Article Text
PDF
Abstract
Background Experience-based codesign (EBCD) is an approach to health service design that engages patients and healthcare staff in partnership to develop and improve health services or pathways of care. The aim of this systematic review was to examine the use (structure, process and outcomes) and reporting of EBCD in health service improvement activities.
Methods Electronic databases (MEDLINE, CINAHL, PsycINFO and The Cochrane Library) were searched to identify peer-reviewed articles published from database inception to August 2018. Search terms identified peer-reviewed English language qualitative, quantitative and mixed methods studies that underwent independent screening by two authors. Full texts were independently reviewed by two reviewers and data were independently extracted by one reviewer before being checked by a second reviewer. Adherence to the 10 activities embedded within the eight-stage EBCD framework was calculated for each study.
Results We identified 20 studies predominantly from the UK and in acute mental health or cancer services. EBCD fidelity ranged from 40% to 100% with only three studies satisfying 100% fidelity.
Conclusion EBCD is used predominantly for quality improvement, but has potential to be used for intervention design projects. There is variation in the use of EBCD, with many studies eliminating or modifying some EBCD stages. Moreover, there is no consistency in reporting. In order to evaluate the effect of modifying EBCD or levels of EBCD fidelity, the outcomes of each EBCD phase (ie, touchpoints and improvement activities) should be reported in a consistent manner.
Trial registration number CRD42018105879.
- healthcare quality improvement
- health services research
- implementation science
- quality improvement methodologies
Statistics from Altmetric.com
- healthcare quality improvement
- health services research
- implementation science
- quality improvement methodologies
Introduction
There is widespread and active involvement of service users, their carers and family members in activities relating to healthcare.1–4 In terms of quality and safety, partnering with service users is required for effective individual care and for healthcare service design, overall governance, policy and planning.4 Active engagement of service users in the planning and development of healthcare is key to effecting change.5 As such, research on codesign and coproduction with consumers in healthcare has a relatively long history.6–9 Evidence from a 2013 systematic review (40 studies) suggests that the patient experience, when robustly collected and analysed, is positively associated with clinical effectiveness and patient safety.10 A more recent systematic review of 65 codesign studies of healthcare suggests that codesign encourages shared goals and might improve service user/provider relationships and communication, subjective health outcomes and service user satisfaction with the service provided.8 However, codesign in healthcare is notoriously difficult to implement. Barriers to its successful implementation include a lack of resources (eg, funding, codesign facilitators) and managerial support, staff turnover, logistical barriers for engaging vulnerable service users and cohort retention.6 Despite these barriers, in the last 5 years there has been an increase in published codesign work.
Experience-based codesign (EBCD) is a relatively newer form of participatory action research that involves service users, first piloted in 2005 to improve the care and treatment experience of patients with head and neck cancer and their carers.11 It integrates ethnographic research and service design methods with the principles of consumer engagement to improve patient care and provider experiences of care. Since the pilot study,11 EBCD has increasingly become a more structured and prescriptive method. Due to the quality improvement nature of EBCD, the stages are viewed as cyclical, continually improving the service or care pathway. According to the Point of Care Foundation (PoCF) toolkit,12 EBCD framework consists of eight stages: (1) observe clinical areas, (2) interview service providers and service users, (3) develop a trigger film (an edited videotaped interview film highlighting themes from the service user interviews), (4) service provider feedback event, (5) service user feedback event, (6) joint service provider and service user workshop(s), (7) codesign groups, and (8) celebration event. Accelerated EBCD (AEBCD) is an adapted method where the codesign process is accelerated by using pre-existing service user experience narratives from pre-existing interviews.
Despite the availability of EBCD toolkits,12–14 there are currently no reporting standards or EBCD-specific quality appraisal instruments to guide the appropriate conduct and reporting of these studies. Using all EBCD stages can be resource intensive and researchers might eliminate or adapt EBCD stages to satisfy time and resource constraints of a project. However, the success and quality of EBCD projects likely rely on how closely they adhere to the EBCD framework (ie, fidelity) as well as adequate scoping of the service provider and service user experience and skilled facilitation of codesign events.15 Despite the increasing number of published EBCD projects, there are currently no systematic reviews describing EBCD use in healthcare services. The aim of this systematic review was to examine the use (structure, process and outcomes) and reporting of EBCD in designing health service improvement activities.
