems related to the development and operation of management information systems. What was surprising was the similarity of problem areas across various types of MISs and within various levels of development. Table 13 depicts and cate- gorizes the types of problem areas perceived by the SSAs. Overall, organizational problems (reported 43 times) predominated. 63 v9 TABLE 13. —MIS problem areas Area Number Area Number Organizational: Technical: Staff limitations 20 Turnaround 12 Organizational obstacles 9 Lack of automation 7 Lack of funding 8 Inflexibility of software 7 Lack of State support 6 Total 26 Total 43 Informational: System design: Too little data utilization 14 Developing systems 3 Poor data quality 9 Poor forms design 3 Insufficient program evaluation Paperwork burden 2 followup data 3 Total 8 Total 26 Staff limitations in terms of number or expertise were reported more times (20) than any other problem. There were 9 reports of States encountering organizational obstacles and 14 reports of problems relating to funding and State support. Informational problems (reported 26 times) and technical problems (reported 26 times) were perceived by the States as other major areas of concern. Systems design problems were reported relatively infrequently (8 times). Since it was anticipated that States might experience different types of problems at various stages of MIS development, the categories of problems reported were examined by level of MIS implementation. Those States which had no MIS or made only minimal use of CODAP data, reported that organizational problems were paramount. Table 14 depicts the number of positive responses to categorical problem areas by the level of systems implementation in the SSAs. Because the number within each level of implemen- tation was different, a ratio based on the number of reported problems and the number of SSAs was developed to facilitate comparisons across groups. Those States which operated estab- lished systems reported the greatest number of organizational problems (excluding "None") with only minor differences between those making major revisions and those in a piloting phase. In terms of the problem-ratio indicator, States with established systems or in a piloting phase had the highest problem ratios. The organizational obstacles reported often contributed to the next largest problem areas: unmet informational needs and tech- nical problems. Fourteen States reported that the data were underutilized; 9 reported that the quality of the data was either questionable or poor (table 13), and others cited specific deficien- cies in their systems such as a lack of data for followup and/or evaluation. The States undergoing major revisions experienced the greatest rate of unmet informational needs. It was interesting to note that the States with no MIS development reported the lowest rate of unmet information needs. Perhaps these States were too busy dealing with their organizational problems to consider using manage- ment information. These data further suggest that the availability of a substance abuse MIS stimulates an awareness of the need for additional information as existing needs are met. The largest number of technical problems related to turnaround time (12), lack of appropriate automated capabilities (7), and the development of software which was too inflexible to meet the changes required (7) (table 13). The greatest relative number of technical problems were reported by States with either established (R=0.85) MISs or ones undergoing major revisions (R=0.75). States in a phase of early development also report a considerable number of technical problems (R=0.50). Overall the greatest relative number of problems were reported by States undergoing major revisions (R=2.6) followed by the "None" (R=2.5) and established systems (R=2.4). The low overall reporting 65 99 TABLE 14.—Number and ratio of categorical problems, by level of implementation Categorical problems Laval of implementation (4) Organizational Informational Technical Design Total Number Ratio Number Ratio Number Ratio Number Ratio Number Ratio Early development (6) . . . .. 2 0.3 3 0.50 3 0.50 0 = 8 1.3 Piloting (7) ............ 6 0.86 4 0.57 1 0.17 1 0.14 12 9.2 Major revisions (8) ....... 5 0.62 7 0.87 6 0.75 3 0.37 21 2.6 Established (13) ......... 10 0.77 7 0.54 11 0.85 3 0.23 3 2.4 None/other (13) ......... 20 1.6 5 0.38 5 0.38 3 0.23 33 25 Total N::..:.05.45 43 —- 26 - 26 = 10 — 105 — of significant problems during the early stages of MIS development seems to reflect a pattern of initial optimism and confidence followed by disillusionment after MIS establishment and a fairly high level of dissatisfaction by the time it was decided that significant changes were needed. Table 15 depicts reported problem areas by the design complexity of State management information systems. States with no MIS reported the greatest rate of organizational problems (R=1.9), followed closely by States operating their own system (R=0.91). The fewest organizational problems were reported by the States operating either combined drug/alcohol (R=0.54) or an addendum system (R=0.50). Although the States operating their own systems reported the greatest rate of organizational problems, they also reported the lowest rate of unmet information needs. Apparently they were successful in dealing with the organizational obstacles which arose. The problem ratios closely paralleled the design complexity of each system with the most complex State system reporting the greatest number/rate of problems, followed by the less complex addendum system. The CODAP/IDARP systems, clearly the least complex of the three, also reported the lowest problem ratio (R=1.6). It appears that the development and maintenance of a substance abuse management information system necessarily includes a number of related problems. A State's ability to overcome these problems and obstacles may be the single most important predictor of the eventual fruitfulness and utilization of a management information system. Table 16 is the final summary table. The data therein compare State management information systems and the SSA environments served across selected variables discussed throughout this report. Certain caveats applicable both to the information in table 16 and the analysis in this report are reiterated: ° Much of these data represent the "best informed judgments" of the SSA agency staff and/or interviewer. ° Some data may appear conflicting or contradictory due to simultaneous and ongoing development in the SSAs. ® Subtle shifts in data interpretation may have developed between individuals responsible for the data collection and those performing subsequent analysis. ° The open-ended structure of the interview methodology left gaps in information from some sources. Nevertheless this report (and the summary data in table 16) represent important baseline trends at a given point in time. 67 TABLE 15.—Number and ratio of categorical problems, by MIS type 89 Categorical problems MIS type (N) Organizational Informational Technical Design Total Number Ratio Number Ratio Number Ratio Number Ratio Number Ratio CODAP/IDARP (14) . ..... 9 0.64 7 0.50 5 0.36 2 0.14 23 1.6 CODAP and addendum (4) . . 2 0.50 3 0.75 2 0.50 1 0.25 8 2.0 State drug systems (11) . . .. 10 0.91 5 0.45 10 0.91 2 0.18 27 2.5 Combined drug/alcohol (13) . 7 0.54 9 0.69 7 0.54 3 0.23 26 20 None (8) .............. 15 1.9 2 0.25 2 0.25 2 0.25 21 1.4 Total N ........... 43 — 26 - 26 10 - 105 - 69 TABLE 16.—Comparisons of management information systems and Single State Agency (SSA) environments ; Number . Use of SSA - Client Drug Type of MIS Level of MIS of MIS Processing CODAP CODAP responsibility | capacity | budget implementation mode coverage staff tapes £ © £2 States g c 3 2 w |B £ = 9 3 z °|E 2 g B gl 8] 8] o 2 lz slg 5 5 5 5 MEIHE = 1 A ZS a e ow | ow 2] =| e =| 5% S| 8 s| 3] E| =| 5% 3| 2 .| 33 E152 <|g|8 S| 5] ZF c|z| | &|lal&| 222s 5082] 5|z| 32a wl 21 213181831 oI ZI] 2] 8] = Sls s|8|w| E|E| S| S|8] 8] 82] 8] E ¢1¢18(2818|alel&ls|s|elels|2|3|8|E]|s|e|s|2|2|Elelelslsglils di C c a ® = = > gp SI&IE12121%21213|3|21a|8|8|8(g(d|S||w]|e]lc|&=2lV]|A]|2|3]|S|3 Alabama ....... X X X X X X |X Alaska . . ....... X X X X X XxX: X |X Arizona . ....... X |X X X X |X X X Arkansas ....... X X X X X X X |X California. ...... X X X X X X X Colorado ....... X X X X X X X Connecticut . . . .. X X X X X X Delaware ....... X X X X X X |X X Florida, « ws: +55 + X X X X X X Georgia . ....... X X X X X X X NOTE: Leaders on the table for a particular State indicate that information was not obtained. 0L TABLE 16.—Comparisons of management information systems and Single State Agency (SSA) environments—Continued i Number . Use of SSA Client Drug Level of MIS Processing CODAP responsibility | capacity | budget Type of MIS implementation of MIS mode coverage ConA staff tapes £ ® Q - States g c 2 = | @ £ E s ‘@ Zz © £ > S = 3 » o £ . © > © o | © ° 23 slg| | 25 ol 2 3 £ | 523 233 JHEHE IH HIRE lg) 588 <|5| 8 S| 5|= 218 clz| 22a |8|2|5| 8g 518 =| | 8121312] ¢|e » a cs] 8 Sla| a < < or a < £ = = 2 0 £ E = 3 a a g z 2 = tei iplaid Ri BlelEld|Z EIR iE cl el a tl818llalwm) 2] 25) 5 | 2| 2] |®|le|®d]|ls|olo|E|8lo|s5|2|5|e|s|=2{s]5|5|&8|~|R|z|&|2|S S| coy V AliIAj2Z| 0 0) | Fj | | EWI Djvlmlieoelx | || V]IAIZ] DOC] Hawaii. ...... X X X X X X Idaho: «wu sms so X X X X X X |X Winois , uu sus su X X X X X Indiana . vs =n 5 4 X X X X X X X |X X OWE: ; vais sus vn X X X X X X X X Kansas .a: cus sa X X X X Kentucky ...... X X X X X X |X Louisiana. . . .... X X X Maine . . ....... X X X | X X Maryland. . . .... X X X X X X X NOTE: Leaders on the table for a particular State indicate that information was not obtained. TL TABLE 16.—Comparisons of management information systems and Single State Agency (SSA) environments—Continued ; Number +: Use of SSA Client Drug Level of MIS Processing CODAP responsibility | capacity | budget Type of MIS implementation of Mis mode coverage CODAP Staff tapes £ 3 = States £ _ c 2 x 5 |B g z S ® Zz gl: $ : i : HEHE o ° o >| 2 >| 2 5 3 = s| = 7 £ |B 2 0 2 ° «| «] 3] © z(g| 3 g| 8 s|2 853 2 +) 2) 3 HEIEIEIR IR s1E|s|.lo]818 c|z|818le|8|2l2|2]2 5|8|l2| | 5| 8] 382] 2|2 sl al sl8(8l3|3]e|Z|S|s|2|&|25 2 82 CE|s|5|23|2|2]3|2|%|E s1 512818885 |5|0|5|5|8|5|28|5|2 5 253/851 | z|8| 8s 518/810 |R|1%91%212|8|38|a|8|0|&8|g|d|S|a|m|o|c|&|=2|V|Alz|3|6|a Massachusetts. . . . | X X X X X X Michigan . . ..... X X X X | X X Minnesota . . . . .. X X X X X Mississippi . . . . . . X X X X X X X Missouri . ...... X X X X X X X |XX X Montana . ...... X X X X X X X X Nebraska... .... X X X X X X X X X Nevada . ....... X X X X X X X X New Hampshire X X X X X X |X New Jersey ..... X X X X X X X X NOTE: Leaders on the table for a particular State indicate that information was not obtained. ZL TABLE 16.—Comparisons of management information systems and Single State Agency (SSA) environments—Continued SSA Client Drug Level of MIS Number Processing | CODAP Use of responsibility | capacity | budget Typeof MIS implementation of Mig mode coverage CODAP staff tapes £ © P= © States E C 3 g 3 | 2 g % 2 : z & £ | 2 @ & 2 2 2 2 81s 5 liz 5 £ g £ El E28 Ts | £ S| 3 2X |¢ 2s ° e =|l=]35|2|%|g E12 S| 38 § 22 |w|5|% BE v2 8 2 |ElxlE|8 c © oo Q = ©» > po D = - oO Q ° Q o = ° ® > 3 a a > E a 3 o z 2 2 | 8 © o | _ = > = Qo @ wl lzlg|2|2l S| C1212 2B |S]: 12 eE|E|£|5]18|8]z|=2|8]|:z ol ool | re 121818 ec [Bla l>l2 5 |8]2 |» 3 s s| e253 EE 9 g S| |w | wv c = £ = 3% |v = ° x 22 | ello] | ® = x 1 = Eo iaote #210 | 0 |8|ojel2|2 (218 285 S§|lef~|R]a|2|2|2 o|0|o / V 2100 |» [O00 |W |g |W |DdD|a|m| lolx |&]|=s v 4% Z2|D2|0|&a New Mexico. . . . . X X X X X X X X New York . ..... X X X X X X X X North Carolina . . . |X X X IX X X X North Dakota. . . . X X X X X X OhiG: ws cus sn X X X X X X X X Oklahoma . ... .. X X X X X X X X Oregon . ....... X X X X X IX X X Pennsylvania X X X xX |X X X Rhode Island . . . X X X IX X South Carolina . . . X X |X X X NOTE: Leaders on the table for a particular State indicate that information was not obtained. €L TABLE 16.—Comparisons of management information systems and Single State Agency (SSA) environments—Continued SSA Client Drug Type of MIS Level of MIS Numer Processing | CODAP ii responsibility | capacity | budget yp implementation mode coverage staff tapes £ ® 2 Stat ® ° < ates £ c 2 a = | 8 E c 2 3 Zz oS |E 3 o ] s| 8| 8] o 83 mr g £ 2 § 8 EI b. > > ° o - - ° ~ @ =|5|% g|8 sl 31 Elles 3| 2 . 123 Elsl8|<|5|8 g§|5|= gg c|z|3|2(a|8|e2|2|8]s slelellzl5]3|2|2|z slalaigi8la|d]le|S|Sl2|8]|Z|>|E 2 s|8||E|E|S5|3|2|28]3(|2|¢2|c¢ oo y - » 2 3 Slo |g |w |v c =| E = ol =| © slo x]=]|¢8 cw lw] «| & S| 3 = g cE &@ @ o oO oO 8 Oo = Z c o ~ o > © ~ ~ ° o 518180 |R|7|R%]|2|3|0|a|8|0|&|g|&|S|(a|m|o]c|&|2[(V|A|[2]|3]|S]|3 South Dakota. . . . X X X X X X |X X Tennessee . ..... Xx X X X X Texas . ........ X X X X X X Utah. ......... X X X X X Vermont . . ..... X X X X Virginia. . . ..... X X X X X Washington . . . .. X X X X X X X West Virginia . . . . X X X X X X X X Wisconsin . ..... “EL TF X X X X Wyoming. ...... X X X X X X X NOTE: Leaders on the table for a particular State indicate that information was not obtained. HIGHLIGHTS AND IMPLICATIONS The project methodology was designed to illuminate patterns of State-level drug abuse management information systems in terms of their organizational frameworks, operational technologies, perceived information needs, and level of output reports produced. Preliminary analyses and tentative findings have been presented. Selected highlights and implications are presented below. Highlights ° More than half of the States surveyed were planning or executing changes in their MISs. Some of these changes were the result of internal decisions to improve the systems or implement additional components. Others were imposed by outside forces such as new Federal or State Government requirements or pressure from community groups. In many cases the required changes conflicted with the original plans for the MIS. ° The SSAs expressed a strong desire to obtain information useful for program monitoring, evaluation, planning, and budgeting. Less emphasis was placed on the use of MISs to collect information relating to clinical activities or research needs. ° The CODAP system was heavily integrated into the State systeins with 84 percent of the States using the Federal forms, 70 percent mentioning usage of the data, and 35 percent using NIDA tapes for automated processing. ° As MISs develop beyond CODAP, the sequence of new forms added begins with expanded admission and discharge instru- ments. Client services information, external indicator data, and expenditures tend to be added next. Progression of data collection reflects the type of information which SSAs desire to receive from MISs. ° Most States are now aware of quality-control procedures and consider them an important aspect of system implementation. ° Almost all States are involved to some extent in automated processing of their data and many now have programers on staff to assist the file-building, maintenance, and reporting procedures. ° Although output reports are now routinely generated by the majority of MISs, the amount of data collected still greatly exceeds the amount of data analyzed and distributed as output reports. ® In terms of system complexity, four major typological cate- gories of State substance abuse management information 74 systems were delineated: (1) States using only CODAP information, (2) those combining additional data with existing CODAP forms, (3) those with State drug abuse information systems, and (4) those with combined drug/alcohol systems. Only a limited number of discrete patterns were found to exist with each typological category when organizational structure, informational needs, level of reporting, and problem areas were considered--thus indicating that systems complexity might not be the best typological descriptor. Although the analyses conducted on the data collected did not precisely "explain" variations in State MISs, patterns related to levels of SSA responsibility, total State drug budget, client static capacity, and extent of CODAP coverage did emerge and were related to design complexity. The most overwhelming developmental problems experienced in the States were organizational obstacles and staff limitations. implications NIDA assistance to the States to enhance MIS development should be directed particularly toward the development of information for planning, monitoring, and evaluation since these were the needs most often expressed by the SSAs. The development of a modular model information system is not feasible at this time due to the wide variation in MIS development, operations and resources, and the lack of consistent developmental/operational patterns. Production and distribution of guidelines and recommendations for the development of MISs could be helpful to the SSAs but cannot be expected to solve the host of problems experi- enced by the drug abuse management information systems. New systems can, however, profit from the experiences of States with more developed systems, and a mechanism for continuous information exchange should be developed to facilitate this process. NIDA should assist States in achieving statewide client coverage with either CODAP or State-developed systems. This should be done in conjunction with the production of CODAP tapes at the State level for more timely turnaround. Systems should remain flexible due to the likelihood of con- tinued mandated and internal requirements for change. It is imperative that MISs be based on a functional manage- ment approach with data collection, analysis, and report generation designed around the decisionmaking responsibilities of the SSAs. 75 Continued efforts should be made to appraise developments in State-level MISs that are focused on this functional manage- ment approach rather than design complexity or technological sophistication. RECOMMENDATIONS OF THE REVIEW PANEL A draft of this chapter was presented to a panel of experts for review on June 1, 1977, at a seminar sponsored by the National Institute on Drug Abuse in Orlando, Fla. This panel included the directors of Single State Agencies, IDARP personnel, and people from the county, city, and clinic level. The panel emphasized the conclusions of the report: The primary purpose of a management information system is to assist administrators in the deployment of organizational resources. In order for the informational needs of the Single-State Agencies to be met, client-oriented data need to be supple- mented with information relating to external indicators, services provided, program evaluation and monitoring, and financial management. Data collection should be minimized, with all data elements reflecting careful consideration of informational needs. Data utilization should be maximized with an emphasis on providing timely analysis, presentation and distribution of reports (supplemented by training in the use of the reports when necessary). The panel also suggested a plan for promoting the development of drug abuse management information systems through interstate cooperation and technical assistance from the Federal level. They recommended: Progress toward common operational definitions of such terms as "management information systems" and "evaluation" (possibly through circulating tentative definitions through the mail for review and feedback). Establishment of an interstate communication network to facilitate exchange of ideas and avoid duplication of effort through: 1. Maintenance and dissemination of up-to-date information on drug abuse management information systems by the National Institute on Drug Abuse (or one of its contrac- tors). 24 Development of a "resource directory" listing the high- lights of individual systems and cataloging the activities 76 of various systems within general categories (i.e., needs assessment and planning). 3. Creation of task forces relating to common issues and problems. ° Strengthening of two-way communication between State/county personnel and the National Institute on Drug Abuse in such areas as: 1. Federal funding policy (e.g., use of "slots" as a basis for reimbursement). 2; Federal reporting requirements (since changes present difficulties with staff training, software revisions, and expense). AUTHORS PADDY COOK BARRY J. ROSENTHAL, M.S. CHERYL DAVIS, M.A. Richard Katon and Associates, Inc. Rockville, Maryland REFERENCES Ackoff, R. Management Misinformation Systems. Management Science, 14:147-156, 1967. Anthony, R.N. Planning and Control Systems: A Framework for Analysis. Boston, Mass.: Harvard University Graduate School of Business Administration, 1965. Davis, S., and Freeman, J. Hospital managers need management information systems. Health Care Management Review, 1:65-74, 1976. Elpers, J.R., and Chapman, R.L. Management information for mental health services. Administration in Mental Health, 1973. pp. 12-25. Gorry, G.A., and Morton, M.S.S. Framework for management information systems. Sloan Management Review, 13:55-70, 1971. Kennevan, W.J. Structuring and managing a management informa- tion system. In: Kennevan and Joslin, eds. Management and Computer Systems. Arlington, Va.: College Readings, Inc., 1973. pp. 365-373. 77 Pool, I. de S.; McIntosh, S.; and Griffel, D. Information systems and social knowledge. In: Westin, A.F., ed. Information Techology in a Democracy. Cambridge, Mass.: Harvard University Press, 1971, pp. 241-249. Weinstein, A.S. Evaluation through medical records and related information systems. In: Guttentag, M., and Struening, E., eds. Handbook of Evaluation Research. Vol II. New York: Maral Decker, 1976. pp. 397-477. Westin, A.F. An introduction: Descriptions of the developing systems by their advocates. In: Westin, A.F., ed. Informa- tion Technology in a Democracy. Cambridge, Mass.: Harvard University Press, 1971. pp. 15-29. Withington, F.G. Five generations of computers. Harvard Busi- ness Review, July-Aug. 1974. pp. 99-108. 78 3. A Case for Management Information Systems Helping the Manager Make Decisions Regarding Difficult Resource Allocation Problems Edward Leibson, Ed.D. INTRODUCTION The managers of drug abuse treatment programs, like the managers of most businesses,’ bear significant responsibility for resource allocation. Indeed, it might be said that the essence of management is the making of decisions regarding resource allocation thereby leading the program toward its goals and objectives. In the drug abuse treatment setting this is an awesome task complicated by a number of factors: ° Scarce resources: Managers must cope with chronic shortages in resources, for example: 1s Staff resources: Trained staff who are willing to work for the typical wages at drug abuse treatment facilities are scarce. The strains of working with a recalcitrant population leads to early "burnout" unless the staff is supported by reasonable schedules, timely and construc- tive feedback, rewards, positive experiences, and so on. 2s Volunteer resources: Volunteers, so vital to the sur- vival of many treatment programs, are notoriously fickle and difficult to plan with, 'It has been said that business and government are alike in all unimportant respects. Business managers are expected to profit personally from their efforts. Government managers are likely to end up in jail if they make a profit. And so the system of motivation varies tremendously between the two sectors. 79 3. Facility resources: Space, equipment, and supplies are always at a premium. 4, Community and political resources: The treatment center must be constantly apprised of the community and political forces that support the program with money, in-kind donations, public relations, services, and referrals because the competition for these resources is keen. ° Uncertain demand: The demand for drug treatment services cannot be easily predicted but fluctuates in response to seasonal factors, drug availability, street attitudes, peer pressure, alternatives to care, and so on. ° Difficult consensus: Because the treatment center works with a chronic population and because its goals and objec- tives deal with intangible and unquantifiable commodities, the program administrator is hard pressed to parcel out the proper "amount" of a resource to achieve a specific "level" of outcome. The manager must handle resources through intricate maneuvers to keep the program alive. The purpose of this chapter is to demonstrate how information-- particularly the information derived from a management information system (MIS)--can assist the clinic administrator to procure, allocate, and redeploy limited resources in the most logical and efficient manner. The ability to understand and use an MIS requires neither a massive training effort, a computer, nor a host of forms and procedures. At bottom, an MIS is a formalized method of collect- ing and reviewing data. A simple bed count or a routine tele- phone debriefing about client status upon discharge may constitute part of the MIS. This chapter will demonstrate how the routine data available to most administrators can help in meeting the resource allocation challenge. THE VALUE OF MIS TO DRUG ABUSE TREATMENT MANAGERS In the ideal situation, clinic administrators operate their programs on the basis of "behaviorally described and measurable objectives." These are objectives which specify what activities (behaviors) are to be performed (e.g., staff will conduct intake interviews) and which also state in explicit quantitative terms how often, how many, and to what extent the activity is to be carried out (e.g., each staff member will conduct 60 intake interviews a month, each lasting 1 hour). The MIS should provide the administrator with the tools to measure the monthly, weekly, daily (and possibly even hourly) investment of resources in each objective and the resulting progress that is made. 80 Frequently NIDA programs express objectives in terms of "static" and "dynamic" capacity, i.e., projected utilization rates. A static capacity is defined as the number of clients in treatment at any point in time; the dynamic capacity is defined as the number of individuals in treatment during a period of time and is a function of the static capacity and the turnover rate. For example, a program with a static capacity of 100 and a turnover rate every 3 months will have a dynamic capacity of 400 per year. Sounds like a pretty specific objective--right? Wrong! This defines only one parameter (i.e., projected utilization rate) of the treatment program. One also needs to specify the type, scope, and quality of the services to be provided; the capability of staff to deliver such services; projected treatment outcomes; community support; and management and fiscal strength. Even without specifically articulated objectives, program adminis- trators may have an intuitive sense of the program objectives and the types of activity that will lead to these objectives. For example, administrators know that a booth at a State fair will increase the community's awareness of the clinic and thus increase referrals. In fact this may be so but unless this technique is compared to another, the administrator will never know whether an investment of the same resources in the local jail or probation department would be more productive. What then are the types of readily available data required by the administrator to measure the investment of resources into specific program goals and objectives? Data needs vary according to specific program objectives, but at a minimum administrators will profit from the data sets described below. Staff Activity The type of staff and the hours required to carry out program activities vary with the type of service provided. In general administrators will find staff activity data useful in developing work schedules, designing training courses, assessing the quantity of care, and assessing staff workloads. For example, the adminis- trator may want to know the level of staff activity for the following cost or activity centers: a division, a department, or a subdi- vision of a treatment program; a group of services; a group of employees; or any other type of subclassification within which functions of an institution are logically grouped for purposes of cost allocation. ° Outpatient services: number and duration of services, by personnel type ° Inpatient services: number and duration of services, by personnel type ® Residential services: number and type of services, by personnel type 81 ° Methadone services: number and dosage of units dispensed as well as the number and type of pharmacy, medical, and nursing services provided, by personnel type ° Crises services: number and duration of phone, drop-in, or other intervention services, by personnel type ° Referral services: number and type of placement services, by personnel type ® Education, information, and training services: number and duration, by personnel type. These include services dedi- cated to informing all types of groups, and providing “ork mary" and "secondary" prevention services. Appendix A contains a detailed list of activities used in the auto- mated information system in Wayne County (Detroit), Michigan, to record staff activities. Figure 1 illustrates the categories of staff activity that may be reported in connection with various services. Client Data In order to allocate resources adequately among the services and cost centers of the program, the manager must know how many clients are currently in treatment, the rate of admission, and other client characteristics. Thus it is important to know (on a monthly, weekly, or daily basis): ° The size of the caseload: The size of the caseload for both the entire program and the cost centers will tell the manager when the program (or cost center) has reached its saturation point. It will also reveal inappropriate allocations among cost centers. For example, assume that the caseload in the intake cost center is at its maximum. The caseload at the methadone cost center has not reached its peak and, although more clients could be accommodated, the program will only be able to admit clients who do not need intake services to complement methadone treatment. The intake service has become the "limiting factor" in the treatment program while available resources in other cost centers lay idle. The size of the caseload can be computed from a data set that includes at least: 1. Beginning caseload: How many people were in treatment at the beginning of the period? 2. New admissions: How many people entered treatment for the first time during the period? 3. Readmissions: How many people were admitted to treatment who have been in treatment before (at this or other clinics)? 82 Wayne County Department of Substance Abuse Services STAFF ACTIVITY There may be a number of discrete staff activities for each type of service, and these may vary somewhat among programs. Each program must divide up (and clearly define) the discrete units by which staff activity is to be measured. For example: UNITS OF SERVICE STAFF ACTIVITY Number of patient days Total direct service hours Hours of therapy Hours face-to-face with clients Doses of methadone Total direct service hours Crisis intervention Manning hours Education Hours face-to-face with clients Referrals or intake Manning hours These data can be aggregated monthly and on a year-to- date basis to derive a ratio of staff activity per unit of service. Similarly, the variance between the actual and planned staff activity can be completed. FIGURE 1.—Categories of staff activity reported in connection with related services. 83 4. Terminations: How many people elected to conclude treatment or were terminated by the program? 5. Closing caseload: How many people were in treatment at the end of the period? 6. Delinquent cases: How many people are still on the rolls but have not presented themselves within x period of time? Caseload Data Such data items as client demography and current treatment status, drug history, criminal justice experiences, etc., can only be collected client by client and then aggregated periodically. For example, the client's legal status may have a significant bearing on the treatment approach (i.e., resources allocated to treatment) and the administrative approach (i.e., resources allocated to establishing liaison with the criminal justice system). At a minimum the manager needs to know how many clients are 1) participating in a pretrial prosecutor's diversion program, 2) released on recognizance, 3) on presentence diversion, 4) deferred to probation and treatment, 5) on early prison release, or 6) on parole. Routinely collected CODAP data provide much of the basic data required as a baseline for trends in client character- istics, and these data categories can be supplemented to reflect unique program concerns. Daily Client Scheduling Data But the number of individuals (unduplicated count) enrolled in the program over time is not sufficient. Suppose that two clinics each have 100 clients enrolled; if clients come to Clinic A three times a week and to Clinic B every day, the implications for resource allocation are quite different. It becomes important, then, to keep a daily count of clients actually seen by different service components. This provides insight into the size of the caseload per staff member and enables a comparison between time allocated among the various services. It is also important to compare the clients seen on a daily basis with the clients scheduled to be seen and with staff potential. This will provide some indication of staff productivity and the adequacy of the scheduling system. The difference between clients seen and clients scheduled can also suggest an examination of client satisfaction. Financial Data: Budgeted and Actual The budget is an essential data set for the clinic manager because it provides broad operating parameters. It can be used to set out objectives and to apportion the available fiscal resources among them. In so doing it creates a bridge between the financial resources and the other resources (human, material, etc.) required to accomplish objectives. 