key: cord-0951317-91g5xn0p
authors: Lambe, Kathryn; Lydon, Sinéad; McSharry, Jenny; Byrne, Molly; Squires, Janet; Power, Michael; Domegan, Christine; O'Connor, Paul
title: Identifying interventions to improve hand hygiene compliance in the intensive care unit through co-design with stakeholders
date: 2021-06-10
journal: HRB Open Res
DOI: 10.12688/hrbopenres.13296.1
sha: 2924305f31a817342965a089d98fe8d49175e16c
doc_id: 951317
cord_uid: 91g5xn0p

Background: Despite the effectiveness of hand hygiene (HH) for infection control, there is a lack of robust scientific data to guide how HH can be improved in intensive care units (ICUs). The aim of this study is to use the literature, researcher, and stakeholder opinion to explicate potential interventions for improving HH compliance in the ICU, and provide an indication of the suitability of these interventions. Methods: A four-phase co-design study was designed. First, data from a previously completed systematic literature review was used in order to identify unique components of existing interventions to improve HH in ICUs. Second, a workshop was held with a panel of 10 experts to identify additional intervention components. Third, the 91 intervention components resulting from the literature review and workshop were synthesised into a final list of 21 hand hygiene interventions. Finally, the affordability, practicability, effectiveness, acceptability, side-effects/safety, and equity of each intervention was rated by 39 stakeholders (health services researchers, ICU staff, and the public). Results: Ensuring the availability of essential supplies for HH compliance was the intervention that received most approval from stakeholders. Interventions involving role models and peer-to-peer accountability and support were also well regarded by stakeholders. Education/training interventions were commonplace and popular. Punitive interventions were poorly regarded. Conclusions: Hospitals and regulators must make decisions regarding how to improve HH compliance in the absence of scientific consensus on effective methods. Using collective input and a co-design approach, the guidance developed herein may usefully support implementation of HH interventions that are considered to be effective and acceptable by stakeholders.

Healthcare-associated infections (HAIs) present a serious challenge to safe, effective, and efficient healthcare. HAIs are of particular concern in the Intensive Care Unit (ICU), where prevalence rates of between 20% and 30% have been reported 1 . Research suggests that more than half of HAIs may be preventable 2 , with appropriate hand hygiene (HH) considered to be the most effective safeguard 3 . Despite the importance of hand hygiene (HH), there are a number of weaknesses 4,5 in the research evidence to guide the implementation of HH interventions:

• Lack of methodological rigour. There is a lack of methodologically robust studies to explore the effectiveness of interventions to increase HH compliance 4-6 ;

• Lack of a theoretical basis for intervention. Safety interventions commonly fail to have a theoretical basis to support the implementation of evidence into practice 7,8 and HH interventions are no exception 6 ;

• Lack of practical guidance in how to apply and sustain good HH practices. There are few descriptions of concrete and practical strategies to translate these guidelines to everyday practice in ICUs 8 , and;

• Inadequate understanding of the complexities of the environment and organisation in which the behaviour is to take place. Changes in an organisation require the consideration of a range of factors interacting at different levels of an organisation 7 .

These weaknesses mean that decisions regarding HH improvement are made in the absence of scientific consensus 4 . The purpose of this study is to take a co-design approach in order to leverage the knowledge of subject area experts and other relevant stakeholders to provide guidance on pragmatic interventions for improving HH compliance in ICU settings. Co-design in healthcare involves the equal partnership of those who work within the system (healthcare staff), those who use the system (patients and their families/carers), and those who design interventions for the system (researchers), with a shared goal of achieving better outcomes or more efficient processes 9 .

The Behaviour Change Wheel 10,11 was used to provide a theoretically grounded and structured approach to intervention development. The Behaviour Change Wheel is a framework for understanding behaviour and developing interventions to target behaviour change. This framework has been used in previous studies of HH [12] [13] [14] . This framework specifies nine intervention functions, which describe ways that behaviour may be targeted and changed 10 . These functions were used in the current study to support the development and analysis of HH interventions identified in the co-design activities.