Methods
This systematic review adheres to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guidelines.16 The research questions were informed by the Donabedian evaluation model (box 1).17
Research questions
Structure-related questions
Where were the studies conducted (country, setting)?
What was the size and the make-up of each stakeholder group?
What was the training or skill set of the facilitators?
What training was provided to the participants?
What resources were used?
Process-related questions
How did the study adhere to the experience-based codesign (EBCD) framework8 (ie, fidelity)?
What were the methods of gathering experience data?
What were the methods of the codesign phase?
What were the dropout rates and reasons from the codesign phase?
How were EBCD projects being evaluated?
How long was the EBCD process from planning to codesign completion?
What were the touchpoints identified in and across the included studies?
Outcome-related questions
What were the deliverables of the EBCD?
What were the outcomes of the EBCD process evaluations?
What were the participant views on the EBCD process?
Eligibility criteria
Due to the expected variations in using and describing EBCD, we defined the minimum requirements to be considered EBCD for this review as including two phases where service users were participants in both phases. During phase 1, relevant service user experience data must have been identified and summarised to identify touchpoints (or equivalent) either using service user data from the local service or using previously developed materials such as the AEBCD. During phase 2, codesign workshop(s) must have included at least one service user participant to develop recommendations or activities that provided professional, organisational and system service improvements.
We included all relevant qualitative, quantitative or mixed methods studies that used EBCD to design a new or improve an existing healthcare service or pathway, or studies that evaluated the EBCD process. We excluded studies where no service design or improvement was evident. Opinion pieces, editorials/letters, government reports and conference proceedings were excluded.
Search strategy
To identify potentially relevant reports of EBCD studies, we searched the following electronic bibliographic databases from database inception to 20 August 2018: MEDLINE, CINAHL, PsycINFO and The Cochrane Library. Searches included combinations of the following MeSH terms and keywords: ‘participatory action research’; ‘shared decision making’; ‘patient decision making’; ‘experience-based co-design’; ‘experience-based design’; co-design*; codesign*; ‘patient engag*’; ‘patient involv*’; ‘narrative design’; ‘co creat*’; ‘health services research’; patient; consumer; ‘patient care planning’; ‘delivery of health care’; ‘service planning’; ‘service design’; disease; and health. There was no restriction by date or language. We also searched Google Scholar using search phrases ‘experience-based co-design’ or ‘experience-based design’, and hand-searched the reference lists of relevant articles such as systematic reviews, and the included articles. The references were managed using EndNote V.X8 (Clarivate Analytics, 2018).
Study selection
Titles and abstracts of articles retrieved from the search strategy were independently screened by two reviewers who assessed the eligibility of relevant full-text articles. Disagreements were resolved through consensus among the two reviewers with third review author as arbiter.
Data collection
A standardised data extraction form of open and closed questions was developed, piloted for two included studies and adjusted accordingly before extraction of the remaining data. Data extraction included closed questions such as size and make-up of stakeholder groups, EBCD toolkits, facilitator and stakeholder training, completion of each stage of EBCD, mode of stage delivery, time to complete EBCD and recruitment and dropout rates. Open questions included author details, stated aims, setting, geographical location, resources allocated to study, EBCD framework details, analysis method for experience data, improvement activities and EBCD evaluations. Data were independently extracted by one reviewer and 100% of the data extraction was checked by a second reviewer. Any discrepancies identified by the second reviewer were checked against the study publications in the first instance and any resulting disagreements were resolved through consensus among the reviewer group.
Critical appraisal
As action research contributed to the development of EBCD,18 we used the draft Guidance for assessing action research proposals and projects 19 which comprises 20 questions used to guide critical reflection. Critical appraisal was not used as part of the eligibility criteria, but to describe the studies. Each study was independently appraised by two reviewers. Percentage agreement was calculated between reviewers and any discrepancies between appraisals were resolved by a third reviewer.
