How older people enact care involvement during transition from hospital to home: A systematic review and model

Abstract Background Current models of patient‐enacted involvement do not capture the nuanced dynamic and interactional nature of involvement in care. This is important for the development of flexible interventions that can support patients to ‘reach‐in’ to complex health‐care systems. Objective To develop a dynamic and interactional model of patient‐enacted involvement in care. Search strategy Electronic search strategy run in five databases and adapted to run in an Internet search engine supplemented with searching of reference lists and forward citations. Inclusion criteria Qualitative empirical published reports of older people's experiences of care transitions from hospital to home. Data extraction and synthesis Reported findings meeting our definition of involvement in care initially coded into an existing framework. Progression from deductive to inductive coding leads to the development of a new framework and thereafter a model representing changing states of involvement. Main results Patients and caregivers occupy and move through multiple states of involvement in response to perceived interactions with health‐care professionals as they attempt to resolve health‐ and well‐being‐related goals. ‘Non‐involvement’, ‘information‐acting’, ‘challenging and chasing’ and ‘autonomous‐acting’ were the main states of involvement. Feeling uninvolved as a consequence of perceived exclusion leads patients to act autonomously, creating the potential to cause harm. Discussion and conclusion The model suggests that involvement is highly challenging for older people during care transitions. Going forward, interventions which seek to support patient involvement should attempt to address the dynamic states of involvement and their mediating factors.


| INTRODUC TI ON
There is long-standing consensus, reinforced by policy-led initiatives, that being involved in one's health care promotes choice and equity and indeed is an individual's right. 1,2 Involvement is seen as an essential tenet for improving both the quality and safety of care. 3 Despite an articulation of a moral obligation and belief in the mechanisms by which involvement in care contributes to better health outcomes, 4,5 there is no clear understanding of what being 'involved in one's own care' actually looks like.
Various bodies of the literature talk around involvement with reference to the importance of the patient-professional relationship 6,7 and influencing contextual factors such as having time and information 8 but the more nuanced interactional and the dynamic nature of involvement has been largely overlooked. This is highly important in the current healthcare climate where there is increasing expectation that patients can, and increasingly more interventions that support patients to, take on more responsibility for their own care needs and decisions about treatment. [9][10][11] Taking on responsibility necessarily involves patients performing 'work' 12 that includes reaching in to a complex health-care system. 13,14 The dynamics of how patients undertake this work may be a key determinant in the success or failure of these policy-led initiatives and interventions.
Older people represent a particularly vulnerable group for whom involvement may be most challenging. [15][16][17] They have complex healthcare needs, frequent hospital stays and high rates of readmission. 18,19 The transitional period from hospital to home, in particular, represents a fragile time for older people. Deconditioned from their hospital stay, they are often required to take on new care regimens alongside re-integrating and coping at home. The individual experiences of older people during this period have been captured in numerous qualitative studies 20-23 but a synthesis that draws out and provides conceptual clarity about how people enact involvement has yet to be performed. This could, among other things, support the development of interventions.
We therefore sought to systematically review published qualitative data to provide greater conceptual clarity about the dynamic and interactional aspects of how patients enact involvement in their own care. Using the lens of older people transitioning from hospital to home, the overarching aim of the current study was to develop a model of patient involvement in care.

| ME THODS
A systematic and empirically driven approach to synthesizing the current evidence was employed to ensure that the concept of involvement was true to the patient experience. For the purposes of this review, involvement in one's own health care was defined as any actions undertaken, as well as thoughts and feelings held in support of pursuing a health-and well-being-related goal.

| Data extraction, analysis and quality assessment
Each paper was read, and findings about involvement, as per our definition, were initially coded (independently by JM and NH) using an existing involvement taxonomy as a theoretical framework (Appendix S2). 1 This framework was chosen over others 24,25 as it provided greater conceptual clarity about different types of involvement at the individual level. Data relating to context (barriers and facilitators to involvement) and inferred consequences in relation to these findings were recorded. As coding progressed, we moved from a deductive to an inductive approach to capture aspects of involvement that did not fit into the theoretical framework. We checked that our interpretations of the findings aligned with each other and with the emergent categories (termed 'types' of involvement). Where required, we revisited the original paper to explore meanings and potential assumptions. To construct the model, we examined findings that reported multiple ways of enacting involvement, to understand how involvement could change within individuals in the context of one care episode or activity.
To ensure that the model accurately represented the original data, the extracted findings from the studies were revisited and compared with the model. The model was subsequently interpreted to provide an overall understanding of the process of involvement of older people during transitions.