84 Throughout the fiscal year, financial status reports indicate progress vis-a-vis budgeted expenditures and revenues. In a sense the difference between the budget and the financial status report is like the difference between the appointment schedule and the daily count of clients: one is a plan, the other is a report of actual activity. Space, Equipment, and Other Material Resources Data Space, equipment, and other material goods must not be overlooked in the discussion of deploying limited resources. And so informa- tion systems should include data on: ° Space assignment per activity (e.g., square feet per cost center). ° A supplies-utilization summary by cost/activity center. ® Communications and duplication charges by cost center or activity. ° An inventory of equipment assigned by cost center. These data compared with program goals and objectives will shed light on the appropriateness of resource allocation. Client Progress Data Improvement during treatment should be recorded according to some systemwide accepted measure for each client. Indicators of educational achievement, vocational rehabilitation, employment, drug abuse, and level of functioning should be collected at intake or soon after and updated on a periodic basis. A number of indicators such as urinalysis results or length of time in treat- ment can also be used to trace change over time. These data will serve two purposes. It will enable the treatment center to track individual client progress, to tailor treatment to the client's changing needs, and to assure continuity and appropriateness of care. Secondly, data about all clients may reveal certain patterns of treatment that result in the most desirable client changes or outcomes and these suggest a direction for overall program policy. INFORMATION AVAILABLE FROM DATA SETS: THE BASIS FOR MANAGEMENT BY EXCEPTION As a means of focusing management attention (resources) on the most important sources of data and reports, the manager should review information that reflects exceptions to or variations 85 from anticipated results. Commonly called "management by excep- tion," this approach requires two basic ingredients: ° A baseline, budget, objective, or other measure of anticipated or planned level of outcome. ° Report formats or information protocols that highlight these variances or exceptions so that the manager need not wade through volumes of irrelevant data to find the important kernels. Routinely collected data can be aggregated in numerous ways to provide an overall understanding of where resources have been and should be allocated. Because the possibilities for generating information are limitless, a few examples are offered as a basis from which the reader can construct information most germane to his/her own program. Example 1: Daily Bed Count As a first example, take a simple area such as bed count in a residential program. Data can be collected by counting the number of occupied beds in the residential unit. To assure consistency and comparability of data over time, certain definitions should be established as to whether a bed is indeed "occupied," e.g., pass and AWOL status, or simultaneous membership in another subunit or program. The bed count should be taken at the same time each day such as 12 midnight. Example 2: Dally Count of the Number of Clients Served This count, compared to the number of people the program planned to serve will indicate whether the program is over or understaffed. A ratio, broken down to reflect the various cost or activity centers of the program, will go even further indicating the appro- priateness of the balance of resources among cost centers. The following formula indicates how it is computed: Actual beds used-planned beds used i 2 Sie Fone ate ood 100=Percent variation Simply subtract the planned occupancy rate from the actual number of occupied beds and divide the remainder by the planned rate. This amount, multiplied by 100, produces the percentage of variation. Thus a therapeutic community with a planned occupancy of 75 beds which actually had only 60 beds filled would have experienced a negative 20 percent variation as shown in the calculations below: 60-75 75 ><100=-20% 86 This means that for a particular day the program was 20 percent below plan. The trends of daily data can be reviewed over time to provide a picture of improvements or declines. Or the daily statistics can be aggregated to provide an indication of when operations are getting out of control. This example can be pursued further to reflect a full month. If the month has 30 days and the program has a planned occupied bed rate of 75 beds each day, this is 2,250 bed-days per month. An actual count each of 30 days revealed between 55 and 75 beds occupied with an average of 70 beds occupied. Using the formula, we see that the bed occupancy was down by only 6 percent for the entire month, a different (and more accurate) picture than the data for a single day: 2,100-2,250 Se Sf 100=- 5.250 100=-6% The clinic manager must review the variances to determine their implications--implications for excess staff, for utilization of equip- ment and supplies, for space requirements and, of course, for the influence on the revenue-generating capacity of the facility. Example 3: Capacity Another helpful concept is "capacity"--the maximum number of services that can be provided with the available resources of staff time, equipment, space, and supplies. The relationship between capacity and planned services is this: ° The planned services should never be in excess of the maximum capacity (by definition). ° The capacity may be slightly larger than the planned number of services to allow for peak contingencies. Managers may want to keep track of the changes in their capacity using the following formula: Actual capacity-planned capacity Planned copacity x 100=Percent variation For example, if the capacity of a therapeutic community is 80 (i.e., there are 80 beds) and several beds have to be removed for repair, the following information about capacity can be obtained: 78-80 30 =<100=-2.5% Obviously this minor negative capacity figure is of little concern if planned services were slightly less than the maximum capacity. The variations should be monitored over time. 87 Example 4: Utilization Statistics On the basis of the information about services provided and capacity, it is possible to compute information about utilization using the following formula: Units of service Capacity for service x< 100 The utilization variance can also be computed using the following formula: Actual utilization-Planned utilization . jgp=percent variation Planned utilization Each treatment center manager must decide upon the tolerable variance levels. Under management by exception, the manager need only be concerned when variances exceed the established levels. Example 5: Financial Data It is a relatively simple process to compute an array of variances of expenditures and revenues to determine what the current and year-to-date financial status is. The figure below shows several items from the budget (planned) and the accounting data (actual) 88 for a single month and for a year-to-date period. It also illus- trates how the variances are computed: Given: Item Actual Planned Expenditures for salaries (current month) (a) $38,920 (b) $37,500 Expenditures for salaries (year to date) (c) 179,420 (d) 75,000 Revenues from title XIX (current month) (e) 23,000 (f) 24,000 Revenues from title XIX (year to date) (g) 27,000 (h) 48,000 Then: Item Current Year to date Percent Percent Expenditure variances for salaries 43.7 245.8 Revenues from title XIX 3.4.1 4-43.,7 'a-b/b ><100. 3e~f/f =<100. 2c-d/d <100. “g-h/h <100. Notice several things about this information: A positive expenditure variance reflects an unfavorable situtation (i.e., the program spent more than it intended to) whereas a positive revenue variance reflects a favorable situation (i.e., the program earned more than it intended to). The importance of knowing not only current but year-to-date information is pointed out by the radically different revenue variances for the current year to date. The program admin- istrator who had only current data might be lulled into complacency by a small variance when in reality the year-to- date picture reflects a very serious revenue short-fall. Because the third party reimbursements come at an unpredict- able rate, the use of this technique can be especially helpful. 89 In conclusion, by having simple variance figures available, the clinic manager will be aware immediately of exceptional situations and will be able to determine whether resources are being consumed in the amounts and for the activities planned and, by implication, whether the program is exceeding or falling short of its goals and objectives. Table 1 summarizes the variances discussed in this section for a single month and a cumulative period. Closer exami- nation of this information will determine if variances are "random" or part of a trend. By reviewing such an array of data, the manager can tell which situations are extraordinary and need immediate attention. In this case, the sharp discrepancy between planned and actual title XIX revenues provide warning that an "exceptional" situation is confronting the program. A similar report can be prepared for any indicator that the manager feels is important enough to control. For example, the unit of service (bed-days) used above is a very simple one and could be substituted for or complemented by such diverse elements as number of patient-days, hours of therapy, doses of methadone, number of crises contacts, number of educational contacts, number of referrals, and so on. When there are unexpected differences between planned and actual results in basic indicators, the various ratios derived from those items should be studied and the year-to- date and the current month's variance compared. Furthermore one must determine the seriousness of such differences by compar- ing ratios. If units of service and capacity are both above or below planned, it is a less serious problem than if unit of service is above planned and capacity is below planned. Similarly, when one reviews expenditures and capacity, a difference is less serious if both variances are both above or below planned than if expendi- tures are above and capacity below. The converse would be in comparisons between revenues and expenditures. If there is an above-plan variance of revenues and a below-plan variance for expenditures, the situation obviously is not serious but fortuitous. In short, meaningful analysis of direction of variance, year-to-date trends, and comparisons between the units of measure can give the manager a substantial basis for decisions. ASSISTANCE THAT MANAGEMENT INFORMATION CAN GIVE IN MAKING RESOURCE ALLOCATION DECISIONS By having an MIS (with planned as well as actual activity levels) each individual staff member can know how many units of service he/she is expected to produce and can review actual data to determine how closely the actual activity reflects the planned or expected activity. A simple log, such as that shown in appendix B, can be used to collect the actual staff activity data. Likewise, staff activity data allow the manager to assess the time the staff members are devoting to activities and if these activities are appropriate. When an MIS is not in place, misallocation of time 90 16 TABLE 1.—Summary of variances for a single month and for a cumulative period Current month Cumulative month to date Percent Percent Unit Actual Planned variance Actual Planned variance Units of service (bed-days) 2,100 2,500 -16 4,188 4,500 -6.9 Capacity (bed-days) 2,370 2,400 =1.25 4,740 4,800 -1.25 Staff activity (number) 2,560 2,029 +26 5,120 4,058 +26 Utilization of bed-days (percent) 88.6 93.7 -5.4 88.3 93.7 -5.8 Salary expenses $38,920 $37,500 +3.8 $79,420 $48,000 +65.5 Title XIX revenue $23,000 $24,000 -4.2 $27,000 $48,000 -43.8 and misuse of professional talents may occur without the knowledge of the manager. This may cause frustration and inefficiency when clinical staff become responsible (de facto) for activities that are not their primary or most logical responsibility. When an MIS is in place, the manager can plan in advance to have appro- priate services and staff skills available, can assign staff to appropriate activities, and can reduce frustration. Even in the smallest program it is difficult for the manager to understand by him/herself what is happening throughout the program; the manager may "miss the forest for the trees"; that is, he/she may fail to comprehend details such as: number and type of services provided, number of staff activities by staff category, client characteristics at a point in or over time, and use of nonstaff resources. By collecting information in a routine and organized fashion, managers can build composite pictures of all aspects of the program. These will help distribute workloads on an equitable basis including planning for peak and slack periods, vacation, sick days, holidays, and emergency situations. For example, seasonal patterns may emerge over time. Knowing that the caseload will drop in the summer, the administrator can authorize vacations without concern or increase outreach and casefinding activities. This information also offers an opportunity to review individual clinical staff performance and can be a spring- board for individual and programwide self-evaluation. The data may suggest a need to reorder individual or program priorities and thus to redeploy personnel and other resources. Promotions within organizations usually depend upon opinions regarding the capabilities of the individual staff members. Few organizations (within either the drug abuse field or the business community) have meaningful staff evaluations. Fewer still conduct these evaluations on the basis of mutually agreed upon quantified objectives or quantified evaluative data. Agencies that offer promotions capriciously discourage competent individuals by denying them recognition and sanction less competent individuals. Medioc- rity is encouraged unless promotions are tied to staff performance. As a manager one can set forth specific quantified objectives toward which each staff member can strive and against which each can be judged. Although qualitative factors cannot be excluded, quantified criteria give the managers information that is frequently absent when promotions are considered. The manager need not then depend solely upon memory to assess and compare the amount of effort each staff member invested in the job. Clinic managers have long been aware that simply assigning equal numbers of clients to each staff member does not result in an equitable distribution of work. Some clients have more problems than others. Some require many more services and some are more dependent or less responsive than others. Clients may not show up for assigned therapy, thereby wasting staff time. The result is that assigning a specific client matrix to each staff member may cause differential staff activity. This in turn produces different levels of productivity. By compiling information about the actual units of service provided to each client and about the activities 92 (workloads) of each staff person, the manager can determine the relative client load for each staff member and the variance in workloads created by different types of clients. He or she can then sit down with each professional or program unit and arrive at reasonable and equitable caseload and other workload assign- ments. If, for example, one counselor is faced with a consistent pattern of clients who do not show up for appointments, one of two things could be operating: the counselor "turns clients off," or the recalcitrant clients gravitate to (or are assigned to) that counselor. The manager can then either redistribute the caseload or assist the counselor to upgrade his or her skills. Either tactic would enable the manager to develop more realistic plans and thus reduce the percentage of "noshows." By collecting data regarding amounts and types of supplies, space, equipment, and other resources allocated to each cost or activity center, the program manager can determine whether resource utilization parallels program objectives. For example, the program manager might find that twice as much space and 50 percent more supplies are being invested in the methadone com- ponent than in the counseling component. If counseling is a program priority--and if it is suffering from cramped quarters-- the manager is alerted that an adjustment is necessary. IMPLEMENTING THE MIS: BARRIERS AND STRATEGIES With such compelling reasons for turning to management information for assistance, why does the installation and use of an MIS seem to many programs to be so difficult? This section discusses some common barriers to MIS and offers suggestions for their solution. Confusion Between MIS and Computers By the time computers had developed to the third or fourth generation of equipment, the "magic box" or the "black box" myth was well entrenched and it became common to hear that information systems should be automated: "Manual processing is strictly from the Stone Age and no manager worth his/her computer-generated paycheck should be tied down by it. It is wise to automate as early, as extensively, and as intensively as possible." Based on such sentiments, managers have been led to equate an MIS with a computer. The fears of technical language and of complex and unfathomable mechanical equipment cause many in the field to avoid contact with this management information tool altogether. This is unfortunate. First, the idea that an MIS must be auto- mated is incorrect. Simple manual MISs can serve as effective and rich resources for management purposes. Second, computers-- when they are appropriate--need not be overwhelming and techno- logically terrifying although managers should be prepared to 93 spend some time becoming fluent with the technical vocabulary and the basic functions of computer installations. The technology for accomplishing the data-processing task should assist in MIS man- agement--not hinder MIS use altogether. (See chapter 6 for a detailed discussion of automation alternatives.) Resistance and Fear When an agency staff agrees on a series of quantified behavioral objectives, they establish a baseline for evaluation. Since the failure to reach these objectives or sometimes even their publication can place the manager, the agency, or the staff in a bad light, all three tend to develop paranoia about objectives. These feel- ings--prevalent in all endeavors--are exacerbated in an environ- ment where the request for professional accountability is considered an encroachment on professional integrity. Indignant responses are often evoked from clinical personnel who are asked to plan the time required for face-to-face contact or to report details of their hourly activity. Moreover, many administrators pride them- selves on their intuitive judgment and are suspicious of "hard data" without which they have made do in the past. Managers must be taught to make effective use of sound data for policy and operational decisions. This requires overcoming their protective instinct to ignore anything that puts their programs in a bad light (e.g., data that criticize existing efforts or point to new, possibly more fruitful areas of endeavor that are outside the administrator's competence or interest). In addition, administrators resist collecting data because they fear that it will be used against them by the existing political structure to negatively evaluate, limit, or even eliminate the program. They then throw up a series of barriers just in case the evaluation turns out less than positive. The resistance phenomenon has been characterized by the following hypotheses: ° The greater the perceived threat to the manager's or client's positive self-concept, the greater the resistance to negative findings. ° The greater the distance between the manager's concept of social reality and the actual data collected, the greater the resistance to further data use. * The greater the salience to the client of the function for which data are assembled, the greater the resistance. The picture of the manager who loves to respond to a crisis and likes to "shoot from the hip" is the picture of a manager who will resist an MIS. Management by MIS requires: ° Established target levels for performance; an acceptable deviance or variance range; and an understanding of the possible responses necessary to correct deviant situations. 94 ° A willingness to face critical evaluation and change. ® A mutual trust and sense of partnership among the various actors in the treatment scene--the clinical staff, the clinic manager, and city, county, State, and Federal drug abuse treatment officials. Questions About the Integrity of the System Since data collection is usually conceived as meaningless or counter- productive work, persons assigned these tasks are rarely given adequate training or supported by positive attitudes. This in turn leads to careless work and to questions about the reliability of the data and the integrity of the entire system. A measure of the interest an agency takes in data collection is the amount of time devoted to training personnel in the process of data collection and the status of the individuals involved with it. Many agencies assign untrained clerical personnel to the task of collecting and tabulating data or require that professional personnel squeeze these "unimportant" tasks into an otherwise full schedule. Since a low priority is assigned to these tasks, little value is placed in the contribution of the data to the well-being of the program and the data-collection work is monitored infrequently. If the same concern was expressed over the payroll, the entire staff would be in an uproar over major errors in data collection and data conversion into information in the form of inaccurate paychecks! Cynicism Created by “Imposed” Systems Because of the way in which the drug treatment industry has evolved, most clinic manager had their first taste of MISs when a system was imposed on them from above by city, State, or Federal officials. These early systems were not intended to meet local treatment center needs but those of public policymakers and funding agencies and so the clinic manager found these systems unhelpful and burdensome. Unfortunately these early (and in many cases continuing) experiences created stereotypes about the nature of MISs in general. All management information became branded as superfluous, unreliable, and impractical. Managers must learn to distinguish between information that has immediate local relevance and information that may have relevance two or three tiers away. Because the latter is not desired does not mean that the former cannot be very useful. Resistance to Paperwork "I am a professional. Taking time from my professional day for recordkeeping is demeaning. Recordkeeping wastes time that should be spent working with clients." This a common lament. 95 Paperwork is equated with busywork; recording professional activity is looked upon as an invasion of professional privacy. In fact though, professional ethics and good client treatment--not to mention third-party payers, and management legal require- ments--demand that professionals maintain individual clinical records. A careful recording of staff activities simply recasts this required information into a more useful format, enabling the manager to treat the individual staff members as a team. It remains for the clinic manager to develop protocols and procedures by which only the most valuable information can be collected in the most efficient manner. The manager can minimize, but not eliminate, the paperwork burden. STRATEGIES FOR ELICITING COOPERATION FROM CLINIC PERSONNEL With all the interest in developing MISs and overcoming barriers to sophisticated management techniques, a repertoire of strategies for overcoming resistance is evolving. A few of these techniques are discussed below; in addition, the reader should review chapter 4 for a more expanded discussion of this topic. Tangible Rewards Managers can build confidence in MISs if the latter are used to bring about tangible improvements in patient treatment and staff working conditions. MISs can be used to strengthen the patient scheduling system: if a review of management information indicates that few clients come to the clinic earlier than 11 a.m. except on Saturday, a decision might be reached about altering clinic hours. This action would have the multiple benefits of 1) increasing staff productivity, 2) expanding the number of clients actually treated by the program, and 3) reducing the amount of waiting time for attention. All of this will go a long way toward reducing tensions within the clinic and improving staff morale. Evidence That the Data Are Used in Decisionmaking There must be continuing evidence that the data are used for decisionmaking. Unless the clinical manager understands and actually uses the MIS, the staff will not be motivated to collect the data. Clinical staff should have evidence (through reports, staff meetings, and bulletins) indicating which data are useful and how those data have been employed. For example, if an ever-increasing number of admissions is documented by the MIS data, staff will expect to see a positive response--either the creation of a waiting list to lighten their caseloads, establishment of a referral system to divert some of these admissions, or an 96 increase in staff. Perpetuation or intensification of an intolerable system--in the face of documentation that the system is intolerable-- will discredit the documentation. Feedback to Staff If the data from the MIS are fed back to staff members in a meaningful way, the data-collection system will improve. Clinic staff must get feedback regarding: ° The perceived accuracy of the data ® The actual or potential utility of the data ° The implications of the data for the overall program goals and objectives If the clinic is made up of several departments, it is extremely important that a total clinic report be produced for review by staff throughout the clinic. For clinics with a relatively small staff, it is recommended that an hour per month be set aside to review the relevance of the MIS ouput. Without such mechanisms for feedback, the staff will not perceive the relationship between the data and the other program goals and objectives. Technical Assistance to the Staff Technical assistance must be provided to the clinical personnel so that they can interpret MIS reports. The staff will need to learn about trend analysis, variance analysis, ratio analysis, and other analyses available by virtue of the MIS. The staff will see the value in the data-collection activities if they understand the significance and implications of the resulting reports. For example, if a counselor reads that there is a minimum variance in the expenditure rates for the month or that the units of service are increased while the staff activity is stable, there will be no cause to "jump for joy" unless the counselor also knows: ° That the program was in severe financial difficulty and it implemented a program to curtail expenditures. The program now appears to be a success which means that the level of service can be maintained and staff salaries will not have to be cut; and ° The relationship between units of service and staff activity. In this particular case the trends indicate increasing produc- tivity with an associated increase in revenue resulting from the rising units of service (of which many are billable). 97 Economy in Data Collection A sure-fire way to make an MIS fail is to build in duplication and overlap. One pitfall to avoid is asking for endless information that will be of little if any use. Each data element contained in the MIS should be needed, not just "nice to know." SUMMARY Introduction of an MIS promises no certain work reduction either for clinical or administrative-clerical staff. It does not remove the manager's critical resource allocation problems, to be sure; it merely provides information on the basis of which the manager can carry out these important responsibilities. In fact, the installation and use of an MIS may produce more work for managers. However it is the thesis of this chapter that management decisions rendered on the basis of informed judgment are more appropriate than decisions rendered on intuitive judgment alone. Investment in an MIS that has meaning for the clinical staff will reward the clinic as a whole, supporting better management of limited resources, improved client services, equitable workloads, and objective criteria for measuring staff and client growth. The achievement of these goals can be accomplished with the assistance of an MIS, a technology which is becoming increasingly necessary for clinic managers in drug abuse treatment centers to insure the very survival of this system of care. AUTHOR EDWARD LEIBSON, Ed.D. Director Wayne County Department of Substance Abuse Services Detroit, Michigan 98 Appendix A Sample Staff Activity Codes From the Wayne County Automated Information System STAFF ACTIVITY CODES FOR ISAMIS' GENERAL INFORMATION Several new activity codes have been developed to reflect units of service and staff activity (manning hours) levels required. by ISAMIS reports for Specialized, Supervised, Methadone, Crisis and Intake-Referral modalities. SPECIAL ACTIVITY LOGS A special Agency Staff Activity Log must be maintained to record patient-days (030 and/or 040) and methadone doses (050). These Activity Logs are identified by filling in a worker ID of 999-99-9999. Inpatient days or methadone doses are recorded in the Client ID field. Number in group and time-code fields should be left blank. NEW ISAMIS ACTIVITY CODES The following are new activity codes to record ISAMIS reporting requirements. SPECIALIZED CARE MODALITY 035 Specialized Care Direct Staff Hours: Report the total direct staff hours that staff spend in providing services in the specialized care modality. The total staff hours should include all of the time that a doctor, nurse or counselor spends in doing activities relating to the services of the patient care or related activities. 'ISAMIS is a manual MIS developed by Touche-Ross for the Michigan Department of Health. The acronym stands for Integrated Sub- stance Abuse Management Information System. Wayne County has automated the system and the codes presented here are Wayne County codes. 99 030 045 040 Inpatient Clients: Report the number of beds being occupied by a patient each day in the hospital or residential care unit. The bed count should not include reserved beds for clients who may enter the hospital or residential care unit at a later time. Or if the client has left the program suddenly, this should be counted as a discharge. The bed count should be taken at the same time each day--such as 12 midnight. Enter this count in the Client ID field of the Staff Activity Log. Supervised Care Direct Staff Hours: Report the total direct staff hours that staff spend in providing services in the supervised care modality. The total staff hours should include all of the time that a doctor, nurse, or counselor spends in doing activities relating to the services of the patient care or related activities. Residential Clients: Report the number of beds being occupied by a patient each day in the hospital or residential care unit. The bed count should not include reserved beds for a client who may enter the hospital or residential care unit at a later time. Or if the client has left the program suddenly, this should be counted as a discharge. The bed count should be taken at the same time each day--such as 12 mid- night. Enter this count in the Client ID field of the Staff Activity Log. METHADONE MODALITY 055 744 050 Methadone Direct Staff Hours: Report the total hours that a nurse or doctor spends in dispensing methadone or activities in providing services to the clients in the methadone modality. Monitoring Urinalysis: Report the time that a counselor or other nonmedical person spends in monitoring urine specimen activity. Methadone Doses: Report the total number of methadone doses regardless of the size of the dosage dispensed each day. Enter this count in the Client ID field of the Staff Activity Log. CRISIS INTERVENTION MODALITY 113 213 Crisis Contact (Face-to-Face): If the contact with the client is in response to his/her request for substance related services, i.e., information or advice that is urgent in nature, this type of contact should be documented in the program's client case files. Crisis Contact and (Telephone): If the contact with the client is by telephone and is in response to his/her request for substance related services, i.e., information or advice 100 056 890 860 861 863 865 866 867 868 869 that is urgent in nature, these types of contacts should be documented in the program's client case files. Crisis Intervention Manning Hours: Report the total direct staff manning hours that staff spent in manning the crisis contact point, i.e., the telephone or walk-in desk. Educational Copresentation Activity: Report only the total staff hours that a copresenter spends in giving the presen- tation or lecture. Do not include the number of contacts, travel time, or preparation time spent. Count only the time spent in giving the presentation. Professional Organization: Activities involved in giving a lecture to a professional organization on a substance related matter. Count the number of contacts and the time spent in giving the presentation. Do not include the preparation time or travel time. Educational Organization: Activities involved in giving a lecture to a school or other educational organization on a substance related matter. Count the number of contacts and the time spent in giving the lecture. Do not include the preparation time or travel time. Social Services: Activities involved in giving a lecture to a social services agency--public welfare--on a substance related matter. Count the number of contacts and the time spent in giving the lecture. Do not include the preparation time or travel time. Hospital: Activities related to giving a lecture to a general hospital staff--public or private--on a substance related matter. Count the number of contacts and the time spent in giving the lecture. Do not include the preparation time or travel time. Citizens' Group: Activities related to giving a lecture to a citizens' group in how to combat or deal with specific sub- stance related problems. Count the number of contacts and the time spent in giving the lecture. Do not include the preparation time or travel time. Model Cities: Activities related to giving a lecture to any model city agencies on a substance related matter. Count the number of contacts and the time spent in giving the lecture. Do not include the preparation time or travel time. Union: Activities involved in giving a lecture to the labor union members on substance related matters. Count the number of contacts and the time spent in giving the lecture. Do not include the preparation time or travel time. Group Meeting of Three or More Agencies: Activities related to meeting with other agencies to discuss how to combat any 101 870 872 873 874 884 problems related to substance abuse. Count the number of contacts and time spent in giving the lectures. Do not include the preparation time or travel time. DMH Agency: Activities related in giving a lecture to any of the Detroit Mental Health agencies on substance related matters. Count the number of contacts and the time spent in giving the lecture. Do not include the preparation time or travel time. Other Organizations: Activities related to giving a lecture to any other organization besides the ones already listed on substance abuse matters. Count the number of contacts and the time spent in giving the lecture. Do not include the preparation time or travel time. Legislative Groups: Activities related in giving a lecture to any legislative organization on substance abuse problems. Count the number of contacts and the time spent in giving the lecture. Do not include the preparation time or travel time. Public Health Agency: Activities related to giving a lecture to any public health agencies on substance abuse matters. Count the number of contacts and the time spent in giving the lecture. Do not include the preparation time or travel time. Clergy: Activities related to giving a lecture to church group members or clergy on substance abuse matters. Count the number of contacts and the time spent in giving the lecture. Do not include the preparation time or travel time. INTAKE/REFERRAL MODALITY 057 Intake/Referral Manning Hours: Report the total direct staff manning hours that staff spend in doing intake and referral activities at a specific location, i.e., telephone or desk. SCHEDULE CARE MODALITY 111 112 Diagnostic and Evaluation Interview: If the contact is for the purpose of diagnosing or evaluating the client. This initial contact has to be with a professional staff member of the agency and its purpose should be to develop a treatment plan with the client. Psychological Testing: If the contact is for the purpose of conducting a written or verbal psychological test with the client. 