Exploring and Understanding the ‘Experience’ in Experience-Based Codesign: A State-of-The-Art Review

Abstract

Enhancing user and provider experience are central tenets of value-based healthcare. Gaining access to personal and distinct experiential knowledge is the first stage of an experience-based codesign (EBCD) approach, underpinning the second stage of codesigned improvement: the codesign itself. This state-of-the-art review synthesised the evolving scope and nature of methods to gather experiential knowledge reported in the EBCD literature. Fifty-three of 64 (83%) scholarly EBCD articles reviewed were published since 2017. Methods are evolving to promote inclusivity of diverse user groups and move more rapidly to codesign. However, omitted steps in the methodology undermined fulfilment of core principles of the EBCD approach which may diminish its value as an accepted form of codesign. Experiential knowledge is crucial for designing user-centred health care. The challenge lies in making healthcare experience methods accessible. This review provides guidance on key steps in the first stage of the EBCD approach and modifications that may overcome barriers while upholding core principles and meeting the objectives of the inquiry.

Introduction

A growing number of participatory approaches co-opt those with lived experience of receiving health care, together with those providing health care, to help design more user-centred outcomes (Masterson et al., 2022). Masterson et al. mapped concepts and terms to common approaches, including co-production (e.g., Ostrom, 1996; Realpe and Wallace, 2010; Whitaker, 1980), co-creation (e.g., Vargo and Lusch, 2004), collective value creation (e.g., Alford, 2014; Ramirez, 1999), and experience-based codesign (EBCD) (Bate & Robert, 2007). EBCD is undertaken in two stages. In the first stage, EBCD aims to understand in depth, the experience of users and providers of the service and it is this first stage that distinguishes EBCD from other codesign or coproduction approaches (Masterson et al., 2022). In the second stage, codesign of process and efficiency reform are undertaken, but are subject to, and intertwined with, the identified priorities for improvement of experience identified in stage one. EBCD may be used in many disciplines, but its application to healthcare is our focus. The emphasis on understanding and appreciating the subjective, personal feelings and experience of users (often patients and carers) and providers (often frontline health practitioners), makes EBCD an important methodology for value-based healthcare. Bringing the knowledge of both users and providers to the table at the outset ensures codesigners of improvement understand the overall experience of a service, and build that knowledge into future systems (Bate & Robert, 2007). As the first stage of EBCD is a point of difference from other participatory healthcare improvement methods, it is important to understand the principles and methods that assure the integrity of the approach.

Lived experience of healthcare can offer a window into the knowledge, emotions, perceptions, sense-making, and actions of an individual during an instance of care or over a care journey. Experience does not ‘appear’, it is lived through (Gullion, 2021) and can be influenced by many variables, including family, friendships, culture, finances, age, grief, joy, language, expectations, and perceptions of helplessness, empowerment, ability and disability, and wellness. An individuals’ experience of healthcare makes an impression that can give it lasting importance (Gadamer, 2004), potentially influencing the way they access health care into the future. Words and stories convey experience, recreating life as it is remembered and as how it is made sense of (Churruca et al., 2020). Healthcare experience as a patient, family member or carer, is deeply personal; a time when an individual’s health, safety and dignity may be compromised, and exigent life decisions may be made. In contrast, the meanings that healthcare providers give to their work are often engrained in a complex interplay of empirical evidence and clinical guidelines, and personal, social, political, and institutional tensions; these can be uncovered by observing their work and asking them about their practice. Deep understanding and accurate representation of experience assures the definitive moments that users and providers value most are prioritised during codesign. Four principles were espoused by the founders of EBCD to guide the selection of aligned research methods for exploring and understanding experience (Text box 1).

Text Box 1. Principles for exploring and understanding experience in EBCD; source Bate and Robert (2007)

Four principles for exploring and understanding user and provider experience in EBCD

Immersive understanding: sought in ethnographic paradigms.

Emic understanding: involves taking the view of the insider and calls for thematising and structuring that is open or grounded in the lifeworld of the subject.

Reflexivity: requires the inquirer to consider their own biases and beliefs and mitigate against these skewing others’ experience.

Empathy: requiring the investigator to consciously take on the role of the naïve observer; listening, inquiring, and imagining what it is really like for the participant.