The intervention functions 10 are:

• Coercion: creating an expectation of punishment or cost;

• Education: increasing knowledge or understanding;

• Training: imparting skills;

• Enablement: increasing means or reducing barriers to increase capability, or opportunity;

• Environmental restructuring: changing the physical or social context;

• Incentivisation: creating an expectation of reward;

• Modelling: providing an example for people to aspire to or imitate;

• Persuasion: using communication to induce positive or negative feelings or stimulate action; and

• Restriction: using rules to reduce the opportunity to engage in the target behaviour.

An intervention-such as those that promote peer-to-peer accountability and support among staff in the ICU-has the function of enablement (increasing means or reducing barriers to increase capability or opportunity). It is also important to point out that each intervention has a number of 'intervention components'. These components are the discrete elements that make up an intervention. Using the peer-to-peer accountability and support example, these components may include encouraging staff to remind one another about HH and the appointing of role models. The specific components of each intervention may be implemented in whatever combination is most applicable to a particular ICU.

The overall aim of the study was to derive a list of possible interventions to enable infection control teams to identify an intervention that will address hand hygiene compliance in their own ICU.

The specific objectives of the study were to:

• derive a comprehensive list of HH intervention components from the existing literature and input from an international panel of researchers;

• synthesise the intervention components into a list of unique interventions, each with a central focus and possible variations; and

• to elicit feedback on the suitability of each of these interventions from a panel of stakeholders, including members of the public, ICU staff, and researchers.

This co-design study was carried out in four stages: 1) a systematic review; 2) an expert panel workshop; 3) synthesis of intervention components; and 4) a stakeholder survey. The methods and results from each of these stages is described below, followed by a general discussion.

The purpose of the systematic literature review was to synthesize the literature describing interventions to improve HH in ICUs, to evaluate the quality of the extant research, and to outline the type, and efficacy, of interventions described. This systematic review has been published previously 15 .

Therefore, only summary information on this systematic review are provided below. and KL, a post-doctoral researcher, a woman, >5 years health services research) served as facilitators to guide the discussion and record the proceedings. The participants in the workshops were known to the facilitators, and no-one else was present at the workshops besides the participants and researchers. There were three parts to the workshop.

1. Introduction (1 hour). The workshop opened with a round of introductions. This was followed by a presentation by the workshop facilitators, outlining the purpose of the project, work achieved to date, including the systematic review data, and an outline of the intervention design task. Consistent with co-design principles 9 , the contributions of the facilitators were short and flexibility was built into the workshop so that the expert panel could set priorities and direct the agenda.

2. Small group intervention development exercise (2 hours). The workshop participants were divided into small groups, each with a workshop facilitator. The goal of these groups was to generate ideas for interventions. Each group was assigned two or three intervention functions from the Behaviour Change Wheel and asked to generate ideas for intervention components under each of these functions based on their professional experience, knowledge of the research literature, and/or clinical experience. To aid in the intervention generation process, the groups were provided with worksheets with the following prompts: describe the intervention component idea; describe the target group; identify the strengths and challenges of the intervention; identify any cost implications; and identify appropriate outcome measures.

3. Discussion and refinement (2 hours). Workshop participants reconvened to share their ideas and comment on the intervention components in a round-table discussion format. The discussion sessions were audio-recorded, and handwritten notes were taken by the facilitators. The worksheets completed by the small groups were also retained for analysis. The audio recordings were not transcribed. The purpose of the recordings was as a back-up to the written notes should anything be unclear or require clarification.