102 114 115 116 117 119 120 121 122 129 130 131 132 133 134 Individual Therapy: If the contact is with an individual client (i.e., one-to-one basis) and is therapeutic in nature. Intake Evaluation (Prospective Client): If the contact is for the purpose of diagnosing and/or evaluating the client when the client is prospective and does not have an assigned ID number. Interview After Closing: Follow up on client after he/she has been terminated from the agency. The contact has to be therapeutic in nature and must be documented in the client's case records. Psychology Intake Orientation: If the contact with the client deals with the psychological aspect of the client's treatment. This must be documented in the client's case records. Vocational Counseling: Individual contact with the client in helping the client with his/her vocational needs. Educational Tutoring--Individual: If the contact is to help the client with his/her educational process. Client Family Therapy: If the contact is with one or more of the client's family and is therapeutic in nature. The client may or may not be present. It must be documented in the client's case records. Number in group should be completed indicating client and family members. Group Therapy: If the contact is therapeutic in nature and involves (a) two or more clients, or (b) two or more clients and others, either family members and/or collaterals. Group therapy is generally characterized by: (1) limited size of the group, (2) predetermined goals or objectives. A roll call must be taken and documented in the client's case records. Cotherapist Activity: Report only the staff time spent in the group session with another therapist as the head therapist. Do not include the number of clients or the ID number of the clients in your staff activity logs. Work Therapy: If the therapy constitutes work-related tasks. Activity Therapy: If the therapy is related to physical tasks. Play Therapy: If the therapy is recreational in nature. Craft-Art Therapy: If therapy is performing activities related to craft-art activities. Vocational Counseling Group: If the contact is to counsel the clients on their vocational needs or problems. 103 135 136 138 211 214 215 216 221 Vocational Testing Group: If the contact is to give a written or verbal test for clients to see in what areas of vocational placement they belong. Vocational Instruction Group: If contact is to give a class in vocational activities. Educational Instruction Group: If the contact is to conduct classroom instruction to help clients with their vocational needs. Diagnostic and Evaluation (Telephone): If the contact is for the purpose of diagnosing or evaluating the client by tele- phone. This must be documented in the client's case records. Individual Therapy (Telephone): If the contact is with an individual client and is therapeutic in nature. If the client has an urgent problem and cannot come into the clinic. This must be documented in the client's case records. Intake Evaluation (Telephone): If the telephone contact is for the purpose of diagnosing and/or evaluating the client when the client is prospective and does not have an assigned ID number; if the client has an urgent problem and cannot come into the clinic. This must be documented in the pro- gram's case files. Interview After Closing (Telephone): Follow up on client by telephone after he/she has been terminated from the agency. The contact has to be therapeutic in nature and must be documented in the client's case records. Client Family Therapy (Telephone): If the contact is with one or more of the client's family and is therapeutic in nature. The client may or may not be present. The contact must be documented in the client's case records. 104 Appendix B Sample Staff Activity Log 105 901 STAFF ACTIVITY LOG 1. Agency 3 code [11] 2-yy Nes 3. Staff Name: ssno. LITICTICTTT] (a5 5. 6. Ts 9. 10. |11. 12. Service/ No. in | Time Agency of Mo. Client ID No. Activity Group | Units Served (20-21) (22-27) (28-30) 31-32)|(33-34)|(35)(36)Y (37-39) 1 2. Period covered (Mo., Yr.) 13. Comments 4. Discipline code I I PP C1] (18-19) 4. People and Data Systems Some Issues of Integration George De Leon, Ph.D. INTRODUCTION Roth management information and evaluation data systems utilize quantitative information for monitoring program operations or assessing program effectiveness; they converge in a common aim--to help organizations clarify and achieve their goals and efficiently deliver the promised services. Often such data systems are not viewed positively by treatment people. Rather than technological extensions that facilitate human services, they are seen as remote substitutions for face-to-face interaction. Thus human factors must be considered in implementing information capabilities. This chapter discusses some issues and strategies for integrating treatment program people with data systems. MANAGEMENT INFORMATION AND EVALUATION Important distinctions exist between research, evaluation, and management information. Each refers to a process of posing questions, gathering information, reviewing the implications of the data, and acting on these implications. Often, these endeavors rely on the same data base but there are important differences in the questions addressed, the depth of the data resourced, and the strategies employed toward resolution. Research is a generic term that refers to a search for basic mechanisms. Frequently it involves experimentation and the intervention with a process by which to manipulate putative variables in order to shed light on the how or why of that process. Little experimentation occurs in drug treatment modalities because of a reluctance to tamper with the treatment process and because of ethical and legal considerations involved in manipulating people. 107 Evaluation is also a generic term that encompasses the several levels of inquiry which assess the value or validity of a program's treatment effort. Five levels are described below: the first two are the primary although not exclusive domain of management information systems. The last three levels usually require a more extensive data system. Demonstration Evaluation Is a model, treatment, or program feasible? A new program should be described in terms of its purpose and conditions of operation; how or why the program works is secondary. If the program is assembled and the conditions of operations met within suitable limits, the demonstration is evaluated as successful. Operations Evaluation Is a program honestly and efficiently operating according to its blueprint? Questions here pertain to the adequacy of staff, space, and equipment; sophistication of procedures, recordkeeping, fiscal administration, and management; and sufficiency of services provided. This type of "nuts and bolts" evaluation assumes that integrity of operations is necessary for achieving program goals. Process Evaluation How does the treatment work? Process questions ask about the relationship between client change and treatment components. In the therapeutic community, for example, the encounter group is a treatment component assumed to facilitate awareness and learning. Does it? Similarly, the psychological mechanism of identification is assured to mediate the influence of role models upon clients. Does it? Outcome Evaluation (Immediate) Is a treatment program effective? Does it reach the goals that relate to its purposes? While operations evaluation monitors the integrity of service provisions, and process studies explore how treatment works, outcome evaluation assesses achievement of goals. Treatment goals may include client changes during and at the termination of treatment stated in terms of drug abstinence, legal drug maintenance, increased employment, reductions in antisocial activity, or positive psychological change. Client outcome status obviously relates to treatment goals as reflected in the labels "graduate", "dropout", "complete," and "expelled." 108 Outcome Evaluation (Long-Term) If a treatment goal is achieved, how enduring or stable is that outcome? Long-term evaluation focuses upon relapse or, conversely, the stability of treatment effects. Followup studies assess client status at some postprogram period in comparison to an earlier status. Comparative outcome evaluation assesses the relative effectiveness and cost-effectiveness of different modalities assumed to be treating the same problem and similar clients. WHY DO PROGRAMS NEED DATA SYSTEMS FOR SELF-EVALUATION? Publicly funded programs with an information processing capacity contain a key mechanism for survival. As businesses, they must demonstrate cost-effectiveness and maintain good public relations. Program-based information can provide accurate pictures concerning effectiveness and cost-effectiveness at the local and State levels which are often the general repositories for funding. Accountability Regardless of the requirements for public accountability imposed from outside, drug programs can benefit from monitoring their own daily internal operations. While maintaining an orderly house is desirable in its own right, discrepancies between proposed and actual operations weaken effectiveness. For example, two clinical directors trying to do the work of five lessens the overall impact and ultimately reduces the number of program successes. Actually, accountability of internal operations is a reflection of the integrity of the program's personnel who after all are the significant mediators of positive therapeutic change. Well inten- tioned but overburdened staff have a particularly raw sensitivity to administrative pressures. Sensing danger from outside evalu- ators and board members, program directors feel compelled to hide or distort matters of recordkeeping, understaffing, census, and treatment and management failures. Although rationalized in terms of program survival or commitment to the disadvantaged, these nondisclosures are nevertheless deceptions which are poten- tially harmful. Absolute honesty and total respect for role models are essentials in the treatment process especially in residential settings. Cumulating increments of deception affect staff morale insidiously and reinforce the residents' characteristic mistrust of the "system." Although these effects are not directly expressed, their covert contribution to staff turnover and "split rates" should not be ignored. 109 Quality of Care Self-evaluation is necessary to improve the quality of care. Clients differ as to their attraction and adjustment to the various treatment modalities, and data which identify such differences improve screening techniques and facilitate successful referral to other treatments. The identification of client differences permits a direct assessment of the match between client and program and aids in optimal treatment. Further, reliable client social-psycho- logical profiles provide an empirical basis for improving client management and treatment. For example, differences between the psychological and behavioral difficulties of hardcore narcotic addicts and polydrug abusers compel variations in treatment plan and in staff assessment. The capability for self-evaluation guides changes in staffing patterns and clarifies treatment philosophies and program goals. For example, differential retention rates may compel programs to reexamine the reality of their treatment goals, to limit the univer- sality of their treatment outreach, and to review the relative contribution of degreed and nondegreed professionals. Staff education is a critical but indirect dividend of self-evaluation. Data invariably stimulate staff to communicate, conceptualize, self-examine, and read. This effect is particularly beneficial for paraprofessional staff who are too busy working in the clinical- management "trenches" and are usually not inclined to interpret their own work. Self-evaluation activities generate a steady flow of input that broadens their perspective. Finally, self-evaluation can help all members of the program staff function better as a team. Traditionally, there has been a per- ceived gap between clinical and management efforts in human services. Training program people in self-evaluation develops rapport between clinical and nonclinical staff and between pro- fessional and paraprofessionals. Teaching staff members to objectify work is an implicit goal of any self-evaluation effort, and this process enhances communication. Feedback between the various staff groups will improve skills at all levels and shape achievable goals. IMPLEMENTING EVALUATION CAPABILITY Evaluation systems can be designed, implemented, and operated in several ways: ° By external professionals ° By lay treatment center staff using manuals developed by outside experts ° By systems/evaluation professionals employed to train the treatment center staff 110 Each alternative has its advantages and each creates its own problems. Externally Developed and Operated Systems Many data systems now in programs were developed and are accessed by teams external to the program. These usually large multiprogram data systems (CODAP, DARP, DAWN, NDAC) are designed to obtain and access information that addresses epidemio- logical and funding questions at the State and Federal level. Such systems can monitor the ebb and flow of changes from a wide perspective and are needed for making broad policy decisions. External professional teams can also develop data systems for use by individual programs.’ Externally developed systems can support treatment center self-evaluation as well as accounting, scheduling, and other administrative functions. There are some specific advantages common to both varieties of externally implemented data systems. Trained professionals can develop and operate a system, review results, and provide informa- tion in the form of reports. The economic and service advantages of this approach are obvious. Program personnel involved in the management and delivery of services are not likely to have the expertise, the motivation, or the time to carry on evaluation functions at any level. Systems developed and processed by external personnel, therefore, offer a unique professional service. They function in a consultant-advisory capacity for program administrators who are the ultimate decisionmakers. In addition, outside personnel foster both an objective and a detached perspec- tive of clinical and management operations which are necessary for improvement and positive change. External systems reflect the strength of broad experience developed and refined from previous trial-and-error applications in other settings. Relevant information forms can be constructed, and time-tested procedures for process- ing and decisionmaking can be instituted. In short, individual programs need not reinvent the wheel in acquiring a systems capability. There are, however, serious disadvantages inherent in the use of systems developed by "outsiders," particularly those that purport to serve the needs of the treatment center. No single management information system (MIS) or evaluation system will address appro- priately the individual differences of programs or treatment modal- ities. Any system must be adapted to the unique features of the program and staff involved. Some limits of externally based data systems are: "The best example of this is the service bureau system discussed in the chapter "Automation Alternatives in the Drug Abuse Treat- ment Setting." ° They are designed to report "epidemiological" data on many programs and modalities and fail to provide sufficient indepth information on any single program. ° Externally developed systems can only approximate the day-to-day processes of individual programs. Continuous changes in clients and services remove external data systems from these changing activities and result in an information lag for the program and modality. ° Forms must be continually revised to be sensitive to changes "in the street." External changes may be unresponsive or sluggish and, thus, data ceases to be current and relevant. ° The useful information contained in data banks is generally not accessible to program personnel. ° External data systems are often perceived by program people as imposed burdens mandated from the outside. Frequently this results in noncooperation at the data-gathering level (and a poorer quality of data). Manuals: Their Limits and Uses While manuals have been designed to teach program personnel the essentials of management information or evaluation, they are usually researched, written, assembled, and distributed by external teams. Nevertheless manuals can be useful if they provide sound information, well-tested procedures, and intelligent formats designed with an aim toward teaching. In any developing area of human inquiry the manual or source book inevitably appears as a key instructional device with several advantages: ° Manuals contain uniform procedures that permit program personnel to learn practices and study information that can be compared with other programs. ° The written material in manuals can be studied and ultimately mastered. Thus they are extremely cost-efficient supplements to slower, more expensive teaching approaches. However while manuals instruct, they do not motivate, and this is a serious disadvantage in drug treatment settings for several reasons. Meaningful utilization of manuals depends upon the skills, resources, schedules, and sophistication of the users. Programs are conspicuously understaffed in research and manage- ment information personnel. Clerical staff who may be competent in gathering and processing data are burdened by other chores that place management information tasks low in priority. Program leaders must be completely committed to the values of information and evaluative systems. Manuals should be distri- buted to all program staff and must be supplemented with orienta- tion by knowledgeable program directors. 112 Nonprogram people generally author manuals in language which is not meaningful for treatment staff. Though intended for the nonprofessionals who need it most, they often serve as training tools for graduate students and professionals. Drug-free programs in particular have fewer academics and even fewer evaluator- researchers. Guided supportive training necessarily precedes the more demanding effort required for self-instruction. Manuals therefore are more likely to be used and appreciated at a later stage of management information capability. In-House Training of Self-Evaluation The implementation and maintenance of systems depends heavily upon the day-to-day workers. All management information and evaluation systems require that data be gathered by frontline people. Moreover the quality of the data--its clarity, completeness, and consistence--is controlled by these same people. Data systems, no matter how elegant, are quite useless if the people who partici- pate in them are not active, contributing, and receiving from the system itself. The direct training of self-evaluation capability overcomes many of the disadvantages associated with the other alternatives. The implementation, operation, and most important, the continuing use of the system depends upon the cooperation of all treatment center staff. ° Program people are more motivated to gather and process information when they have close control over instruments, definitions, system specifications, and interpretation of data. ° People in the human-services industry trust each other more than they trust outsiders such as scientists or technicians. When program people teach other program people the why, how, and what of self-evaluation, there is a greater likelihood of acceptance. ° Once program people are trained in self-evaluation and learn how to use information for constructive change, the motiva- tional seed is planted for catalytic efforts toward innovation and treatment change. The most serious drawback of this approach is money. It is very time consuming and expensive to train the entire treatment center staff in self-evaluation and using data systems. The expense is however more than justified in that this training emphasizes the human factors in implementing data systems. 113 IMPLEMENTING SELF-EVALUATION AND DATA SYSTEMS THROUGH TRAINING Implementing self-evaluation and a data system is a challenging and time-consuming job. Many programs never get past first base because they fail to follow some simple guidelines during the two stages of MIS implementation: preparation and maintenance. Stage 1--Preparation Program leaders, convinced of the need for and the importance of the system, set the attitude of respect and need for self-evaluation. ° Initially, leaders should officially call meetings of all lower staff for a full discussion of a data system, its purpose, promise, relevance, calendar, staffing, paperwork, and other demands. ® These sessions must elicit the honest reactions and questions of the staff. Full education in each staff role must be offered without implying recalcitrance or incompetence. ° The unique advantage of in-house information must be stressed. Program people are best able to gather data. This in turn must be translated back into a meaning and purpose which gives their work special value. ° A special task is to allay the personal and interpersonal fears and difficulties that data systems introduce in terms of job loss, the acquisition of new skills, and competition from newer often more educated personnel. ° The staff must contribute directly to the modification of any aspect of the implementation phase. The explicit relationship between the input of work needed and the outcome product must be emphasized by teaching the staff to identify with the role of the decisionmaker. This identification process is often helped by "reverse identification": Having the decision- maker temporarily carry out some of the tasks of lower staff, i.e., coding, "cleaning," and completion of forms. Identifica- tion fosters a cooperative rather than an imperial attitude among workers at different levels. Whenever possible the self-evaluation operation must be separately budgeted so as not to utilize program operation or service funds. Even a small program with limited evaluation staff and modest computer costs should consider developing an independent proposal to funding sources other than those providing service-delivery money. This tactic is essential to minimize the fiscal strain between treatment and management information or research objec- tives. Initially a management information and/or evaluation team need not consist of more than one full-time professional and a 114 capable assistant drawn from the program ranks. Thereafter staff and resources can be augmented commensurate with the levels of evaluation. The proposal for funding a data-system department should be developed by the professional coming on board. After the utility of the system has been demonstrated, the evaluation staff can then consider asking the program for fiscal support. Human Factors There are fundamental differences between management information and academic evaluation people, and treatment program staff. MIS and evaluation people utilize numbers, employ logical models of operation, and generate concrete products or outcomes as criteria of performance. As spectators rather than participants in service delivery, their perspective on organizational and treatment proc- esses is problem focused and decision oriented; and finally, their vernacular is technical and different from that used by clinicians, counselors, and frontline workers. These differences between the roles, goals, and languages of data people and those of program workers are often the basis for polarity, mutual suspicion, and organizational distance. Unfortu- nately the gap is even greater in human-services programs and particularly in drug treatment settings. There is no good rule for selecting data people (or any other type, for that matter) on the basis of personality type. Rather they must be made aware that they can be perceived either nega- tively as intruders who are at best a necessary evil and whose work is costly and often irrelevant, or positively as helping specialists. Their personal integration with others is crucial for the acceptance and success of data-system capacity. Integration is facilitated if data people are made aware that they often exert subtle but positive influences on clients as well as staff: ° Data people reflect a different lifestyle even though they are otherwise often similar to clients and staff. Their 9-to-5 presence tends to transmit a "normalizing" or stabilizing effect and to stimulate interactions. Mutually perceived differences between client, staff, and data people are lessened if these two groups are encouraged to talk to each other. ° Data personnel implicitly transmit a message to clients and staff that the program is self-critical and struggling to improve. This message fosters a sense of credibility which heightens the value of the program as a special endeavor. ° Specific evaluation procedures such as psychological testing move clients toward involvement and inquiry. Assessing their attitudes, intellect, and personality increases interest in themselves and in the process of change. Not infrequently 115 clients who have participated in research and evaluation projects request private meetings to discuss further their individual protocols. ° Weekly checks (edits) of the quality of the data that the staff is providing reinforces the need for quality data. The data-gathering staff should regularly be invited to attend and participate in meetings where decisions are made on the basis of system reports. ° Displaying the outputs of evaluations in public places in the form of easily interpreted graphs and charts will stress that the data are used on a timely basis and will let the staff know the results of their collection effort. ° The evaluation and systems staff should periodically drop in to the offices where data are being gathered to discuss problems and frustrations and to show concern for the people who support the system. ° Continuous training in the techniques of data analysis will increase staff ability to comprehend and benefit from the data they generate. Finally, self-evaluation activities employ and train clients in such areas as interviewing, test administration, simple data analysis, and writing. These job functions enhance verbal skills, open new intellectual challenges, and offer previously unconsidered vocational options. Stage 2--Maintenance The essential mechanism for maintaining data systems is that of feedback, both immediate and programwide. While the first involves small increments of information given at frequent intervals to the data-gathering staff (and program leaders), the second involves larger and more fully developed reports delivered less frequently to the entire program. Shared elements of both include descrip- tions of the program populations; client change during the treat- ment process; success, failure, and improvement rates; and in particular, how data jibe with clinical impressions. Such feedback is often experienced as inherently interesting and is easily related to by all staff. Changes in program policy or goals resulting from data analyses are exciting but demanding upon the program. Such changes should be introduced in small increments that can be easily installed and quickly evaluated. Immediate feedback should be built into the system from the beginning so that people cannot fail to recog- nize the importance of the system and their role in it. This can be done through large programwide formal sessions which are primarily tutorial and scheduled at least three times a year. The central aim of the feedback seminar is communication 116 during which any stereotypes of management information, evalu- ation, and clinical staff can be attacked. The large seminar should be conducted in a common language and be based upon brief reports with easily understood charts and simple statements of results. These reports should routinely cite the names of helpful individuals. In addition reactions both written and oral should be requested from all personnel in the program. Written visual reports are also tangible feedback products. These can be developed for publication or for the general program seminars. A CASE STUDY The classical therapeutic community (TC) is historically and philosophically removed from an established health-care institution. It evolved from quasireligious roots and its founders and directors are very often former addicts, alcoholics, or criminal-offender "paraprofessionals" who neither speak the language nor appreciate the rigor of behavioral science. Establishing a data-system capability in these settings illustrates that the challenge is not of integrating people with abstract data systems but rather of training new role relations and changing interactions toward mutual support among diverse clinicians, administrators, and evaluators. In Phoenix House, a classical TC, the strategies of integration parallel the changing focus of management needs and level of evaluation in progress. For each purpose the ingredients for maintaining integration can be identified, e.g., ongoing feedback, tangible products, and strategies for translating data into program policy. In the early stages the management information advantages of data systems were emphasized. The first data-system team consisting of one hired professional and an assistant from the program met routinely with clinical and administrative directors. At these meetings new intake forms were developed collaboratively. The staff were encouraged to support the data gathering process and they became convinced that data from good client records and file systems would permit easy monitoring of the composition and size of the population and would help the program stay within budget and help monitor workloads and retention. The clinical staff came to be convinced through their experience with weekly computer printouts of the help that such a data system could be in facilitating quality client care. After jotting down a number on a form or checking a box, a clinician was then free to carry on more important everyday clinical management functions. This positive "forgetting" experience tended to weaken resistance to intake and progress forms and reinforce a policy of good recordkeeping. 117 Data had to be translated for program relevance. Scheduled meetings focused upon issues of retention, admissions overload, "split rates," and understaffing. A bird's-eye view of program operations from the computer printout quickly clarified the "reality" of problems. Hard information often validated staff perceptions of program obstacles. Rather than being threatened, the staff could then appreciate its overextended efforts in generally underprovided conditions. Numbers actually minimized defensiveness and time- wasting arguments from impressionistic disagreements. Energies were directed toward solving problems which tended to reinforce staff competency. Two MIS products derived from this stage were financial reports and the program brochure--a 20-page monograph which described the sociodemographic composition of the program's population. The brochure was also a collaborative effort of staff and residents. Assembled with easy-to-read figures in an attrac- tive design (by the Phoenix Graphics Department), this collective expression helped reinforce program pride, provide educational material for staff and residents, and was also a community public relations resource. Based upon the management information capability, several limited- process studies emerged. These required psychological data measuring client change which was an additional strain upon personnel and procedures. A programwide 2-hour tutorial seminar was utilized to discuss the purpose and demands of the expanded evaluation. Staff and residents were encouraged to criticize and question the meaning, costs, and utility of the effort. These tutorial sessions educated, permitted "ventilation," strengthened the public-speaking skills of staff and clients, and increased interest and participation. In particular, the success of the project was seen to depend upon good data obtained with honest and complete client cooperation. Another positive dividend was the full data team interacting with all other personnel. Published articles were the main products of the process studies. These were not only useful to others in the field, but provided tangible reinforcers for those of the in-house staff involved in the effort for a year or more. Again, the programwide seminar provided feedback to residents and staff on the results and conclusions of the studies. Percent- ages of dropouts, emotional changes, or the rare relapse to drug use in treatment were provocative findings stimulating reflection and self-examination. Remarkably these "unsophisticated" audiences displayed the skepticism characteristic of science. Profound questions of interpretation surfaced which led to fundamental discussions of methodology, the relevance of measuring instruments, and the validity of clinical research. The tutorial seminar was found to be probably the single most effective means for integrating people into data systems and self-evaluation. Discussions may produce questions, disbelief, or agreement but the exchange facilitates involvement, dissolves 118 stereotypes, and maintains the staff involvement necessary to assure quality data. Clients and workers are more cooperative, helpful, and motivated when their interest is sustained through participating in all phases of the evaluation effort. Currently Phoenix House is conducting large-scale studies of process and long-term outcome. These projects require an expanded evaluation staff and greater program cooperation and commitment. Indeed assessing the central question of treatment effectiveness demands special courage from program founders, directors, administrators, and staff. To integrate outcome evaluation at the program level, discussions focused upon a balanced view of such efforts. The theme of the discussions was that the "truth" is good and it emphasized the following points: ° Programs equipped with honest success, failure, and improve- ment statistics can be more accountable and credible to the funding agencies; ° When programs understand which treatment works best for which clients, treatment plans can become more precise and treatment more successful. Also staff "burnout" from working with clients who do not benefit can be reduced; and ° The realistic appraisal of immediate outcome and long-term success reduces the sense of inferiority which underlies overextended claims for service. Determining the number of individuals who complete a program permits the program to assess whether it is delivering what it claims to deliver. No service is maximally effective for the entire spectrum of clients. Clarifying the long-term stablility of success allows a program to place its contribution to health-care in per- spective. The final products of the outcome effort at Phoenix have yet to emerge but already there have been notable changes in program policy and procedure. For example, the importance of measuring the client's status over time has been recognized. These studies have repeatedly shown client differences in relation to length of stay which stresses the importance of measuring "time in program" when evaluating the effectiveness of a residential treatment. More important, the findings have made clear the distinction between clinical success and statistical success. Graduation (clinical success) depends upon clinical criteria that satisfy the philosophy and experience of clinical workers. But statistical success (measurable positive change) is more frequent than gradua- tion and indicates that program influences may be more extensive than previously thought. This finding has reinforced faith in the clinical effort and raised staff morale. Moreover criteria for graduation are changing to accommodate shorter lengths of stay and to reflect individual differences in the rate of client change. 