A qualitative formative evaluation of a patient facing intervention to improve care transitions for older people moving from hospital to home

Abstract Background The Partners at Care Transitions (PACTs) intervention was developed to support older people's involvement in hospital to improve outcomes at home. A booklet, question card, record sheet, induction leaflet, and patient‐friendly discharge letter support patients to be more involved in their health and wellbeing, medications, activities of daily living and post‐discharge care. We aimed to assess intervention acceptability, identify implementation tools, and further develop the intervention. Methods This was a qualitative formative evaluation involving three wards from one hospital. We recruited 25 patients aged 75 years and older. Ward staff supported intervention delivery. Data were collected in wards and patients' homes, through semi‐structured interviews, observation, and documentary analysis. Data were analysed inductively and iteratively with findings sorted according to the research aims. Results Patients and staff felt there was a need for, and understood the purpose of, the PACT intervention. Most patients read the booklet but other components were variably used. Implementation challenges included time, awareness, and balancing intervention benefits against risks. Changes to the intervention and implementation included clarifying the booklet's messages, simplifying the discharge letter to reduce staff burden, and using prompts and handouts to promote awareness. Conclusion The PACT intervention offers a promising new way to improve care transitions for older people by supporting patient involvement in their care. After further development of the intervention and implementation package, it will undergo further testing. Patient or Public Contribution This study regularly consulted a panel representing the local patient community, who supported the development of this intervention and its implementation.

The discharge period can be especially risky for people who may be deconditioned and still recovering, 12 and yet are required to resume selfcare activities at home. Discharge is therefore a stage in a transition of care not an end-point of care. Evidence on the effectiveness of transition interventions in reducing hospital readmissions is mixed. [13][14][15][16][17][18][19][20] Further, clarity about the active ingredients of such interventions is challenging to decipher because of the multiple and variable components that can be included. 13,14 Interventions can span across, or be restricted to, the predischarge period (e.g., patient education, medicines reconciliation and/or predischarge planning), the bridging period (e.g., transitions co-ordinator and/or patient-centred discharge instructions) or the post-discharge period (e.g., follow-up and/or patient hotlines) periods. [13][14][15][16][17][18][19][20] However, there is some evidence that patient education with or without selfmanagement may contribute to reduced readmissions. 21 More precisely, interventions that seek to 'enhance patient capacity to reliably access and enact post-discharge care' 21 may contribute to better patient outcomes.
The exact mechanisms for this, however, remain unclear.

| The rationale for patient and family involvement
Transitions involve knowledge transfer between multiple professionals, services and individuals. 19 These transitions have been described by some authors as structural safety 'gaps', which heighten the potential for failures in communication and other safety failures. 22,23 Patients and their carers are the only constant throughout these transitions and so have opportunities to limit the impact of these safety gaps and support the safety of their care, by identifying or preventing mistakes through coordination across settings. 24 However, opportunities to help close these gaps at transition depend on taking a more active role in their care in the hospital. Taking an active role is not without challenges, however, as patients can often 'passively' receive care and avoid asking staff questions. 25,26 Further, busy staff may not prioritize engaging in conversation with patients, 25,26 meaning that patients can sometimes return home ill-prepared to manage themselves, their care, and condition. If the potential benefits for patients and families in 'plugging the safety gaps' across transitional care are to be realized, we need ways to support patients to be more active, recognized, and supported partners in their care during the hospital stay.