While flexible application is a feature of EBCD, proponents claim its replicable process and ontological and epistemological roots enable it to stand alone as an accepted Quality Improvement methodology (Blackwell et al., 2017). Ontologically, phenomenology is a philosophy of experience. Epistemologically, research methods for direct contact, observation of the environment and for gathering personal accounts must be sensitive to the ethical, social, and systemic challenges of user and provider cohorts, honour the authenticity and generosity of experience shared, and apply analytical frameworks that place the words and stories in the larger social, political, or institutional issue. The intent of the experience discovery process is to appreciate how users make sense of their experience or construct meaning, rather than their individual views or perspectives (Bate & Robert, 2007). Bate and Robert proclaimed the specialist skills in ethnography and contextual inquiry required for these activities: “Without these methods, what often poses as experience research is actually little more than a conversation that anyone might have had, and words and stories without analytical frameworks do not speak for themselves,” (p.32). There may be many ways to explore, understand and represent experience, but in research philosophy, commitment to common principles through alignment of methods to ontological and epistemological roots and replicable process are important.

Methods in the First Stage of EBCD

In the first stage of EBCD there are four steps:

1. Initially, the researcher or non-participant undertakes observation of the clinical environment to understand how the interface between the users and providers is shaped (Bate & Robert, 2007).

2. The researcher conducts interviews with providers to contextualise the picture of service provision and some of the pressure points.

3. Interviews with users of the service are undertaken to understand how people make sense of their experience of that service and the points in the care journey that need most attention.

4. During data analysis, shared narratives of definitive moments, termed ‘touchpoints’, crystallise what it is like to be cared for and work within a particular service.

Identification of touchpoints is considered central to EBCD (Bate & Robert, 2007) as they can act to reduce residual power gradients and become the foundation for patients, family, caregivers, and staff, to undertake codesign of improvement in the second stage, ostensibly as equals (Blackwell et al., 2017). Traditionally, a film that portrays user experience is played to providers in conjunction with analysed touchpoints (Bate & Robert, 2007; Donetto et al., 2014) to draw them from their clinical orientation around objective process pathways into the lived experience of users to ‘trigger’ momentum for change. While methods may vary somewhat, the intention of touchpoint analysis is to achieve a picture of the whole user journey under scrutiny, with the component parts clear from the user and provider view, so that both have a shared vantage point. The first stage of EBCD thus ensures the experiences of care (subjective pathways) are known and can be interwoven with process improvements (objective pathways) in the second codesign stage, that might otherwise have focussed primarily on efficiency and effectiveness (Figure 1).

Figure 1.

The two stages of EBCD: Subjective knowledge of ‘experience’ gained from Stage 1 is interwoven with process knowledge in Stage 2 to codesign user-centred improvement.

Purpose of This Review

This review aims to understand the scope and nature of methods to gather experiential knowledge reported in the EBCD literature, and how these methods align with the core principles of the EBCD approach.

Since its inception in 2005, two reviews (Donetto et al., 2014; Green et al., 2020) found there has been variation in the way the EBCD approach has been applied and reported. Both reviews found non-participant observations had been underused. Existing qualitative data had sometimes been used to represent user experience, and Green et al. found archived video had supplemented or replaced filmed user interviews to create a trigger film. While these reviews determined EBCD was still considered an engaging and worthwhile approach, the cost, length of time to complete the whole EBCD approach (estimated at 9–12 months), and human resources required, had proven challenging, driving modification of methods (Donetto et al., 2014; Green et al., 2020). For example, use of existing qualitative data had moved project teams more rapidly to codesign. This ‘accelerated’ EBCD (AEBCD) approach (Locock et al., 2014) was considered to have provided ‘good enough’ (Green et al., 2020) representation of the cohort under study. Democratising healthcare so that users influence change that matters to them may require technocratic research processes to be modified; this may be an acceptable trade-off, particularly to include diverse and sometimes marginalised user experience (Williams et al., 2020).

The emphasis on subjective experience is relatively new in healthcare research, and methods to include diverse users in small or large-scale healthcare improvement where they feel safe and can most meaningfully impact change are evolving (e.g., Francis-Auton et al., 2023). It was unclear to the authors how EBCD methods were currently being applied and on what basis methods might be adapted to meet the objectives of the inquiry as well as the needs of participants, all of which contribute to the integrity of an accepted and replicable research approach. A state-of-the-art literature review is considered a helpful strategy for tracing the historical development or turning points of thinking in relation to a body of knowledge (Barry et al., 2022). We therefore elected to undertake a state-of-the-art review of peer-reviewed EBCD articles to create a subjectively informed summary of contemporary thinking about EBCD methodology and inform our implementation of a robust EBCD approach (Cheek et al., 2023).

Methodology

We applied a state-of-the-art literature review method (Grant and Booth, 2009; Barry et al., 2022) that incorporated 1) a systematic search strategy, to identify peer-reviewed studies that specified the use of EBCD, and 2) an interpretive synthesis, of the methods used to understand and represent experience beside the four principles espoused by Bate and Robert (2007) to reflect evolving knowledge in this field. Guided by Barry et al.’s (2022) approach to state-of-the-art reviews, our synthesis involved becoming familiar with the literature noting similarities and differences across articles referring to core principles, discerning drivers for modification of methods and where adaptations or omissions veered from the principles espoused, noting exemplars that may be useful for others and future directions of research.

