Using Situated Interviews to Engage Frail Older People in the Co-production of Improved Emergency Healthcare

ABSTRACT

Including frail older people in the development and improvement of healthcare is a topical issue and co-production represents a common approach. How to practice co-production effectively, however, remains challenging, particularly when including vulnerable populations. This paper provides methodological highlights from a project designed to improve care pathways for frail older people. The project applied a co-constructive approach to co-production using situated interviews. We make four recommendations for practice and two linked conclusions – that situated interviews represent a flexible and accessible method for engaging vulnerable populations and that conceptual clarity is essential to the delivery of effective co-production.

KEYWORDS:

• Frailty
• situated interviews
• coproduction
• emergency care
• healthcare

Introduction

Commitment to the co-produced delivery and improvement of healthcare services is currently a popular management practice that infers a particular approach to public service organisation within which the inclusion of service users is paramount (Radnor et al., 2014). The practical application of co-production, however, is not always easy (Batalden, 2016). The inclusion of vulnerable populations in co-production work represents one such challenge. This is a particular challenge in healthcare settings owing to the added uncertainty that being a patient can bring. The inclusion of frail older people in the co-production of healthcare services is a particularly pertinent and topical issue within healthcare organisations delivering care in nations experiencing ageing populations. In such countries, frail older people tend to represent a ‘small but challenging’ patient group owing to the multiplicity and complexity of the health conditions experienced and the potential for recurrent admissions that this can cause (Reeves et al., 2018). Consequently, frail older people have been considered ‘hard to reach’. Indeed, patient perspectives are largely lacking in the development of clinical scores and measures to assess frailty (Rahman, 2018). In order to show how frail older people can be included in the improvement of healthcare services this paper presents methodological highlights from a study that aimed to co-produce baseline data to inform quality improvement work addressing care pathways for frail older people accessing a busy emergency department in the UK.

The project discussed explicitly sought to include frail older people in providing a baseline for quality improvement work. As well as their hard to reach and often excluded status, the inclusion of frail older patients in this work was particularly important due to increasing critique of the clinical use of the term frailty (Grenier et al., 2017; Pickard, 2014; Warmoth et al., 2016). Frailty is a lay term that has been appropriated into clinical practice to aid patient centred care of typically older people with multiple co-morbidities (Cluley et al., 2020a). A growing number of studies have highlighted resistance to frailty among those so labelled; problematizing frailty as an unwanted and potentially harmful clinical term (Warmoth et al., 2016).

It is important to note that while frailty has a shared ‘lay’ meaning, clinically it is generally agreed that frailty has multiple manifestations due to a combination of factors rather than a single disease (Fried et al., 2001). Indeed, frailty is a condition and experience that is now increasingly reported in the UK. In England, 1.8 million people over 60 and 0.8 million people over 80 are living with frailty (Banks et al., 2019). Additionally, 65% of people over 90 are clinically determined to be frail (Banks et al., 2019). It is predicted that by 2030 the number of people aged 85 and over will increase by two-thirds, while the general population in England will grow by just 10% (Banks et al., 2019). While older people are generally considered to be a vulnerable group (Allen, 2017), frail older people represent the most vulnerable of this population group. Living with multiple co-morbidities can result in a situation whereby the individual is dependent on others and/or objects for support with basic tasks, such as using a frame to aid walking or having home carers to assist with self-care. Frailty is also generally experienced as cumulative decline. Individuals move from being fit towards severe frailty as their functional state deteriorates due to cognitive and physical impairment (Rockwood et al., 2005). Consequently, frail older people can be considered to be vulnerable for a number of reasons: they may not be able to care for themselves fully; they may not be able to comprehend easily what it means to take part in research; they may be coming towards the end of life; and they may not be able to cope with adverse situations.

To exclude the voices of frail older people from the co-production of healthcare delivery and improvement due to their ‘vulnerability’, however, could be detrimental to the provision of services designed for this population group. As stated, a number of studies have identified that for patients so labelled, frailty can be an undesirable label that is often resisted (Britain Thinks, 2015; Warmoth et al., 2016). Based on this, recommendations have been made to avoid the use of frailty in so-called frailty services (Britain Thinks, 2015).