The data collected from the workshops was entered, collated, and organised using Microsoft Excel 2016. The data collected from the workshop resulted in 16 intervention components. The audio recordings of the workshops were not used as the facilitator notes as the worksheets were sufficiently detailed (see Underlying data 17 ). The expert panel emphasised the social environment as a key point of leverage for improving HH compliance (n=7 interventions), particularly advocating a collaborative approach among staff and infection control teams to goal setting and action planning (n=4). Suggestions were also made around improving team working in specific difficult clinical situations, assigning role models, and "charming nagging" to encourage compliance among peers. The importance of good role models and well-chosen HH "champions" to shape the approach to HH on the unit was also emphasised; these individuals should be well-respected, approachable, and encouraging. The interventions were largely generated under the Environmental Restructuring function (n=5). Other interventions included providing feedback on compliance and carefully using emotional or fear-based messaging to emphasise the consequences of poor compliance.

The purpose of this third stage was to collate the intervention components derived from the first two stages of the study and to synthesise them into a detailed list of interventions, which could be feasibly presented for evaluation to the stakeholder survey group in the final stage of the study.

The intervention components derived from the literature review (102 components) and the expert panel workshop (22 components) were collated, categorised using the intervention functions of the Behaviour Change Wheel, and synthesised into 91 unique intervention components (see Underlying data, 17 ). An iterative process was used to group similar components. For example, the components 'public display of compliance rates', 'unit-level feedback', and 'monthly report cards emailed to chiefs of service' were collapsed into a single intervention entitled 'monitoring and feedback at unit level'. The synthesis process was carried out with the following goals in mind:

• it was determined by the authors that the number of interventions should not be so large as to be unwieldy or unusable in the context of a toolkit to improve HH compliance; and

• the specificity and detail of the individual components should be preserved insofar as is practical, to ensure that the final interventions remain useful and actionable.

For each intervention, a summary description was prepared of the core intervention concept, possible variations where appropriate, strengths and challenges, and cost implications. This process was carried out using Microsoft Excel 2016 by two members of the researcher team (POC and KL), and then independently reviewed by another member of the team (SL). These summaries were composed based on learning from the research literature and expert panel discussion, the experience of the research team, and the unique components that were collapsed into each intervention.

The iterative synthesis process outlined above collapsed the 91 unique intervention components into a final list of 21 interventions (see Table 1 and Underlying data for a more detailed description of each intervention).

The purpose of the stakeholder survey was to systematically elicit input from a range of stakeholders on the utility of the 21 HH intervention described in Table 1 using the 'APEASE' criteria (affordability, practicability, effectiveness, acceptability, side-effects/safety, and equity). This is a set of criteria set out in the Behaviour Change Wheel guidance to assist intervention developers in evaluating behaviour change interventions 10 .

Stakeholder survey: methods Recruitment and participants. A purposive sampling strategy was used to recruit stakeholders by directly emailing potential participants to ask them to complete the survey. Everyone who was contacted agreed to participate. Representatives were recruited from three specific groups: public representatives, ICU staff, and health services researchers. Members of the public and ICU staff received a €100 voucher for their participation. The final sample (n=39) included 11 members of the public, 11 ICU doctors (mean 6.7 years of experience), 10 ICU nurses (mean 16.4 years of experience), and seven health services researchers (mean 13.2 years of experience). The number of respondents was pragmatic, based upon the number of participants we were able to recruit in the time available, and for which there was funding for incentives. Of the respondents to the survey, two health service researchers and one member of the public also participated in the expert panel workshop.

Ethics statement. Ethical approval was received from the National University of Ireland, Galway's Research Ethics Committee (ref: 18-Sept-17) and written informed consent was obtained from all participants to participate and use their data in the study via a tick box.

Procedure. The survey was carried out using SurveyMonkey, and online survey software (Google Forms is a freely available alternative). The participants were provided with brief background information on the study in a five-minute video introduction. The 21 interventions outlined in Table 1 were then presented in a random order to each participant. For each intervention, the participants were provided with the intervention summary description (see Underlying data 17 ).

For each intervention, the participants were asked to rate their agreement with each of the following five APEASE dimensions on a 0 (strongly disagree) to 100 (strongly agree) slider scale:

• Affordability: intervention can be delivered within an acceptable budget.