119 Currently Phoenix House has established task forces to integrate the staff into a programwide self-evaluation effort. In these task forces all staff are involved in processing information, reformulat- ing program policy, and developing special projects. Committees deliberate upon issues such as retention, special needs of minorities and women, expanding the effectiveness of community relations, and developing educational and training efforts for residents. Each group develops a position paper based upon the literature, clinical experiences, and hard data. The relevance of this model for integrating people with data systems and self-evaluation is obvious. First, data are presented and used within proper constraints and interpretations. Second, committees are a forum for exchange, homogenizing language, reinforcing mutual support, stimulating critical reflection, planning, and initiating collaborative writing. Based upon the position papers, specific proposals for new demon- stration projects emerged and task forces are now committed to evaluating all new projects. This is prima facie evidence of a change in consciousness of program people toward self-assessment and data systems. Since new projects are considered tentative and contain some risk of disrupting operations, their worth must be assessed expeditiously. In this sense, the essential lesson of data systems has been learned: new ideas and old practices must be questioned and answered. In a larger sense, the consciousness has changed from one of program survival to thriving. A healthy program knows what it is doing, understands why it works (or does not work), and identifies for whom it works best. Management information systems are a part of this consciousness. AUTHOR GEORGE DE LEON, Ph.D. Phoenix House Foundation New York, New York 120 5. Computer Software Bridge to Information Utilization Clyde B. McCoy, Ph.D., Anne C. McCoy, M.Ed., and James E. Rivers, M.A. INTRODUCTION Commercial computers made their appearance only a quarter of a century ago, yet their impact on our society has been as dramatic and pervasive as any tools or machines of the technological revolu- tion. Moon landings and space exploration were dramatic events made possible by computers but probably more important in terms of their impact on society is the involvement of computers in the daily activities of individuals. The routine activities of daily life such as telephone communications, receipt and payment of bills, credit purchasing, grocery buying, etc., increasingly involve computers. Developments on the horizon such as computer termi- nals integrated with home television sets are part of a continuing trend in computer applications which makes it imperative that every individual's education include some knowledge of computers. Computers differ from many technological innovations in that their impact is not limited to specific activities, industries, or businesses. Certain fields such as business and science were quick to recognize and utilize their advantages. However, other fields such as social services and health care delivery agencies have been much slower. Although health research scientists have made marvelous use of computers, those agencies which deliver social and health care services have been reluctant to utilize the full capabilities of them. An example is the field of drug abuse treatment and research. Drug abuse researchers have used computers extensively to investigate trends and relationships in the etiology of drug abuse. Treatment evaluation researchers are beginning to use computers more frequently to make comparisons and to assess the effective- ness and efficiency of treatment. But administrators of drug treatment programs who must make difficult resource allocation decisions and who must satisfy the increasing demands for public accountability have not taken advantage of computers with their ability to provide needed timely and accurate information. 121 Drug abuse treatment, like many other social and health care delivery services, has in recent years grown tremendously in the scope and complexity of its activities, in its dependence upon tax dollars for its operation and, consequently, in its need for infor- mation to improve efficiency, determine effectiveness, and account for expenditures. The scope of these information needs varies according to the scale and complexity of the organizations and client populations, but in most cases the data required to produce the needed information are of sufficient quantity and complexity to require the use of a computer. The requirement for computers is basically one of efficiency, but it also involves timeliness, a very important consideration when management and funding deci- sions are at issue. Further, the availability of computer services can often make the difference as to whether or not important functions are performed at all. Why have social and health care service agencies been so slow to utilize computer capabilities? There are several possible answers including the bias against computer technology and quantified data by human-service workers who perceive these as antithetical to a humanistic perspective. This perspective has resulted in the failure to train human-service workers and administrators in computer utilization. These individuals, as do people in general, see computers as mysterious and frightening with a technical complexity dissuading many from attempting to learn even a little about them. Too often those who could benefit considerably from computer usage continue to reject computerization when the admis- sion of ignorance followed by minimal time and effort to learn something about them could lead to greatly improved information, knowledge, decisionmaking power, and effective and efficient utili- zation of resources and cost-effectiveness. The purpose of this paper is to demythologize computers for potential social and health care delivery users." We will not focus upon the tremendous advances in recent years in computer hardware (the machines themselves), but rather upon computer software (the language and logic the machine uses). It is these latter advancements in particular which permit those people with minimal knowledge about computers and computer programing (the providing of instructions to computer machines) to conduct very sophisticated investigations and analyses and also to obtain timely information and reports. This chapter begins with the observation that there is a great lag between the potential capabilities of the computer and its actual utilization in the health care and social service fields. It is assumed that this utility lag represents a knowledge gap which is rooted in the mystery which surrounds these miraculous machines which are presumed to be able to accomplish great feats at the push of a button. While it is true that these electronic wonders "Those who have need to use a computer for any purpose regard- less of how remotely involved with the actual detailed operation of the computer system. 122 have the capability of performing astounding tasks, it should always be kept in mind that human beings are required to do the thinking and to provide precise (programing) instructions to the machines. Without programed instructions, the electronic machines are incapable of performing. In this paper we impart a basic knowledge of how computers work, emphasizing the necessity of the human element in computer programing. We refer to all of these human instructions and programing as software in contrast to the hardware--all the electronic components, metal cabinets, levers, buttons, terminals, tape drives, magnetic tape, and other storage devices. In the most general terms, the appropriate combination of hardware and software is a computer system. We emphasize software because it is the primary tool which has so greatly improved the user's ability to handle large quantities of data and to perform very complicated and varied tasks with speed, accuracy, and efficiency without extensive technical knowl- edge. In effect, software provides a bridge between the user of computers and the highly technical world of computer operation. Although "engineering" (programing knowledge, time, and expense) is required to construct the bridge, a knowledge of software is the simplest path to effective computer utilization in that it permits one to cross over the technical abyss. Such software bridges provide computer access to larger numbers of users than ever before. SOFTWARE DEVELOPMENT Although the hardware devices that could improve the efficiency and effectiveness of health and social service activities in the areas of policy formation, planning, budgeting, and service delivery, exist today (and have existed for some years), the necessary software development which would permit their utili- zation by these agencies still lags behind. Software developinent requires that the appropriate knowledge be applied to "program" the computers to perform the desired applications. Science and industry, possessing both the resources and competitive nature, have developed not only sophisticated and elaborate hardware but also large quantities of varied software to meet many of their desired applications. Much less a development of software has occurred for application to the health and social service areas presumably because of a lack of funds and the scarcity of knowl- edgeable and interested people. Since computers do not operate at all without software, much of it is developed and provided by the manufacturers and vendors of computer hardware. Typically a bundled package of hardware- software is sold together as a complete computer system; this type of software generally is a systems package. Practically all computer systems come with such software as the operating system (although some users modify the manufacturer/vendor-developed system or develop their own operating system). Unfortunately, hardware manufacturers/vendors, in their eagerness to sell expensive hardware, have often taken advantage of the buyers and users of their computers (particularly government agencies) by misleading them as to what the computers can do. They have failed to inform the users that the hardware-operating software will not perform many desired tasks and specific user applications without a great deal of instruction (programing) by very knowledgeable and experienced people. Indeed, programing or software has today become considerably more expensive than hardware. To appreciate the nature of software, we now elaborate on how software is developed and the components that are required before desired information can be extracted from the computer. As indicated above, the manufacturer/vendor typically supplies the operating system (or executive-monitor) which permits and controls all other operations; it is the general framework within which all other programs operate. Since the computer system requires input/output devices (ways to get information into and out of the computer) so that one can communicate with it, the operating system also provides programed machine commands to control the activities of these input/output devices. Computers follow human instructions but they can only "read" and "understand" machine language which is composed of special types of symbols (binary code). An elaborate system of software pro- grams has been developed for the purpose of translating other types of more easily coded programing (source) languages (such as COBOL, FORTRAN, and RPG) into the machine language. These software translators are known as program compilers and they too are usually provided by the manufacturer/vendor. Other important software generally provided along with the hard- ware permits the programer to detect and eliminate programing coding errors as well as to ascertain whether the program actually did what was intended. Most applications performed by the computer are initiated through programers--people knowledgeable in the standardized programing languages. Another general type of software structure is classified as utility programs. These are of several varieties and are used as "finish- ing tools" to assist the programer or operator in sorting files and presenting data files? in different ways for processing. Most large computer operations employ a variety of utility programs to facilitate data conversion, transfer, file construction, etc. Because there are general functions common to many computer systems, it has been possible to develop much standardized soft- ware. The research and development expenses for new software are typically high and the developer wishes to recover, at minimum, Any set of organized data where each specific piece of information has a known location on some machine-readable form such as punched card, magnetic tape, or other storage device 124 these R. & D. costs. Therefore the software developed is usually broadly applicable to make it attractive to as many users as possible. By distributing the software as widely as possible, the developer maximizes the chances that costs will be recovered and that the users will obtain software at affordable prices. Software written for wide distribution to multiple users is known as a software package--a specific computer program (or set of programs) designed to perform one or more well-defined functions which the developer designed with several potential users in mind. The package is made available in "canned" form with associated documentation and maintenance and is offered either free or at a fixed price. The usually modest price of such standardized software should be weighed carefully against the not insignificant probability that it will have to be modified to meet the user's specific needs. Any required modification to the standardized software necessitates that programers be knowledgeable about both (1) the standardized software program to be modified and (2) the specific application desired. The services of such knowledgeable programers can represent considerable expense to the user. This expense is particularly great when a special computer program application must be written for a single user who must bear all the modifi- cation costs. There are a large variety of software packages on the market today but all of them can be classified into two basic categories: systems packages and application packages. Systems packages are programs or sets of programs that make it possible to use a computer more conveniently or operate it more efficiently. Examples are the operating systems, assemblers, compilers, input/output control routines, translators, and debug- ging aids mentioned previously. Systems software is still largely the domain of the computer manufacturers but independent suppli- ers are entering the market at an ever-increasing rate and achiev- ing some noteworthy successes. Although only about 25 percent of the currently available software packages fall into the systems category, these systems packages account for more than 50 percent of total revenues. Application packages are programs or sets of programs that A — specific data-processing or computational tasks. After getting off to a comparatively slow start, both the computer makers and independent suppliers are now extremely active in the development of packages to handle a broad range of business applications. Payroll packages are by far the fastest selling type to date, but packages for accounts receivable, accounts payable, general ledger, inventory control, production scheduling, and other common business data-processing functions are also being widely accepted. On the engineering and scientific front, appli- cation packages are comparatively old stuff: packaged routines to handle matrix inversion, multivariate statistical analysis, transcen- dental functions, and many other common but lengthy and tedious computational tasks have been in widespread use for many years. 125 Two common types of software programs which may cost up to a quarter of a million dollars are (1) data-base management systems and (2) financial management systems. Since all computer users are required to create, store, maintain, process, and retrieve blocks of data, data-base systems are usually employed at the large centralized computer operations of corporations or govern- mental facilities that seek to accomplish multifunctional tasks for various departments or divisions as well as for the central office of the organization. When their data requirements are large and varied, data management systems provide an organized file struc- ture and language routines to facilitate the utilization of massive data files. Financial management systems are specialized data management software designed to handle multiple aspects of financial data including payroll, accounts payable, accounts receivable, general ledger, inventory, tax computations, payroll deductions, etc. These large packages offer applications suitable for some health and social service agencies, including drug abuse treatment pro- grams, particularly when these agencies are part of a large munici- pal, county, or State administration. However, few if any health and social service agencies have been given sufficient resources either to develop new or utilize existing software to meet their informational needs in a timely fashion. Since these agencies typically depend upon large computer centers for their services and such centers attempt to serve many types of users, the question of priorities: programing time, computer access, turn- around time, etc., becomes crucial. Most of these computer operations b_.gan with accounting departments or functions and have added other users, e.g., police and water and sewage departments, which they service until each of these can justify their own systems. Typically the larger users with more routine applications receive the higher priorities. This situtation is not unique to health and social service agencies. A similar problem has been noted in business, industry, and education. As the demand for computerization has increased, certain departments, divisions, and individuals are denied the access, technical assistance, and other services required to accomplish their specific needs. The increased demand for pro- graming time for writing specific applications has been particularly acute. This demand results basically from the lack of existing software for the numerous needed functions. However, this demand could be diminished were there more users (or potential users) who were aware of software and knowledgeable enough to use it without the assistance of programers or other computer staff. Programers themselves could reduce the demands on their time by being more knowledgeable of existing software. Although software systems have the potential to narrow somewhat the gap between the capabilities of computers and their actual utilization, there still remains a great need to make computers accessible and usable to more types of individuals. Many more people should be able to utilize computers in their own special 126 areas of interest by acquiring a minimal knowledge of programing and software utilization. Such individuals, by stimulating the development and utilization of appropriate software, can hasten the greater acceptance and utility of the computer by: ° Reducing reliance upon technical computer personnel. ° Reducing costs of application programs. ° Making access to and provision of needed information more timely. ° Stimulating greater interest and knowledge in overall utiliza- tion of computers. Many potential users are not prepared to utilize software because they do not know the specific organization of their data (format) and the minimal set of instructions needed to apply the software to it. Until such knowledge is commonly available, the initiative to utilize software must be taken by administrators and certain other knowledgeable agency personnel. However, there is also a continuing need for computer personnel to appreciate the informational needs and required applications of the users. It is especially important that these computer personnel be able to communicate effectively with the user concerning his needs. Unfortunately it appears that computer personnel themselves have fostered much of the myth and mystery surrounding computer operations. The ignorance of most people concerning computers does not allow the external evaluation of computer personnel, their applications, and productivity using standard criteria. Their unique relationship and control over the operations and production of the computer have led to technical elitism and imperi- alism. The effect has been the accruing of considerable control of user organization functions, information production, and per- formance to computer personnel. User organizations are typically dependent upon computer person- nel by virtue of the former's ignorance. By failing to communicate with the user, by relying on computer jargon and failing to document technical activities in understandable form, computer personnel have assumed the guise of practitioners who possess mysterious power over others. Given a desire to retain this autonomy and power, computer personnel see their vested interest in keeping the masses (of potential and actual computer users) in ignorance. The naive user can be made to believe that s/he is obtaining extensive services from computer personnel by receiving voluminous output from the computer. In fact, however, many tasks of computer people are accomplished by short, simple instructions to software programs which represent small amounts of programer time. The user who knows nothing about computers or software is in no position to argue with data-processing personnel who deny his/er request for an application by saying it cannot be 127 done. Many such denied requests could in fact be fulfilled by pro- grams that the programers do not wish or do not have the compe- tence to write. The scarcity of technically specialized computer personnel (particularly high-level programers capable of designing software) has made them an expensive and powerful professional group on whom users are most dependent. The potential users of software are less dependent upon computer programers, as a result of software development. Some user applications can be accomplished by using software requiring minimal knowledge of computers and programing, thus reducing the user's reliance upon his/her own computer staff (should s/he have one) or upon the programing staff of large computing centers (e.g., city, county, or State government computers, university computer centers, or computer service bureaus). No software system yet permits very refined applications to a large variety of restricted tasks but such a development seems only a matter of time. Several existing packages are quite versatile. As suggested earlier, the impetus for using computer power to maximum advantage must frequently come from professionals-- administrators rather than computer personnel--who do not have the appropriate knowledge, substantive interest, and financial motivation in health and social service areas to provide the needed stimulation. Unfortunately the great advancement of software in financial, scientific, and statistical areas has not been matched in the health and social service areas. Nevertheless by utilizing existing software and writing minor modifications or interfaces, many software developments can be utilized advantageously by these professionals and administrators. REQUIREMENTS FOR UTILIZATION OF SOFTWARE Although software is designed to be transferred to different hardware configurations, one cannot assume that any specific software package will run on any computer. Compatibility of the hardware and software is determined by several requirements: ° The programing language(s) of the software and the capability of the computer system to handle it. ° The size requirements: How much computer storage space and processing space are needed to run the software? ° Accessibility: How may the software be accessed (card reader, terminal, tape, disks, etc.) by the computer and does that capability exist? ° Storage media: Storage for software may be cards or magnetic tape which will require operator assistance, or they may be on some media such as disk that requires little or no operator assistance when accessed by remote devices such as terminals. Once compatibility has been determined, the user must be aware of the various requirements of the software in order to utilize it, such as: ° The format of the data (the location and size of each piece of data on the type of storage on which it is located). ° File structure (how the data were stored, i.e., sequential, indexed sequential, fixed, or variable length records). ° Instructions: When software is designed, it must be deter- mined how the user will be able to utilize the package. The designer decides what and how many instructions the user will have available and how the user must interact with the computer to receive the desired results. Each software package is designed separately and most likely by different people so the method for using one package may not be transferable to other software. DISADVANTAGES IN THE CONSIDERATION OF SOFTWARE Software programs represent much research and development with concomitant costs and, hopefully, sophistication and utility. They represent savings of time and finances to the user. However, there are certain disadvantages to be considered. The disadvan- tage of all software packages is that they are limited to general types of functions, and any specific requirements not met by the package requires extensive and expensive modifications. Many of the large and complex software systems are so complicated (and produced by so many different programers) that modification is extremely difficult or impossible. Although designed for easy use, such software generally requires a knowledgeable computer staff to maintain, modify, and correct errors in such programs. The major disadvantages of utilizing software packages are sum- marized below: 1. Amount of space required for storage and computer process- ing. Each program requires a certain amount of computer space. This is limited on all computers, and space require- ments translate into computer expense. The more options and tasks that can be accomplished by the software, gener- ally the more space and time and, hence, the more expense to run it. If the user is only required to use a small number of the supplied options, then the time and space needed to process the entire software package represent a waste of computer 129 time and space. The user is trading on the advantage of not having to develop his/her own programs for the disadvan- tage of running a software package which is not totally utilized each time it performs a job for the user and, hence, is more expensive and time consuming than having a single economical program to do each specific task for each run. But many individual programs written specifically for a task can also be inefficient because there are many programing means to accomplish the same task and some of the routes taken by programers can be lengthy, time consuming, error ridden, and inefficient. 2. Although one of the major strengths of software is the capability to perform preprogramed tasks with minimal instruc- tions and codes from the user, this strength can become a disadvantage when tasks other than those which have been preprogramed are needed. The modification of software can be time consuming and expensive. In some instances it is more efficient simply to write new programs to do the desired tasks although this is not always feasible when the additional tasks depend upon or interface with the routines of the standardized software package. This is particularly true of large data-base managment and financial management systems. 3. Other disadvantages concern the purchase and maintenance problems of some software. Since there is much invested in the research and development of software (and potentially much to be earned), some software is expensive to purchase and maintain. Also many software suppliers will not supply the source documents (original programing languages) so that user modification is impractical or impossible. SOFTWARE DEVELOPMENT AND UTILIZATION IN DADE COUNTY The many needs for the use of computers in Dade County in the area of treatment, program evaluation, management, planning, and research have made us fully aware of the great lag between the potential capabilities of computer systems and users' actual utiliza- tion of them. Our experience in having to devise a computerized data system to handle large and varied data files has led us to utilize many software programs developed by others as well as to devise some of our own software much of which was developed to take fuller advantage of existing software. The gap between the capabilities of information systems and managers' actual use provided the incentive for us to devise a computerized data system to handle a large variety of data files. Consequently we developed an interactive system which illustrates the advantages and capabilities of software systems. Our software makes it possible to handle simultaneously many different data files with large N's and a large number of variables with a minimum 130 of effort and knowledge. The heart of our system is what we call "The Quick Interactive System." Although the use of the Quick System is very simple and requires little user knowledge or skills, one should have a competent programer to set up and modify the system as is required by most software packages. However, programing time spent on the system is minimal considering the time the user can actually access and analyze the data rather than having to rely on a programer to submit each request for information. Our system was developed to free our programer's time to design new systems rather than having that time used just to submit program runs. In addition the system gives the user the much desired capability of quickly manipulating data in a variety of ways. Some preliminary assumptions were made which we do not discuss in detail here but they must, however, be mentioned. The system is designed for use on an interactive terminal to interface the basic computer. Any data must be designed, collected, and implemented to assure an appropriate format for computerization and a capability of being transferred to magnetic tapes and disks. Also the data variables and values need to be clearly defined and labeled for best results (e.g., following SPSS? procedures). Not only does this facilitate and improve output but it hastens access from the terminal. In other words, the system assumes that the preliminary data processing has been performed in such a manner that permits quick and easy identification and formatting of all data items. This is of course extremely important for setting up, maintaining, and expanding any data-based information system (manual or computerized), i.e., that it will take full advantage of what is already available as well as make possible easier develop- ment of one's own programing-software routines. Among the major strengths of the Quick System is the capability of working on large and varied data sets which can be stored, updated, recoded, merged, and manipulated very quickly--a capability not always possible with other statistical software processing. Another very important aspect of the Quick System is the full utilization of another very frequently used software package, SPSS, which is a package of statistics ranging from simple per- centages and frequency counts to complicated algebraic prediction equations. However, Quick operates without many of the incon- veniences of SPSS such as the need to resubmit control cards. Quick also has the added capability to execute numerous runs simultaneously, even of different statistics. Because of the complete development of SPSS in meeting most of our needs for tabular and statistical analyses, we have developed our system utilizing its full capabilities. But one would not necessarily be 3Statistical Package for the Social Sciences, an example of an applications package discussed earlier. 131 limited to such a particular system although SPSS seems to have great advantages over other software packages because of its ease of use and complete structure for retaining and maintaining data files. Familiarity with SPSS requirements also aids one in the design stage of constructing and collecting data items to assure the fullest utilization of a computerized data set. The Quick System extends more fully the capability of computerization by permitting easy and fast manipulation of many and varied data sets within the same terminal session. The basic advantages of the Quick software package are the following: ° Allows different data sets to be investigated in one terminal. ° Allows two (or can be expanded to more) separate runs to be initiated on the same data set in one terminal session. ° Allows data files and software to be combined in an interactive system that can be completely accessed through a single terminal (at any of the usual computer terminal sites, or leased for one's own research site). ° Requires use of security key to safeguard confidentiality of the data base. ° Permits dynamic troubleshooting of itself. ° Accepts all SPSS procedures (or could include other software packages or one's own package). ° Allows incorporation of permanent recodes categorization. ° Most important, allows easy and quick expansion of data bases and expansion of the interactive system itself. The following example illustrates the use of the Quick system by showing one of the many possible interactions between the user and the computer via a terminal. x EXECUTE QUICK QUICK INTERACTIVE SYSTEM READY FOR USER THE FOLLOWING DATA FILES ARE AVAILABLE FOR ACCESS 01 ne 03 Ou 05 Ob 07? 