We searched Cinahl, Medline OVID, PsychInfo and Scopus databases for titles or abstracts that specified the EBCD approach, using the search terms ‘experience based codesign’ OR ‘experience-based codesign’ OR experience based co-design OR experience-based co-design. The search was conducted in January 2023, and identified 513 articles for consideration, after duplicates were removed. We reviewed abstracts and included empirical articles in English; no time or research field limits were set. We excluded non-empirical papers (e.g., protocols, editorials, commentaries, reviews, conceptual papers) and evaluations of EBCD studies as we specifically wanted to understand the methods used and the rationale for adaption, key inclusions in peer-reviewed studies. Full text review of 175 papers was conducted independently by six authors (NH, EA, LVB, MS, NR and RCW) and data was extracted into a pre-designed pro forma in Microsoft Excel. Each of the papers was then double reviewed by two authors (EFA and CC). Conflicts were resolved by EFA and CC through discussion based on the inclusion/exclusion criteria, yielding the final review set of 64 papers. The Prisma flow chart (Page et al., 2020) summarises the results of the search process at each stage (Figure 2). As our intention was to review and describe the current state, a quality assessment of reviewed papers was purposefully not undertaken (Grant and Booth, 2009).

Figure 2.

Overview of results of search process at each stage.

Results

Appendix 1 shows the 64 full-text papers that were examined and included in this review, describing 60 studies (Isobel, Wilson, Gill, & Howe, 2021a; Isobel, Wilson, Gill, Schelling, & Howe, 2021b; Blackwell et al., 2017, 2018, 2021; Waroonkun 2019, 2020, separately report aspects of the same EBCD project). The methods used to collect and analyse non-participant observations, user, and provider interviews, to generate touchpoints and to represent experience, were extracted. Where explicit, methodological challenges and adaptations used to overcome these challenges in completing stage one of EBCD were also collated. EBCD first appeared in the peer-reviewed literature in 2012 (Figure 3), used predominantly in studies in the United Kingdom (UK), (n = 28). Since 2016, there has been wider international application with studies in Australia (n = 13), Canada (n = 4), United States (n = 4), Netherlands (n = 3), Spain (n = 2), Thailand (n = 1), Sweden (n = 2), Norway (n = 1), South Korea (n = 1), South Africa (n = 1). All studies adopted EBCD with the intention of involving service users and providers in improvement.

Figure 3.

Growth in publication of peer-reviewed EBCD studies.

EBCD has been reported in healthcare settings except for Girling et al. (2022) who aimed to improve the outcomes of young people with mental health problems in forensic services, and de Waardt et al. (2021) who designed a survey to collect public perspectives on implementation of the new Compulsory Mental Health Act. EBCD was mostly undertaken in acute hospital settings (Figure 4), as a standalone approach and in association with other methodologies. For example, Baltaxe et al. (2022) combined EBCD with iterative Plan-Do-Study-Act cycles to deploy a codesigned prehabilitation service, and Knowles et al. (2018) combined AEBCD with a ‘future workshop’ approach.

Figure 4.

Area of application of EBCD in papers included in this review.

Adherence to all four principles of immersion (through non-participant observations), emic understanding (through application of ethnographic and grounded methods), empathy (through non-participant observations and user and provider interviews) and reflexivity (through non-participant observations) was not evident in the first stage of EBCD in all studies. Overall, those papers that report exploration and understanding of experience using all of ethnographic observations and user and provider interviews as the first stage of EBCD were published in earlier papers; examples include Tsianakas et al. (2012a), van Deventer et al. (2016); Blackwell et al. (2017, 2018, 2021), and Raynor et al. (2020). The majority of papers omit at least one of non-participant observations, user, or provider interviews. Film was most often used to represent user experience; film creation ranged from a compilation of actual filmed interview snippets to a whole additional production using film producers, script writers and actors.

Non-Participant Observations

Thirteen out of 60 studies conducted observations of the care interface, ranging from 2 to 80 hours, with some studies, Blackwell et al. (2017), Raynor et al. (2020), Tsianakas et al. (2012b) conducting observations for significantly longer periods of time (150 hrs, 190 hrs, 219 hrs, respectively). Two, Gustavsson (2014) and Robinson et al. (2020), out of the 13 studies that undertook observations did not report the duration of observations. Where conducted, these were generally described as being used in conjunction with interviews to develop touchpoints. Few studies provided justification for the omission of observations altogether, but those that did cited limited resources – time or human (Table 1). Observations may also have been omitted for other legitimate reasons. For example, projects that were designing online applications or surveys were arguably designing a new user/provider interface. Those published from 2020 - 2022 may have been impacted by limited access to clinical environments during the COVID-19 pandemic.