| Partners at Care Transitions intervention
A full description of the intervention development is presented elsewhere 27 but summarized briefly here. To develop an intervention to support greater in-hospital patient engagement in their care, we first explored the experiences and perspectives of healthcare professionals and older patients 26,28 and brought these together using Functional Resonance Analysis Method 29 to model transitions of care from hospital admission to 30 days post-discharge. 30 This approach enabled us to identify four key activities that patients and families are responsible for (to varying degrees) following discharge from the hospital: In our resulting underpinning theory of change, 30 we hypothesized that supporting patients to 'know more' and 'do more' in the hospital would better prepare them for being at home, thus improving transitional care outcomes. The Partners at Care Transitions (PACT) prototype intervention was co-designed through a series of stakeholder workshops (patients, carers, healthcare staff and design team) and through the consolidation of evidence from a range of sources. 27 The prototype PACT intervention comprises the following core patient-facing components: 1. An information booklet for patients, encouraging them to be more involved in their care by retaining independence, signposting and offering suggested questions; 2. A stand-up 'question card' for patients to write and display their questions to staff, and promote communication between staff, patients, and families; 3. A 'hospital record sheet' for patients to record events or conversations regarding their health; 4. A ward induction leaflet to orientate patients to ward routines and reduce the disorientation that can happen on hospital admission; 5. A patient-friendly discharge letter based on the four functions, for staff to complete and provide to patients at discharge. These materials were combined within a purpose-designed envelope and provided to patients shortly after admission to the ward. The patientfriendly discharge letter was given at discharge. The intervention was designed to allow flexibility and local adaptation in methods of supporting patients to undertake the activities, which is in keeping with emergent complex intervention development approaches, [31][32][33][34] for example, supporting patients to dispense their own medicine or having conversations to understand the purpose of their medications

| Research aims
In line with Medical Research Council guidance 35 and more recent framework guidance 36 that recommends iteration during complex intervention development, this study aimed to explore the acceptability of the prototype intervention and identify areas of improvement. We further aimed to identify implementation strategies. Our research aims were threefold: 1. Explore the acceptability and usability of the intervention for patients, caregivers and staff. Within the framework of acceptability, our study aims fit most closely with the constructs of intervention coherence (user understanding of the intervention), affective attitude (user feelings about the intervention) and burden (perceived effort required to use the intervention); 37 2. Identify implementation strategies for the intervention; 3. Identify modifiable areas for improvement in relation to usability, usefulness and acceptability of the intervention.

| Study design and implementation
We conducted a formative evaluation using multiple qualitative methods, to explore the acceptability of the prototype PACT intervention and its initial implementation. We focused on gathering user perspectives (in-patient older adults, carers, ward-based practitioners) whose views were crucial in this developmental and early implementation phase.
We designed this formative study to have two phases. In the first phase, shortly after consent was provided, researchers introduced patients (n = 9) to the intervention, including an explanation of what it is and how to use it. This phase was designed to explore how best to deliver the introduction and to use this understanding to develop materials for staff to support them to undertake this introduction. This learning was then formatively integrated into the second phase, wherein the ward staff undertook the introduction of the intervention to patients (n = 16). Short multidisciplinary ward team briefing sessions were provided throughout the study to maintain staff-level awareness and promote opportunities for them to support the intervention.

| Setting and sample
We approached patients and staff on three wards (two older adult, Eligible patients were: aged 75 or over; likely to return to their own home; English-speaking and; an in-patient for at least one night.
Patients were excluded if: they resided over 30 miles radius of the hospital; were at the end of their life; or were unable to give informed consent. Carers or relatives were also invited to participate in the study alongside the patient. Ward staff were purposively selected for interviews based on their involvement with the intervention.
Interviewees were provided with verbal and written information and sufficient time to ask questions, before their providing written informed consent. Observations were conducted with verbal information and consent.  38 This model aims to support intervention and implementation development by identifying appropriate techniques for targeted behaviours. In this study, we focused on the core components of the model (i.e., capability, motivation, and opportunity) that influence our initial targeted behaviour of interacting with the intervention. Data collection occurred within three specific settings: mainly across the selected wards, in participants' homes and on one occasion intermediate care settings. This design allowed us to 'follow' patients and their use of the intervention as they transitioned from hospital to home. All data, including reviewed documents, interviews (which were audio-recorded where possible) and observations were documented on a semi-structured 'contact form' after each data collection contact-(prompts included intervention use; feedback; suggested changes).

| Interviews-Patients and carers
With consent, and before being introduced to the intervention, we conducted interviews with patients to explore their current and desired involvement in their care, so we could later understand their intervention use in context. A second patient interview was conducted 7-14 days post-discharge from the hospital to explore barriers to intervention use, how it was used and areas for improvement (usability, readability, content, aesthetics). See Supporting Information: Files S1 (outline of patient data collection), S2 (baseline interview topic guide for patients) and S3 (follow-up interview topic guide for patients).