To engage frail older patients our project specifically applied a co-constructive approach to co-production as a conscious methodological decision, using situated interviews to enact this approach. In this article, we provide a critical, reflective account of the experience of using this method, which is novel in this clinical context and with this patient group. In this way, the arguments we present here contribute to both the literature addressing co-production in healthcare settings and also the methodological development of the situated interview process.

Participants were asked to talk about their perceptions of frailty and their experience of healthcare provision for frail older people in order to provide baseline information for the hospital to use in further quality improvement work. Indeed, the methodological experience discussed here details the beginnings of a wider quality improvement journey regarding the improvement of care pathways for frail older people accessing emergency care in an NHS trust hospital in England. The hospital involved is committed to a patient centred approach to improvement. The baseline information is now complete, a formal report has been shared with decision makers at the hospital and a presentation of the findings is due. This information will be used by the hospital to inform further quality improvement work.

While the study was conducted in an NHS hospital in England, our findings are transferable to co-production projects addressing aging populations in other countries. The challenges associated with frailty are experienced similarly internationally (Gwythner et al., 2018) and co-production is an internationally used approach to healthcare service provision (Batalden et al., 2016). Two transferable concluding statements are made regarding the engagement of vulnerable populations in co-production work; first that situated interviews provide a flexible and accessible method for engaging vulnerable groups in the co-production process, and second that co-production requires conceptual clarity before use in order to ensure the delivery of effective co-production. Before the use of situated interviews is reflected upon, we first outline our specific approach to co-production in order to provide necessary context to our methodological choices and to ground the study in the co-production literature.

What is co-production?

Co-production has a varied theoretical background and is a term that is often used differently in different contexts. This has resulted in the development of a range of definitions and frameworks that can get lost in practice (Voorberg et al., 2015). In healthcare settings, co-production is an increasingly popular approach. However, it is also often used without conceptual clarity (Batalden et al., 2016) and is frequently used interchangeably with other concepts such as co-creation, co-innovation and co-design. This collective approach is often carried out without much consideration of what any of the terms really mean (Voorberg et al., 2015). This is problematic in itself in that these are not conceptually interchangeable terms, they each have their own theoretical background and underpinning principles. Consequently, Voorberg et al. (2015) call for increased attention to conceptual detail and clarity to ensure effective co-production practices.

Within public service management co-production is largely associated with the production of public value. Public value has been defined variously; however, it generally refers to the value that public services add to society or the common good (Moore, 1995). Public value, moreover, is often seen in terms of societal and also individual benefit (Meynhardt, 2009) perhaps explaining why the provision of public value and individual public value experiences through the co-production of services now represents a common approach to healthcare delivery, and indeed, an approach that is often implicitly assumed to be beneficial (Voorberg et al., 2015).

Conceptually co-production employs a public service logic that embraces value creation from the perspective of public service organisations, their service users and fundamentally, the interaction between them (Osborne et al., 2016). A number of scholars have attempted to clarify co-production to reflect the unique nature of public services by developing models, matrixes and frameworks (see Batalden, 2016; Cluley & Radnor, 2019; Osborne et al., 2016; Voorberg et al., 2015). In order to work towards a clear conceptualisation of co-production in public service organisations (PSOs), Osborne et al., (2016) provide a 2 × 2 matrix based on a combination of service management and public management theory. Here, co-production is defined as ‘the voluntary or involuntary involvement of public service users in any of the design, management, delivery, and/or evaluation of public services’ (Osborne et al., 2016). Key to their argument is the assertion that the public service user and their unique experiences are fundamental to the value co-production process. The 2 × 2 matrix depicts four types of co-production that can result from the fluid and changeable relationship between the service user and the service provider. These include pure co-production, where the service user and provider co-produce the service and its outcomes; co-design, whereby service users are consulted on issues of service design, implementation, and evaluation; co-construction of service experience, whereby the lived experience of the service user and the service experience interact to produce the overall lived experience of the service; and co-innovation where citizens and services come together to create innovative services. Common to these ideal types is the involvement of both the service user and the service provider. Indeed, the theory and practice of co-production and its associated practices is predicated on the inclusion stakeholder voices in the creation of public value experiences.