Table 1.

Methodological Issues in the First Stage of EBCD Raised by the Authors of Reviewed Articles.

Methodological issue	Papers	Health domain	Adaptations	Limitations/risk identified
Project team’s lack of existing knowledge about subject area or research methods	Blackwell et al., 2017; Wright et al. 2018; Wright et al. 2021	Palliative care Palliative care	Conceptual framework developed for project team to identify and classify key experiences Support for staff moving from clinical to research role	Research team lack of understanding of health domain Clinical staff lack of understanding of research
Difficulty recruiting user participants	Tsianakas et al., 2012a; van Beusekom et al. 2021; Cooper et al. 2016; Crowther et al. 2022; Kynoch et al. 2022; Girling et al. 2022	Cancer, mental health, primary care – asthma, postnatal care Youth offenders in community detention facility	Interviews with forensic staff and researchers/epistemic experts Meta-ethnography (qualitative synthesis) of scholarly articles reporting offending youth experiences	Lack of diversity of users Change oriented around larger clinician group
Limited capacity of users to participate Service users sometimes act as gatekeepers for contacting informal carers	Gustavasson, 2014; Heerings et al., 2022	Neonatal care Supported living – older people/intellectual disability/mental illness	Neonates - recruited parents instead Intellectual disability/long term mental illness – clients, carers and epistemic experts informed content for 42 trigger films, engaged inclusive theatre company to produce films Used with non-participant observations to inform touchpoints. Recruited family support workers to provide their experience.	Over emphasis on parent experience Lack of diversity if only able users recruited.
Filming vulnerable groups - offers no protection from identification and preserves testimony indefinitely.	Larkin et al. 2015; Blackwell et al., 2017	Mental health Palliative care	Did not show film Did not film interviews
Personal accounts powerful and potentially overwhelming to present, or may be embarrassing to usersPotential distress of users	Larkin et al. 2015; Springham and Robert, 2015; Isobel et al. 2021a, 2021b	Mental healthMental health	Careful selection to retain core principle but not overdramatise eg did not include data about serious eventsOngoing user peer support groupSafety plan incorporated agreement not to share traumatic or distressing details without prior consent of focus group members.
Participants not open to filmed interview	Boudioni et al. 2015Cooper et al. 2016; Kenyon et al. 2016; Blackwell et al., 2017; Kynoch et al. 2022	Cancer careMental healthCaesarian surgeryPalliative carePalliative care, postnatal care	Film constructed by production companyAudio-recorded participant interviewsFilmed narrative interviews an optional extra	May not represent participant voiceStaff declined filmed interviewsRecruitment still limited
Difficulty getting patients with exigent immediate issues to think collectively	Moser et al. 2021	Cancer care
Lack of provider participants	Heerings et al. 2022	Supported living – older people/intellectual disability/mental illness
Time-consuming and resource-intensive process.	Robinson et al. 2020; Kenyon et al. 2016; Chudleigh et al. 2022; Cooper et al. 2016; Kenyon et al. 2016; Crowther et al. 2022Springham and Robert, 2015	Surgery - Ear, nose, throat, CaesarianParents of babies receiving positive newborn bloodspot screen resultsMental health; Caesarian surgery;asthmaMental health	Accelerated EBCD using existing qualitative dataFilm archives for trigger filmWeb-based methodsDid not complete non-participant observationsBuilt extra time into project to allow for filming of user interviews	Better when there are verbatim accounts.Local participants may not relate to itRelied on project teams’ previous experience of service
Organisational changes (e.g., staff turnover).	Girling et al. 2022; Kynoch et al. 2022	Youth offenders, postnatal care
Tension between service providers/users	Kenyon et al. 2016	Caesarian surgery	Collaboration, engagementHighlight the emotional highs as well as lows in touchpoints	Space of time between immediate crisis and EBCD activity
Lack of equity if not filmed like users	Blackwell et al., 2017	Palliative care	Service providers interviewed after project – film retained for teaching	Refusal

Interviews About Experience

Of the 60 studies, two omitted user interviews, two cited unreported user interviews, and seven omitted provider interviews. Accelerated EBCD was proposed by Locock et al. (2014) whereby film archives of existing user experience were used to halve the time taken for the overall EBCD cycle from 12 to 6 months. The authors retained non-participant observations and in-depth interviews with local providers. In the papers reviewed, three studies used existing literature and/or unreported user experience to get to codesign, omitting all of non-participant observations, user and provider interviews in their study: Barker et al. (2021) developed ‘touchpoints’ from findings of previous unreported qualitative interviews of users of a similar service. The touchpoints were then presented for discussion at stakeholder (mostly health provider) feedback events, and analysis of the minutes of these meetings were used as a record of stakeholder experience; Morris et al. (2021) used existing evidence of primary care patient experiences to inform ‘trigger discussion’; Walter et al. (2022) drew from unreported user experience to inform codesign. Knowles et al. (2018) used an AEBCD approach drawing from an archive of 38 previously collected narratives. The authors justified their AEBCD approach as allowing a more rapid process and proposed this approach might also be less threatening to local professionals than an approach that critiqued ‘their’ service.