| Interviews-Staff
The guide (see Supporting Information: File S4) for staff interviews was similar to the patient guide, being informed by the COM-B model of behaviour change. 38 Interviews further explored staff views on the intervention, including areas for improvement and ideas for developing implementation tools. Interviews were conducted with a range of multidisciplinary ward staff who engaged with the intervention.

| Observation
Patient-and ward-level observation was used to explore how patients, carers and staff interacted with the intervention. We undertook observation of a number of critical points in the process of staff delivering and patients receiving the intervention. These points included the introduction of the intervention to patients by staff, routine care interactions (e.g., ward rounds, dispensing medicines), visiting times and discharge. Short conversations with patients or staff involved in these observations helped contextualize what was observed, how the intervention could help staff and how they could use it in their role.

| Physical intervention components
On visits to patients' homes for post-discharge interviews, we looked at and collected data about the intervention components given to patients, to explore how they had been used.

| Data analysis
Given the practical orientation of our overall research aim, template analysis 39 was used to orient the analysis to the three research questions. In keeping with a formative evaluation approach, data analysis was iterative, starting during data collection, with findings used to inform subsequent data collection. The main unit of analysis was the contact form (see Supporting Information: File S5). Each form was reviewed independently by one of three researchers (T. M., N. H., R. S.), annotating key data. These key data were then grouped by research aim, then sometimes split or moved following discussion, and developed into draft findings. This process was repeated across the study period, with emergent findings discussed at regular meetings with the wider team (including the programme manager, CI and research nurse staff) contributing to the process and final findings. Disagreements were resolved via consensus discussion.
There were two additional steps specific to research aims 2 and 3. To address the second research aim (development of potential implementation tools), the team first listed all the desired behaviours of both ward staff and patients/carers. These included delivering the intervention (e.g., staff giving it to patients), identifying the barriers to this behaviour (e.g., not perceiving the booklet to be useful to patients) and creating suitable responses to each barrier. We did this by systematically thinking through staff and patients' capacity, opportunity and motivation 38 for intervention engagement (e.g., training to understand why the intervention is important and how to verbalize this to patients). To address the third research aim (improvement of the intervention components), we used a research design technique that allows intervention developers to concisely convey the importance of the intervention and merge it with the core message to share with users. This method involves condensing the information into four levels (five words; one sentence; one paragraph; two paragraphs). We applied this to the overarching aim stated within the booklet and to the four functional activities.

| FINDINGS
We approached 57 patients and 25 consented to take part, along with 6 carers ( Table 1). The primary reasons for declining participation were illness and disinterest in the research. During Our analysis was focused on the three research questions of acceptability, potential implementation strategies and modifiable areas for improvement. We present the results of our analysis below, with reference to each research question in turn.
3.1 | Research aim 1: Explore patient, caregiver, and staff acceptability and usability of the intervention Most patients or their families read the booklet and a minority wrote in the booklet or question card, while the 'hospital record sheet' was not used. Reasons why patients and families did or did not appear to demonstrably using the intervention, included: the number of physical components; patient capacity or willingness to engage (either with their care or with the intervention); physical and mental health status (e.g., feeling unwell or unable to do more); not being 'a reader'; feeling it might be helpful for 'other' patients, and a preference to look at the intervention at home. Where families engaged, this was often as a replacement for the patients (e.g., if the patient was unwell). The patients and carers who used it most were already actively involved in their care, the intervention therefore having a greater 'fit' with their usual behaviour. While there was a general consensus among staff, patients, and families that patients can and should be more involved in their hospital care, the booklet did not encourage this strongly enough.
Many found the question card difficult to write on and found it too 'strong' or formal an action to take to engage with staff. Patients would often reveal their uncertainties to researchers, but did not seem to view these as specific 'questions' for ward staff. Some

| Research aim 3: Identify modifiable areas for improvement in relation to usability, usefulness and acceptability of the intervention
Having identified the key messages to be conveyed in the booklet we condensed the information into four levels (five words; one sentence; one paragraph; two paragraphs; Table 2). These messages, as well as data gathered about patients' and staffs' capabilities, opportunities, and motivations to use the intervention guided proposed intervention changes. The intervention was revised with the aim of improving usability, usefulness, and acceptability. Implementation tools were developed to provide supportive prompts and minimize staff burden (Supporting Information: File S6, summary in Table 3).
Changes were consulted with designers, the patient and public panel, and the programme management group.