Synthesis of results
Frequencies of closed questions from data extraction were calculated to provide descriptive information about studies. Studies were first synthesised to address the use of EBCD (ie, structure and process-related questions17) relating to aims and settings, resourcing, participant characteristics and methods used in the included articles. We further examined the use of EBCD by exploring the fidelity of the included studies against the eight-stage PoCF EBCD framework. We identified the 10 activities as these related to each stage of the PoCF EBCD framework and calculated how closely each study adhered to the framework (EBCD fidelity).12 Each study activity scored 1 (completed) or 0 (not completed or unclear) per activity and calculated as mean EBCD activity score × 100%. Outcome-related questions relating to EBCD deliverables, strengths and weaknesses and participant views on the EBCD process were reviewed narratively. Where possible, improvement activities were categorised using the framework as defined by Locock et al 20 into: small-scale changes; process redesign at the team level; process redesign between services; and process redesign between organisations (adapted from Adams et al [21]). We examined reporting of EBCD studies by identifying whether each activity as outlined above was clearly reported in the publications.
Results
The search strategy yielded 647 records, of which 38 full-text articles were reviewed. We excluded 11 articles predominantly for being the wrong publication type (online supplementary table 1). We identified 27 articles reporting 19 completed and one ‘in progress’ study that met eligibility criteria and were included in this review (figure 1).
Supplemental material
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram of the search strategy.
Critical appraisal
Critical appraisal was completed by two reviewers with 93. 5% agreement (online supplementary table 2). All critical appraisal items were satisfied by two studies, with 10, five and three studies meeting at least 80%, 60% and 40% of the criteria, respectively. Only half of the studies adequately described the relationship between the researchers and participants. Twelve studies (60%) either reported ethics approval or discussed ethical issues relating to the project. Thirteen studies (65%) reported funding to support the project as well as successfully completing the project without issue. Thirteen studies (65%) discussed the extent to which the aims and objectives of each stage were achieved.
Supplemental material
Structural characteristics
Settings
Structure-related characteristics are summarised in table 1. Most studies were conducted in acute hospital settings in the UK. Healthcare areas using EBCD were mostly mental health (five studies), cancer (six studies including one study of cancer and intensive care unit), paediatrics (three studies), emergency departments (ED; two studies including one study of the geriatric palliative care experience) and one study each in palliative/end-of-life care, maternity, geriatric outpatient services and primary care for service users with multimorbidities.
Structure-related experience-based codesign elements of included studies in chronological publication order
Stakeholder groups
Of the 20 included studies, 12 described the project team including descriptions of advisory committees, key stakeholders or site personnel. Stakeholder involvement was not always clear as participant groups often changed after experience data were collected and analysed. Where reported, service provider experiences were represented by nurses, doctors and allied health with some including ‘managers’, clerical staff, receptionists and other ‘staff’. Service user experiences were represented by patients, caregivers, family and/or service user advocates.
Facilitation, training and resources
Of the 20 included studies, 17 described the facilitators (table 1), 11 of which described facilitator training and/or qualifications. No study reported training the EBCD participants. Financial support was acknowledged in 16 studies, two of which were specific to travel costs to attend EBCD training or conduct non-participant site observations.22 23 Half of the studies reported using an EBCD toolkit.
Process characteristics
EBCD duration and fidelity
The EBCD activities as they relate to each stage of EBCD are described in figure 2 and the process-related data are presented in table 2. The EBCD studies, from stage 1 to 8, took a median (range) of 9 (8–19) months and AEBCD took a median of 8 (4–8) months. EBCD fidelity (figure 2) across all studies was median 75% (25%–100%) with only two studies achieving 100%. The stages most often omitted or lacking description were stage 1 (observation) and stage 8 (celebration event). Where celebration events were held, EBCD participants as well as additional stakeholders external to the project were involved. Due to the inconsistent reporting of outcomes, we did not evaluate the effect of fidelity on implementation activities.