08 09 10 11 12 CDP CLIENT SURVEY -- '73 CDP CLIENT ADMISSION -- '7Y4 CDP CLIENT ADMISSION -- '?5 - '?7 CDP CLIENT PROGRESS -- '75 CDP CLIENT TERMINATION -- '?3 - '77 JMH EMERGENCY ROOM ADMISSION RECORDS -- '72 - '77 JMH EMERGENCY ROOM INTERVIEWS JMH EMERGENCY ROOM FOLLOWUP SCHOOL SURVEY ARRESTEES INTERVIEUWS TASC FOLLOWUP NEIGHBORHOOD TRACTS -- DRUG RATES AND SOCIAL VARIABLES WHICH FILE NUMBER DO YOU WISH TO EXAMINE? * 03 THANK YOU FOR SELECTING FILE O3 WouLD You 01 oe 03 oy 05 LIKE TO: MAKE AN SPSS RUN ON THE RAW DATA IMAGES? MAKE AN SPSS RUN ON THE SAVE FILE? CREATE A SPECIAL SUBFILE AND RUN SPSS ON IT? USE INTERACTIVE CODE BOOK ROUTINE? USE INTERACTIVE GROSS TABS ROUTINE? ¥. User Response No *. Computer Response 133 * oe WHERE DO YOU WANT THE PRINTOUT SENT? 01 - MEDICAL SCHOOL D2 - COMPUTER CENTER 03 - TERMINAL 03 HOW MANY COPIES DO YOU WANT? 2 DO YOU WISH TO USE LEVEL 5 OR LEVEL b OR SPSS? b ARE THERE ANY PREBUILT RECODES FOR THIS FILE THAT YOU WOULD LIKE TO USE? Bl - YES 02 - NO D1 WHAT IS THE NAME OF THE RECODE? 005-DRUG LIST» DO10~-3EX ANY OTHER PERMANENT RECODES? gl - YES 02 - NO oe ENTER ALL PROCEDURE CARDS NOW {AT END- TYPE @ EOF. THEN HIT CARRIAGE RETURN} CROSS TABS VAR 05 BY: 010 STATISTICS ALL @ EOF {NOTE: YOUR RUN FILE NUMBER IS 003114k1l FOR CHECKING YOUR RUNZ DO YOU WISH TO REENTER THE QUICK SYSTEM? 134 0L - YES 02 - NO x oe EXIT QUICK To further illustrate how software packages are being used in Dade County, we will demonstrate how they answer certain manage- ment and administration questions. One demonstration of software utilization comes from seeking an answer to a seemingly simple question: What are the communities of origin for drug clients: Where are the drug clients located; or, what is the spatial distribution of the clients in treatment in the drug programs? The answer to this question has been used for many purposes in Dade County such as to determine whether clients attend clinics in or near their own neighborhoods. Are there, for instance, clusters of clients where no treatment facilities exist? Administrators from the city of Miami wished to know before allocating any funds how many and what type of clients lived specifically within the city as opposed to the rest of Dade County; and the Dade County administrators wanted to know whether the drug clients were from specific neighborhoods in which other county services were located. Data collected from drug programs provided the data elements which were processed into requested information through the utilization of computer software. In order to construct numbers and rates of drug clients for specific neighborhoods, each client's file contains some geographic designation of residence. In many cases that designation is not the exact address: whenever there is a need for the respondent to remain anonymous, exact addresses are replaced by some geographic location such as the nearest street corner. So as to allocate each respondent to meaningful geographic units such as police districts, catchment areas, zip codes, census tracts and blocks, etc., certain computer software routines are used. Each address (or substitute) is processed through two software programs designed by and available from the U.S. Bureau of the Census (Ad-Match and DIME reference programs, Bureau of the Census, 1970). This processing allocates specific geocodes to the address so that each individual's computer- ized file may be identified with several geographical units in Dade County. We have found the census tract to be most useful for our purposes. Following the geocoding of data files it is possible to analyze two types of data: (1) individual characteristics, and (2) aggregated data which may be grouped according to the geocoded units designated in individual files. We aggregate the data by census 135 tracts so as to utilize the extensive data collected by the U.S. Census which represents a rich source of interview data on the total population available on computer tape. Because of the geocodes which are recorded via the software onto each individual's record, it is possible to convert other variables into aggregated data which represent geographical units. Through the use of other software (sort-merge routines) the entire client population of one clinic or an entire comprehensive program is sorted into the respective geographical areas. By sorting the client files by geographical areas, we then create geographical files which indicate the number of clients, etc., for each specific geographic unit. The geocodes used and the created files are constructed such that other data like health and crime statistics (as reported by other agencies) will be added for com- parisons with the drug program's client statistics. The computerization of data in such a fashion by the use of software programs permits descriptions of the data by mapping and graphing as well as by statistical descriptions. For example a computerized mapping system being developed at present by our staff permits the data to be displayed on detailed maps by the computer. Comparative analyses of the distribution of the different study populations will reveal the differential spatial and social patterns of the comparative groups. Detailed descriptions of the characteristics of the neighborhoods will also permit accurate profiles of the various neighborhood environments in relation to different rates of drug use. SUMMARY All organizations are accountable for an output whether it is the production of goods or the delivery of services. The quality and quantity of such output is determined by the organization of resources including the personnel and the tools with which the personnel work. Most business and scientific organizations have concluded that the automation of data provides necessary tools for meeting their goals. Similar decisions confront service organiza- tions today, and the future will exert even greater demands upon health service delivery organizations to computerize. However, computerization itself does not insure increases in organizational effectiveness, efficiency, or timeliness. Increases in organizational effectiveness are only accomplished through compatability between organizational structure and the organization of the computer system. Critical to such compatibility are the factors of (1) knowledge, (2) time, and (3) expense. Optimization of each of these factors by organizations requires effective utilization of software. Few organizations can afford computer staffs to accomplish all their needs. And even with knowledgeable computer personnel, maximization of accessibility, timeliness, and appropriate computer 136 application require that other personnel have some access to and control over the computer and its functions. Therefore, software becomes the bridge between the people with minimal computer knowledge and computer utilization. Time more than any other resource requires full optimization in an effective organization. Not only does it determine how much is accomplished, but also the quality. If time is inadequately orga- nized and utilized, then services are performed poorly and, in many cases, not performed at all. A properly designed computer system including user-oriented software permitting most or all personnel access to the computer to accomplish all types of paper- work and information processing is potentially the greatest of "timesavers" and it also allows for the providing of other services not possible without computerization. Regardless of satisfaction or dissatisfaction with "paperwork," most service organizations spend as much or more time in some type of paperwork and information processing as they do in actual contact with clients or other activities. The amazing processing speed of computers permits maximization of time if certain conditions are met: ° Concomitant speed and timeliness must be available in getting information in and out of the computer processor. ° Personnel must have sufficient access and knowledge to utilize such input/output procedures. ° The output of information must be in a form and language understandable to the users of the information. Meeting these conditions requires effective utilization of software. The allocation of the resources of an organization must include a sizable investment in information processing of all types regardless of how such information processing is accomplished. Expense is represented by both the type of personnel and the time allocated to the particular required functions. Manual and semicomputerized operations probably utilize more personnel and time than is realized, and preclude many activities that could be accomplished with computerization. The release of staff time from certain information processing functions to other needed functions would be one great benefit of computerization in addition to enabling the accomplish- ment of tasks not possible before either because of constraints of time, knowledge, or capability. Of course computerization itself represents a great expense to the organization although the actual purchase of hardware by health and social service agencies has come more within the realm of possibility in the last few years with the advent of minicomputers. Also competent and economical computer service bureaus allow organizations to share the costs of computerization with most of the advantages and few of the disadvantages of having one's own large system. However the operation, maintenance, and utilization of the computer system represent the greatest expense. A great part of that expense is programing the machine to meet exact organizational needs and to maximize the efficiency of both computer time and personnel time. Optimization of resource allocations relating to computerization and its accomplishments requires the utilization of software developed either by the organization or using software developed by others. Having to write specific programs for the computer for each specific task is expensive, inefficient, and slow. With the availability of an effective software network, the utilization of personnel, knowledge, and expense is more efficient, timely, and effective. AUTHORS CLYDE B. McCOY, Ph.D. Department of Psychiatry University of Miami Miami, Florida ANNE C. McCOY, M.Ed. Department of Electronic Data Processing Miami Dade Community College Miami, Florida JAMES E. RIVERS, M.A. Department of Research and Evaluation Dade County Comprehensive Drug Program Miami, Florida 138 6. Automation Alternatives in the Drug Abuse Treatment Setting Herbert M. Birch, Jr. and Kerry G. Treasure More and more drug abuse treatment managers are turning to computers to assist them in operating management information systems and to direct their programs. Yet many managers are uninformed about the advantages (and disadvantages) of automation and often do not know how to go about selecting an appropriate system to meet their particular needs. This article is intended to shed a ray of light on automation at the treatment center level. It provides a basic definition of data processing, a set of evaluative criteria for assessing automation alternatives, the major data- processing options, and a straightforward strategy for selecting and implementing an MIS. WHAT IS DATA PROCESSING AND HOW DOES IT RELATE TO DRUG ABUSE PROGRAM TREATMENT MIS? Regardless of the type of organization (e.g., drug treatment program, meatpacking plant, carwash), its data-processing function is made up of five separate interrelated steps (Awad 1970): ° Origination of data » Input of data into the system ° Manipulation of the data ° Output of the reports ° Storage of information for future use Figure 1 presents an overview of these five data-processing steps. It shows the process of admitting a new client into the program; the same steps would be followed for a billing transaction, for preparing a payroll, or for a client encounter. Each of these 139 v1 Client encounter form Utilization Utilization File tally report J — J Men pin? % J 1% Origination Input Manipulation Output Storage Step 1 Step 2 Step 3 Step 4 Step 5 Feedback FIGURE 1.—The data processing cycle steps may be accomplished through a variety of methods. In any case the objective is to complete each step in the quickest, most accurate, and most cost-effective manner. Depending upon the nature of the organization, the most appropriate method might be manual, semiautomated, or fully automated processing. The remainder of section 1 elaborates on each of these steps. Section 2 describes the various processing alternatives and their relative applicability to the drug treatment setting. Origination of Data The first step in the processing cycle requires capturing data about an event on a source document. This step is rarely affected by the medium used to process the data (i.e., manual or automated). That is--and this point should be stressed--there is no way to avoid having the provider, the counselor, or some other individual employee record the transaction or activity on a source document. However, the amount of information recorded for each event may be affected by the processing medium. For example, a fully automated system may encourage the tendency to collect as much data as possible at the point of occurrence since the automated system can accommodate more efficiently large data bases at steps 3, 4, and 5 (discussed below). Figure 2 portrays several types of source documents upon which drug abuse treatment programs usually rely. Input of Data into the System Input simply means that the data recorded on source documents are entered into the MIS. This input step can be accomplished by manual, partially automated, or fully automated means.' To illustrate, consider two methods available for inputting data; each of these examples represents an extreme on the continuum of complexity: ° Manual Input: Input into a manual system is accomplished simply by people, paper, and pencils. The input data are derived from the source documents. The input device used to record the transactions is an ordinary pencil. Figure 3 illustrates the relationship between the input and the input device. Once data are input into the system (e.g., the expense ledger) they are available for processing during steps 3, 4, and 5. ° Fully Automated Input: At the opposite extreme the fully automated system may employ a terminal as the input device. Again, the data contained on the source documents are the 'In a highly complex automated system, steps | and 2 may be com- bined if, for example, source materials are entered directly into the terminal of a computer rather than being recorded first as hard copy. 141 vl Discharge record Treatment record . Requisition Admission form form Vendor invoice Staff time report FIGURE 2.—Examples of source documents evi Manual system Automated system Source document Source document Input device Input device Data processing system FIGURE 3.—/nput step source of input data. The data are then displayed on the screen, edited, and stored in the system for processing during steps 3, 4, and 5. This approach, becoming more cost-effective as the cost of terminals decreases, provides an excellent method of insuring accurate data input. Since data processing can be no more accurate than the input of source data, this is a critical feature. Manipulation of Data Manipulation is the process by which data are converted into meaningful information. There are five basic subfunctions (Awad 1970) of the manipulation step: classifying, sorting, calculating, recording, and summarizing. An example from the drug abuse treatment setting will illustrate each of these subfunctions; the means (automation or manual processing) by which these subfunc- tions are accomplished will be discussed later. ° Classifying: Grouping source documents into a logical order so that the data can be easily retrieved throughout the processing steps. In drug abuse treatment settings, this usually requires that the source data be arranged alphabet- ically or numerically. When data are filed this way, it is easy for staff to quickly locate treatment, admission, dis- charge or financial data about a particular client simply by knowing the client's name or assigned number. ° Sorting: Depending upon the purpose of the manipulation step, the source data are further categorized into meaningful subclassifications. For example: 1. If the purpose of the manipulation is to produce patient and third-party bills, then the source documents would be sorted by type of third-party payer. 2. If the purpose of the manipulation is to compare utiliza- tion and client characteristics by modality, then the source documents would be sorted into subclasses that distinguish among the modalities. 3. If the purpose of the manipulation is to compare trends in utilization over time, then the client data would be sorted into chronological subclassifications. ° Calculating: This is the process in which data are added, subtracted, multiplied, divided, or in which more complex mathematical procedures are performed. Calculations result in such "quantities" as percentages, totals, ratios, costs, statistical measures (Chi-square), and so on. For example, calculations performed on client financial records will generate information concerning the account balance of that individual. Calculations in which the client utilization levels are compared to the planned utilization levels will result in ratios of planned to actual utilization. And calculations performed on an 144 employee's timesheet will result in an amount of gross pay for the period. Recording: Sometimes the calculations (performed above) do not lead to a complete answer but only provide intermediate information that will in turn be used to produce the desired end product. When this occurs the results of the calculations are recorded or stored so that they may be used later. Two examples are helpful: 1. The accounts receivable card for each individual is updated daily through a series of calculations. The new subtotal for the day is recorded on the accounts receivable card from which a total is drawn and a bill prepared at the end of the month. It is the intermediate recording of daily subtotals that enables the billing clerk to compute the monthly balance due. 2. The treatment center that wishes to develop unit cost information must perform three separate calculations, record the answers from each calculation, and then combine the recorded results to reach the unit cost. Specifically the treatment center must compute infor- mation regarding: Be The direct labor costs associated with the service unit. b. The other direct costs associated with the service unit, Co The overhead and administrative costs associated with the service unit. As each piece of information is computed, it is recorded for eventual addition to the other pieces of information. Summarizing is the task of setting out the results of the manipulation step in a clear, concise manner. The precise format of the summary will depend on the audience to whom the report is addressed. For example the unit costs may be summarized: I. By modality (with the three cost categories combined) if the report is being used to evaluate relative costs among the various modalities. 2. By type of cost category (direct labor, other direct and overhead) if the NIDA project officer is concerned about the ratio of overhead costs to direct costs. 3. By unit of service if the report is to be used to prepare bills. 145 Figure 4 illustrates the various ways in which the information about unit costs might be summarized. It is easy to see that although each of these subfunctions can be conducted manually, a computer can be very helpful in the more tedious data-processing steps. More quickly and accurately than the human hand, the computer can classify, sort, manipulate, report on, and summarize large volumes of data. Hence there is enormous value to the automated approach to data processing. Output of the Reports This step is to a large extent human-based and requires that the reports and summaries produced thus far be communicated to the appropriate individuals, be used to make decisions, and create feedback into the system when additional information is required. Although the computer can assist in getting the reports to the appropriate individuals (i.e., by address labels or instructive headings), it cannot affect the use of the information per se and, once again, the human element controls the data-processing sphere. Storage of Information for Future Use Data and information processed through the system thus far can be stored in manual or automated files for future use. Whether manual or automated the files should make provision for easy access (e.g., with indices) and for periodic purging (i.e., eliminat- ing information that has no further use). With this understanding of the nature of the data-processing task, the manager is now ready to assess the data-processing requirements of his/her own treatment program and review these against the advantages of the various data-processing alternatives: in-house computers, service bureaus, and manual systems. Each of these options is discussed in the following section. WHAT ARE THE ALTERNATIVES FOR DATA PROCESSING? The drug treatment program has a range of data-processing alternatives available to it. Unfortunately there is no "pat" approach that will meet the needs of all programs equally well and each treatment center must review all alternatives and select for itself the method most appropriate to its needs. This section will assist the manager by providing: ° A general distinction between hardware and software. ° An analytic framework of criteria that should be considered in evaluating any data processing alternative. 146 Cost factor analysis Service and Direct laboratory Other direct Overhead total cost Dollars | Percent | Dollars | Percent | Dollars | Percent Admission $23.15. cv» 16.00 69 3.00 13 4.15 18 Counseling $36.98. .... 25.50 69 5.25 14 6.23 17 Modality cost report Service Outpatient Residential Methadone Admission . . .. $18.50 $23.18 $17.24 IN Counseling . . . . 26.00 35.17 21.78 Urinalysis . . . . . 5.00 7.21 3.79 Fee scale Service Fee Counseling swoon sss 30 TE sR FREE» 58 sn $36.98 NN ADMISSION, oon usnswnsnsnspsusus arn: vw 23.15 Urinalysis. . o.oo oii 6.42 Meals. . ...... 1.25 FIGURE 4.—//lustrative output formats 147 ° A discussion of the relative characteristics of three major data-processing approaches: in-house computers, service bureaus, and manual systems. Hardware and Software: A Clarification of Terms A decision to procure an automated data-processing system requires that the program obtain both hardware and software. The distinc- tion between these two wares can be illustrated by analogy to a player piano: ° The piano itself is equivalent to the hardware--the computer. It is simply a machine that without some instruction sits mute. ° The punched tape which drives the piano corresponds to the computer software. It tells the hardware which keys to activate and which pedals to depress. Without the piano the music tape (like the computer program or software) is incapable of making music. The hardware and software are mutually dependent. To carry the analogy further, if the bar in which the piano is to be installed stresses a modern tempo, the piano must be driven to play contemporary songs. However, if the proprietor of the bar is only able to find music tapes from the 1930s, the piano will be of little value in entertaining the clientele. Or if the piano can accommodate tapes 14 inches wide and contemporary songs are produced on tapes 12 inches wide, the proprietor of the bar faces an equally difficult situation. In the same way, the computer hardware and software must be compatible. A Framework for Evaluation: Twelve Systems Evaluation Criteria As the treatment center begins to explore data-processing alterna- tives, it should be concerned with 12 critical characteristics of each option. ° The system must be available as needed: Can the system be available when and as often as needed? Is there a waiting time to purchase the system or to obtain assistance in imple- menting the system? Is the system still in the developmental stages and, if so, is a reasonable deadline likely to be met? If the system is to be shared by more than one organization, will the drug treatment organization be able to exercise priority on use of the system? Are the foreseeable timelags tolerable given the data-processing needs of the treatment center? ° The system must be reliable at all times: Is the processing system reliable? (For example, if equipment is used, does it 148 function smoothly? If manual processing is employed, are the individuals dependable?) Is there a backup process that can handle the work if the equipment or individuals are out of commission? Are the facilities for repair competent, respon- sive, and timely? The system must stress security of information: Given the confidential nature of drug treatment records, are the process- ing procedures designed to safeguard sensitive information? Do the processing procedures comply with the Federal confiden- tiality regulations? What routine precautions are taken to protect information from damage due to fire, flood, blackouts or other disasters? The system must be able to turn around information in a reasonable time: Turnaround time is the length of time between step 2 of the data-processing cycle (data entry into the system) and step 4 (output). Does the system produce routine reports within a reasonable turnaround time? Does the system have the capacity to turn around nonroutine reports on an as-needed basis with little delay? The prepackaged software should be thoroughly suited to the treatment center needs: Many treatment programs will find that prepackaged software systems (the piano music tapes) are well suited to some of their needs. Does the treatment program really understand its needs? Does the prepackaged software meet these needs? Are mismatches between program needs and software capabilities tolerable considering the expense of building a software package from the ground up? Can the software be tailored in any way to make it more suitable? Can the treatment center obtain proprietary rights to the software (i.e., buy the software rather than lease or rent it) so that changes can be made in it or so that the treatment program will be protected if the software company goes out of business? The automated system should be accompanied by considerable programing consultation services: If the treatment center entertains the thought of purchasing hardware, what assist- ance does the hardware vendor offer in writing programs (software) for the computer? What is the cost of these services? What type of staff resources will the treatment center require to develop and maintain its own software? What type of assistance can be expected after the initial startup period for such services as modifying or adding programing capabilities? The system should be expandable enough to grow with the treatment program over coming years: What is the maximum volume of transactions that can be handled by the system? How is the cost of additional transactions accrued? What additional hardware costs will be incurred to support expanded capacity? 149 ° The system should be flexible enough to accommodate changing requirements: Can the software be easily modified to produce special reports? Does it have report-generator capacity? Will the system accommodate changes in billing formats, Federal and State reporting requirements, and planning issues? Can the system address idiosyncratic evaluation issues? ° The system should be relatively easy to install: If the system is to be installed on the premises of the drug treat- ment center, what renovations and physical accommodations will have to be made to the environment? How much space will be required and what functions will be displaced in allocating space to the new computer facility? ° The system should be fairly easy to implement: What time schedules can be predicted for the various implementation phases? What level and type of staff resources are required to implement the system? What types of training will be necessary to get staff "on board" the system? What types of forms will have to be designed for data origination and input? ° The system should be simple to operate: What type and how many trained staff will be required to operate the system on a daily basis? What type of staff will be required to upgrade or modify the system over time? ° The system should be cost-effective: Although these last questions are more difficult to answer, one must at least ask what are the true costs of the system: hardware, software (initial procurement or development), processing costs, staff salaries and training, installation and startup costs? Are there any "hidden" costs? What are the guarantees that the costs will not increase rapidly? How do the costs of the alternative systems compare with the costs of the data- processing (probably manual) system currently in place and are any increases in cost justified? In short, this list gives the treatment center manager a point of departure. By asking the questions, the treatment program can begin to get an idea about the relative attributes of the alternative systems. As a primer on the various alternatives, the following sections discuss--in broad brush strokes--the merits of four data-processing alternatives. It must be kept in mind however that within these broad categories the particular characteristics of each system may vary considerably. Nothing can substitute for a penetrating analysis by the treatment center prior to committing substantial sums of money to a data-processing alternative. 150 In-House Small-Business Computers: The Wave of the Future The evolution of computer technology has been rapid, and computers have grown in the direction of wider applications and user orienta- tion. Until recently, however, computers remained sufficiently complex and costly that their utility in small businesses was severely limited. In the early 1970s, the industry branched in a radical direction applying advanced technology to a line of computers appropriate for small-business use.? For the first time the possi- bility of a drug abuse treatment center buying or leasing its own computer has become reasonable. What is a small-business computer? In the most gross sense, a computer has five important components: ° Input mechanisms: Recall that step 2 in the data-processing cycle requires that original data be entered into the system. Thus the first important part of the computer is the input mechanism. There are two major types: 1. A keypunch machine which is not physically attached to the computer and which produces batches of punched cards that are fed to the computer at a later time. 2. The more typical input mechanism for small-business computers is the input terminal which resembles a typewriter. This mechanism is attached to the computer, either occupying the same room or being connected by telephone or wires from a location. This is known as online data entry. ° Central processing unit (CPU): This is the "guts" of the computer where the classifying, sorting, calculating, report- ing, and summarizing take place. ° The instructions (software): This part of the computer corresponds to the music tape for the player piano and can take one of three forms: tape, disk, and diskette. The latter is most commonly used by small computers and resem- bles a phonograph record. ° Storage units: Data files, like instructions, can be maintained on disks, diskettes, or tapes. ® Printer: This part of the computer translates machine language into "hard copy" reports and documents. In addition the computer can project images on a cathode ray tube (CRT) which resembles a television screen. 2Small business is usually defined in the computer science literature as being between 50 to 150 employees and having not more than $5 million in annual revenues. Clearly most drug abuse treatment programs fall within this definition. 151 Figures 5 and 6 illustrate several popular small-business computers demonstrating the compactness of the five components described above. There are many other types of hardware available from other vendors; these examples were selected randomly. Small-business computers can be used for a variety of purposes in drug abuse treatment settings. Some of the applications can be supported by prepackaged software programs; others require special programing efforts: ° Financial management applications: Prepackaged programs include payroll, accounts payable, cash disbursements, accounts receivable and billing, general ledger, and cost accounting. Special programs can be developed to determine unit costs and for financial planning. ° Client management applications: Fewer prepackaged software programs exist to support client management in drug abuse treatment programs. However, software can at least theoret- ically be developed to perform scheduling, medication monitor- ing, client demographic characteristics analysis, utilization analysis, productivity and outcome studies, Federal and State reporting, and so on. Such developments must usually wait on clear definitions of these terms--not a problem with the more generally understood terminology of accounting. Obviously, all small-business computers do not have identical capabilities. Some companies (e.g., Basic Four, Datapoint, and Digital Equipment Corp.) specialize in small-business computers. Larger companies (e.g., IBM, Honeywell, Burroughs, Singer, and NCR) have a diversified product line of which small-business com- puters comprise a relatively small part. Despite this diversity, it is possible to make a general assessment of the characteristics of small-business computers in terms of the 12 criteria introduced earlier. ° Availability: Because the treatment program has its own machine and does not compete with other organizations for its use, availability is rarely a problem and is in fact one of the most compelling reasons for obtaining one's own computer. ° Reliability: These machines have proved to be generally reliable. Because periodic failures of input or output com- ponents are to be expected, the treatment program should assess vendor responsiveness to (and charges for!) making necessary repairs. The treatment center should also try to identify comparable equipment in the area that might be "borrowed" during protracted "down" periods. ° Confidentiality and Security: Because the data never leave the premises, confidentiality of client records is assured. As with any information system, however, data can be abused internally and so the treatment program should develop a written plan to restrict circulation of data and prohibit unauthorized access. 152 Figure 5. Example of a small computer . 