| DISCUSSION
In general staff, patients, and carers felt there was a need for, and understood the purpose of, the PACT intervention, thus it appears to have high coherence, which is a prerequisite for successful implementation. 40  and practice, and we will deal with these in turn.

| Implications for transitional care theory and research
There is emerging evidence that patients and their families are part of a healthcare system's potential for adapting to changing conditions, often stepping in when care fails or is suboptimal. 23,41 Examples of this support include chasing medications and appointments, querying medications and side effects and presenting to health services when they judge their condition needs to be escalated. 23,[42][43][44] Patients and families also 'step in' to maintain the safety of their care at points of transition between healthcare services and settings. 23 Our intervention was specifically designed to support this previously less acknowledged role of patients and families in improving the safety of transitional care. Where previous transitional interventions have acknowledged this role, it has usually been limited to information giving and encouraging 'self-management'. 16,18 In recognizing that the ability of older adults to undertake activities in their homes is directly impacted by the degree to which they undertake these activities within the hospital, we have shifted the focus on the transitional intervention from one that is about discharge to one that is situated predischarge. This study indicates that this shift in focusgiving patients a role in, and responsibility for, maintaining their own safety-is a recognized need.
The need for clinical teams to prepare adequately for discharge is well established empirically. 45 This preparation has hitherto largely included a handover to clinical and social care teams, and other agencies, with less focus on the preparation of the patient or family themselves, outside of the (often rushed) discharge conversation. Therefore, focusing our intervention on supporting older adults to effectively 'practice for being back at home' throughout the hospital stay, is more novel. However, the recent interest in transitional care interventions that embrace healthcare as a complex adaptive system, 46 suggests an extension of the adaptive potential to include older adults and their families would not be unwelcome.

Intervention message Description
Central message: Active patient role Patient to have an active role in preparing to be at home, by 'doing' (e.g., practising) and 'knowing' (e.g., questioning).

Secondary message: Upstream/ downstream effects
Knowing that in hospital events have an effect at home, for example, poor preparedness to manage in hospital could lead to poor ability to manage at home.
Secondary message: Balancing risk Acknowledging that there are risks and challenges to achieving an active patient role in the hospital, for both patients and staff, but the inability to address these risks in the hospital could mean they are transferred into the patient home.
Secondary message: System visibility System processes are rarely articulated to patients, but by doing so we could help patients know what to expect next, respond appropriately and potentially save staff time.

T A B L E 3 Intervention changes and implementation tools
Intervention changes Implementation tools Components and content were condensed to increase patients' capability to use the intervention.
Prompt cards to increase the ability of staff to introduce the intervention to patients and deliver key messages.
Core messages strengthened to increase patients' motivation to be more involved in their care.
Handout to promote awareness in staff and motivate them to use/ interact with the intervention.
The mechanism for asking questions was made simpler and less 'challenging' to increase opportunities for patients to ask questions.
Visual prompts to raise staff awareness and increase opportunities for interacting with the intervention.
Patient-friendly discharge letter was streamlined to reduce staff burden and increase motivation and ability for them to complete it.
Defined staff roles and activities to increase staff capability and promote opportunities for using/interacting with the intervention.
Alternative media were created to communicate core messages to increase the capability of patients to interact with the intervention.
Guidance for flexible local adaptation to increase ability and motivation for using the intervention.