The eight stages of experience-based codesign (EBCD) (developed by the review authors as informed by the Point of Care Foundation8 EBCD toolkit).
Reporting and completion of experience-based codesign activities, duration and fidelity of included studies in chronological publication order
Data collection methods
Site observations were conducted for 5–20 hours per site with the exception of Tsianakas et al,24 who observed two service areas for 219 hours in total. The individual experiences of service users and service providers were collected in all 20 studies. The predominant method used was stakeholder interview with median 15. 5 (5–40) service users (14 studies) and 24 (4–54) service providers (13 studies). Joint or stakeholder-specific focus groups, workshops or meetings involved median 14 (6–38) service users (three studies) and 7 (5–17) service providers (five studies). Three studies used national archived service user interviews (ie, AEBCD) with one study supplementing archive data with local service user interviews.
Data analysis and touchpoints
Fourteen of the 20 studies systematically analysed experience data. Analysis methods varied, including thematic analysis (one study), colour-coding themes (one study), interpretative phenomenological analysis (three studies), framework analysis (three studies), qualitative content analysis (two studies) or thematic discourse analysis (one study), constant comparative method (two studies) and burden treatment theory (one study). Touchpoints were identified by 13 studies although these were often presented as summaries with only examples provided.
Twelve studies created a trigger film of video or audio-recorded interview excerpts. Other formats used to ‘trigger’ discussion (eight studies) during the joint workshop included touchpoint lists and experience maps of service user experiences (six studies). Interview quotes (three studies) and lists of improvement areas (one study) of service provider experiences were also used.
Stakeholder feedback events (used by 16 studies) included median 7 (4–39) service users (reported in 16 studies) and 17 (3–64) service providers (reported in nine studies). Improvement priorities were identified by participants (16 studies), researchers (one study) and not reported in two studies.
Joint workshop
Nineteen of the 20 studies had completed EBCD to at least the joint workshop stage (one incomplete) although only one study described the framework used to run their workshop (The MAXimising Involvement in MUltiMorbidity -MAXIMUM framework).25 Workshop delivery was face-to-face for all studies, with 12 studies reporting between 2–15 service user participants and 2–16 service provider participants per meeting (ratio of three service users to every four service providers), and one facilitator to every five participants.
Small codesign teams
Half of the included studies described using the small codesign team stage of EBCD. The number of teams formed, and the number and mode of meetings, were highly variable and largely dependent on the number of improvement priorities identified. All but one study used mixed teams of service users and service providers.
Dropout
It was often unclear whether the same participants were involved in both the data collection and the codesign workshops. Throughout the codesign workshops researchers often emphasised voluntary participation, resulting in a small core group (usually service providers) with others participating on an ad hoc basis. In two studies the protocol was amended to recruit an additional cohort for codesign, which was attributed to the transitory nature of the service users, high service provider turnover or time delays between EBCD stages.
Outcome characteristics
EBCD deliverables
Studies aimed to improve a service or care pathway (12 studies), evaluate the EBCD process (2 studies) or reported both improvement and evaluation (6 studies) (online supplementary table 3). Only two studies predetermined EBCD outcomes: (A) improving informational and educational resources or (B) the number of formal complaints on a specific ward. Project costs were only evaluated in one publication, which compared the cost of AEBCD with EBCD, and reported that AEBCD was cheaper than EBCD at £8289 GBP vs £30 485 GBP, respectively.20 26
Supplemental material
The studies that listed the improvement activities (11 studies) indicated 1–38 improvement activities per site, service or care pathway (online supplementary table 3) were generated by EBCD. Where improvement activities could be categorised, most were attributed to a redesign within team (six studies), small-scale changes (four studies) or redesign between services (one study) and one study had an even distribution of changes across categories.