• Practicability: intervention can be delivered with minimal disruption to patient care.

• Effectiveness: intervention is likely to improve hand hygiene compliance.

• Acceptability: intervention will be considered appropriate by staff in the ICU.

• Side effects/safety: intervention will not have any unwanted side-effects or unintended consequences. A free-text box was also provided for each intervention for any additional comments. Statistical analysis was completed using IBM SPSS version 21. For each intervention, the mean of the responses to each of the five APEASE dimensions was calculated. An overall mean APEASE score for each intervention was then calculated by deriving the mean of the dimension scores. A total of 2.4% (n=119 responses) of the data was missing. No imputation methods were used to account for the missing data. One-way ANOVA was conducted to examine differences in scores for each intervention across the three participant groups. For each of the intervention functions, a mean APEASE score was also calculated (e.g., for enablement, the mean of the scores for the three enablement interventions was calculated).

The mean APEASE scores for the 21 interventions ranged from 53.5-81.3 on a scale of 1-100, and are shown in descending order in Table 2 , with the individual APEASE scores for each intervention provided in Underlying data 17 . The highest-scoring interventions were 'ensuring availability of essential supplies', 'providing strong hand hygiene role models within professional groups', and 'comprehensive active education and training'.

The intervention function with the highest mean APEASE score was Modelling (74.0), although only one of the included interventions fell under this intervention function. The next highest were Enablement (73.6, three interventions) and Education/training (71.0, five interventions). Environmental restructuring (62.9, three interventions) and Coercion (73.2, two interventions) received the lowest mean ratings.

No significant differences were found in mean APEASE scores between the three groups of participants. The free-text responses to the interventions were mixed. Concerns that were most frequently raised included that the interventions may create a negative atmosphere of scrutiny or be difficult to implement in a fair, transparent way. A summary of the free-text responses about each intervention is provided in Underlying data 17 .

While regulators prioritise improvement in HH and specify 'what' standards must be achieved, there is also a need for practical guidance on 'how' these standards can be achieved 18 . The relative lack of guidance on how to improve HH compliance is exacerbated by the paucity of strong evidence to identify what interventions are likely to be effective 15 , and how potential interventions should be implemented to meet the needs of specific ICUs. However, despite the lack of robust research evidence, there is still a need to provide advice and guidance to hospital managers and clinicians on how best to invest limited resources to support improvements in HH compliance. With that advice in mind, the purpose of this study was to use the available literature, expert and stakeholder opinion to identify potential interventions for improving HH compliance in the ICU, and to give some indication of the suitability and acceptability of these interventions.

Ensuring the availability of essential supplies for HH behaviour was the intervention that received most approval from stakeholders. Interventions involving role models and peer-to-peer accountability and support were also well regarded. Education/training interventions were commonplace and popular. Punitive interventions were poorly regarded. It is perhaps unsurprising that the need to ensure the availability of essential supplies for HH behaviour was the highest-scoring intervention in the stakeholder survey. The World Health Organization recommends the use of alcohol-based hand rub at the point of care 19 . The availability of essential supplies is the most fundamental HH intervention, and any other interventions are likely to have limited effectiveness if essential supplies are unavailable. The lack of supplies is less of an issue for high-income countries, such as Ireland. However, the same may not be true in lower income countries. A recent systematic review of levels of HH compliance in ICU settings found that healthcare practitioners were 65% compliant in high-income countries, as compared to 9% compliant in low-income countries-at least partially due to poor resourcing 20 . The limitations of the products used for HH, such as alcohol-based handrub (e.g. drying time, odour, skin irritation, ease of use) have also been identified as barriers to HH compliance 21 , and this is suggested as an area of future research and development. It is further suggested that regular audits are conducted in ICU units to ensure that HH supplies are always available for ICU staff, visiting healthcare teams, and members of the public/patients.

