Doing involvement: A qualitative study exploring the ‘work’ of involvement enacted by older people and their carers during transition from hospital to home

Abstract Context Being involved in one's care is prioritised within UK healthcare policy to improve care quality and safety. However, research suggests that many older people struggle with this. Design We present focused ethnographic research exploring older peoples' involvement in healthcare from hospital to home. Results We propose that being involved in care is a dynamic form of labour, which we call ‘involvement work’ (IW). In hospital, many patients ‘entrust’ IW to others; indeed, when desired, maintaining control, or being actively involved, was challenging. Patient and professionals' expectations, alongside hospital processes, promoted delegation; staff frequently did IW on patients' behalf. Many people wanted to resume IW postdischarge, but struggled because they were out of practice. Discussion Preference and capacity for involvement was dynamic, fluctuating over time, according to context and resource accessibility. The challenges of resuming IW were frequently underestimated by patients and care providers, increasing dependence on others post‐discharge and negatively affecting peoples' sense and experience of (in)dependence. Conclusions A balance needs to be struck between respecting peoples' desire/capacity for non‐involvement in hospital while recognising that ‘delegating’ IW can be detrimental. Increasing involvement will require patient and staff roles to be reframed, though this must be done acknowledging the limits of patient desire, capability,and resources. Hospital work should be (re)organised to maximise involvement where possible and desired. Patient/Public Contribution Our Patient and Public Involvement and Engagement Panel contributed to research design, especially developing interview guides and patient‐facing documentation. Patients were key participants within the study; it is their experiences represented.


| INTRODUCTION
Shorter hospital stays, which support patient preference to be at home and reduce strain on acute care resources, can result in people needing ongoing care, often requiring multi-agency input. 1 Unfortunately, it has been estimated that one in 10 patients experiences an adverse event in the immediate post-discharge period. 2 Alongside the stress associated with being hospitalised, 3 harm is also caused by systemic issues, such as inadequate care provision across care boundaries, including across settings. 2 Williams et al. 2 (p. e829) suggest that 'there is great potential for significant reduction in harm from even small improvements in this process (of transition from hospital to home)'.
Patient involvement has been suggested as a way of improving the quality and safety of patient care, 4,5 particularly by contributing to enhanced system functioning. 6 This is especially relevant when care delivery and system functioning are challenging, for example, when delivering care across settings, boundaries, and at transitional moments such as when patients return home after a hospital stay.
Following Murray et al., 7 we consider involvement to be nuanced, dynamic and relational, changing over time and influenced by context and interaction. In this way, patient involvement has the potential to operate in multiple ways to influence system functioning. For example, Schubert et al. 8 suggest that by navigating a 'fragmented system', patients/caregivers can 'identify and prevent mistakes from happening, and participate in improving their care' by enabling care co-ordination across multiple settings and providers. Likewise, O'Hara and Lawton 9 argue that patients have the potential to act as 'information conduits' across settings, thereby improving safety and reducing harm. However, despite being the highest users of the National Health Service (NHS), research shows that older people, in particular, struggle to be involved in their care, [10][11][12] therefore minimising potential contributions towards patient safety. Moreover, little is known about the desired involvement of older people from their perspective across the transition from hospital to home, especially over time.
Within this study, we explored what older people (aged 75+) understood by 'involvement', how they 'did' involvement and where there were opportunities for enhanced involvement, during and after a hospital stay, in ways that were acceptable to them.

| METHODS
We undertook a longitudinal focused-ethnography 13 exploring the involvement and experience of older patients from hospital admission to 3 months post-discharge. 14 This enabled us to explore 'involvement at transitions' as a specific phenomenon through inquiry and engagement with older people in everyday life and over time, something limited within the current body of literature. We adopted an inductive, pragmatic approach, with analysis being data-driven and interpretive. Specifically, we aimed to move past individual accounts of experiences and perceptions to identify 'underlying ideas, assumptions, and conceptualisations' 15 from the corpus of data, including from multiple participants, while remaining rooted within individual accounts and experiences.