Recently, Osborne (2018) has suggested that conceptually co-production does not provide the best explanation of how public value emerges through public service delivery. Osborne (2018, p. 225) suggests the concept co-creation, defined as ‘an interactive and dynamic relationship where value is created at the nexus of interaction’. This theoretical progression is based on service dominant logic whereby value is thought to be co-created ‘though the combined efforts of suppliers, employees, customers, stockholders, government agencies, and other entities’ (Vargo & Lusch, 2008, p. 148). Pivotally, Osborne (2018) stresses the importance of lived experience in this relationship and the varying circumstances this will result in regarding the public value experience. We have further developed this concept of co-creation (Cluley & Radnor, 2019, 2020) to argue that public value is a heterogeneous experience that will be different for different actors depending on a diverse and changeable range of factors.

This ongoing debate is not to say that co-production is now a redundant term, rather it shows that co-production can mean different things and that various approaches to its practice can be adopted depending on the conceptual framework that is chosen. It can be seen, moreover, that co-production really is not something that can be meaningfully and effectively used without some explanation. It is important that the most appropriate form of co-production is chosen based on individual project aims. While we, the authors, have previously contributed to the theoretical and conceptual development of co-production and co-creation (Cluley & Radnor, 2019, 2020; Cluley et al., 2020b), this paper focuses on our experience of implementing such theories in our methodological practice. The point we want to stress is that when doing work that focuses on public value, as with all methodological choices, practitioners must be mindful of the variety of concepts and frameworks available and choose one that will best suit the task in hand. Engaging vulnerable population groups is one such task that requires the consideration of the various approaches. Indeed, it is this argument that informed the conscious approach to co-production adopted in the study reflected upon here. The engagement of frail older patients in a constructive approach to co-production forms the focus of this paper. We draw on our experience of using situated interviews to do this, focusing particularly on the benefits and challenges encountered. An overview of this study and the approach to co-production used are now provided.

The study

The study worked to two linked aims:

1. To explore how key stakeholders understand and make sense of frailty

2. To explore key stakeholders perceptions of healthcare provision for frail older people accessing emergency care.

As outlined, the purpose of the project was to provide baseline information regarding stakeholder perceptions of frailty and healthcare provision for frail older people accessing emergency care to support further quality improvement work in this area. The desire to provide baseline information detailing stakeholder perceptions was based on findings from other projects that suggest a ‘top-down’ regulatory approach to addressing frailty may not necessarily improve outcomes for frail older people (Gobbens et al., 2010) and that the ability of health carers to provide care that matters to this population depends on their understanding of their patient’s personal and social construct of frailty (Dworkis et al., 2016). In order to ensure that frail older people experience effective and efficient health care within emergency care settings; moreover, knowledge of how those involved make sense of and understand frailty was considered imperative.

Approach to co-production

Based on this starting point and the purpose of the project we adopted the co-construction approach to co-production outlined by Osborne et al. (2016). The co-construction approach to public value co-production focuses on the service system as a whole, including the service user’s lived experience of the service (Osborne et al., 2016). As Osborne et al. (2016, p. 647) outline ‘this is about how the service experience integrates with their overall life experience. It results partly in their personal experience and satisfaction with the service, but also more fundamentally in how the service experience impacts upon their own life at an emotional and personal level’. Important to this approach to co-production is the acknowledgement that lived experience will directly affect how service users experience and engage with services. Pivotal to the co-construction approach to co-production, moreover, is the belief that individual and collective lived experience is brought to the service, having the effect of shaping and impacting the service itself. It is this knowledge that the project specifically sought to access in order to provide a patient centred foundation for further improvement work.

Method of engagement

Situated interviews were used to enact the co-construction approach to co-production. Situated interviews combine the flexible focus of semi-structured interviews with the open and situated approach of ethnographic research (Gale & Sidhu, 2019). A situated interview, moreover, takes place, or is situated, in the site under investigation; indeed, the environment and the participant’s interaction with it is considered relevant observational data. In addition to including the relatively flexible approach to questioning found in traditional semi-structured interviews, situated interviews also take account of the goings on of the setting meaning that additional information regarding the interview can be included as important such as environmental, sensory and emotional factors. In this way, they are in practical terms similar in style to ‘ethnographic’ interviewing but are not necessarily embedded within the epistemological aims of a traditional ethnographic investigation of (medical) culture.