| Patient and public involvement (PPI)
We convened a PPI session with six members of our existing panel patient to explore how they interpreted a selection of extracted quotes from the included studies. The group comprised older people (aged 70 and over) and two of their carers, all with experience of emergency hospital admission and discharges within the previous three years. The group sorted the provided quotes initially into 'involved' and 'not involved' and then into our suggested subtypes.
Their sorting agreed with ours and the types of involvement, as defined by our research process, very much resonated with their experiences.

| Quality assessment
The quality of included studies was assessed using an adapted version of the consolidated criteria for reporting qualitative research (COREQ). 26,27 The tool consists of 30 items (with a total possible score of 60) covering the research team and reflexivity, study design, setting and data collection, data analysis, findings and ethics. Two researchers (JM and NH), independently screened the studies and discrepancies, were resolved by discussion and revisiting the papers.
Studies were not excluded on the basis of this assessment.
The protocol for this study is registered with PROSPERO No. CRD42017058696.

| FINDING S
Three thousand and sixty publications were identified, which through screening ( Figure 1) provided sixteen studies that contributed to the development of the model. [20][21][22][28][29][30][31][32][33][34][35][36][37][38][39][40] The studies collectively included 303 participants with 170 patients in 12 studies and 133 caregivers in eleven (see Table 1). One study specifically sought to include patients from ethnic minority backgrounds including gypsy travellers 30 ; other studies did not specify the make-up of their study population. A broad range of admitting conditions were reported across the studies, and the types and extent of information on social support for patients varied greatly. The experience of involvement in transitions was a specific focus within four studies. 29,30,34,40 The remaining studies were concerned with general experiences of care and transitions. Of note were three studies that, despite aiming to explore the general experience of transitions, reported extensively on involvement. 34,38,39

| Study quality
All studies met more than half of the 30 quality assessment reporting criteria. Studies scored least well in relation to reporting about the research team, reflexivity and some aspects of the research design, scoring better on areas such as data analysis, findings and ethics.

| Summary of types and subtypes of involvement
Four types and 12 subtypes of patient-determined involvement were identified (Table 2). We also identified three types and seven subtypes of professionally mediated patient involvement along with a number of other contextual factors that appeared to influence involvement including, for example, having a supportive family and experiencing emotional problems.

| Patient-and caregiver-determined involvement
Types of involvement included 'non-involvement', 'information-acting', 'challenging and chasing' and 'autonomous-acting'. Non-involvement represented a state in which people became passive recipients of care and even absent/failed care. The absence of patient and caregiver involvement was evident in all studies. Even where non-involvement was 'desired', patients appeared to hold assumptions about the standards of care that they would receive. This was demonstrated through showing disappointment when expectations about care were not met. 32 Resigned non-involvement was reported alongside highly influential contextual factors such as low mood and ill-health and was arguably the most debilitating subtype of non-involvement. 28,40 The second type of involvement, information-acting, could be active or passive. The literature showed that being more active often failed because health-care professionals did not appear to 'consider' or understand expressed desires or know how to respond. 21,[37][38][39] This resulted in patients and caregivers moving between states of involvement in attempting to resolve a single aspect of care. Challenging and chasing, as the third type of involvement, highlighted the work and effort required to question staff and source information. 28,30,33,34 Examples of chasing were seen exclusively in caregivers. Challenging and chasing often came about through dissatisfaction, anxiety about the future and distrust of the system but was facilitated by interaction with a service that appeared willing to listen. The final type of involvement, autonomous-acting, was often a consequence of non-involvement in care, mediated through feeling excluded by professionals.

| Professionally-determined types of involvement
Patients and caregivers alluded to three ways in which they felt professionals mediated involvement through 'exclusion', 'informationseeking/information-giving' and 'consultation'.
In general, patients and caregivers suggested that care providers hampered their efforts to obtain information. Being busy, appearing unapproachable and authoritarian, and being focused on discharge, meant that patients felt unable to pose questions. Patients suggested that professionals did not listen, avoided eye contact, demonstrated little insight into the family circumstances and did not see patients as individuals. Even where patients felt able to approach staff, nurses appeared unable to answer their questions, deferring to absent doctors.
Nobody tells me (about leaving hospital). I asked them (nurses) but they don't even know themselves). 40 Despite not always having the answers to questions, there was evidence that some staff did seek to obtain information to the extent of chasing. For example, Andreasen 28 reported on how one member of staff 'phoned God and everybody' on the patient's behalf only to be told that they would have to wait until the following week for the essential item of toileting equipment. 28 This could represent a form of staff exclusion from the services that they work in but also challenging and chasing, similar to that observed in patient-determined involvement. A more extreme version of patient perceived exclusion was observed in two studies where health professionals appeared to 'close the door' by overtly declining requests for help. I told them I couldn't manage at home and needed to stay a few more days. But the doctor told me there was no place at all for me on the ward or in hospital. 39 Staff also mediated involvement through information-giving. While this could be useful, it could equally be unidirectional, lack consultation and tailoring, and be inappropriately timed. Patients did, however, indicate that information, if given in the right way, could encourage involvement. Finally, a more positive approach to encouraging involvement through professional consultation was described as including activities such as formal discharge planning meetings, home visits or more informal routes such as a bedside consultation approaches. 21,29,31,38