| Research question
Can older people be more involved in their care? If so, how and in what ways?
This was the first study in our programme of work seeking to improve the quality and safety of care through development of an intervention designed to increase patient involvement, specifically in older populations.
Consequently, inherent in the programme design, and this study's research question formulation, was a theoretically informed assumption that older people can be more involved in their care and that being so will have a positive impact on that care. 4 | 1937 older peoples' experiences and preferences around involvement and so we also aimed to explore: To what extent do older people feel involved in their care? What are their perspectives on this?
Where are the opportunities for older people to be more involved in their care, should they desire this?
To what extent do older people feel able to be (more) involved in their care? What has, or would help them to, feel able to be (more) involved in their care?
We felt that being open to different experiences and perspectives about involvement, while being committed to making improvements, would increase the likelihood of developing a person-centred intervention, sensitive to the lived experiences and preferences of those we were seeking to help.

| Sample and setting
Community-dwelling adults aged 75+ were the target study group as they are most likely to experience variability in care at transitions.
End-of-life patients were excluded because they tend to have dedicated post-discharge care pathways. Likewise, people being discharged to live in residential care were excluded as they were likely to experience different care at transitions to their communitydwelling peers due to readily accessible postdischarge support.
A total of 32 patients aged 76-99 years were recruited from six hospital wards across multiple specialities from three hospitals within two NHS trusts in Yorkshire, North of England (see Table 1). A total of 18 family members were also recruited. We purposively recruited a diverse group: individuals of different ethnicities (with translation assistance), people with and without relatives performing a 'carer' role, and a variety of ages including the 'oldest old' (aged 85+).
Patients with cognitive or language impairments and those lacking capacity to consent were included, provided they had suitable support.
Recruitment and initial interviews/observations and early followups were completed in hospital, during which staff were also spoken to informally during observation work. Further contacts with patients took place in intermediate care settings and in patients' own homes.

| Recruitment
Patients were recruited shortly after admission to the hospital. Initially, decisions about which patients to approach were opportunistic; sampling became more purposive as the study progressed.
Senior ward staff helped identify eligible patients and made initial approaches. Researchers discussed the study with patients and their family, if present; all those approached were given a participant information sheet and the opportunity to ask questions as they considered participation. All participants provided written informed consent and were assigned pseudonyms to maintain their anonymity.
The study was approved on 8 March 2017 by Wales 7 NHS research ethics committee (17/WA/0057).

| Data collection
Semi-structured interviews were the primary means of generating data, supplemented by observations, 14 'go-along interviews' 14,16,17 and relevant contextual information from patients' care records. We looked at care records after initial interviews to explore the extent to which people knew and understood the reason for their admission.
We also looked at care records when people moved to new care facilities and were unable to recall information about transfer dates and next steps. This facilitated accurate data capture and enabled timely follow-up. All patients consented to this access.
Each contact with patients was recorded as a field visit (FV). One hundred and sixty FVs were conducted in total (by authors N. H., R. S. and L. H). The fewest number of FVs with a participant was three; the highest was nine. The timing of FVs varied according to patients' care journeys, but broadly occurred at admission, before/at discharge, shortly after discharge, several weeks post-discharge, and 3 months post-discharge and/or at readmission.
Interviews were audio-recorded, where possible. Observations were recorded through field notes. Researcher interpretation and key reflections were noted after FVs; these were used to provide contextual information during analysis.

| Data analysis
Tacit analyses were done throughout the period of data collection by authors N. H., R. S. and L. H., who each reviewed their own data and met regularly throughout the project to discuss key ideas. H., with regular input and sense-checking from R. S. and L. H. to ensure that the identified themes represented the whole data set.
Key ideas were organised by N. H. into categories and subcategories, followed by identifying patterns and relationships between these categories. Similarities and differences between categories were used to construct themes and subthemes and the relationships between them (see Figure 1). Additionally, comparison of themes across FVs for each participant was done to explore continuity and change in perspective and experience over time.