Participant’s perception of EBCD
Process evaluation data were available for eight studies, with evaluation for Gustavsson’s neonatal and diabetes studies reported together.27 Both service users and service providers had positive views of the EBCD process26 28 29 and reported that specific, measurable, achievable, relevant, time-bound goals reflected their service improvement needs.29 Wright’s geriatric ED study22 found that staff had changed their personal practice and had developed ongoing multidisciplinary team collaborations as a result of EBCD.30 In the Cheshire and Ridge palliative care study,28 31 commissioners had commented that EBCD was run as a change management process that felt more engaging and less tokenistic in service user participation. Participants in the Tsianakas cancer study24 stated that the collaborative nature of EBCD gave service users a greater sense of direct responsibility for the work and its outcomes built a strong relationship between service users and service providers and noted a higher level of clinical engagement in the improvement effort than is usually observed in other projects. Service user participants from the Gustavsson’s neonatal and diabetes studies27 reported that the diversity of views, when presented face-to-face, resulted in a common perspective of patient processes. Participants also noted that the power relationship between professionals and patients was more equal in the EBCD than in actual care relationships. In contrast, service provider participants in Piper’s ED study32 study found it difficult to balance EBCD activities with other work commitments despite being positive about the EBCD approach.
Discussion
We identified 19 complete and one ‘in progress’ published EBCD projects aimed at improving healthcare services. As expected, the largest uptake for EBCD was in its country of origin (UK) and there is an increasing application of this method with most studies published after 2014 (15 studies). Despite the recommendation to complete all stages of EBCD,12 15 our review indicates that EBCD fidelity remains less than 100%. This might be attributed to authors’ perceptions of the flexibility of the EBCD framework,12 barriers to implementing codesign (ie, lack of resources and managerial support, staff turnover, logistical issues, cohort retention, information asymmetry),6 8 or the lack of evidence demonstrating that higher fidelity leads to better service user experiences (a limitation of the wider healthcare service codesign literature).6 8
Palumbo’s systematic review of coproduction in healthcare8 indicates that conflicting priorities and beliefs between service providers and service users as well as information asymmetry to be major barriers to codesign. The PoCF EBCD framework12 attempts to overcome these via site observations and sharing experiences during the joint workshop. Both methods provide insight into the healthcare service, help contextualise the touchpoints raised and move preconceptions about the service experience from what should be to what is. To this end, formally presenting the service provider experience during the joint workshop in addition to that of the service users could mitigate information asymmetry. However, few studies presented the service provider perspective, potentially contributing to conflicting design priorities and limiting engagement.
Similarly, few studies implemented site observations, none of which were carried out by the service providers. The PoCF12 encourages service providers to undertake observations so that they gain insight into the day-to-day delivery and experience reality of health services. However, making time for service provider observations without managerial support might limit service provider engagement in EBCD, as they are often required to volunteer time in addition to their existing workload expectations.6 Issues that were otherwise unreported by participants during interviews and focus groups (stage 2) might have been missed in studies that failed to complete observations,12 especially when researchers were not familiar with the service area. Where completed, observations were conducted in person by the researchers so the reliability of observation data was dependent on the method of data recording, coding scheme, observer experience and training, and the nature of the work environment.33 The effects of selectivity and observer-related factors (eg, fatigue, inattention) could be lessened by using pairs of observers or video-recording EBCD activities could be considered as a means of obtaining comprehensive and consistent data. This method facilitates greater flexibility in the time and duration of data collection.
In this review, we argued that any EBCD studies must at least involve two key phases, namely experience gathering phase and codesign phase with patient participation in both phases. Nevertheless, any non-adherence with activities or stages outlined by the PoCF EBCD framework could potentially compromise the extent of participation of service users and other stakeholders, and the quality of the experience gathering and codesign processes. Future studies should explore the relationships between fidelity as prescribed by the PoCF EBCD framework and service user experiences.