Where experience knowledge was gathered from user and provider cohorts, the size of the cohorts varied from n = 3 to n = 63. A range of methods have been used to capture experience including interviews (face-to-face, telephone and on-line), focus groups/workshops, non-participatory and participatory observations, survey, archived film, archived interviews, photographs, service user diaries, emotional mapping exercises, and creating personas. Recruitment of user participants was cited as a challenge (Cooper et al., 2016; Crowther et al., 2022; Girling et al., 2022; Gustavasson, 2014; Heerings et al., 2022; Kynoch et al., 2022; Tsianakas et al., 2012b; van Beusekom et al., 2021), that sometimes led to small sample numbers or omission of participant data (Table 1).

Over one third (22) of the 60 studies that conducted user interviews published from 2012 through to 2022, filmed the interviews. Authors reported reluctance among participants to be filmed (Table 1). Three studies noted that the use of filmed interviews afforded users little anonymity, the legacy of the film preserved their testimony and so dissemination needed to be carefully controlled, or the content of the film was considered overwhelming and a potential source of distress. This was most pronounced for mental health cohorts. For example, Larkin et al. (2015) reflected on ethical considerations, including consent, anonymity, confidentiality, and data ownership, mental health stigma, and the potential for re-traumatisation following retelling or reliving difficult experiences. Springham and Robert (2015) raised the issue of power as a particular difficulty within mental health projects, where professionals can impact upon an individual’s civil liberties. These researchers carefully censored their inclusion of over-dramatic experience that may cause harm, while Isobel et al. (2021a, 2021b) developed a safety plan before focus group interviews, whereby traumatic events were not shared with the group unless explicit permission of the group was received at the time. Due to the length of time between interviews and implementation of improvement, sustaining engagement with users and providers was also challenging.

Data Analysis and Touchpoints

Of 60 studies, 52 articulated the frame of analysis used - thematic analysis (34 studies), content analysis (6 studies), grounded theory (2 studies), inductive analysis (3 studies), framework analysis (1 study), theory-informed analysis (3 study), interpretive phenomenological analysis (2 study), and a combination of content analysis and thematic analysis (1 study). The level of detail regarding the analysis process varied with some studies noting the type of analysis only, often without reference to methodological techniques, while others included details in line with the Consolidated Criteria for Reporting Qualitative Studies (COREQ) criteria (e.g., who undertook the data analysis, description of coding tree, software used). Eight out of 60 studies did not clearly articulate an analytic approach. Some authors did not mention data analysis at all, while others provided vague details. For example, Hackett et al. (2018) p. 2, stated that “participants’ experiences were gathered and analysed for touchpoints”.

The term ‘touchpoint’, the critical moments experienced in relation to experience of the service, was used by most, but not all, studies. Hill et al. (2018) referred to these as ‘contact points’, and van Beusekom et al. (2021) as ‘significant moments of interaction’. However named, the process of identifying and prioritising areas of concern for improvement were variously considered part of stage one, experience gathering, or stage two, codesign, and may therefore be reported in subsequent reports. Where touchpoints were reported, the process for selection included building experience or emotional maps (e.g., Hackett et al., 2018; Gustavsson, 2014; Brady et al., 2022) or playing all interviews to the investigating team or a Steering Committee who chose touchpoints (Cooper et al., 2016; Heslop et al., 2019). Some presented summaries of data, participant quotes, or compiled film for participants to choose or validate touchpoints (e.g., Boudioni et al., 2015; Hackett et al., 2018; Springham and Robert, 2015), and in others the selection process was unclear.

User experience was represented by a ‘trigger film’ compiled from snippets of filmed interviews in 21 studies. The film compilation was generally directed by the researcher or non-participants where the snippets were chosen to represent the selected touchpoints (e.g., Boudioni et al., 2015; Springham and Robert, 2015). Three studies employed non-researchers such as journalists to conduct interviews, and two employed theatre directors, playwrights, and actors to create trigger films. For example, Heerings et al. (2022) employed a theatre company to produce the trigger films in redesigning care for long term care settings for people with intellectual disabilities or serious mental illness, as gathering experience posed significant design challenges. The project team followed a systematic process to choose touchpoints based in their observation data, and reflective of the interview data gathered from carers and providers. It was reported the playwright felt participants’ narratives were too lengthy and lacked poetic use of language, so changed most of the participant phrasing to engage the audience with greater impact. The researchers revised the texts with the playwright to prevent a loss of thematic content.