The provision of strong HH role models within professional groups was the second highest-scoring intervention and is related to peer-to-peer accountability and support (ranked fifth).

Ownership and leadership were the most notable enablers of HH compliance mentioned by Irish HH policy makers in a recent study 8 . Healthcare workers have reported that they frequently adjust their behaviour to match those that they see in clinical practice, and can feel strongly influenced to abstain from compliance by negative role models 22 . This suggests that positive role models can also have the same effect. This positive motivation could be achieved by encouraging senior staff members, particularly consultants and senior nursing staff, to act as role models for junior staff members 23 . However, it is important that in such interventions the role of different levels of leadership (e.g., senior clinical staff, management) and styles of leadership are clear if the intervention is to be effective 24 . Given the positive disposition of the stakeholders to these types of intervention, as well as the relative lack of research on this approach 15 , it is suggested that the efficacy of this approach should be examined in future research.

Five of the 21 HH interventions identified in this research were concerned with the education and training of healthcare practitioners. These interventions received favourable ratings from the stakeholders. Given that 79% of studies in our systematic literature review were concerned with education, and 69% with training interventions, it is clear that these approaches are ubiquitous in HH improvement programmes 15 . However, ICU staff frequently report having the requisite knowledge and skills to carry out HH, which may mean they are resistant to mandatory HH training 12,25 . Therefore, it is suggested that there is a need for targeted HH training in which ICU staff receive individualised and direct feedback on their HH performance at the bedside. Individual feedback is supported by a number of studies 26, 27 . Therefore, it is recommended that, at least for experienced ICU staff, that there is a shift from an approach of standard training delivered to all staff, to one that is tailored to the needs of specific units. To illustrate, in a recent study carried out across three Irish ICUs, it was found that nurses were more likely to engage in HH than other healthcare professionals, and the ICU staff observed were 2.6 times more likely to comply with HH if the indication for HH was self-protective rather than patient-protective 28 . Future research should examine patterns of behaviour across professional groups, shift times, settings, and indications, and report their findings in detail, so that guidance on how to tailor HH training and education for specific units and groups can be developed.

Two of the interventions that were identified and reviewed in this study involved coercion or reprimands-proactive corrective action and warning letters. Both of these potential interventions were unfavourably rated by the stakeholders, and these punitive approaches are likely to be ineffective. In fact, they may lead to a tendency for individuals to then cover up any errors that are made and are counter to the 'just culture' most healthcare organisations wish to foster. A just culture approach recognises that healthcare professionals make errors and may take shortcuts or fail to follow protocols 29 . There may be good reasons why staff do not follow procedures (e.g., not following HH protocols during an arrest due to the urgency of delivering care). In a just culture, it is recognised that there is a need to understand why healthcare professionals make errors and the importance of encouraging honest reporting from healthcare workers as to why things may go wrong. Identifying issues with the system is an important precursor in order to develop a tailored HH intervention; future research into co-design and collaborative approaches to doing so would be extremely advantageous.

The strengths of this study are that it utilised a multiple-methods approach with input from the research literature, experienced researchers, ICU staff and members of the public. However, there are a number of limitations of this research that should be acknowledged. First, the stakeholder survey participants were all based in Ireland, which may lead to questions about the generalisability of our findings to other countries. However, as Ireland's HH compliance figures are broadly in line with those in other high-income countries 28 , we believe that the findings can be generalised to ICU settings in other high-income countries. Second, the recommendations and assessments of the HH interventions are based upon opinion rather than upon scientific evidence for their effectiveness. This is certainly true, and can be attributed to the absence of robust HH research that can be considered sufficiently rigorous to support the identification of effective interventions 15 . However, in the absence of sufficiently rigorous studies, our findings provide some direction on the HH interventions that should be considered based upon the collective opinions of the international experts and stakeholders that participated in this study. Finally, and related to the previous limitation, we did not evaluate the effectiveness of any of the HH interventions. This was beyond the scope of co-design study. However, future research is needed to focus on assessment of intervention effectiveness.