| State-change model of involvement
By exploring findings which reported multiple types and subtypes of involvement, we were able to observe that the process of enacting care is not static or necessarily a trait-determined approach. Rather, people change their 'status' depending on their interactions with services and other contextual factors. Thus, in constructing a model that represents this dynamic interactional process, we have oriented 'types' of involvement into 'states' (Figure 2).
Findings reported up to four state changes. The model commences with information-acting and desired non-involvement.
Despite the fact that some findings started at the point of professional exclusion and even challenging and chasing, it is likely that patients and caregivers were at least passively information-receptive at a prior point. Professionally mediated involvement influenced the next step of patient-determined involvement, where exclusion, leading to non-involvement, could move into a state of autonomous-acting or challenging and chasing. In the literature, there were frequent F I G U R E 1 PRISMA Flow chart Papers identified in database search (EMBASE, CINAHL, MEDLINE, PsycINFO, ProQuest) and Google Scholar(n = 3060) Title and abstract screening (after de-duplication) (n = 2275) Papers excluded (n = 2245) Full paper screening (n = 30)
Papers identified from reference lists, citation searching (n = 10) Papers remaining after exclusions (n = 16)  Information-Acting Passively receptive/ seeking 22,29,32,33,37 Willingness to receive and give information that may be unexpressed or acted out through waiting for the 'right time' (with potential health consequences)

| D ISCUSS I ON
'I would just love to be informed….' 32 Actively seeking/ giving 20,22,29,32,33,36,38,39 Taking or creating opportunities to ask questions. Most often in response to perceived failures in care delivery such as absent information 'We were pulling it (looking for information) on our own because otherwise it was just a black hole…you're kind of thirsting for information that whole time ' 22 Challenging and Chasing No subtypes 22,30,32,33,38 Challenging decisions that fail to take their wishes into consideration or chasing support when services are unresponsive 'We rang up several times on the ward but they don't bother to answer or anything. Then two o'clock in the morning I rang up, I said "What's happening, why can't you inquire more,"' The caregiver subsequently found out that his wife had been moved to intensive care 30

| Implications for research and practice
Interventions aiming to support older people to transition from hospital to home have been the subject of numerous systematic reviews. [49][50][51][52][53] Self-management and/or education, as a way of empowering individuals to be involved in and take control of their care, was the second most common component of these multicomponent interventions. 49,50,53 The contribution of self-management to outcomes is challenging to disentangle, but there is some suggestion that inter-  54 Evidence of training, however, is not apparent in existing transition interventions for older people, [49][50][51][52][53] and in the studies in the current review, use of such skills was not apparent. Reasons are likely to vary, but the hospital setting itself is likely to be a factor.
System pressures emphasizing patient flow may limit the opportunities for relationship-building. The ethos of hospital care is to manage acute illnesses rather than support maintenance of long-term conditions. Establishing meaningful involvement with patients under these circumstances may be particularly challenging and resource intensive.
There are current improvement drives towards greater 'patient activation' 54 in relation to facilitating involvement and self-management, but these are at risk of labelling patients by trait and fail to acknowledge the dynamic nature of involvement that fluctuates in response to various compelling contextual factors. The ultimate aim of supporting involvement would be to create a space that enables patients to shift from passive information-acting to actively voicing their concerns in such a way that does not fundamentally seek to change their way of being and that effectively meets people 'on their own turf'. 55

| CON CLUS IONS
Previous studies reporting older people's experiences of involvement during hospital stays indicate that patients want to be involved in their care. 15,56,57 This review and interactional model supports this and shows that non-involvement is not a desired state for most patients but a consequence of system-level forces and other contextual factors that act to erode efforts to become involved. Future interventions require a more nuanced approach that supports staff to recognize all states of patient involvement as valid, to reflect on how their behaviours can influence involvement and to understand how these can impact on patient safety and experience. For those who desire non-involvement, a greater understanding of the factors that perpetuate this state will need to be explored. Respecting the wishes of these individuals while countering against the potential to widen health inequalities will be a fine balance for such interventions.

CO N FLI C T O F I NTE R E S T
There are no conflicts of interest.

DATA AVA I L A B I L I T Y
Data sharing is not applicable to this article as no new data were created or analyzed in this study.