Methodological Adaptations

Thirty-seven of 64 papers described adapted methods, summarised in Table 1. Common limitations remain an inability to recruit users representative of the whole cohort (lack of diversity and numbers) and an inability to recruit providers (lack of time, attrition). Authors attributed the requirement for filmed interviews to have limited user participation and Crowther et al. (2022) felt the subsequent lack of user voice may have oriented change around the larger provider cohort. As well as limiting recruitment, filming of interviews was thought to be a potential cause of distress, and risked identification of users beyond the current project (Table 1). User experience was represented in other ways including having a patient talk about their experience at a codesign event (Taylor et al., 2021), creation of hypothetical patient cases (Hill et al., 2018; Knowles et al., 2018), story boards or emotional maps (Bradway et al., 2020; Robinson et al., 2020).

There were methodological limitations in the context of mental health, cancer and palliative care, and paediatric cohorts. Issues regarding the potential for (re)traumatisation, consent, particularly regarding the legacy of film recordings, privacy, confidentiality, poor recruitment, power imbalances, and data “ownership” were raised. Adaptations were used to overcome issues encountered. For example, Coy et al. (2019) explored the emotional/experiential aspects of moderate to severe burn injuries in children and interviewed parents only so as not to retraumatise the children. Larkin et al. (2015) chose to film provider experiences to address the inequity in only filming users; the authors argued that service-users (mental health patients) and carers were potentially disadvantaged when numerous nursing staff and managers have experiences that illustrate the dominant structures and discourse of the health system. The authors acknowledged that filming users offered no protection from identification, with film indefinitely preserving their testimony, their state of health, how they chose to identify themselves (e.g., as a user), and the experiences they had. Springham and Robert (2015), identified the need for peer support for mental health participants in the overall EBCD process.

Some authors reflected on the quality of relationships built between project teams and participants for genuinely collaborative research. For example, Carr et al. (2021), developed a physical activity and lifestyle program with Australian Aboriginal families living in remote communities. Non-Aboriginal researcher efforts to learn and converse in the languages spoken in each community were perceived to have been well-regarded and to have strengthened partnering relationships and trust. However, the authors found that genuinely collaborative EBCD takes time and can be costly with long distance remote travel, accommodation costs and wages.

Discussion

This review builds on previous reviews of EBCD (Donetto et al., 2014; Green et al., 2020) and coproduction approaches (Masterson et al., 2022) by focussing on one stage of one approach to scrutinise in more depth methods to explore and understand user and provider experience, and the constraints that have driven adaptation. Exploring human experience in dynamic health environments can be complex. The steps of non-participant observations of the user/provider interface, user interviews, provider interviews, and associated data analysis in the first stage of EBCD musters deep experiential knowledge but it is also resource intensive and contributes to a lengthy improvement approach. Recruiting and retaining participants for the whole program of work has sometimes been difficult. Modification of methods to gather experience knowledge may be necessary to promote inclusion, to overcome limited participation, to move to codesign more rapidly, or to adapt to different contexts. While EBCD is known for its flexible application, greater adherence to core principles is required lest there is a misconception that ‘anything goes’, potentially undermining EBCD as a credible form of codesign. The challenges lie in making EBCD accessible to both researchers and quality improvement teams; not all teams have available time and in-depth expertise in relevant phenomenological techniques and clinical knowledge, and improvement also needs to keep pace with complex adaptive health care environments. To provide guidance, we have summarised the current state of knowledge of the steps in the first stage of EBCD, some of the challenges encountered, and drawing from reviewed articles, made pragmatic suggestions for project teams undertaking this form of inquiry-led codesign that are in keeping with the core principles of the approach.

Non-participant Observations

In the original formulation, EBCD includes substantial non-participant observations of the care interface allowing researchers to explore and understand how providers “go about their business in real time” (Bate & Robert, 2007, p. 88). Consistent with reviews by Donetto et al. (2014) and Green et al. (2020), we found that non-participant observations remain an underutilised method: only 20% of studies in this review included non-participant observations. When they did, they ranged in duration from 2 to 219 hours; it was unclear how the quality and quantity of non-participant observations were determined, or how the observations were used or impacted the non-participant standpoint.