1 54 Figure 6. Example of a small computer Turnaround time: Turnaround time is not usually a problem with an in-house computer if the treatment center has formal- ized processing priorities and if the computer has adequate capacity to meet routine programing needs. Turnaround time can also be affected by the data input mode: both batch input and online input are common to small-business computers; the latter method, although more expensive and technologically complex, results in speedier response times and can accomplish tasks that a batch processing system cannot. For example: L Immediate status determination: A registration clerk can query the computer files via a terminal to determine if the client is enrolled in the program, is scheduled for an appointment that day, or is eligible for a partic- ular payment plan. 2. Medication monitoring: The computer files can provide immediate feedback (via the terminal) regarding the appropriate dosage for medication, the approaching expiration date of the prescription, delinquent urinalysis, or contraindications for medication. Packaged software programs: Prepackaged software programs for small-business computers are generally limited to standard procedures such as payroll, accounts payable, and general ledger. It is unusual to find more precisely tailored packages that support client management activities. In "shopping around," the treatment center may find a package that appears to meet many of its needs. Caution must be exer- cised to assure compatibility between the software package and the hardware of choice. Additional programing is very expensive. Programing service and technical support: Most of the small-business computer vendors offer "turnkey" systems. These imply vendor responsibility for smooth operation of both hardware and software. Because of the competition among vendors, the consumer can usually insist that the system operate satisfactorily as a condition of contract accep- tance. However, industry surveys indicate that post- installation services leave much to be desired. Therefore as part of the contract the treatment center should specify the amount and type of ongoing technical support to be provided and, if possible, identify particular individuals who will perform these ongoing services. It is considered desirable even to draw up two contracts--one covering hardware and the other software. Expandable: Most small-business computers can be expanded considerably by adding storage capacity or a faster printer. If the treatment center plans to grow and expand the computer function, it should consider (1) the cost of the original equipment and (2) the cost of expansion equipment. It may be that Computer A is less costly in the short run but that Computer B (with the larger initial capacity) is more cost- effective over the period of growth. 155 Physical installation: By definition these small-business computers are compact and relatively easy to install. How- ever they are likely to require physical plant modifications such as air-conditioning, circuit breakers, fire extinguishers, and electrical system upgrading. The costs of these modi- fications must be part of the evaluation of the alternatives. Ease of implementation: For turnkey systems implementation is relatively straightforward if the vendor provides adequate operator and user training. For unique programing applica- tions, implementation is far more difficult and will require commensurate investment of time and staff resources to design program specifications, test and debug the system, and implement it entirely. Ease of operation: Most small-business computers have been designed for operation by clerical-level staff. This is a significant attraction of these computers. However, if the system needs to be modified, the treatment center will need continuing access to a programer. Cost-effectiveness: Small-business computers (hardware alone) range from $5,000 to $100,000; software development costs may run up to three times the hardware costs and increase proportionately with the level of programing effort. The fixed costs of an in-house computer are relatively high and include the monthly purchase or lease payment, salaries, and space. Variable costs, on the other hand, are generally low being limited to punch cards, printer forms, disk packs, utilities, and miscellaneous supplies. Cost-effectiveness, then, depends upon whether the system capacity is used fully enough to justify the high fixed costs. Computer Service Bureaus A service bureau is an organization that owns a computer and that (1) leases or rents time on the computer, or (2) uses its computer power to perform specific tasks for clients. In contrast to the small-business computers described above, these companies usually employ the largest and most sophisticated equipment available as shown in figure 7. Service bureaus are sponsored variously by: Proprietary companies: The majority of service bureaus are for-profit companies. Some offer a wide array of services and others are highly specialized, offering only a single service (e.g., patient billing). Drug abuse coordinating agencies: Although not usually considered as service bureaus, single-State and single-county agencies often own or have ready access to computers which can perform data-processing tasks for drug treatment centers. Insofar as considerable data are already submitted to these 156 ! | | Jd J r “ i | ¥ | | I | Figure 7. Example of a service bureau agencies, ? it makes sense to explore the possibility of complet- ing the data-processing cycle through this mechanism. Service bureaus have two major modes of operation (in fact many service bureaus offer both modes as options): Online processing: This mode requires that the drug treat- ment center have a terminal on its premises that links the center directly to the computer. Data are entered into the service bureau computer via this terminal and, similarly, data can be retrieved by issuing a request through the terminal. This mode greatly enhances the responsiveness of the service bureau but is a rather costly feature and should be carefully justified before the expense is incurred. Batch processing: With this mode the treatment center forwards (by mail or hand delivery) batches (packages) of input documents to the service bureau according to a regular schedule (e.g., monthly). This requires that the treatment center collect its source documents (step 1 in the data-processing cycle) in a standard acceptable format that can be easily processed by the computer. Service bureaus like any service industry vary considerably in the quality and cost of their product. Nonetheless the following generalizations can be made about them: Availability: With the larger service bureaus availability is usually not a problem. However if the service bureau has one or two large clients (and the treatment center is not one of them!) the center may experience difficulty gaining access to the facility. Reliability: Because service bureaus tend to use modern equipment and to have adequate backup computing power to cover periods of machine failure, reliability is generally good. Confidentiality and security: Although security leaks are more apt to occur whenever data leave the treatment center, service bureaus are usually contractually obligated as well as being dedicated just from a good business point of view to maintaining a reputation for security and, thus, can be expected to comply with any reasonable safeguard plans. Treatment centers should review the Federal, State, and accrediting agency (if applicable) confidentiality regulations with the service bureau to determine how compliance is to be assured. Turnaround time: With the proprietary service bureaus, treatment centers can usually negotiate any length of turn- *FMIS and CODAP for example. 158 around required. In contrast, service bureaus based in coordinating agencies are notorious for their lengthy turn- around times. Overall, turnaround time with a service bureau is dependent upon: 1. The data entry mode, e.g., online or batch processing. 2. The relative priority of the treatment center in the hierarchy of service bureau clients. 3. The capacity of the service bureau computer to handle all of its clients. 4. The quality of the input data provided by the treatment center. Incorrect data (rejected by the system through a series of edits) will have to be corrected by phone or mail thus increasing the turnaround time. Packaged software programs: Packaged software support is one of the most important criteria in selecting a service bureau and in this respect service bureaus vary most widely. It is in the interest of the service bureau to assist the treatment center in locating software packages that meet treatment center needs and that are compatible with service bureau hardware. (The problem of hardware and software incompatibility is not as great a problem in this environment-- in contrast to the in-house computer--because software pack- ages are made to run on the kinds of computers typically found in service bureaus.) Although the number of packages available for operation in the service bureau setting is large, the treatment center will find that these prepackaged systems relate primarily to financial management needs and do not address the largely statistically oriented client management requirements. Programing services and technical support: Most service bureaus offer programing services and technical support to encourage new clients. These services can be used to modify existing programs, develop new software, train staff, draft input forms, and so on. The treatment center should negotiate a fixed-price contract for specific assistance. Moreover if the treatment program pays to develop a new software package, it should obtain ownership rights so that a shift to an in-house computer or other service bureau will not require redevelopment of the software. Expandable: The expansion capability of a service bureau is almost unlimited. However, the costs of the service bureau's services expand in almost direct proportion to the transaction volume. Hence a treatment center should monitor its costs closely to determine when the fixed costs of an in-house computer are justified by large volumes. Ease of installation: Ease of physical installation is a major advantage of the service bureau approach. At the most, 159 terminal(s) will be installed and this is as simple a process as installing a telephone. ° Ease of implementation: Although the need for training staff in data collection and interpretation remains necessary, the treatment center that selects the service bureau approach will not need to train its staff in computer operations. However during the early system design stages, implemen- tation can be difficult depending chiefly on the originality of the software package. ° Ease of operation: The ease of operation is another strong point of the service bureau approach. The service bureau usually develops data input records in machine readable form, processes all data, and prepares all finished reports. This approach obviates the need for computer operators at the treatment center. ° Cost-effectiveness: The fixed costs of a service bureau are usually nominal being limited to a small monthly hookup fee. (Conversely, the installment fee can be charged at the beginning of the contract in a lump sum.) Service bureaus will usually require a contract for a minimum period of 1 year to cover the variable costs associated with one or more of the following: 1. Number of cards punched 2. Amount of data entered into the system 3. Number of transactions processed 4. Number of documents produced Because service bureaus compute charges in a large variety of ways, the treatment center should obtain very detailed estimates for all types of charges possible lest it be saddled with unanticipated "hidden" costs. Manual and Machine-Assisted Systems: Traditional but Reliable At the simplest end of the data-processing continuum lie manual systems. Manual data-processing systems--like their automated counterparts--must complete the five steps in the data-processing cycle: data origination, data entry, data manipulation, data output, and data storage. Certain of these steps can be assisted by simple business machines or techniques that greatly enhance the efficiency of manual systems at a reasonable cost: ° Data Entry: There are two ways to economize in the data entry step. The first involves a manual one-write system; the second is an automatic one-write system. These methods record data in different records simultaneously, with a single 160 entry. Usually, one-write systems (both manual and auto- mated) are associated with accounting functions, but they can be used with equal effectiveness for client recordkeeping applications. ° Data Manipulation: Data manipulation in a manual system can be aided by calculators, accounting machines, and special filing systems that speed retrieval and sorting activities. ° Data Storage: Manual files can be organized in a number of ways that increase access and retrievability, such as color coding, tickler files, cross-files, mail-sort files, and others. As the advent of computer-supported data-processing systems in small organizations becomes more likely, these organizations tend to overlook the substantial advantages of manual or machine- assisted information systems. Manual systems are as available as are competent clerks; they are as reliable as those clerks are competent. They are extremely easy to operate (if somewhat tedious) insofar as they generally require only clerical skills. And most important, manual systems are cost-effective for a great majority of small service organizations: Treatment centers can harness nonscheduled time of onboard staff members to run the system; they need make little if any capital investments; and they will incur only nominal variable expenses. Despite these compelling advantages, there are significant short- comings to manual systems. Turnaround time is notoriously sluggish. Although very flexible during initial periods of growth, manual systems can not be expanded appreciably without becoming inaccurate, inefficient, and generally overwhelming. Also users of manual systems (more so than users of automated systems) tend to overlook the need for documentation--an oversight that penalizes the treatment program (1) during audits and fiscal reviews, (2) when stafff turnover creates the need for retraining, and (3) if it hopes to conduct longitudinal analysis of the data generated by the system. WHAT TASKS ARE REQUIRED TO SELECT A DATA-PROCESSING ALTERNATIVES? ° "The computer firm didn't give us what we wanted!" ° "The staff simply wasn't prepared to take over where the computer firm left off. We are going to have to hire a programer instead of a desperately needed additional coun- selor." ® "The machine was obsolete by the end of the first year." lol ° "The data are so riddled with errors that I still have to rely on intuition to make management decisions." These common complaints are not only the cry of drug treatment agencies but of every type of organization that leaps into data processing without a careful analysis of its own information needs and the many alternatives available for satisfying those needs. This final section of the chapter lays out a logical plan by which treatment centers can launch a data-processing project. Figure 8 presents the sequence of these tasks. It should be noted above all that nothing can substitute for deliberate cautious planning and analysis by the treatment center manager. Conduct an Informal Survey of Experiences and Alternatives The drug treatment center should not begin its investigation of data processing alone or at ground zero. Many similar service organizations have gone down the data-processing road and there are eager vendor "guides" to assist the treatment program in its journey. The treatment center manager should take advantage of the free advice available from a number of sources. ° Other service organizations: Organizations with successful data-processing histories are usually generous with the details of that success. And the victims of data-processing "horror stories" can be counted on to recount the events that gave them trouble. ° Vendors: Vendors offer a range of presale services and gimmicks from which the objective treatment center manager can profit. Being careful not to make a premature commit- ment, the manager can participate in seminars, receive instructive literature, obtain free cost estimates, and get assistance in conducting an information needs assessment. ° Trade associations, Single State Agencies, parent organizations, and funding agencies: These organizations all have a vested interest in the fiscal and administrative integrity of the treat- ment program and so frequently offer technical assistance, information system guidelines, model systems, workshops, and other services that may assist the treatment center in its preliminary search for a data-processing alternative. ° Public accountants: Almost every treatment center engages an accountant (at least on a part-time basis); this resource-- who is intimately familiar with the accounting information needs of the treatment program and well versed in several alternative accounting system approaches--can usually provide a valuable source of guidance to the treatment center. ° Objective consultants: Occasionally the treatment center may wish to employ the services of an information system consultant. Managers who seek such assistance should be certain that 162 £91 Task 1: Informal survey Task 4: Develop detailed systems specifications Task 2: Select a team Step 1: Define systems framework Task 3: Determine systems objectives Task 5: Step 2: Complete a systems specification format Solicit and evaluate vendor bids FIGURE 8.—Task sequence the consultant is familiar with the drug abuse environment and try to make sure that s/he has no vested interest in one particular data-processing alternative or system. Select a Data-Processing Team One of the most important initial planning steps is to select the data-processing team. The word "team" is used deliberately: the project will probably span several years, create numerous political and technical hurdles, and influence every functional area of the treatment center. Thus the project team should be comprised of carefully selected individuals. The team leader should be an individual with a systems orientation, a broad familiarity with the management requirements of the treatment center, the authority to complete the assignment, and have the full backing of top center management. Team members should be drawn from every depart- ment or discipline within the treatment center to assure a multi- dimensional focus of the developing system and to help "sell" the system throughout the treatment center. Determine System Objectives Once a firm commitment to the data-processing system project has been made, the treatment center must develop specific objectives and priorities for the project: ° Objectives state clearly what the treatment center intends to accomplish by implementing the system. Wherever possible, the objectives should be stated in measurable terms so that (1) there can be no disagreement about the "meaning" of the objective, and (2) so that progress toward the objective can be monitored. ° Priorities acknowledge that there are limited resources avail- able to implement the objectives and that ultimately some objectives are more important than others. Ranking or clustering the objectives according to their importance, although difficult to achieve in practice, will further communi- cate the overall intent of the data-processing project. The best approach to developing the systems objectives is the "question and answer" technique. During a meeting of the project team, one should attempt to reach agreement on answers to questions such as: ° What can the system do for us? I. How will it support the service delivery process? 2. How does it provide information for evaluating that process? 164 3. How will it facilitate billing third party payers and clients? 4. How can it help us improve counselor productivity and reduce wasted time? 5. How can it guide the resource allocation process? ° How much can we spend for the system? 1, What percentage of the total budget can be committed to the project? 2. How many additional specialized people can we afford? 3. What investment can be made of management time? 4, What will be the ongoing costs? 5. What will be the cost savings? ® What are the key considerations in evaluating alternative systems? What is the priority among these considerations? 1. Reliability 2. Availability 3. Confidentiality 4, Turnaround time This meeting should produce a list of systems objectives that will form the basis for systems development and communication among all players. These objectives are not the final word: they will be refined or changed during the remaining tasks. Develop Detailed System Spacifications The previous section stressed the importance of evaluating data- processing alternatives as they relate to treatment center needs. To do this the treatment center manager must understand the nature of the evaluative task. For example, assume that the program is faced with three alternatives: Systems A, B, and C. The first impulse is to compare A to B, B to C, and A to C to determine the "best" system. But this approach misses an impor- tant point: The evaluation should be made not between the alternative systems per se--but between X (the treatment center's own definition of system requirements) and A, X and B, and X and C. In other words, the evaluator must have a clear under- standing of his/her own needs. S/he must have information about the alternative systems and must also have a firm and comprehen- sive understanding of the treatment center's information needs. These needs should be documented in a format that facilitates 165 communication between the treatment center and the system vendor. The process by which the information needs are cataloged is called system specification.® A system specification ("spec") is defined as follows: The activities to be supported by the MIS, the outputs required to support those activities, and the processes necessary to produce those outputs. In short the system specifications define the treatment center's needs at each step in the data-processing cycle. They are the cornerstone of the MIS and must be very detailed and compre- hensive. A METHODOLOGY FOR DETERMINING INFORMATION NEEDS AND FORMULATING SPECIFICATIONS Broadly Define the System Framework It may seem axiomatic but the first step is to define the manage- ment information system. There are as many definitions for MIS as there are users of MIS. Consequently, major problems often arise from basic disagreements concerning the parameters and ingredients of the MIS. Does client management and quality assurance impact on the MIS? Is accounting part of the MIS? Will the MIS perform evaluation functions? These and other basic questions must be answered explicitly prior to selecting a process- ing alternative. One way of defining the MIS involves the concept of system "modules" introduced in chapter 7. Each system module is a self-contained minisystem or subsystem. Like building blocks, the modules combine to form a larger system. In applying the module concept to drug abuse treatment centers, one approach is to develop modules around the major management decisionmaking areas: planning and budgeting; client treatment; financial manage- ment and accounting; and evaluation. Each treatment center can define the MIS to include, exclude, or reorganize these functional modules. As an illustration of the amount of detail required in an MIS definition, the following functions may be included: The Planning and Budgeting Module The planning and budgeting module provides forms for and describes how a program can: ‘System specification as used here is not to be confused with software program specifications that are utilized by computer analysts. 166 ° Identify appropriate goals and objectives and formulate a program plan. ° Use the program plan and historical data from the other MIS modules to develop a quarterly or monthly revenue/expense budget. ° Use this budget to meet external reporting requirements (e.g., to NIDA). ° Use the budget throughout the year to monitor the financial status of the treatment center. ° Identify unit costs for each type of service provided by the program. The module stresses that detailed planning and budgeting be done once a year but that the plan and the budget themselves become important tools for program management throughout the year. The Financial Management and Accounting Module As the funds available for drug abuse treatment become more limited, a premium will be placed on programs that can demonstrate tight internal control and efficiency with monies. The financial management and accounting module is comprised of four functions: ° Accounts receivable management ° Billing ° Payroll ° Voucher processing (i.e., paying bills) These four functions are carried out in a way that facilitates cost accounting or cost allocation and allows management to identify costs and revenues according to the type of activity with which they are associated. The Client-Treatment Module The client-treatment module is comprised primarily of a set of input forms in the client record. These forms relate to the four submodules of the client-treatment cycle: ° Admission ° Treatment planning ° Service delivery and progress evaluation 167 ° Discharge This is the most critical module in the MIS because the information collected throughout the client-treatment cycle is used throughout the MIS for such purposes as staffing, billing, reporting to NIDA, planning and budgeting, and evaluation. The Report and Evaluation Module The evaluation/monitoring module does not involve the collection of any additional data but, rather, tabulation and analysis of data collected by the other three modules. Three types of analysis are performed by this module: ° Evaluation of individual client progress; ° Financial analysis, e.g., comparison of actual to planned expenditures and ratio analysis; ° Program analysis of two types: Assessment of the success in meeting program goals and objectives; Review of routine program indicators, such as staff-to-client ratio, turnover of client population, and percentage of successful discharges. The evaluation/monitoring module discusses how information col- lected by the other modules can be summarized, tabulated, and displayed to be of most valuable assistance in the decisionmaking processes. Figure 9 shows the relationships among the four MIS modules outlined above. Complete System Specification Worksheet for Each Module A system doesn't do what you want . . . it does what you tell it. So goes the cliche. Figure 10 illustrates a practical and concise format for telling the system what to do. This format will enable the treatment center manager to specify MIS requirements by module in an organized manner. The elements in the worksheet relate directly to the steps in the data-processing cycle. For example, the specification sheet for the client treatment module would be approached by first defining the functions that the module is to support: admission, treatment planning, service scheduling and delivery, progress evaluation, and discharge. The output for the system (i.e., the information required to 168 691 Planning budgeting module Report and evaluation module Client treatment module Financial module RELATIONSHIPS AMONG THE 4 MIS MODULES: EXAMPLE Review Data bank iannad/ on clients > p » for future actual planning Data bank P —» CODA on client report el characteristics Jo me: sr sss ms. sss on: ss om: ame rd — — — —————————— = —— a ————— Client admitted Staff planning schedule Accounts receivable ledger card FIGURE 9.—Functions of the 4 MIS modules 0LT a | — | Client | Utilization encounter | | tally MY Origination Input N\ Utilization report Storage Ne Function Origin of form us, vs Prosssing Output Storage Dimensions 1. Admission Intake/ Counselor Admission Name Sort by age and Age and sex Retain admission | 500 admissions Interview Form Age sex and key breakdown by statistics for per year Sex drug problem drug problem 2 years 10 percent of Date of Calculate percent | New admissions Retain client clients will be admission by age and sex by month and data for 2 years billed upon Drug problem groups drug problem after discharge admission Modality Calculate Medic- Admissions by Counselor aid charge for modality, by number admission month Medicaid Admissions by number intake counselor ° Bill to Medicaid ° ® 2. Client treatment Physical examina- re] Physician Encounter J cn | Forme TT FIGURE 10.—Specification worksheet—client treatment module / perform the decisionmaking tasks of each function) correspond to the information requirements as developed in chapter 7. From there the manager (and the project team) can fill out the rest of the form. This is basically a trial-and-error, advise-and-consent process. Throughout many items will be entered and then dis- carded (and possibly reentered!) as the treatment center staff struggle to define their management processes. But the struggle will pay off. There is no substitute for careful system planning and this set of specifications provides a good foundation for the entire project. Solicit and Evaluate Vendor Bids "Let the buyer beware" is not a particularly comforting credo for drug abuse treatment centers in the new world of automated MIS but it is a necessary one. Because treatment centers can, we believe, profit from a dose of old-fashioned competition in evaluat- ing and selecting among the data-processing alternatives, a well- structured specific Request for Proposal (RFP) which includes the systems objectives and specification worksheets can assist the treatment center to communicate its systems needs to prospective firms. It can also provide those firms with a helpful framework within which to describe their systems. In summary, this chapter has provided a look into the world of data processing from the point of view of the drug treatment program manager. We hope that he or she will now be able to analyze additional information on his/her own. Because the field is so technical, the potential for costly mistakes is great. Hence we have stressed planning and obtaining objective expert help. To us the advantages of automation are so compelling that we feel planned risk is clearly justified. AUTHORS HERBERT M. BIRCH, JR. KERRY G. TREASURE Birch and Davis Associates, Inc. Silver Spring, Maryland REFERENCE Awad, E.M. Automatic Data Processing: Principles and Procedures. 2nd ed. Englewood Cliffs, N.J.: Prentice-Hall, 1970. 171 7. Fundamental Considerations in Developing an MIS Willie Davis and Kerry G. Treasure When the words "system" and "management information" come to mind, they often conjure up thoughts about computers, sterile and tedious statistics, and endless paperwork and procedures--none of which have anything to do with the client or with providing services to him/her. But management information systems need not fit this undesirable mold. The purpose of this chapter is to pre- sent an overview of a Management Information System (MIS) in which: ° The client is the key ingredient. ° Relevant client data, statistics, and financial data are included. ° Paperwork and procedures are specifically designed for efficiency and compatibility with the routine procedures of service delivery. WHY IS MANAGEMENT INFORMATION NECESSARY? Managing any business--and a drug abuse treatment center is certainly a business of sorts--requires that the program manager make routine decisions that affect the operations of the center. In fact a prevailing definition of management holds that ". . . the art of management is simply the process of making decisions." Assuming that this is true, it is reasonable to expect that decisions be made on the basis of informed judgment: ° The manager provides the judgment ® The MIS helps him/her stay informed At the drug abuse treatment program level, management information is necessary for four purposes: treatment, financial management, evaluation and monitoring, and planning and budgeting. 172 Figure 1 portrays the management cycle and purposes for which information is required. Although these purposes are different, they may only require different ways of looking at the same information. Essentially all the information required as a basis for decisionmaking centers on seven basic management questions: ° Who? ° Provided what? ° To whom? ° When? ° Where? ® What happened? ° How much did it earn/cost? If the data needs are similar for most management and clinical users, it follows that the information can be collected once in a format that will satisfy all intended users. An efficient system will collect as little information as possible and use it for as many management purposes as possible. WHAT IS A MODEL MANAGEMENT INFORMATION SYSTEM? A model management information system is a set of data, input forms, processing procedures, and reports that support decision- making in all phases of the management cycle. This chapter will illustrate the applicability of a model MIS to a drug treatment program by using a modular approach. Planning Module A planning module aims at the following objectives: ° Identify appropriate goals and formulate a program plan. ° Use the program plan and historical data from the other MIS modules to develop a quarterly or monthly revenue/expense budget. ° Use this budget to meet external reporting requirements (e.g., to NIDA). 173 LI Planning Budgeting e Develop goals and objectives Pl. Develop service programs —P . Develop rate structure e Evaluate alternatives e Assess funding sources e Develop operating budgets e Negotiate funds Initiate operations Participating «p> providers Interaction with environment Client population Evaluate and adjust to current environment Summarize and report < on operations Account for and record interaction FIGURE 1.—Management cycle ° Use the budget throughout the year to monitor the financial status of the treatment center. ° Identify unit costs for each type of service provided by the program. Although the emphasis is usually on delayed planning once a year, a well-designed planning module will function as an effective management tool throughout the year. The Client Treatment Module The client treatment module is seen as comprised primarily of a set of input forms in the client record. These forms relate to the four phases of the client treatment cycle: ° Admission ® Treatment planning ° Service delivery and progress evaluation ° Discharge Information collected throughout the client treatment cycle can be used for such purposes as staffing, billing, reporting to NIDA, planning and budgeting, and evaluation. The Financial Management and Accounting Module As public funds available for drug abuse treatment become more limited, programs will be increasingly required to demonstrate tight internal controls and efficiency with money. The four functions of the financial management and accounting module can be carried out in a way that facilitates cost accounting or cost allocation and allows management to identify costs and revenues according to the type of activity with which they are associated: ° Accounts receivable management Ki Billing ° Payroll ° Voucher processing (i.e., paying bills) The Report and Evaluation Module The evaluation/monitoring module does not necessarily involve the collection of additional data but rather tabulates and analyzes data collected by the other three modules. Three types of analysis can be performed by this module: 175 ° Evaluation of individual client progress ° Financial anaylsis, e.g., comparison of actual to planned expenditures, and ratio analysis ° Program analysis of two types: 1. Assessment of success in meeting program goals and objectives. 