In conclusion, despite the large financial and logistical investments required to implement a HH intervention, hospitals and regulators must make decisions regarding infection control policies in the absence of scientific consensus on what is effective. However, by using the collective input from a range of stakeholders, it is hoped that some guidance can be provided as to the HH interventions that are at likely to be suitable and acceptable (or not) to stakeholders, and encourage the rigorous evaluation of these HH interventions in future studies. This project contains the following underlying data:

-Upload to open access v3.xlsx (references of the papers described in the systematic review (stage 1), the collated notes from the expert workshop (stage 2), synthesis (stage 3) and the anonymised questionnaire data (stage 4)).

-Intervention and ratings.pdf (description of the 21 interventions, APEASE ratings, and stakeholder qualitative comments). This is a very interesting study which addresses an area of critical importance in healthcare -hand hygiene and infection prevention control. A co-design approach is taken with mixed methods employed to identify which components of the Behaviour Wheel Change might be most effective in improving hand hygiene compliance in the ICU. It describes an approach that allows interested parties to understand the gap between "work as imagined vs work as done", and by describing the barriers in a systematic way, to attempt to overcome these barriers.

The study is well thought out and builds on previous work carrying out a systematic review of the area. The working of the expert panel is well described and in keeping with co-design; the stakeholders for the survey are a good mix of public, ICU staff and researchers. A minor point, but perhaps to be considered in future such work -there might be a benefit in the inclusion of the following in co-design workshops: representatives of regulators and manager; the former because it is easier for regulators to mandate changes without understanding the barriers to achieving them, and the latter because theoretically, they control the resource to enable delivery.

It is an interesting finding that the simplest things matter a lot in terms of influencing behaviouractual supply and availability of PPE. It would be interesting to see reference to the PPE situation at the start of the COVID-19 pandemic in the discussion but we appreciate that this paper precedes that timeframe.

Introduction -definition of co-design -researchers are just one example of those who design interventions to improve the healthcare system. certainly use your findings in my own work and imagine many others working in the fields of implementation science, improvement, clinical practice (whether IPC or ICU) will too.

Your review identified existing interventions to improve HH in in ICU and you did this using the intervention functions of the BCW. I wondered. How did you find this approach? Was it difficult? Any particular problems with this approach. Did they all fit the framework?

Your approach is interesting and valid, I wondered why you chose to categorise to intervention functions rather than behaviour change techniques (an approach I have seen others take)?

Your co-design workshops looked really thorough. Did the patient representative have any unique views or ideas? With the intervention development exercise -I see you didn't share the did you intervention components designed/published in the literature.

In table 1 I see you have one element that wasn't identified in the literature -tailoring of training according to group, which I thought was both novel and insightful. I also liked the way you applied the APEASE criteria.

When did you conduct your groups? (Sorry if I missed this -couldn't see it). Was it during COVID-19 or pre? I wondered because the top intervention was availability of supplies. I thought we had nailed that challenge in most of the Western world -I appreciate you refer to other countries too.

On that note -you talk about the need for tailoring; how might that be done with your intervention components?

What are you as a team going to do with the results of this study? It would be great to see you test these interventions in practice.

A final note -love the term "charming nagging". Whoever came up with that is a genius!

Is the study design appropriate and is the work technically sound? Yes

Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes

Systematic review of studies on compliance with hand hygiene guidelines in hospital care

The authors wish to thank Dr. Kiran Reddy and Ms. Ann Conroy for their assistance with recruitment for this study

No competing interests were disclosed.

We confirm that we have read this submission and believe that we have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. 

School of Health Sciences, Birmingham City University, Birmingham, UK Thank you for approaching me to review this study. I very much enjoyed reading this, I will Competing Interests: No competing interests were disclosed.Reviewer Expertise: Implementation Science, Behaviour change, Improvement Science, Infection Prevention I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.