Echoing Donetto et al. (2014) and Green et al. (2020), we emphasise the importance of conducting non-participant observations for understanding and contextualising experience. Going further, we suggest that observations offer rich insights into how and why things work (or not), and how they might be judged or redesigned, and satisfy all four core EBCD principles: ‘immersive’ and ‘emic’ understanding, ‘empathy’ and ‘reflexivity’. This process underpins deeper understanding of the experience of users and providers. Even for time-pressed project teams within the healthcare sector, this work may illuminate tensions that are disclosed when exploring user experience, expectations, and provider objectives. Non-clinical project teams are afforded valuable insight into the healthcare domain. When reported, this information will also help those undertaking an accelerated EBCD approach to understand how the experience gathered may be applicable in another setting.

Provider Interviews

To recruit providers for interviews the provider group must value the improvement activity and/or be influenced by a creditable provider champion. As a relatively homogenous group, thematic analysis is well suited to most provider interviews and can be undertaken quite readily. However, frames of analysis are also useful in positioning the experience in the context of inquiry. An analytical framework may include the experience or process maps derived from non-participant observations or may be a framework more directly related to the reform purpose, such as work analysis, ergonomics, or health action process. Blackwell et al. (2017) provided an example of a conceptual framework developed specifically for their context of palliative care to guide a non-clinical team. Better reporting of frameworks will build knowledge in this area.

User Interviews

Recruiting users has proven challenging, perhaps exacerbated by the requirement for interviews to be filmed. The nature of some user cohorts, for example persons with intellectual disability, children, or reluctant participants, has driven adaptation of novel methods. Adaptations have included interviewing carers, epistemic experts, and reviewing existing literature to provide a picture of user experience. Where explicit, thematic analysis was predominantly used to analyse the user interviews in reviewed papers to categorise themes emerging from user data. In reducing data to common themes, where the sample is homogenous or data saturation is achieved, this form of analysis can be readily undertaken as demonstrated by its popularity. Other methods may be more suitable where the user group is more heterogenous, for example, in a diverse user participant group, the particular is of interest, or where the analytical team have identified axiology that may skew gathered perspectives in an unwanted way. Methods that anchor analysis in the lived experience of the individual may more appropriately satisfy the principle of ‘emic’ understanding and ‘empathy’. An example is constructivist grounded theory (Charmaz, 2017), where preliminary descriptive codes are assigned line-by-line to transcribed data, paying attention to the individual, life context and emotional tone, and ‘in-vivo’ codes are used where possible to preserve participant voice. However, this process can be quite lengthy. Interpretive effort is still required to situate how users construct meaning of their experience within the overall quality improvement agenda.

Narrative techniques can be used to weave together the crucial links between individuals, communities, and the language(s) they use (Falconi & Graber, 2019). Hermeneutical phenomenology is another interpretive method that more closely focuses on interpretations of meaning (Frechette.et al., 2020) and has also been used to understand the impact of illness or healthcare on individuals. Using this method, the crucial role of the researcher is to uncover the hidden meaning rather than that which shows itself (Suddick et al., 2020). Emerging qualitative techniques to explore the meaning-making process of vulnerable groups include Qualitative Vignettes (Cheah et al., 2023) and Rapid Qualitative Analysis (St George et al., 2023). In the latter, the use of action-oriented themes aims to deliver information for issues that require a swift response and may be used in conjunction with longer traditional methods. Development of techniques to identify and represent user experience more quickly is an opportunity for further research.

Research teams may have resources to undertake robust exploration of user experience and analyse and report the results that are not available to hospital-based quality improvement teams. Lived experience that is explored, understood, and conveyed in ways that readers can value as trustworthy makes re-use possible and ethically responsible. A novel example was provided by Girling et al. 2022; unsuccessful in recruiting young people who offend, the authors considered many different avenues for inclusion before conducting a meta-ethnography of qualitative research studies exploring young people’s direct experiences of mental health in youth justice services. The data included verbatim accounts and direct information about a broad range of experience, allowing identification of an initial set of touchpoints.

Methods that offer nuanced understanding of experience take considerable time and more specialised attention, which may not suit all projects or project teams. Those that have the resources ought to undertake and report these methodically so that they may inform others. Where the resources are not available to undertake robust exploration of experience, where participants are limited, or where there is a need to move to codesign more rapidly, an accelerated EBCD approach (AEBCD), may provide a more time- and resource-efficient cycle of activity.

Accelerated EBCD

AEBCD re-uses existing data that is reflective of the context. What constitutes ‘good enough’ in terms of representation of the current context needs clarification. When reporting qualitative studies, information including the characteristics of participants, what research questions generated the data collected, how the data was collected and analysed, will assist others undertaking EBCD to understand how well the previously acquired data accurately represents the socio-cultural features of the present population, or its limitations. Authors of AEBCD found the use of verbatim participant quotes in existing data helpful. Such flexibility in EBCD may accommodate a rapidly changing health service environment better, allowing researchers and participants to come together to reflect on an issue, and move forward to codesign in real time (Rapport and Braithwaite, 2020).