Consistent with the PoCF EBCD framework, interviews were often used in favour of focus groups to gather participant experiences. EBCD facilitators have reported that individual interviews engage service providers and enhance their commitment to the EBCD process.34 Compared with focus groups, individual interviews require fewer participants and data collectors per data collection event, are easier to schedule and take less time to organise and transcribe.35 As such, focus groups might not have been adequate to identify all relevant touchpoints. However, we were unable to evaluate the effectiveness of focus groups versus interviews in generating touchpoints due to limited touchpoint data. The generation of touchpoints and interview analyses varied across studies and was not always conducted in a systematic way. This could be due to the lack of guidance in the PoCF EBCD toolkit (among others).
For codesign to be successful in healthcare there must be cohort retention and a reconfiguration of the power dynamic between the service users and service providers. Codesign studies with formally engaged and funded facilitators are more likely to maintain momentum, engage and retain participants and generate improvement priorities.6 Although the majority of included studies used a facilitator, facilitator training was apparent in just over half. Facilitation is particularly important during the codesign stages (stages 6 and 7) as they are pivotal for successful EBCD as it is during these stages that improvement priorities are set and activities are designed. However, although all completed EBCD studies included stage 6, only half of the studies completed EBCD to stage 7; often reporting that service providers experienced difficultly balancing EBCD with work commitments. Consistent with previous reviews,6 8 the authors cited a lack of funding, support and time as barriers to codesign workshops and teams. Although participant views on involvement in EBCD were generally positive with service users reporting a more equal power dynamic than exists in the care relationship,27 the significance of these stages should not be understated as they allow for power relations between service providers and service users to equalise over time. By omitting workshops and/or small codesign teams, the voice of service users is less likely to be heard, and service providers remain expert providers rather than working as partners in a codesign process. Therefore, emphasising to participants the flexible nature of attendance and level of participation in the stage 7 might enhance involvement and reduce dropout.34
There was a lack of consistency in the reporting of EBCD projects which may be due to no standardised reporting guideline. Many studies failed to report the project outcomes (ie, touchpoints and planned improvement activities) and recruitment and dropout rates or ability to maintain participants when transitioning from experience gathering to codesign phases. Therefore, it is challenging to identify how the varied use of EBCD affected its success. Often projects were reported across multiple publications, and published data were limited and needed to be supplemented with reports in the grey literature to understand the method. Future reporting should include adequate detail so the reader can evaluate quality.
Our review demonstrates that EBCD has predominantly been used for service improvement in local settings. With the increasing expectation of service user engagement in healthcare, we recommend explorations of extending the use of EBCD to the development or redesign of healthcare policy. This would require adequate resourcing and the involvement of healthcare executives and policymakers throughout EBCD, especially during stages 6 and 7 of the process.
In light of the increasing recognition of engaging consumers and end-users in research design, EBCD could be a useful method for designing complex research interventions36 37 and maximising both person-centredness in healthcare and the likelihood of successful use.1–3 Within the Medical Research Council complex intervention framework,38 the EBCD method could be used to design components of a complex intervention as a phase I study, which can subsequently be tested in phase II–IV studies. The authors are aware of only one study where EBCD was used to design a complex intervention to improve breast and lung cancer services. Tsianakas et al 39 reported that although the touchpoints were shared across diagnoses, they translated into improvement priorities that were specific to the healthcare service. This emphasises the importance of stages 6 and 7 where service users and service providers discussed priorities for improvement.
Strengths and limitations
As far as the authors are aware, this is the first systematic review to evaluate the use and reporting of EBCD for the design or improvement of healthcare services. This complex review was informed by multiple frameworks (ie, PRISMA,16 Donabedian model,17 Guidance for assessing action research proposals and projects 19) presenting a comprehensive overview on this increasingly used method. However, publications relating to codesign activities likely exist in design or codesign journals and the grey literature was not abstracted to the major healthcare-related databases used in this search. Given this review is limited to the published literature, we recognise that some publications may have been missed.