| Involvement work
For health care professionals (HCPs) to deliver care to patients, work needs to be done. That is, decisions need to be made, activities undertaken, and tasks need to be completed-often by multiple people, over time, and within different contexts. Care delivery work is, predominantly, visible and acknowledged as work-HCPs are employed to carry out these tasks, for example. However, our study also found that receiving care required patients (and/or their relatives) to do 'work' too, including decision-making, undertaking activities, and completing tasks. We consider the work that people do as, and on behalf of, patients to be the 'labour' of involvement, conceptualised in this paper as 'involvement work' (IW).
We propose that IW has three dimensions-cognitive, emotional and instrumental (see Table 2). These dimensions often coexisted and were experienced and/or enacted simultaneously. IW also operated along a continuum, with people moving between states of 'Noninvolvement' at one end and 'Involvement' at the other. Being involved was a dynamic, interactional, and relational process. For example, during her first admission, Pearl (91) wanted HCPs to 'look after her'-she was tired and felt that being in hospital provided respite from doing everyday tasks that she normally managed (patientdesired non-involvement). However, HCPs became concerned that Pearl was unable to manage independently at home because she was both reticent and struggling to mobilise independently on the ward. Hospital processes seemed to tacitly support these expectations by promoting and prioritising care delivery work, and minimizing or failing to support opportunities for involvement.
Second, we found that enacting IW required resources, which was also largely unacknowledged by staff and patients, although it was frequently articulated as an issue by relatives. Because the provision of and access to resources were variable and/or unequally distributed, doing IW was sometimes challenging, even when desired.

| Entrusting IW
This involved actively or passively minimizing participation in one or more types of involvement and took two forms: outsourcing and delegating. Both forms were common within our cohort of patient participants. For some people, outsourcing and delegating were enduring experiences; for others, they were transient experiences, usually adopted during episodes of acute ill-health. was treated matter-of-factly and often remained unacknowledged by both patients and their family members.

| Outsourcing
Entrusting IW to others was the most frequent and desired form of involvement in hospital and was supported by standard hospital care processes, within which hospital staff undertook many cognitive and practical aspects of involvement for patients as part of caring for them. However, desire to entrust IW to others was not static and it varied, primarily according to time and location (see Table 3).

| Maintaining IW
Not all patients wanted to outsource IW during their hospital stay.
However, those wishing to retain autonomy frequently had to resist hospital processes; this often required undertaking additional work.

| Resisting processes
Although there was some variation between wards, standard pro- had to become a more active 'refusal'.

| Resources for IW
Involvement was often resource intensive, frequently requiring knowledge and information, social support, and material resources (see Table 4). Some patients had limited access to resources such as informal support, and this resulted in an increased reliance on health and social care services; interestingly, patients relying on formal care provision often struggled to resume IW post-discharge.
At times, resources needed to be externally provided (e.g., information about medication or expected post-discharge care, and equipment) and it was problematic when these resources were not Early in admission, outsourcing decisions and care-related activities were often done because patients were not able, or did not want, to do these for themselves. Some expressed relief that staff were undertaking activities for them, experiencing their hospital stay as respite As discharge planning continued, many patients became more interested in being involved in decision-making; being able to decide place of residence was a concern across the sample. Some patients felt able to contribute to discussions themselves. However, many patients preferred to delegate their involvement to relatives. Alongside resources provided by others, the capacities and capabilities of patients themselves could be resources to doing or resuming IW. In particular, feeling confident communicating with HCPs seemed to be a key facilitator for active involvement for both patients and relatives, especially when seeking information, challenging/resisting processes that minimised involvement opportunities, and influencing decision-making.
Overall, we found that patients with access to multiple resources, that is, both their availability and the means to utilise them, were often more effective at maintaining or resuming IW according to their preferences and were more likely to have a positive influence on their care when they did choose to be involved.