Purposeful inclusion of users, through considering carefully where their contribution best suits the overarching project aims given the context and research processes (Edelman & Barron, 2016), protects against ideological and moralistic inclusion of users, unnecessarily consuming their time and their potential exposure to research harm. It may be that moving directly to design is the best approach for the intended improvement project. If so, there are other codesign and coproduction approaches that do not include such emphasis on experience knowledge as a first, structured, stage (see Masterson et al., 2022).

Development of Touchpoints and Trigger Films

The methods used for developing touchpoints varied considerably among projects and was often obscure. The lack of shared methodology may reflect the lack of observation data. Workflow and processes contextualise touchpoints, a key output of the first stage of EBCD. Experience maps were used by some studies, where user and provider experience were mapped onto a schematic patient journey. This work achieves a picture of the whole user journey under scrutiny, with the component parts clear from the user and provider view, ready to move forward to codesign.

Filmed user interviews continue to be used as a humanising mechanism, drawing providers from their clinical orientation to the human life world of receiving care. This has been reported as exceptionally powerful in catalysing intention for change (Donetto et al., 2014). It is understandable that authentic filmed experience has been pursued, as conveying experience stories to others requires interpretive effort. Such interpretation has historically been considered ‘unscientific’, but this needs to be weighed against ethical and moral considerations. Limited recruitment, potential for unnecessary distress, identification, perpetuity of film and use outside of the current project were identified as risks by some authors of reviewed papers. Adams et al. (2015) reviewed perspectives of filmed narrative two years after the original EBCD project and found interpretive editing of film made patient narrative ‘fragile’; the legacy of film beyond the immediate project could subject it to changing organisational agenda. Films purporting virtue and life’s higher purpose are increasingly a strategy used in advertising to influence positive brand attitude changes (Hamby et al., 2022). The manipulation of participant data by persons outside the research team to create a film raises other important methodological and ethical questions. Clearly defining the purpose of the film, for example, whether to educate or to influence, ethical and moral considerations for the user cohort, and how to de-identify participants in the film (pixelating facial features, e.g.) might assist in deciding whether to pursue filmed real life, or whether simulation is preferable.

There are other ways to convey experience including hypothetical personas or stories derived from interpreted experience. Advanced technology offers alternatives to real people including avatars and other simulated characters, where true participant experience can be represented without identifying individuals. While a very exciting area for development, the stories and personas must be credible, and stand on accurate methodical techniques that can be scrutinised.

Strengths and Limitations

This inquiry was undertaken by a research team to inform the design elements of an EBCD project. Our perspective was heavily influenced by a research lens seeking robust research methodology. Some EBCD projects are undertaken by quality improvement teams in healthcare settings who may have limited access to specialist research skills and undertake more pragmatic design. That we limited our search to peer-reviewed articles means EBCD project reports in the grey-literature are likely to have been missed. While this was a limitation in the overall search, the subject of our review was the methodology used in capturing and reporting experience; it is expected that this methodology is reported in scholarly articles and that the peer-review process will assure core research principles are reported. The comprehensive nature of EBCD may make reporting all components difficult within the word limit of a journal article. It may be that each study in the experience-gathering stage may have been reported as separate qualitative studies. Nevertheless, if the abstract stated the study was conducted within an overarching EBCD approach, the paper(s) would be included in this review. Where there was more than one paper reporting the same study, the article(s) that described the first stage of EBCD was included, aligned with the subject of the review. Exclusion of evaluations occurred as the methodology for the first stage was often not explicit, but this may have eliminated important discussion about methodological challenges and strengths. We excluded articles that were not in English or those where the full texts were not accessible through library subscriptions which may have eliminated some works that warrant further consideration.

Conclusion

EBCD is a two-stage approach to system improvement where in-depth exploration and understanding of user and provider experience is the first stage, and codesign is the second. Our review sought to understand the ‘experience’ gathering component of EBCD. Inclusion of those with lived experience is increasingly mandated by health services and health professional bodies, to guide health system reform that better meets consumer needs. Growing recognition of the importance of healthcare that meets the needs of both consumers and providers is reflected in the recent rise in scholarly EBCD articles. User engagement must be underpinned by research principles that assures the integrity of their shared experience. For those gathering experience knowledge, there is opportunity to broaden and mature the methods used to accurately collect and represent experience and to more fully report this methodology in scholarly articles. Use of existing data that is trustworthy may alleviate resource constraints in other projects and allow improvement design more aligned with the fast pace of system change.