No method is epistemologically neutral (Cassell, 2005), but like semi-structured interviews, situated interviews are epistemologically flexible and do not presuppose the long-term participation in a study site required in ethnographic practice. The overall purpose of a situated interview is to allow the interviewing process to not only take place in the research site but also to take account of the research site or environment, and to capture ‘situated sense-making practices’ (Housley & Smith, 2011). Similar approaches include ‘walking’ or ‘go-along’ interviews, although these may presuppose mobility (Butler & Derrett, 2014; N. Gale & Sultan, 2013; Garcia et al., 2012; Jones et al., 2008). The researcher is urged to be mindful of and, upon analysis, reflective of the objects, noises, smells, emotions, sensations, and sights of the interview and is encouraged to take detailed field notes. In this way, both the interviewer, the participant, and the space the interview takes place in construct the interview ‘talk’. This triple prioritisation of space, senses, and talk, results in a holistic interview process whereby all detail is rendered important (N. Gale & Sultan, 2013). Methods that encourage participants to co-construct accounts of emplaced experience and enable reflection on action-in-context enhance co-production of qualitative data (Balbale et al., 2016; N. Gale et al., 2019; N. K. Gale et al., 2019).

In addition, situated interviews destabilise the tacit ‘rules’ that are sometimes associated with traditional semi-structured interviews, such as the ‘one hour rule’ whereby interview quality is judged by interview duration and the expectation that interviews should be conducted in a space that creates an ‘unbiased’ relationship between the interviewer and the participant (e.g., Jacob & Furgerson, 2012). In contrast to these restraints in accounting for the setting of the interview as part of the process and indeed the talk, the situated interview allows the researcher to take account some of the lived reality of the participant. It also allows this lived reality to dictate the parameters of the interview, in that the interview can stop and start again at another time and it can be shorter or longer than traditional interviews.

The situated interviews were conducted over the summer of 2018 and were carried out by the first author in a busy NHS trust emergency department in the English midlands. The study was granted ethical approval by the London – Brighton and Sussex, NHS Research Ethics Committee. The stakeholders involved in the study included healthcare professionals involved in the delivery of emergency care for frail older people (n40), frail older patients (n30) receiving emergency care and their friends and family (n30). Situated interviews were used to engage with all of the stakeholders based on their ‘hard to reach’ (busy work schedules for the staff) and ‘vulnerable’ (frail) status in order to explore the lived experience inherent within their perceptions of frailty.

The interviews focused on two key issues to reflect the project aims– experiences of frailty and experiences of healthcare provision for frail older people. Consequently, the interview schedule operated two question sets (See box 1). Questions were left deliberately open to allow the participants to lead the content of their talk and to allow the interview to freely explore particular issues raised by the participants. Participant information sheets detailing the study were provided and informed, written consent was obtained. The researcher explained the aim of the project and answered any questions regarding participation. Participants were informed that they could withdraw from the process at any time.

Analysis methods were selected that complemented the diversity of the data set and the dual aims of the project. The participants’ talk was analysed using discourse analysis (for experiences of frailty), specifically using Wetherell and Potter (1988) ten step guide. The framework method (N. K. Gale et al., 2013) was used to analyse talk about experiences of care and improvements. The first author carried out the analysis using Nvivo Pro 12, first organising the interview talk into the two categories, perceptions and experiences of care, and then further organising the talk using Nvivo nodes. The nodes for each category were then used to search for interpretive repertoires and themes across the data. The first author’s field notes for each interview were added to the top of interview transcripts before they were inputted into Nvivo. In this way, field notes were included in the process of analysis, allowing the first author to situate the interview talk back into the emergency department. This facilitated a closeness with the interview talk that allowed a more nuanced approach to analysis whereby the interview setting and the impacts of this could be included. A written report was shared with decision makers at the hospital and a presentation to the executive board is due.

Benefits

As previously stated, in order to allow for the engagement of frail older people in the project situated interviews were specifically chosen for their flexible and holistic approach. Using situated interviews allowed the researcher carrying out the empirical work to conduct the interviews at bedsides while the frail older participants were waiting. Indeed, the emergency department setting was of particular importance to this project because of the focus being how key stakeholders perceive frailty in emergency care settings. Regarding the engagement of frail older patients in the co-production work, this form of interviewing brought with it a number of practical and outcomes related benefits.