| Consequences of entrusting or enacting IW
Although maintaining IW was challenging during hospitalisation, it appeared to enable people to resume involvement more effectively postdischarge than those who entrusted IW to others throughout their stay.
Outsourcing, in particular, seemed to contribute to loss of confidence and deconditioning in undertaking activities. This sometimes meant that People who wanted to maintain their IW were more likely to seek information than people who outsourced their care to others: 'I always ask, I'm a great believer in asking, asking questions, and they may not know the answers but they'll get to know the answers for you, you know? So I find that it's like, life's less complicated that way'. (Diana, 78) Information about postdischarge care was often limited and most patients had few means of accessing additional or correct information when it was missing or inadequate. Access (or not) to resources, such as a computer or internet access, was sometimes instrumental in being able to resolve issues. For example, Doris (99) received a letter asking her to call a telephone number to book a clinic appointment. Unfortunately, the number provided was no longer in service and Doris had no means of contacting the clinic to book an appointment. Compare this with Ray (76), who, when faced with a similar situation, was able to source the correct clinic number using an internet search engine. He not only called to make his appointment but also alerted staff to the error on the letter, who assured him they would change the incorrect information.
Likewise, 'delegates' were also likely to be active information-seekers: Lack of resource provision could also impact on post-discharge experiences by minimising opportunity for involvement in hospital, with visible consequences for patients and staff-missing mobility aids increased reliance on staff to help patients mobilise, for example.
However, some problems did not become apparent until patients left hospital and as such were likely to remain invisible to hospital staff, who may not be aware of issues unless problems were significant enough to trigger readmission. For example, having no understanding of new medication is not necessarily a problem while HCPs are dispensing it in hospital. However, once patients become responsible for this post-discharge, lack of understanding can lead to unintentional medication noncompliance.
Alongside consequences for patients, enacting or entrusting IW had an impact on both care providers and families. For HCPs, active involvement could save them time and reduce care delivery work. For example, patients going to the toilet independently meant that staff time was not needed to help patients. However, some peopleespecially relatives-felt that they were treated as a 'nuisance' by staff when they enacted IW or when they sought out resources for IW. Relatives, in particular, said that because information provision was minimal, they frequently had to seek this out. However, doing so often meant interrupting HCPs during tasks and some people said that staff tacitly communicated displeasure at such interruptions.
Information flow was largely controlled and dictated by HCPs and people frequently struggled with this, feeling limited power to effect change or have an influence unless this was facilitated by HCPs.
However, the power dynamics of IW should not be seen as unidirectional. By having the decision on whether to enact or entrust IW to others, patients were able to impact both positively and negatively on the care delivery work that HCPs did and the IW that relatives undertook. The power to say 'no' to doing IW was particularly potent; HCPs wanted to (and were also duty-bound) to care for patients, and family members often felt obliged to provide time, energy, and any financial cost it took to undertake IW on their behalf. IW, then, is always an interplay between people and is often negotiated relationally and interactionally.