2. Review of routine program indicators such as staff-to- client ratio, turnover of client population, and percen- tage of successful discharges. The evaluation/monitoring module demonstrates how information collected by the other modules can be summarized, tabulated, and displayed to be of greatest assistance in the decisionmaking processes. Integration of the Four Modules: An lilustration We now show how two of the most common occurrences in the treatment center generate data that are employed by each module within the MIS. Client Admission to the Program A client is admitted for treatment at a methadone program: ° The client management module: The admission form is com- pleted following a decision on the appropriateness of admitting the client. This same information aggregated over many patients and a significant period of time is used to plan staffing patterns and levels within the client management function. ° The financial management and accounting module: The information gathered on admission is also used by the financial management and accounting module to establish an accounts receivable ledger card. ° The report and evaluation module: The information gathered on the admission form is used to complete the CODAP report for submission to NIDA; it then becomes part of the data bank for future evaluation of: 1. Changes in the client's status from time of admission. 2. The characteristics of the client population. Information about this particular client will be combined with similar information about all clients to produce aggregate reports. 176 ° The planning and budgeting module: The fact that an admission was made will be compared to the planned timetable for admissions to see if the program is on target for the year. And the information in the data bank will be used as the basis for planning budget levels. Client Receives Treatment The most important activity in the treatment center is the providing of services. This is also the activity that has the most significant impact on the MIS. The discussion below illustrates how the collection of information at a single point in the process interfaces with the entire MIS. ° The client treatment module: When a client receives treat- ment, a record is made of the encounter; it describes who did what to whom, when, and where. On the basis of this information the staff can provide a record of client progress and response to treatment, decide on the course of future treatment, and schedule working hours and assignments. ° The financial management and accounting module: This same information serves as a notice to the financial management and accounting module that: 1. A bill should be prepared for providing "x" service to a particular client. 2; A cost has been incurred for providing "x" service to a particular client. ° The report and evaluation module: The information in the progress notes allows clinic supervisors to make an objective assessment of the client's progress and the clinic's pattern of service delivery. When aggregated, this information yields evaluation reports. The information can be used to analyze productivity, utilization, outcome of the treatment process, and other aspects of the program. ® The planning and budgeting module: Aggregate data from the encounters will help program administrators determine if services are being provided in the quantities planned. This information will become the data base on which future program plans and budgets are based. Information collected in the process of serving clients is simply "borrowed" to provide the input to the other MIS modules. 177 HOW A MANAGEMENT INFORMATION SYSTEM IS DESIGNED An MIS is usually developed in four steps: ° Identify management information needs. ° Develop a consistent and compatible set of definitions for the collection of data. ° Design forms for data input and display. ° Determine the most efficient flow of information. The purpose of the MIS is to help manage a program. It follows, then, that the first step in developing the MIS is to ask, "What information will help me operate my program more efficiently and most effectively?" This question is based on the most elementary principle of information science: Do not collect information unless you can anticipate a use for it. In the case of a small or understaffed drug abuse treatment center, a corollary is: No matter how "useful" the data may potentially be, do not collect it unless you will have the capacity to proc- ess it on a timely and routine basis. Much of the data required to operate the average drug abuse treatment program are specifiable and consequently can be incor- porated into a model MIS. Included is information about client characteristics, staff activities, and cost/revenues. In addition each program has unique information needs to which the MIS must respond. Information needs are determined by: ° The size and complexity of the treatment center: In a very large program, the need for formal communications (e.g., between counselor and supervisor) will be greater than in a small program where word of mouth will usually take care of most routine communciation. “ The scope and nature of quality review and client progress assessment: If cases are routinely monitored by other than the principal therapist, client records must contain enough information to allow for independent and objective assessment. ° Sophistication of the planning system: If planning and budgeting are carried out in the fashion alluded to earlier in this chapter, a historical data base will be needed that describes patterns of utilization, size of the client load, staff productivity, etc. 178 ® The type of program evaluation and monitoring: Programs vary considerably on the amount and type of evaluation and monitoring they perform. There are at least four possibilities that should be considered in the designing of the MIS: 1. Evaluation of individual clients. 2. Routine program monitoring through program indicators such as percentage of positive urines, staff-to-caseload ratios, and number of successful discharges. 3. Progress toward treatment goals. 4. Financial analysis such as ratio and variance analysis. Identifying information requirements is the most difficult part of designing an MIS. In attempting to develop a list of information needs, several fundamental questions must be answered: ° What are the major activities or decision areas for which my staff and I are responsible? ° What information is necessary to conduct these activities or make decisions? ° How will this information help me conduct these activities or make these decisions? The elements listed should be carefully selected to insure consist- ency and compatibility. ° Consistency: Many of the data elements on the list will be used to serve more than one purpose. For example, "number and type of services provided" will be used for planning, determining unit costs, preparing bills, conducting produc- tivity studies, and assessing the quality of care. The program manager must be sure to collect the information in such a format that it will meet each of these needs. ° Compatibility: Often data elements can be combined with each other to produce additional data and summary reports. For instance the "number and type of services provided by personnel type" may be combined with the "salary levels by personnel type" to determine the direct labor cost per type of service. There must be uniform or compatible definitions for "personnel type" if these two types of information are to be merged. On a day-to-day basis, each of the four modules operates more or less independently. Ultimately, however, they all feed into the same decisionmaking process and so comprise an integrated MIS. The diverse information needs must now be organized into an information classification structure, a compatible and consistent set of data definitions. 179 The following sections describe how the Information Classification Structure is developed and how it is used by each module of the MIS. Three tasks must be completed to develop an ICS: ° Identify the most significant categories of information require- ments. ° Define a list of mutually exclusive and exhaustive elements within each category. ° Assign a number or code to some data elements. Most information required for decisionmaking is in response to seven basic management questions: ° Who? ° Performed what? ° For whom? ° Where? ° When? ° What happened? ° How much did it cost/earn? Virtually every type of information necessary to make routine management decisions answers one or more of these questions. Therefore data categories can be identified. The ICS categories that respond to the questions in the model MIS are shown in figure 2. The figure also shows the phase or module for which this category of information is especially important. Because the four modules of the MIS act independently, yet are concerned with many of the same data or information types, it is necessary that each module speak the same language. The Informa- tion Classification System provides a common set of choices for each of the seven basic management questions (or categories of data). Thus the ICS assures that the data collected by each module is consistent and compatible with the data collected by the other modules. In selecting the data elements for each of the categories, it is important that the elements be mutually exclusive and exhaustive. The following sections describe the considerations that should be made in selecting the data elements for each ICS category. Resource--Staff This category simply lists all of the staff positions currently existing (or planned) at the treatment program. This list should be as short as possible but make relevant distinctions between 180 181 Management cycle phase MIS module Management ICS Plan and Client Financial question category budget management | management Evaluation Who? Resource--staff X X X X Service X X X X Provided what? (billing) Resource--nonstaff X X Client identifier X X For whom? {billing Demographic Where? Cost center (location) X X X When? Date (shift) X X X What happened? N/A How much did Fund X X X it cost/earn? FIGURE 2.-Basic management questions staff types. For example if a variety of people perform counseling, the ICS could simply identify them all as counselors. Or it could distinguish among them according to: « Their level of responsibility: senior counselor, junior coun- selor. ° Their level of education: degreed counselor, paracounselor. ° Their area of expertise: family counselor, vocational coun- selor. The amount of detail selected is wholly dependent upon program needs for detailed management information. Resources--Other This data category lists the types of supplies and other goods, materials, and services routinely consumed by program operations. The list of data elements for this category can be prepared by reviewing the budget and the accounting records (checkbooks) for past time periods and simply noting the expense items. The list should be reviewed critically to determine whether tabs should be kept on each item individually or whether some items can be grouped (e.g., pens, paper, and staples grouped as office sup- plies). Service It is important for (1) billing, (2) recording staff activity, and (3) evaluation that everyone in the treatment program share a common name for each type of service offered. The elements in this category should reflect the service name in current use in the program or the names used by the major third-party payers. Client Identification A single number or name must be used consistently for each client to assure that his or her various records are completed properly. In addition to client identifiers, certain information should be collected about each client such as age, sex, legal history, employment status, and so on. Because this is the most voluminous category and also the one from which many reports are prepared, special care should be taken in listing its data elements. Particular attention should be paid to the standard external reports (e.g., CODAP) that must be submitted. Cost Center Frequently, managers need to plan, budget, manage, and monitor discrete sections of their programs on an individual basis. These 182 discrete sections may correspond to locations, modalities, floors, environments, or some other variable, and are called "cost centers." These centers should embrace a portion of the program that incurs expenses more or less independently of the rest of the program. Usually these centers are under the direction of a single supervisor or manager. Time/Date The time and date are self-explanatory. Particular attention should be paid to this data category when: ° Third-party billings are prepared. All records must bear the same data and thus provide an audit trail to substantiate claims. ° The program operates on shifts, and management requires accounting or client data broken down on this basis. Fund Often funding organizations require that accounting and utilization records describe how the money they provided was spent. The data categories in the ICS should reflect the various funding sources for your program. General Ledger The general ledger (which corresponds to the Chart of Accounts) is the basic control mechanism for recording financial transactions and probably will not vary significantly from treatment center to treatment center. In sum, the data elements in these categories define almost all of the activities that will take place in a treatment program. They are used by all four modules to simplify the data input process and to assure that all data collected by the MIS are consistently defined and compatible from module to module. Often, it is useful to assign a short code or number to the most frequently used data elements for two reasons: ° Codes save time. If two digit numbers are assigned to each cost center, it is easier and faster to write "OI" than to write "Outpatient Drug Free Center, Main Street Location". ° Codes facilitate the combination of data elements from the various categories. Figure 3 shows that codes can be com- bined as long as there is an agreed-upon order for each category. For example the General Ledger Code is always followed by the Fund Code, the Cost Center Code, and the Resource Code, in that order. The second half of the 183 ¥81 710 01 01 02 General ledger Fund Cost center Resource 700 Expenses 01 NIDA grant C1 Outpatient 01 Physician 710 Salary expense 02 Medicaid 02 Residential 02 Counselor ° ° ° ° ° ° ® ° ° ° ° ° ° ° ° ° Example: Salaries for the counselor are paid from the NIDA grant; the counselor worked for the outpatient program. The codes for this transaction would be: 710-01-01-02 These code numbers correspond to the data input forms used in the other modules of the MIS. When the counselor completed his/her staff activity log, he/she noted the correct position (resource) and cost center. FIGURE 3.—Computer code combination figure illustrates how this arrangement facilitates the recording of accounting information. Not all items in the ICS will be coded of course; only when coding makes implementation and operation of the MIS easier and more efficient should it be employed. Once data elements have been identified and defined, the appro- priate format for capture and display of each data element must be determined. Forms design is challenging and often considered the most complicated part of developing a system. Forms simply provide a list of questions that elicit the required information in an organized (consistent and compatible) fashion. Hence what information will be gathered by which form must be thought out carefully in advance. Here are some guidelines: Keep the number of forms to a minimum. Do not collect information on two forms if one will suffice. Carbon paper or self-duplicating sheets should be used to obviate repeated recording of the same information to be used by different modules. Preprint most of the questions (and possibly the limited set of responses from the ICS) on the forms. This will: 1. 2. Assure that the important information is not omitted. Reduce the amount of writing that the staff person will have to do. Provide information in a standard format for easy sum- mary and tabulation. Develop the summary report formats at the same time the data-collection formats are developed: 1. Some decisions do not require information about indi- vidual clients. Because the information aiding such decisions can be accumulated, develop the summary tables to parallel the data-collection forms. Also if you find that some data are not used for any summary purpose, this may be a clue that the data are not particularly valuable. Reevaluate these data elements to be sure that they do indeed have a use. Provide some mechanism for controlling the distribution and collection of the forms: 1. You may want to preprint sequential numbers on some forms (similar to the preprinted numbers in your check- book) so as to call attention to missing forms and aid in tracking them down. 185 2. Keeping some data in bound jounals and/or logs will keep pages from becoming lost. 3. Cross-references from one document to another or from one file to another will help you retrace steps if infor- mation must be verified or updated. ® Plan ahead if you anticipate converting to an automated system. Design the forms even for a completely manual system so that they can be keypunched or machine read in the future. If you do this, you will not have to retrain the service delivery staff when you make the change to automa- tion. Indeed, it might be noted that a large proportion of the gain in automating a data system is in the clarity of thinking required, e.g., careful coding, forms design, attention to reporting requirements, etc., and has nothing to do with electronics. ° Pretest the form before implementing it throughout your treatment center. No matter how logical and complete the form may seem to you, the users of the form will be able to find something that can be improved. Remember, it is the staff that will be using it. If the forms are pretested, you will avoid confusion and possibly a costly printing bill. The forms when developed and pretested form the backbone of the management information system. Decisions must be made on who will handle each form and when each will be used. These decisions will result in a "data flow" in which the following con- siderations are relevant: ° The point at which the information is required. Obviously data must be collected before it is required for decisionmak- ing. ° The time during which it is easiest to collect the information. For example clients may resist providing sensitive information during their first visit to the center. Unless the data are absolutely needed then, delay the data collection until the client is more comfortable. ° The time and skill required to collect the data. Generally most data, particularly of the demographic kind, can be adequately collected by clerical personnel. When the data flow has been determined, it should be depicted on a simple flow chart. Figure 4 presents simple flow chart symbols and an example of their use. This introduction has spelled out the four major steps in designing a management information system: ° Identifying information needs. 186 ® Action ® Procedure ° Form ® Worksheet ® File ® Decision ® Alternative > ® Information flow FIGURE 4—Example of the use of simple flow-chart symbols 187 ° Organizing information needs into a consistent and compatible framework called the Information Classification System (ICS). ° Designing the forms for data collection, tabulation, and display. ° Determining the most efficient data flow. THE DEVELOPMENT OF A MANAGEMENT INFORMATION SYSTEM IN A DRUG ABUSE TREATMENT PROGRAM Planning and Budgeting Module Figure 5 emphasizes that the first (and often shortchanged) step in the planning process is the development of a concise and specific statement of the goals and objectives of the treatment center. Objectives should be stated in quantifiable and measurable terms: ° Quantifiable: Stated in terms of numbers or percentages. For example, an objective might be to assure that 25 percent of the clients obtain employment and hold a job for at least 6 months. In this case, the 25 percent and the 6 months are the "quantified" terms. ° Measurable: Stated in terms about which information can be collected and analyzed. For example, the objective might be to increase client "happiness" by 25 percent. This is not stated in measurable terms because happiness as such cannot be assigned a number on any scale of measurement. The activities should be stated as clearly and specifically as possible. There are three planning assumptions implicit in the statement of activities: ° The most desirable or effective services. Selection of the service mix is dependent upon the philosophy of the treatment program and its staff, Federal and State regulations, research findings, and prior experience. ° The size of the caseload and the portion of the caseload that will be provided each type of service. This is based on a perception of the number of "potential" clients in the commu- nity (i.e., the demand for services) as well as the fiscal and physical constraints on the program. ° The frequency of service or the expected typical treatment plan. Although each client's treatment is planned on an individual basis, it is necessary for planning purposes to determine the average frequency of use for each service type. 188 681 Step 6: Compare expenses to reve- nues and (1) initiate opera- tions or (2) revise Step 1: Establish program goals, ob- jectives, and activities Step 5: Estimate the revenue avail- able for providing services Step 2: Determine staff resources necessary to reach goals, ob- jectives, and activities Step 4: Estimate the cost per unit of service and establish fee scales Step 3: Determine the amounts and types of ‘‘other’” resources required NOTE: —Regardless of the activity or program involved, planning follows a basic “cycle” FIGURE 5.—Planning cycle Even the most modest treatment program has administrative needs in order to make service delivery possible. Many programs are in addition concerned with research, public relations, and staff development, to name a few. Because of the wide variety of enterprises that may take place in a drug treatment center, it is difficult to develop a standard terminology that will fit all of them. In developing support and ancillary objectives, the following questions should be asked: ° What tasks are to be accomplished? ° Who is responsible for accomplishing each task? ° When is the task to be accomplished? ° Where is the task to be accomplished? The answers provide an adequate basis upon which to estimate resource requirements. The resource requirements, in turn, will be used to prepare the budget. Once the objectives and activities have been defined, the program plan can be completed by estimat- ing the resources required. In a labor-intensive industry such as drug abuse treatment, personnel resources are by far the most significant. Since staff requirements comprise so great a proportion of all resources for program operation, they must be planned carefully. Staff resources fall into two categories: service delivery staff and support and ancillary staff (to be defined later). In addition, nonstaff resources must be considered (e.g., rent, supplies, utilities). Because drug abuse treatment is a relatively new field--and an innovative one at that--there are few standards on which to base ratios for treatment programs. However, unless there is a drastic change in the goals and objectives of the orga- nization or some evidence that the planning assumptions developed through the process to be described are erroneous, many of the ratios (e.g., number of counselors per 100 clients) can be used in subsequent planning periods. Most service delivery staff needs are "variable," that is, the staff requirements are directly proportionate to the number and type of activities to be performed. A framework for determining the relationships between size of caseload and types of services and the number and qualifications of staff required must be established first. The service utilization should be developed for each type of service or activity provided. The frequency of service provision and number of clients obtaining each type of service are taken into consideration. This estimate is made on the basis of previous experience tempered with new trends or research findings. It is critical that the estimate reflect attainable levels of performance 190 because all subsequent steps in the planning and budgeting process are based on this figure; an error in this number will be echoed throughout the planning and budgeting assumptions. Once we know how many minutes of service delivery each staff type will provide to carry out each activity and also the amount of service delivery time that can be expected from each full-time employee, the total number of employees needed to accomplish the activity can be determined by dividing the former by the latter. This figure represents the total full-time equivalents necessary for one staff category. The number of full-time equivalents required will rarely be an even number. Most likely, the number of full-time equivalents will be a fraction such as 1.2 RNs or 3.8 counselors. Comparing summary sheets for each cost center may indicate areas in which staff members can be shared between cost centers, thereby reduc- ing the requirements for part-time employees. For example, if 1.8 RNs are needed by one cost center and 3.2 RNs by another, the program manager can consider hiring 5 RNs, one to be shared by two cost centers subject, of course, to scheduling and logistical constraints. Because of the diversity and unpredictability of many adminis- trative and support activities, it is difficult to judge staffing requirements. Nonetheless it is important to establish some means of documenting staffing decisions related to support and ancillary objectives and activities. Documentation will enable programs to review planning assumptions to determine if the original expecta- tions will, in fact, meet the objectives of the program. When the summary has been prepared for each cost center, it may prove useful to accumulate all staff requirements into a single summary for the entire program so that the requirements can be compared to the existing personnel roster and hire or terminate decisions can be made. Recognizing that it may not always be possible to hire the precise number and type of staff "required" or to lay off "overage" staff, the summary worksheet may have a right-hand column to indicate the final staffing decision. Although staff resources comprise the majority of the budget in most drug abuse treatment programs, there are other goods and services essential to execution of the activities and necessary to reach program objectives. These resources fall into two cate- gories. ° Type I: Resources that are related to the overall operation of the program. Things such as rent, utilities, and commu- nications are not planned for each activity but rather reflect the needs of all program objectives and planned activities. ° Type II: Resources that are directly related to a specific objective or activity. For example, urine bottles are needed only for conducting urinalysis. 191 Because there are so many types of resources that may be required, it is difficult to establish an exact formula for estimating all of them. It is possible however to analyze both types according to how they are used: ° Variable resources: These are directly proportionate to some other factor in the treatment program such as number of clients, number of staff, or number of billings. For example: I. Type I: The amount of office supplies required is determined more or less by the size of the staff. The size of the staff is determined not by a single objective or activity but by all objectives and activities combined. 2. Type II: The amount of methadone required is directly proportionate to the number of clients to be treated with methadone. The methadone requirements are directly related to the single activity "providing metha- done treatment." ° Fixed resources: These are needed regardless of the number of clients, the number of billings, or the size of any other factor related to operations. Fixed resources are required simply to open the door of the treatment program. Although they may fluctuate over long periods of time, they are not responsive to day-to-day changes in program operations. For example, when fluctuations occur in the size of the staff or the caseload, the building size remains constant (unless, of course, you move to another building). ° Incremental resources: Incremental or semivariable resources are those that are adequate to sustain a given level of operation but that become insufficient if that level changes even slightly. The best example is equipment. If the client load becomes so great that one refrigerator cannot stock the food, the program must buy a whole refrigerator even if the extra food only takes up half of it. You cannot buy refrigera- tors in "increments" smaller than one. The second aspect of the planning and budget module is the development of a program budget; more specifically, an expense budget. A budget is simply a statement of: ® The costs (expenses) of program resources necessary to meet program objectives and activities. ° The revenue to be earned if program activities are carried out and objectives are reached. The Worksheet for Preparing the Operating Budget: Expenses shown on the following page (figure 6), is the framework for accumulating all the information and assumptions necessary. The following steps will complete the worksheet: 192 WORKSHEET FOR PREPARING THE OPERATING BUDGET: EXPENSES WORKSHEET # COST CENTER PLANNING P RESOURCE ESTIMATION BAS BASE SALARY [WEEK GEN.L. |RES. # | DESCRIPTION PER PERIOD NO. TOTAL SALARIES FRINGE BENEFITS = % OF SALARIES €61 OTHER RESOURCES ESTIMATION BASIS/WORKSHEET # FIGURE 6.— Worksheet for preparing the operating budget This Two ment Can how List each staff resource type and record the appropriate Information Classification System code. These codes are the same as those used in the accounting module and using them will help retrieve accounting data for use in projecting expenditures. Enter the base salary for the upcoming planning period. Place the number of full-time equivalents per period developed earlier onto the Worksheet for Preparing the Operating Budget: Expenses. Be sure to record the Worksheet number to create a source reference for these numbers. To calculate the salary cost per period, multiply the base salary by the number of full-time equivalents and enter the dollar amount into the corresponding period. Calculate the fringe benefits and enter in the appropriate row. List each "other" resource developed earlier. Make note of the ICS codes at this time and record the worksheet number on which the original assumptions are made. For each resource, estimate the anticipated cost. completes the budget by cost center and planning period. tasks remain to complete the expense budget for the treat- program: Sum up the expenses by cost center Total these amounts to arrive at the total expense budget for the entire treatment program you imagine General Motors selling a car and not knowing much the parts and labor cost or the difference in cost between a Vega and a Monte Carlo? It is becoming equally impor- tant that the drug abuse treatment center know the cost of provid- ing each type of service. The objectives of cost finding are: To provide cost information as a basis for establishing billing rates or for evaluating existing rates. To provide information for reports to boards of directors, governmental agencies, and other groups. To provide a basis for negotiating reimbursement rates with third-party payers. To provide data for decisionmaking and program evaluation. A unit cost is the amount of money it takes to provide a service once. Three factors contribute to it: 194 ° Direct labor ° Direct supplies ° Overhead Figure 7 shows the combination of these factors. The unit cost can be determined from the information already accumulated in the program plan and the expense budget and there are numerous ways to compute it. Basically three tasks are required--one task corresponding to each type of cost shown in figure 7. Direct Direct Overhead Total cost labor supplies items per unit 10 minutes Methadone Rent /heat Total cost per unit FIGURE 7.