Of all of the frail older people the researcher approached, all took part in the research. Practically, in terms of recruitment, the need to explain research over the telephone or in a letter was avoided. The researcher was able to speak to the participants in person and was able to tailor her explanation based on how she perceived the situation, using the environmental, emotional, and sensory detail of the setting to do so (see Figure 1 for an example of research notes detailing this information). This detail is often lost in/absent from letters, telephone calls, and recruitment posters (Holt, 2010). For this reason, the researcher was able to plainly explain the project, the level of involvement required and was also able to answer questions there and then.

Figure 1. Post interview observations

Conducting the interviews in the emergency department itself at the very time that the participants were receiving emergency care created a ready and replenishing audience. The participants did not have to move from their beds. This was beneficial in that for some frail older people attending an interview, even if in a convenient place for them, would likely necessitate the involvement of others and therefore require much more preparation and planning. Conducting the interviews in the emergency department also allowed in situ thoughts, emotions and perceptions to be shared. If the participants had been approached after their stay in the emergency department some of this detail may have been missed due to loss of memory.

Additionally, the situated nature of the interviews allowed the researcher to take stock of the situation before approaching potential participants. If a patient was asleep, being treated, or incapacitated, the researcher either waited for a better time to approach the patient or decided not to approach them at all. This decision could be made quickly based on a visual assessment of the goings on. This environmental, sensory and emotional detail of the patients’ situations also allowed the researcher further understanding of some of the everyday reality of the patient. For example, some patients attended with carers from their care home, some attended with numerous family members, and some used walking frames to aid mobility. In taking account of all of this wider detail the researcher was able to tailor and adapt her interviewing style as seemed appropriate.

The flexibility of the situated interview worked very well for the engagement of this particular group, in that breaks could be taken at any time depending on individual circumstances. In a number of cases, the interviews were interrupted by healthcare professionals coming assess/treat the patient. In these cases, due to the researcher being in situ, the interview was able to be returned to once the health professional had finished. When interviews were interrupted, the interviewer added detail to the field notes for the interview outlining where and why the interruption had occurred to allow the interview to be picked up again later. In one case, a patient was moved onto a ward which resulted in termination of the interview but in the majority of cases the interview was able to picked up relatively quickly without too much disruption to interview flow. In addition to this, the interviews were not constrained by a particular time frame. If a participant appeared to be tired or if the participant had simply exhausted what they had to say on the topic of frailty, the interview could be stopped without the researcher feeling like she needed to continue in order to achieve a particular length of interview.

In taking account of the emotional aspect of the interview as integral to the interview ‘data’, the primary researcher was also able to factor herself into the process. As Edwards and Mauthner (2002, p. 15) tell us, the researcher is ‘a central active ingredient of the research process rather than the technical operator that can be inferred by professional ethical codes’. In taking account of body language, tone and the reactions of the patient and their friends and family, the researcher (first author) was able to adapt her interview style as needed. The researcher was also aware that her own body language, tone, demeanour and reactions would impact on how the participants responded to the questions being asked and so worked hard to actively listen to the participants talk and adjust her behaviour accordingly; showing compassion where needed, sharing her own experiences if useful and also being careful, as appropriate, not to show emotions that could be perceived negatively, such as embarrassment, sadness or shock. On occasion the interviewer could smell certain odours such as urine and faeces, some of the participants appeared visually very unwell, and participants shared stories of loneliness and loss of independence. Added to this, a number of the patients initially thought the researcher was a doctor bringing news for them. The researcher was mindful of this and the potential disappointment it could create and introduced herself from the outset. Owing to these factors the interviews were conducted informally and the flow was chatty rather than formal; the researcher felt connected to the participant in a way that allowed her to explore the detail of their talk.