Key recommendations and rationale
According to the findings of this review, several recommendations have been outlined in box 2 in relation to future use, reporting and use of EBCD studies. While we recognise the resource, time and engagement-related feasibility issues of conducting EBCD with 100% fidelity, at least two phases are required to ensure that any codesign is based on the experiences of service users and service providers. First, experiences should be scoped via site observations (stage 1) and collecting individual experiences of service users and service providers (stage 2). Second, the design of improvement activities needs to be a collaborative effort between service users and service providers based on the data collected in the first phase, preferably using codesign teams (stage 7) or a more accelerated approach during the joint workshop. However, we would argue that the exclusion of any EBCD stage would mean the minimum requirements to be considered EBCD were not met and could compromise study quality.
Key recommendations
Recommendations for using experience-based codesign (EBCD)
Prefer individual interviews over focus groups when gathering experience data from stakeholders.
Consider supplementing service user experiences with those of service providers during the joint workshop to minimise information asymmetry.
Limit the time between information gathering phase and codesign phase to minimise the risk of dropout.
Recommendations for reporting
Provide an adequately detailed report so the reader can evaluate quality. Common areas lacking information in the report include:
The relationship between the researchers and participants.
Details on project management.
How the project was funded and supported.
The length and timetable of the project.
List outcomes for each phase of the project (ie, touchpoints and improvement activities) and dichotomise them as locally relevant or generalisable.
Publish a complete EBCD results paper and refer to the published research protocol (if relevant).
A better way to improve feasibility could be the adaptation of EBCD activities to limit resource use. For example, video cameras can provide a means of obtaining comprehensive observation data. However, researchers should strike a balance between objectivity and engagement. By removing the presence of the researcher observer, the project would be less visible to the service resulting in a lost opportunity for engagement. Irrespective of the method of observation, ethical considerations remain and researchers must consider additional consent and privacy concerns, as well as having a clear analytical plan.40 Individual interviews might prove more resource effective than focus groups in terms of time to arrange although it is unlikely to reduce analytical time.35 Individual interviews with service providers could improve stakeholder commitment in the project and minimise dropout, particularly in the transition between the experience gathering and codesign phases. Irrespective of the method, the facilitators should be well trained in the approach being used to ensure the best outcomes from the project and participants.
To ensure the EBCD process is representative of all stakeholder views, trigger films should be supplemented with data from the service provider analyses after the feedback events. Feedback events (stages 4 and 5) serve to emphasise service user and service provider autonomy by allowing self-censoring aspects of their interviews or trigger films and correcting misinterpretation of their data, similar to a member checking process.12 Included studies often reported issues with maintaining their participant cohort from the experience gathering phase to the codesign phase. The time to complete the EBCD process and gaps in moving from stage to stage need to be as short as possible to overcome issues relating to transient participants (particularly unwell service users), high workforce turnovers and other improvement activities detracting attention from EBCD.
Reporting is particularly important as there is variability in the use of EBCD in these projects, and adaptations often occur as the project progresses. Although more generic reporting frameworks exist for quality improvement work in healthcare (eg, Standards for Quality Improvement Reporting Excellence II41), it appears that no studies are using this guide to report EBCD. An EBCD-specific guideline would improve the quality of reporting and would ensure studies are easily understood, comparable and able to be replicated.42 Such a guide could also serve to inform the design of EBCD projects. Until an EBCD reporting guideline is established, researchers need to publish adequately detailed reports and should consider publishing a protocol paper prior to conducting the study38 followed by one EBCD publication once the study is completed.
Conclusions
When conducted well and properly resourced, EBCD might enable effective codesign. EBCD is a useful tool for service redesign and has potential to be used for design of interventions in the research or policy development setting. A reporting guideline needs to be established to encourage researchers to conduct and report EBCD projects in a consistent manner, comparable with other research which would enable replication.
References
Footnotes
Twitter @@rayychan
Contributors RJC and TG conceptualised the review. RJC, TG, AB and LT developed the protocol which was reviewed by all authors. LT conducted the search and all articles were screened and LT, TG and RJC reviewed the full text to determine inclusion of studies. Data extraction was conducted by LT and checked by HYD. All authors completed quality appraisal and contributed to and approved the manuscript.
Funding This review is supported by the School of Nursing, Queensland University of Technology.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available in a public, open access repository.