—Factors contributing to unit cost Direct labor costs are defined as salary and fringe-benefit costs that are incurred in order to provide client treatment and thus to meet the service delivery objectives. These costs typically comprise the largest portion of the budget and, consequently, of the unit cost. One must first determine the labor value of each service unit. This requires the assumptions that: ° If 20 percent of staff time is spent on service y, ° Then 20 percent of the salary expense is attributable to service y. 195 In addition, ° If 20 percent of the salary expense is attributable to service y,» and ° If 300 units of service y are provided, ° Then one three-hundredths of the 20 percent of the salary expense is attributable to each unit of service y. Using this assumption as a guide, the percentage of time it takes each staff type to perform each type of service can be calculated, thus enabling the applicable labor cost to be derived. By dividing each direct labor cost by the projected utilization, the direct labor cost per unit of service is produced. ° Food and kitchen equipment; direct costs in providing meal service. ° Methadone and medicine cabinets; direct costs in dispensing methadone. ® Psychological testing supplies; direct costs in conducting the intake interview. The cost of these resources can be reduced to a unit cost by dividing the total resource cost by the anticipated utilization. Overhead costs include at a minimum administrative activities, possibly research activities, followup, staff development, public relations, and other such endeavors. Typically no fees are charged to participants in these support and ancillary activities and so the cost of conducting these activities must be charged to the services--the "revenue generating" activities of the center. Once overhead costs are incurred to support the provision of service, it follows that the service could not be provided without them. Consequently, each service must share responsibility for the overhead costs; that is, each service must absorb some pro- portion of these costs. However, each type of service does not bear an equal relationship to the total overhead costs: some services require more supervision, more supplies, more space, more typing, and so on. The task of the program manager is to select a basis by which each service can be most logically charged or "allocated" its fair share of the overhead. Selection of the allocation basis rests on the following assumption: ° If y services comprises x percent of some variable, @ Then x percent of the overhead should be charged to y service. Overhead costs are distributed to each service category utilizing a number of allocation techniques or bases. The objective in each technique is to determine the logical relationship between each 196 service and the overhead costs. Once this relationship is deter- mined, allocating overhead becomes a simple task of associating percentages of the total overhead cost to each service. The unit overhead cost is then calculated by dividing the projected utiliza- tion into the overhead cost of each service. The total unit cost is calculated by adding each of the three unit costs above together. As with expenses, the revenue budget is based on assumptions about: ° The number of clients the program expects to serve ° The type and frequency of services to be delivered ° The rate of payment available from various sources for providing x service to x clients; the rate of payment must be estimated separately for each of the major payment sources: 1. NIDA grants and contracts 2. State and local funds 3. Third parties and client fees 4. Donations 5. Other sources of revenue The first two assumptions derive directly from the program objec- tives. The third is based upon program expenses and projections. The General Ledger Code and the Fund Code must be used for each source of income thereby linking the revenue generation process to the management account structure. The Federal share of the total operating costs of a drug abuse treatment program will depend on: ° The type of grant or contract ° The number of years of previous funding ° Federal funding policies such as the maintenance of effort strategy and the allowable cost per slot ° The overall size of the Federal appropriation ° The number of approved client slots ® Program ability to meet local match requirements Estimating revenue from State and local governments is done in much the same fashion as estimating Federal support. Programs must have a current and accurate understanding of the amount of money to be distributed by these governments and the mechanisms for making a request. The assumptions made and/or the award 197 documents or letters of agreement should be taken into considera- tion. Each third-party payer must be identified by the program and a specific amount of revenue associated with it. This is best done by identifying the basis upon which reimbursements will be made and calculating how the criteria will be addressed through the program's provision of specific eligible services. The total pro- jected revenue from each source is then added up. Client Management Module The Client Management Module includes all functions within a drug abuse treatment center which either directly or indirectly support the client's treatment at the center. Since these functions include a wide variety of activities, this discussion will not cover them all but, rather, highlight the conceptual framework within which the activities occur. NIDA, recognizing that detail cannot be entirely overlooked, has initiated several efforts geared toward presenting specific activities within the client management process. These efforts include suggested client recordkeeping systems, central intake monographs, confidentiality procedures, and a number of other client management products. A specific listing of such advisory documents can be obtained by contacting the National Clearinghouse for Drug Abuse Information or the appropriate program representative. Examples of activities which are commonly associated with client management consist of scheduling counseling sessions, dispensing methadone, information gathering, completing and processing forms, and collecting urine specimens. In addition to these are hundreds of other tasks which together form the client manage- ment process. The objectives of preparing a model for the client management module of a program's operation are first, to organize these many activities into their most efficient and effective sequen- tial relationships; second, to insure that the information needed by staff and management is produced by the process and is compatible with other program operations; and third, to fulfill minimal standards established by Federal, State, and/or local authorities. Consequently the client management module presented here has been structured to correspond with the broad functions existing within all drug treatment programs. In addition, proc- esses shown are intended to reflect information made available within other modules and felt to be essential functions of client management. There are a number of ways in which the client management functions can be expressed in a management information system. The functions which probably best describe the key client manage- ment operations within all types of drug treatment programs include the client's: 198 ° Entry into the program ° Treatment plan development ° Treatment documentation ° Disposition from the program ° Followup activites Each of these functions is a major topic for discussion in itself. Clients enter drug abuse treatment programs from several different sources either voluntarily, under court order, or through referrals from other agencies. Despite the client's source of entry, all share a common destiny: admission processing. This refers to preadmission activities, admission activities, and activities per- formed by a Central Intake Unit (CIU). These central intake unit activities, however, will be discussed from the perspective of the drug abuse treatment program's interface with the CIU. In other words, specific functions which take place in the CIU will be presented only if they will occur within the treatment center. A description of CIU operations can be obtained from the National Clearinghouse for Drug Abuse Information. In some instances programs will locate clients requiring treatment through the use of outreach activities. In such cases it is essen- tial to incorporate the following procedures: ° Link the outreach activities to the program goals, objectives, and admission criteria. ° Define the target population sought. ° Outline a specific plan to systematically penetrate the maximum stated proportion of the defined target population for specific time periods. ° Monitor results versus plans frequently. With these concepts in mind one should develop an outreach worksheet which will enable the identification of significant mile- stones achieved, problems encountered, adjustments needed, and revisions made. Entry into a drug abuse treatment program is probably a totally new experience for the client; that is, being queried on every aspect of his/her life and lifestyle while at the same time being thrust into a structured setting. For this reason the entry phase of the client management process must be streamlined to lessen redundant activities which would add to the client's anxieties. An example of such a counterproductive and inefficient process is obtaining demographic information both when completing the CODAP form(s) and when gathering routine intake data on program- developed instruments. 199 Whether clients enter the program from a CIU or directly, several admission processing operations are required: ° The conduct of an initial interview to gather needed informa- tion on the client's: 1. Personal history (including legal, psychosocial, employ- ment/vocational, and educational histories). 2+ Medical history. 3. Drug-use history. ° The completion of the required CODAP forms. ° The completion of other needed documents and/or applications for food stamps, Medicaid, private insurance, and public assistance programs. The above operations are required of all NIDA-funded drug abuse treatment programs either as a condition of the Federal funding criteria or in response to the program's need to seek every possible source of additional funding. The provision of a satisfactory treatment regimen for the new client rests in the ability of a drug abuse treatment program to: ° Correctly pinpoint a client's problem(s). ° Determine a course of action for solving the problem(s). ° Provide treatment in accordance with the planned course of action. The concept of treatment planning is thus not complex. Yet in everyday practice it is difficult to follow because of the many distractions involved in providing treatment to clients. Such distractions must be recognized and placed in their proper per- spective if an adequate scope of treatment is to be provided. In developing the formal written treatment plan, program managers must realize the important function the document serves. The success or failure of each client's treatment rests upon its ade- quacy. It is indeed the focal point of the documentation of client treatment. It has three main components: ° Statement of problem(s) ° Statement of goal(s) ° Statement of proposed therapy The Treatment Plan Form designed for the Client Record System also notes the date problems are identified, target dates for accomplishing goals, and dates problems are solved. A treatment 200 plan may be initiated any time a problem is identified. The goal and therapy for that problem are entered as they are formulated. The treatment plan may be temporary, and initiated after the initial interview and medical examination on the first day. On the other hand, it may be a complete treatment plan formulated at the conclusion of the appropriate in-depth interviews. It may even evolve into a different "complete" treatment plan as new problems are revealed. The treatment plan is updated from the ongoing progress notes as new problems emerge and old problems are clarified and resolved. It therefore becomes the single most important section in the Client Record, providing current documentary evidence that: ° Problems are being identified ® Goals are being established ° Treatment is being formulated ° Progress is being achieved In sum, the treatment plan is a constant reference point--a secur- ity against losing track of problems and their planned resolution in an ever-changing therapeutic regimen. The sole purpose for which drug abuse treatment programs exist is to provide treatment to clients. Treatment is the single most critical activity that takes place within the program. Treatment itself will take on as many unique forms as the wide range of programs in existence. Despite this disparity, programs do share one thing in common--the need to summarize the client's treatment program and progress. This summary is performed in the client's clinical record by updating the treatment plan and recording progress notes. The information placed in the clinical record enables the therapist to obtain an accurate picture of the client's original status and goals, current status and progress, and future plans and objectives. In addition the information contained in the clinical record enables the drug abuse treatment manager to accomplish his/her evaluation activities (discussed in Module 1). Properly documenting each client's history, problems, treatment, and progress in the clinical record should be a routine and painless activity. In many instances however, it becomes quite painful due to the lack of organization, principles, and procedures for setting up and maintaining the clinical record. An organized and uniform record which embodies generally accepted principles is essential within any sound client management system. Without these principles the extraction of critical information such as client contacts, type of therapy provided, and/or client respon- siveness to treatment is virtually impossible. 201 Essential ingredients of a sound clinical record include: ® A comprehensive "data base" ° A treatment plan ° Concise and accurate progress notes ° A discharge summary ° Followup activities Programs which have complete clinical records generally also have sound client management system and evaluation mechanisms. In developing a clinical record system, several decision areas will be encountered by program managers: ° What type of numbering system should be used? ° Should records be filed by client name or number? ° Should a Problem-Oriented Clinical Record System be imple- mented? ° How long should records be kept in an active status? ° Should retired records be microfilmed? Answers to these questions will depend upon the specific needs of the program, services being provided in addition to drug abuse treatment facility available, financial resources, and program objectives. Clients being discharged from treatment require the completion of a discharge summary. This is a one-page form which captures information required by Federal regulations and the JCAH stand- ards. It documents the following specific information all of which is also required on the CODAP Discharge Form: ° Dates of admission and discharge ® Time in treatment ° Reason for discharge ° Employment status at discharge ° Educational status and current enrollment at discharge ° Drug use at discharge In addition the form provides for a narrative summary of the discharge which is also required by JCAH. This narrative should be structured utilizing the problem-oriented format of the Treat- ment Plan and Progress Notes. Problems or treatments mentioned 202 should be referenced precisely to those mentioned in the Treatment Plan and/or Progress Notes. Brief summaries should also be written using the SOAP' method. Any plans or recommendations for followup care should be carefully detailed and explained to the client. The form should be completed by the primary counselor within a week of discharge. The related CODAP Discharge Form can then be prepared by clerical personnel, preferably the client records clerk. Followup activities, that is, client contact after discharge are at once the most difficult activity to carry through successfully but, when they are, serve to provide management with valuable informa- tion concerning the adequacy of treatment techniques. Specific followup activities should be defined by program management as part of the program evaluation functions. The key concern is establishing a means by which data can be gathered from clinical records and compared to client's current status outside the treat- ment environment. The extent to which followup activities are performed is very much dependent upon program resources. Accounting and Financial Management Module As funds for drug abuse treatment become more scarce (and as costs increase) all drug abuse treatment programs must move toward self-sufficiency. This means that program managers must provide services in as efficient and effective a manner as possible and earn maximum revenue. It also means that funding agencies must become more selective in deciding where and to what degree their funds will be applied. To fulfill these new demands, program managers and funding agencies need timely information about the financial status of the organization. Thus, a financial management and accounting module provides management information for two purposes: ° Internal management control: Recall that the last of the seven basic management questions asked "How much did it cost/earn?" In order to make decisions about future opera- tions, expenditures, and revenue policies, the program manager must have information upon which to evaluate past operations, expenditures, and revenue policies. ® External accountability: Every treatment program is respon- sible to some outside group or agency for its funds: govern- ing boards, the Internal Revenue Service, Federal and State grantors, third-party payers, and so on. These parties will use accounting data to determine the "credit worthiness of the program and its cost relative to other programs (i.e., competitors). 'SOAP is an acronym for: Subjective (patient) observations, Objective (therapist) observations, Assessment Plan; the steps in the Weed system of problem-oriented charting. 203 An accounting and financial management system that meets the first (preeminent) purpose will invariably meet the second because the information required to fulfill internal requirements is much more detailed than that required by external parties. Accounting is a body of generally accepted principles, practices, and standards among which are: ° Double-entry accounting ° Cost accounting ° Fund accounting ° Cash, accrual, or modified basis of accounting ® Consistency ° Conservatism ° Simplicity, accuracy, and flexibility The following sections describe each of these concepts. A sound accounting and financial management module should use a double-entry system of accounting. This approach to the "book- keeping" aspects of accounting simply records the two sides of every financial transaction: the "debit" side and the "credit" side. Each time a debit is recorded, there must be a correspond- ing credit entry or entries of equal value. The double-entry system is self-balancing and self-correcting and thus serves as a safeguard against fraud or error. Figure 8 summarizes the basic rules for the use of debits and credits in the asset and liability accounts. This presentation assumes that the reader has a familiarity with the basic mechanics of the double-entry system. If the program does not have an accountant, it should anticipate getting outside assistance in setting up the procedures and books by a professional accountant. By answering the question "What did it cost?" the financial data gathered by this module is one of the most valuable tools available to management. These financial data must at a minimum reflect the overall financial status of the program. But the cost informa- tion should also describe the expenses related to the other basis management questions: ° Who? ° Provided what? ° To whom? 204 S502 Assets Liabilities Stockholder's equity Debit Credit Debit Credit Debit Credit Decrease Increase Increase Decrease Increase Decrease Expense Revenue Debit Credit Debit Credit Increase Decrease Increase Decrease FIGURE 8.—Summary of basic rules for use of debits and credits in the asset and liability accounts ° When? ° Where? It is theoretically possible to break down each financial transaction to answer each of these questions. Practically however, this is a cumbersome and tedious process. As a compromise programs will have to be selective about the level of detail in the accounting system. Information needs must be reviewed carefully to select only the few most meaningful categories. Consider: ° The relevance of the categories for routine management decisionmaking. For example, if specific individuals are responsible for cost centers, each must have information about the costs of his/her own operations. ® The degree of difficulty in collecting the information. For example, because hundreds or thousands of services are provided each week, it may be infeasible to record the costs associated with each one. ° The alternatives for obtaining similar information. For example, if utilization statistics are available from the client treatment module, estimates can be derived without having to accumulate all costs by service. Using these criteria, the model accounting and financial manage- ment module will employ only three categories of the information classification system: the cost center code, the resource code, and the general ledger code. Combined, this system of coding will tell us: ° Where each cost was incurred (the cost center code). ° The type of expense (the resource code). ° The asset or liability account affected by the expense (the general ledger code). This system is called "cost accounting" and enables one to (1) to prepare meaningful cost reports, (2) compare the planned to actual expenses, and (3) identify problems and assign respon- sibility for their remedy. Fund accounting is a concept peculiar to nonprofit organizations. It arises from the grantor-grantee relationship wherein: ° The grantor requires accountability for the particular money provided to the treatment program. ° The grantor limits the uses to which the grant is put, i.e., the granted funds are restricted. Fund accounting is accomplished by assigning fund codes from the ICS to each financial transaction. This enables the preparation of 206 expenditure reports according to fund and the comparison of actual to planned revenue by source. One factor that must be considered in operating an accounting system is timing; at what point will financial transactions be recorded in the books? There are three alternative methods: accrual, cash and combination. Figure 9 illustrates the debit and credit entries required by each method. Under the accrual method of accounting: ° Revenues are identified with the specific time period in which they were earned (i.e., billed). ° Expenses are identified with the specific time periods in which they were incurred. The recordings are made regardless of the timing of the actual receipt of money or payment of debts. Note in figure 9 that the accrual basis appears slightly more complex than the cash basis. Nonetheless the accrual basis is generally considered to provide the most accurate picture of the true financial status of the organization. This approach also facilitates comparisons between financial statements from prior periods and from other organiza- tions. Under the cash method of accounting: ° Revenues are recorded at the time cash is actually received regardless of when it was earned ° Expenses are recorded at the time the bills are paid regard- less of when the bill was incurred As seen in figure 9, the single most important advantage of the cash basis is its simplicity; each financial transaction is recorded only once. However, depending on the timing of the cash flow, the cash basis can result in dangerously misleading financial reports that (1) overstate the cash "available," or (2) understate the revenue "earned." To the extent that these occur, the financial information generated by the AFM module will be less helpful as a management tool. Two methods are commonly used that combine the accrual and cash bases. ° Part-cash, part-accrual method: Under this approach, also known as the "modified" or "hybrid" approach, most financial transactions are recorded on a cash basis. Only those large transactions that are certain to come to fruition are handled on an accrual basis. ° Cash basis accounting--accrual basis reporting: Under this basis of accounting, the treatment program keeps its books on a cash basis throughout the year but makes the necessary adjustments to record the accruals for yearend reporting. 207 Assume that a patient receives services from a drug abuse treatment center on December 20, 1976. For the purpose of this example, the reporting period is a l-month period, December 1-31. The patient is charged $10 for the visit. The patient does not pay cash on the day of the visit but is billed for $10 at the end of December. Upon receipt of the bill on January 1, 1977, the patient remits the full $10. UNDER THE ACCRUAL BASIS OF ACCOUNTING eo During December 1976 the accounting transaction would be: Debit: Accounts receivable $10.00 Credit: Patient fees $10.00 e The December 1976 income statement would show a $10 amount for income. (The amount is earned in December 1976.) eo In January when the patient pays, the accounting transaction would be: Debit: Cash $10.00 Credit: Accounts receivable $10.00 e During January no effect is made on an income account, only on two asset accounts. UNDER THE CASH BASIS OF ACCOUNTING eo During December no entry would be made. e During January the accounting transaction would be: Debit: Cash $10.00 Credit: Patient fees $10.00 e The January 1977 income statement would show income of $10. FIGURE 9.—Debit and credit entries under accrual and cash methods of accounting 208 One of the chief functions of the accounting system is to aid comparisons of financial status from period to period. These comparisons enable the program manager to tell if he/she is "doing better or doing worse" than in the past. The accounting principle of consistency assures that financial reports and records are truly comparable from period to period. In striving for consistency it is wise to remember that: ° The accounting system procedures should be clearly docu- mented. Written documentation serves three purposes: 1. Eases transition when there are changes in accounting staff. 2. Facilitates audits and assures that auditors interpret records appropriately. 3. Provides materials for training when the system is implemented. ° The need for consistency should not rule out implementation of a new system if the old system has proven inadequate. Usually accounting is perceived as a "black and white" activity. But actually there is considerable room for judgment in recording financial transactions. The concept of conservatism in accounting refers to the concept of not overstating income and owner's equity. It may seem self-evident to state that the accounting system should be simple, accurate, and flexible. But often these guiding principles are lost in the confusion of designing and implementing a new system. ° Simplicity: The accounting system should be as simple as is consistent with the principles of good management. In most instances the degree of simplicity that can be achieved is dependent upon the size of the organization and the nature and volume of its financial transactions. Simplicity should not be promoted as a goal in itself and should not be exer- cised at the expense of some of the more important concepts discussed here. ° Accuracy: An accounting system should be set up in such a way that there can be no question as to the accuracy of the event or transaction that is recorded and subsequently reported. ° Flexibility: Since it is difficult to predict with accuracy the nature and volume of financial events and transactions an organization may have to handle from time to time, it is essential that the accounting system used be flexible enough to accommodate the various contingencies and eventualities. 209 There are basically two types of financial transactions: ° Those which deal with money coming in ° Those which deal with money going out Consequently the accounting and financial management module could be divided into transaction types: ° Income and accounts receivable (the forms and procedures related to money coming in) ° Payroll ° Invoice processing (relating to all other [nonsalary] expen- ditures) These three are tied to each other and to the larger system by two additional subdivisions: ® The general ledger (the final accumulation of all financial information in the accounting system); ° Cost accounting (accumulates financial data according to the relevant categories of the ICS and thus is the source of the management cost reports). To assure that all service encounters generate the appropriate charges and bills, income and accounts receivable must interface closely with client treatment. The normal course of events that occurs when a client enters the program and receives treatment is summarized below: ° The admission form provides a cue to the accounting office that a client accounts receivable ledger card is to be prepared. ° The encounter form (a normal offshoot of the client treatment process) is "shared" with the accounting office and prompts the following events: 1. The charges are recorded in the accounts receivable ledger card and the new balance is computed. 2. The patient bill is prepared. 3a The billing journal is updated. ° Weekly or monthly the billing journal entries are posted to the general ledger using the following entries: Debit Accounts receivable: Medicaid Credit Income: Medicaid 210 Additional entries are, of course, required to account for payment or for writing off the debt as uncollectable. One of the most expensive and important recurring functions of any accounting system is preparation of the payroll. Procedures should be devised that minimize the time required to prepare the payroll and that minimize the possibility of error or fraud. The following procedures are generally required: ° Each employee's status (or change in status) must be indi- cated on the payroll change slip which prompts the creation (or change) in the individual pay card. ° The time and attendance report or staff activity log generated during the client treatment process provides the basis for biweekly or monthly: 1. Update of the individual pay card. At this time the gross pay, deductions, and net pay are computed. 2. Update of the payroll journal. 3. Update of the leave register. ° The entries on the payroll journal are summed and totals are entered into the general ledger using the following notations: Debit Salary expenses Credit Cash Credit Withholdings ° The salary expenses must then be summarized according to the relevant categories in the ICS and posted to the cost accounting ledgers. Considerable caution must be exercised in the drawing and distri- buting of checks. Regardless of whether the cash or accrual basis of accounting is used, the voucher system is an effective method of establishing control over expenditures. Properly utilized the voucher system will insure that all: ° Invoices are authorized and reviewed prior to payment. ° Purchase discounts are optimized. ° All expenditures are handled in a consistent manner and can be readily reviewed if questions about payment arise. Illustrative transactions as they relate to an accrual system are highlighted below: 211 ° Receipt of an invoice prompts review of the receiving report to determine that the goods and services have been received. ° The invoice is checked for arithmetic accuracy. ° A voucher is prepared, authorized, and entered into the voucher journal. ° Biweekly or monthly the totals in the voucher journals are posted to the general ledger using the following entries: Debit Expense or asset account Credit Accounts payable Adjusting entries are required (from the cash disbursements journal) when the account is paid. ° The expenses must be coded with the appropriate ICS codes and posted to the cost accounting ledger. The first step in designing and implementing an accounting and financial management module is to create a general ledger (or chart of accounts). The second step is selection of the ICS categories that will comprise the cost accounting ledger. Only when these two "submodules" have been confirmed will an agency be able to develop a module which is responsive to the unique needs of the program. Program Evaluation Module One of the chief purposes of an MIS is to collect statistical data that are relevant to the decisionmaking process in the treatment center. In addition to making day-to-day decisions, the program manager must periodically look back over past operations to review what actually happened and determine: ° If program goals and objectives were reached. ° What changes should be made in the future. This program review completes the management cycle and provides a "feedback loop" that connects the prior period with the coming period. Program evaluation is the process of collecting, tabulating, and analyzing data to answer questions relevant to program goals, objectives, and activities. It usually compares actual program operations and results with previously planned program operations and results. This simple definition should not be allowed to mask the complexity of evaluation. Evaluation is a topic that is difficult to discuss unless all parties share a common set of definitions and concepts. 212 rE pa Cd The next three sections describe three key (and often confused) characteristics of program evaluation. The purpose of program evaluation is to measure actual against planned performance. But the performance or accomplishments of a drug abuse treatment center can be defined in any of three hierarchical ways: © Process: The most rudimentary measures of success are the processes, activities, or tasks performed. Process measures correspond to specific quantified activities. A process evaluation question might be: "Did we provide methadone maintenance services to 100 clients 3 times per week?" @ Outcome: A more significant measure of success is the outcome or results of the activities. An outcome evaluation question might be: "Were we successful in reducing the level of unemployment to 25 percent?" ° Impact: The ultimate measure of success is the impact of the activities and the outcomes. Impact measures correspond to the broad goals of the program and tell whether the program made any significant contribution to the target population. An example of an impact evaluation question would be: "Has the amount of drug abuse in the community decreased by 10 percent?" These three types of evaluation each serve a valid--but different-- purpose. They form a hierarchy of importance and are connected by the following assumptions: ° When measuring process, the assumptions are that: If we complete certain activities, then we will reach our objectives and, hence, our goals. ° When measuring outcome, the assumption is that: If we reach our objectives, then we will reach our goal. ° When measuring impact, no assumptions are necessary because impact studies are addressed directly at the question of goal attainment. In the long run, only the latter form of evaluation tells with certainty that the goals have been met. But as a practical matter most routine evaluation is designed to measure process or outcome. Another reason why people have difficulty communicating about program evaluation is because of its diverse nature. Evaluation is an entire discipline, the various types of which fall along a continuum. Such a continuum has characteristics which are defined in relation to specific goals and objectives and which change over time. Evaluation projects themselves are a form of "business activity" and so can be described by common terminology regardless of 213 their specific purpose. The evaluation cycle comprises the follow- ing steps: ® Deciding what is to be analyzed. ® Identifying the measurement criterion, standard, or value against which success will be judged. ° Identifying the information needed to perform the analysis. ° Determining the procedures necessary to collect, tabulate, and analyze the data. ° Collecting, tabulating, and analyzing the data. ° Reaching conclusions about the data and deciding what changes, if any, are necessary. Process evaluation is usually less difficult than impact or outcome evaluation but all types require the same basic steps. Evaluation projects can be very complex or relatively simple and as the complexity of the project increases, the information require- ments become more voluminous and complex. A general purpose MIS cannot meet all of these complex information needs. One of the major operational objectives of the MIS is that it can be efficient and simple and provide information necessary for routine decisionmaking. Process Evaluation is easily established in an MIS on a continuous basis because all information required for process evaluation is necessary for other purposes in the account- ing and financial management and client treatment modules. The information generated routinely by the MIS may assist in outcome or impact evaluation but it is rarely sufficient by itself. Special evaluation studies will usually be required to respond to outcome or impact evaluation questions. By suggesting that the outcome evaluation capabilities of the MIS are limited, we do not suggest that the other realms of evaluation are less important. Rather, outcome is ultimately one of the most revealing approaches of program evaluation and should be addressed through more sophisticated and flexible research efforts. AUTHORS WILLIE DAVIS, M.B.A. KERRY G. TREASURE Birch & Davis Associates, Inc. Silver Spring, Maryland 214 Y¢ U.S. GOVERNMENT PRINTING OFFICE : 1979 O—300-022 U. C. BERKELEY LIBRARIES TURE RAED COL98143kL8