Writing field notes that represent a constitutive part of the situated interview process to allow the sensory and physical setting to become part of the interview also had the beneficial effect of creating a more nuanced approach to the process of analysis. As outlined earlier two methods of analysis were used to explore the project’s dual aims (discourse analysis for perceptions of frailty and the framework method for experiences of frailty care). Adding the field notes for each interview to the top of the transcripts allowed the researcher (first author) to re-situate the interview talk in the emergency department. This had the effect of re-establishing a closeness with the data that can be lost with time. In the case of the framework method, the field note observations were also able to be included in the Nvivo nodes to facilitate the identification of themes. Nvivo nodes using the field notes included: clinical objects, sensory observations (smell, noises, visuals), furniture, lighting and atmosphere. Field notes were not included in the discourse analysis because this method was specifically chosen to prioritise patient talk when exploring their perceptions of frailty.

The practical benefits addressed allowed for the relatively easy engagement of frail older people in the situated interview process that was part of the co-constructive approach to co-production adopted for this stage of the project. In addition to this, the participants talk about frailty and the improvements suggested were rich and detailed. This was particularly important to this project in that the findings were to be used as a baseline for further improvement work. Participants talked about how frailty feels both physically and emotionally and talked about instances when they have felt frail and how their frailty affects their everyday lives, as seen in the extract below:

Respondent:

I mean try doing what you could do in your 40s in your 60s and you’re in trouble, try doing it in your 70s and 80s and you’re in here! I can still cook my meals and that, I can walk to the shops not far and all that but I can’t lift heavy things, I can’t be out in the garden doing it for the day, I can’t do that sort of thing.

Interviewer:

How does that make you feel?

Respondent:

Well, frustrated of course. Really frustrated. I feel I should be able to do these things, I want to … . but you know you just sort of get on with it all or you think, I’ll just try this.

Interviewer:

Try and do it for yourself?

Respondent:

Yes, I hate to rely on anyone. I don’t like putting on people you see.

This interview then continued with the interviewer and the participant talking through the feelings of frustration and growing dependence the participant now feels regarding his declining physical capacity.

In summary, the flexible and informal nature of situated interviewing coupled with the inclusion of sensory and environmental observations resulted in rich, descriptive accounts that allowed a detailed understanding of patient and carer perceptions of frailty.

Challenges

While the position of the researcher is identified as a benefit above, this also brought challenges to the interview process that require some reflection. The researcher was very aware of the fact that the patients were undergoing emergency care and the additional ‘vulnerability’ this added to their already vulnerable status. The majority of the patients had not expected to be spending the day or more in hospital and the researcher was mindful that this could be frightening and uncertain for the patients. While this was sometimes based on the talk or demeanour of the participant/s, it was primarily a personal perception held by the researcher based on her own lived experience. For this reason, the researcher avoided questions she perceived might create further fear and uncertainty in the participants (see Figure 2 for an example of this). This may not have been the case if more traditional semi-structured interviews that tend to be conducted outside of the research setting had been chosen. In this case, the participants would have had time away from the fear and uncertainty they may have been experiencing at the time of the situated interview. For this reason, situated interviews may result in more guarded questioning and potentially more extreme responses.

Figure 2. Post interview observations

In addition to this, a pattern began to emerge as the researcher conducted the interviews. As sated, the interviews tended to begin slowly with the patients providing short answers without a lot of detail. The first interviews began with the question ‘how would you describe a frail older person?’ This question was designed to be an icebreaker that would encourage to the participants to begin to think through their perceptions of frailty. It was hoped that the participants would provide rich descriptions and perhaps use examples to provide emphasis. Instead, the participants tended to provide short answers that lacked in detail such as ‘someone that is old’ and when probed for further detail provided very little elaboration such as ‘just someone that is old’. Thinking this indicated a problem with the opening question, the researcher changed the opening question to ‘tell me about your experience here today’. This question yielded similarly short answers such as ‘it’s been fine’. After time, all participants did begin to provide increasing detail. It is thought that this muted beginning to the interview could be a reflection of the interview location. The participants had not expected to find themselves taking part in an interview, they expected to interact with health professionals and to focus on their acute health condition. For this reason, when conducting situated interviews in a setting where the participants did not have prior knowledge of the possibility that they might engage in an interview, regardless of how informal it is set up to be, those doing the interview should be mindful that the interview talk may take time to develop.