| Implications for implementation science theory and research
Another key issue emerging from our study is the need for the implementation of the PACT intervention to be adaptable at the point of delivery. We designed this study with minimal implementation support, specifically to examine what might be needed in our subsequent feasibility trial. Indeed, the mixed uptake of the intervention in this study can to some extent be explained by having an incomplete implementation plan. Although staff wanted tools (e.g., laminated prompt cards and handouts) to support the delivery of the fixed intervention components, our findings suggest that prescribing rigid implementation processes to be followed faithfully is unlikely to be successful. Indeed, while flexible implementation might increase perceived intervention complexity (posing its own implementation challenges), variation in the way healthcare practitioners responded to the intervention in our study reinforces the case for allowing adaptation to occur across different contexts (e.g., which staff groups introduce the booklet, answer patient questions, support patients to 'do more' on the ward etc.). Further, as noted by Penney et al. 46 in their recent systematic review of interventions to reduce readmissions, 'interventions that had an adaptive element were more successful, whether it be through allowing local self-organization among individuals in the system, or recognizing that implementation is an evolutionary process that requires change over time'.
There is an increasing recognition that to be effective, interventions need to be considered 'events in systems', where complexity represents as much about the implementation context as the intervention itself. [31][32][33] In our study we tried to address this in two key ways. First, from the outset, we explored and understood the intervention within its target context-in this case, wards with a high proportion of older adults. Second, we explored and sought to develop implementation approaches and tools that support engagement with the broad 'functions' of the intervention, while being clear to not necessarily prescribe rigid 'forms' that achieve these functions. 32,47 Our study provides further support for the need for 'hybrid' interventions that can be adapted at the point of delivery, to achieve change in complex healthcare systems.

| Implications for practice
In our study, ward managers with the trust of, and positive influence over, their teams were found to improve engagement with the intervention. Despite this, we still found variation in the way that healthcare professionals supported the intervention. Such variation understandably exists in a staff group that experiences limited time and changeable teams. Although much of this is outside the control of the research team, we anticipate that staff engagement will be aided in future iterations by outlining specific roles-that is, what staff need to do or achieve. Who or which professional groups fulfil these roles will vary to suit the context of their ward team and a subsequent feasibility trial will explore the effectiveness of this. 48 The issue of staff having to balance risk against potential patient benefits does, however, represent a real and very challenging obstacle for implementing an intervention that seeks to better prepare older adults for discharge. Despite their best intentions, healthcare professionals can be wary of well-intentioned practices inadvertently resulting in harm, for example, increased patient mobility in wards contributing to falls. However, if these risks are not addressed in the ward environment, they are simply 'kicked down the line', with older people facing potentially greater risk within their community-dwelling due to compounding frailty, deconditioning, and disorientation. Put simply, hesitancy to 'balance the risks' in hospital, effectively reduces the risk for hospital staff but raises it for discharged older adults and their families.
In theory, the hospital could-and arguably should-be an environment in which staff can support patients to safely practice skills for home.
However, it is difficult to see how hospitals could take on more of this risk given an already overstretched health service. 49 Although increasing adoption of 'discharge to assess' models 50 acknowledges that hospitals are not the safest place to assess patients' longer-term care and support needs, hospital staff and management also need to understand risk, not as something to be managed within the discreet boundaries of service but rather as distributed across services, settings, time, and people-with the biggest risk to the older adults themselves.
While our study findings have allowed us to make a set of recommended changes to the PACT intervention to increase its acceptability, and a further set of recommendations regarding implementation, the current situation in healthcare services presents a whole range of challenges that are out of our control and not limited to this intervention.

| Study limitations
This was a small qualitative study that aimed to generate learning to facilitate improvements in the intervention and inform the development of an implementation strategy. The study therefore necessitated participants who could read and speak English and while we hoped for more ethnic diversity in the sample, it did not happen.
Further, the study was limited to two specialisms within a single hospital. These two factors may make the findings specific to the context of the patient group and wards involved. However, we have confidence that the findings apply more broadly, based on significant knowledge gained from two preceding studies of the patient and staff experience across many other National Health Service organizations 17,18 and the involvement of professional and patient stakeholders.

| CONCLUSION
The PACT intervention offers a novel means for improving care transitions for older people leaving the hospital, by supporting patients to be involved in their care during their hospital stay to SHANNON ET AL. support them to manage when they return home. This qualitative formative evaluation suggests that while there are a number of important challenges for supporting the PACT intervention, it was found to be acceptable for many patients and staff, with necessary changes and staff-facing implementation tools identified, which aim to increase its usability. The intervention will be further developed and tested in a subsequent feasibility trial.