| DISCUSSION
The findings of the study suggest that most participants were not actively involved in their care in hospital. While non-involvement was largely desired during this time, it was also tacitly promoted by hospital processes, which automatically assumed responsibility for most tasks people normally engaged in. 'Non-involvement', then, was often a kind of 'collaborative accomplishment' between HCPs and patients: from the point of admission, many patients wanted and expected to 'outsource' and 'delegate' their IW to others, while care delivered in hospital often failed to enable active patient involvement-even when desired-by doing 'IW' on behalf of patients. At times, this seemed to benefit both patients and care providers-patients wanted to be 'looked after' and staff wanted to care for patients in ways consistent with hospital processes, which implicitly supported non-involvement. At times, shortages of care resources, especially staff, also lent itself to patients being uninvolved; some patients sensed that the most helpful role they could play was as a passive patient. Non-involvement, then, was sometimes a type of 'collusion' between patients and HCPs. This may provide some short-term 'benefits', but as noted, it can also result in longer term consequences, especially post-discharge. These were infrequently anticipated by care providers or patients. It is also important to recognise that 'good' patient involvement will not look the same for all people and that for some, non-involvement (e.g., outsourcing) or delegated involvement may be the preferred approach. In these cases, facilitating greater involvement of relatives, where possible, could be beneficial. For other people, 'passive' forms of involvement, such as an understanding and acceptance of care and treatment plans, may be 'adequate'. 22 This means that information-sharing between HCPs and patients and their families is vital-not just as a means to active involvement, for example, shared decision-making, but as the means of involvement itself. Importantly, HCPs may need to take the lead with regard to providing information, as many older people are 'information receptive,' but not active in seeking information. 7 Active information provision may also be required if patients are to be 'information conduits' between parts of the system across the transitional journey. 4 Patient expectations have been identified as a key factor in determining participation, with patient desire (to participate) proposed as a prerequisite of participation. 25 We too found that when patients expect and desire others to do IW on their behalf, they are less likely to participate in their care. This, combined with services geared towards passive patients, creates a 'perfect storm' of non-involvement, much of which may be desired by patients and tacitly welcomed by service providers. Carman argues that organisational characteristics, policies, and practices can (positively) influence patient participation. 23 However, despite potential system benefits when patients engage in IW, services appear to be predisposed towards noninvolvement. This is often a consequence of the way work is organised, rather than deliberate exclusion by HCPs.
Alongside 'desired non-involvement', 7 approaches towards more active types of involvement and ability to undertake IW appear to be mediated through access to, and ability to leverage, multiple resources and are therefore subject to unequal distribution. Importantly, these include peoples' capacities and capabilities. The concept of 'patient activation' has been used to describe the 'knowledge, skills and confidence a person has in managing their own health and healthcare'; higher levels of activation are promoted as a means of improving health-related outcomes. 26 Greene and Hibbard 27 go as far as to say that 'patients should be more active and effective managers of their health and health care'. In some ways, our study supports such a proposal. We found that the patients in this study most effective at exerting influence and enacting IW in ways meaningful to them were proactive; had existing relevant knowledge; were confident talking to HCPs; and/or were able to resist, challenge, or work around problematic organisational processes within multiple settings.
However, we also found that few participants felt able to do these things, especially when unwell. Consequently, it may be useful to consider concepts such as 'activation' as a resource-which people may or may not have access or ability to leverage at a given timerather than an attribute of an individual. In agreement with Sinding et al., 28 it is important to acknowledge the potential barriers that people can face when attempting to be involved. Otherwise, uncritical promotion of increased patient involvement may serve to aggravate existing health and social inequalities.
Moreover, an individual's (high) level of activation may not be sufficient to enable them to actually undertake more active forms of IW. them. Likewise, the IW people can engage in-especially within institutional settings-is likely to be particularly influential on someone's subjective experience of independence, regardless of how reliant they are on others to do things for them.
For some people, then, lack of involvement and reliance on others will be detrimental to independence, while for others, reliance on others will be consistent with being independent if they can retain autonomy over the things that are important to them. Therefore, 'meaningful' involvement is not 'one-size-fits-all'; rather, it requires a person-centred approach that takes account of a person's desires, 'psychological make-up, biography, social context and cultural heritage', 31 alongside the resources they have access to.

| LIMITATIONS
As with all research, our study has limitations. We recognise that observational methods and data generated within shared environments may introduce bias into findings. For example, people may change their behaviour while being observed. However, by spending extended periods of time on wards and with participants, we feel that people became comfortable around the researchers, enabling us to capture naturally-occurring behaviour.
We also believe that data generated by observational methods provide insights that go beyond verbal accounts and therefore have utility despite these limitations. 32 We are also aware that hospital wards offer limited privacy for personal conversations and for patients to express their care experiences, thereby introducing the potential that accounts are limited. However, we took every effort to use private rooms where possible and build trust and relationships with participants, which we believe encouraged honesty. Also, by following up with people postdischarge, we also provided opportunities for them to share experiences in private spaces, outside of care environments.
We acknowledge that qualitative research is rarely representative, and our findings are therefore not generalisable to all older people transitioning from hospital to home. However, we believe that the in-depth nature of the work, comprising multiple data generation methods, provides findings that are credible, dependable, and contribute to research in this area. Moreover, our project patient and public involvement groups have repeatedly reviewed the findings at various stages of data collection and analysis. They felt that we captured important themes and perspectives, many of which mirrored their own experiences of being community-dwelling older adults, all of whom had experienced transitions of care from hospital to home, suggesting transferability of our findings.

| CONCLUSION
Receiving and being involved in care often require patients and families to engage in 'work' that remains largely hidden and unacknowledged. Multiple factors influence the involvement that people desire and enact, including patient characteristics, relational dynamics, resource availability, interactions with others, and system processes. However, in hospital, many people 'entrust' IW to others and struggle to resume activities. This can result in increased reliance on people and services postdischarge, alongside a diminished sense of independence. Enhancing involvement could contribute to positive patient experience and safety. Doing so will require encouraging IW when people are reticent, facilitating IW when people show willingness and desire, and resourcing IW to ensure that burden is minimized and inequalities are not aggravated.