Conclusion and recommendations

This project was carried out specifically to include frail older people in the provision of baseline knowledge on which to ground further quality improvement work addressing care pathways for frail older people accessing emergency care in an NHS trust hospital. The hospital is committed to delivering patient centred, co-produced improvements. Owing to the marginalised status of frail older people’s voices in the development of approaches to assessing clinical frailty and also mounting evidence to show that frailty is an unwanted label among those so labelled, this project was specifically conducted to access frail older patients perceptions and experiences in order to provide more of a bottom up approach to co-produced service improvement. The co-construction approach to co-production was chosen as the most appropriate method of co-production and situated interviews were chosen as an accessible and inclusive approach to enact this.

Our use of situated interviews, in the context of a co-constructive approach to co-production, has demonstrated that concerns about the practicalities of working with frail older patients can be addressed. Indeed, the quality of the data that we were able to co-construct with participants demonstrates how valuable the views of users of a service are to quality improvement work. The use of situated interviews, as opposed the more common ‘semi-structured’ interview which tends to be conducted in a neutral place some time after the event under discussion, facilitated sensitive and flexible engagement with extremely vulnerable participants. It allowed what have previously been considered ‘hard to reach’ voices to be included.

The in situ nature of situated interviews had benefits both in terms of process and outcome. In terms of research process, the situated interviews allowed for the needs of the participants and the demands of the setting to be accounted for flexibly. Participants were not under any pressure to perform for a fixed period of time, conversations could start and stop numerous times as needed. The researcher was also able to sit alongside the participant during their hospital experience, rather than being in a position of asymmetric power dictating the space and time of the interview. In this position the researcher was also subject to the temporalities and characters of the hospital, such as having to stop the interview when the healthcare professionals were doing their work. The field notes made as part of the situated interview process facilitated a closeness with the data at the stage of analysis that is often lost with time. In the case of the framework method, the interview setting was accounted for as interview data and used in the Nvivo coding process, allowing the emergence of nuanced themes.

In terms of outcomes, situated interviews enabled the co-construction of high-quality data that allowed in depth exploration of the research questions. Regarding the project aims (to explore how key stakeholders understand and make sense of frailty and to explore key stakeholders perceptions of healthcare provision for frail older people accessing emergency care) situated interviews revealed that patients considered to be clinically frail understand frailty in terms of a disruptive and stigmatised way of being. This adds to the growing body of critical literature addressing the clinical use of the term frailty. In addition to this situated interviews also allowed us to identify synergies and comparisons between staff and patient suggestions for improvement.

Based on this experience, we conclude, that:

1. Situated interviews offer a practical approach to facilitate engagement with vulnerable population groups. This can enhance ethical practice by ensuring that vital voices are not excluded from co-production work based on their ‘hard to reach’ status. In addition to this, the combination of place, sensory, and emotional detail that is part and parcel of the situated interview process both adds depth and context to the interview findings and allows the interview to become a continuously reflective space which allows for a multi-layered connection between the interviewer and participant.

2. The choice of a particular approach to co-production allowed the project aims to be realised. The co-constructive approach to co-production frames lived experience as fundamental to the co-production process; a level of detail that was required for this project in order to explore real-world perceptions. The explicit acknowledgement of this particular approach to co-production allowed the project aim and purpose to be realised.

Based on this the following recommendations are made for those considering the use of situated interviews with vulnerable groups in a healthcare setting:

• From outset of research/quality improvement projects aiming to co-produce healthcare service experience, co-production should be positioned as a methodological approach that brings with it a particular service ethos rather than a value in itself (Voorberg et al., 2015).

• Those embarking on planning a co-produced approach to healthcare delivery and improvement should be mindful of the theoretical basis of co-production and the variety of ways of doing co-production that exist. In relation to this, as with all methodological considerations, the most suitable form of co-production should be chosen for the task in hand.

• Situated interviews represent a practical and robust way of including vulnerable populations in the co-production (particularly a co-constructive approach) of healthcare services in that they are premised on the understanding that the situation under investigation is part and parcel of the experience and in situ reflections can enhance interview talk. We, therefore, recommend that those undertaking co-production work with ‘vulnerable’ patients groups consider situated interviews as a viable option for engagement.

• If using situated interviews the researcher should be mindful of their position of power and potential relative strangeness within the setting and the impact this could have on vulnerable participants.

Disclosure statement

No potential conflict of interest was reported by the author(s).