Transitional care for patients with acute stroke—A priority‐setting project

Abstract Background The scope of this priority‐setting process is communication and collaboration in transitional care for patients with acute stroke. Actively involving persons with stroke and their family caregivers is important both in transitional care and when setting priorities for research. Established priority‐setting methods are time‐consuming and require extensive resources. They are therefore not feasible in small‐scale research. This article describes a pragmatic priority‐setting process to identify a prioritized top 10 list of research needs regarding transitional care for patients with acute stroke. Methods A pragmatic priority‐setting approach inspired by the James Lind Alliance was developed. It involves establishing a user group, identifying the research needs through an online survey, analysing and checking the research needs against systematic reviews, culminating in an online prioritization of the top 10 list. Results The process was completed in 7 months. A total of 122 patients, family caregivers, health personnel and caseworkers submitted 484 research needs, and 19 users prioritized the top 10 list. The list includes the categories ‘patients and caregivers’ needs and health literacy’, ‘health personnel's common understanding’, ‘information flow between health personnel and patients and caregivers’, ‘available interventions and follow‐up of patients and caregivers’, ‘interaction and collaboration between health personnel and caseworkers across hospital and primary healthcare’ and ‘disabilities after stroke’. Conclusion This paper outlines a pragmatic approach to identifying and prioritizing users' research needs that was completed in 7 months. The top 10 list resulting from this priority setting process can guide future research relating to communication and collaboration during the transition from hospital to the community for patients with stroke. Patient and Public Contribution Members of three stroke organizations participated in the advisory group. They gave feedback on the scope and the process, distributed the surveys and prioritized the top 10 list. Persons with stroke and their caregivers submitted research needs in the survey.


| INTRODUCTION
Depending on functional and cognitive disabilities, patients may need long-term rehabilitation after discharge from the hospital to ensure the best possible outcome. [1][2][3] The concept of transitional care includes both hospital and posthospital interventions that promote a safe and timely transfer of patients between levels of care. [4][5][6] However, while there is no consensus on what constitutes transitional care after stroke, engaging both patients and healthcare personnel is essential, and interventions may include discharge planning, patient and family education, follow-up care, rehabilitation and the transfer of information between care providers. 4,[6][7][8][9] Healthcare reforms in Norway have led to more responsibility for rehabilitation being transferred from hospitals to primary healthcare (PHC) in the municipalities, 10 resulting in earlier discharge from hospital. 11 Lack of involvement and communication, for patients in need of follow-up, can lead to uncertainty and a feeling of lack of control. 12 After hospital discharge, patients may experience a variety of unmet needs, including needs for more service, information and therapy, 13,14 indicating that there is insufficient communication between the patient and health personnel, as well as between health personnel across care levels.
There is an increasing focus on patients' and users' rights to participate in research, and legal and health policy documents support these processes. [15][16][17] Inviting users to set research priorities can reduce the described mismatch between patients' needs and researchers' agendas, 18 and reduce research waste. 19 In addition, since users may feel a greater sense of ownership of the research and its results, this could have an impact on the implementation of the research results, as the research is perceived as more relevant. 20,21 Priority setting processes focus on involving users to identify their research needs and to prioritize these needs. 18,22,23 There are few publications dealing with priority setting processes in the stroke context, and none on transitional care from hospital to PHC. Two large priority setting processes in the United Kingdom resulted in four top 10 lists: Pollock et al. 24 prioritized a top 10 list relating to life after stroke, Franklin et al. 25 a top 10 list for people with aphasia, while the Stroke Association prioritized two 10 top lists relating to stroke prevention, diagnosis, prehospital and hospital care and stroke rehabilitation and long-term care. 26 There is no gold standard method for setting priorities for research. 21,27 Because well-described priority setting approaches require resources, are time-consuming and often aim to set priorities at the national level, 23,27 they might need to be adapted to fit smaller research projects. 21 The aim of this article is therefore to describe the pragmatic priority setting process used to identify a prioritized top 10 list of research needs regarding communication and collaboration in transitional care for patients with acute stroke (TracStroke).

| METHODS
The TracStroke project is part of the Bridge Builder Initiative at Oslo Metropolitan University 28 and is affiliated with Akershus University Hospital. The university hospital has the largest stroke unit in Norway, serving 24 municipalities with approximately 560,000 inhabitants. 29 The TracStroke project is based on the Bridge-Building Model developed at the university, 30 which entails developing needs-led research approaches. The first author has attended a needs-led research course at the university, while the last author has previously participated in needs-led projects. The scope of the TracStroke project is examined in a Norwegian context, where increased responsibility for patients' care and rehabilitation was transferred to the municipalities through the Coordination Reform of 2012. 10 The scope is derived from a clinical question raised by clinicians at the hospital's stroke unit, as recommended in the Bridge Building Model. 30 As preparations are recommended when developing an approach, 20 the first author contacted health personnel and organizations for persons with stroke and requested their feedback on whether the scope of research needs relating to communication and collaboration in transitional care for patients with stroke was interesting and whether they wished to participate in a prioritysetting process. Since they responded positively, the project's scope was defined as 'communication and collaboration in the service transition from hospital to municipal rehabilitation after stroke'.
The method used for priority setting in the TracStroke project was strongly inspired by the James Lind Alliance (JLA) Guidebook, which explains the methods and principles of its priority-setting partnership (PSP), 23 as well as by frameworks described in other needs-led research. 20,31 The JLA is a systematic process in which patients, caregivers and health professionals are invited to take part in an equal PSP to identify and prioritize research needs. 23 The process includes appointing a steering group and establishing a PSP and identifying the users' research needs, before processing and checking them against systematic reviews. The final step is a prioritization process that includes both an interim process and a workshop to arrive at a top 10 list of prioritized research needs.
The JLA is a well-known method, 32 but it requires resources and supervision by a JLA consultant 23 and is time-consuming. 27 This means that it was not feasible for the TracStroke project and the research group to use. Our approach did not include the roles and responsibilities of the JLA resources or the steering group, 23 and the research group carried out and led the project. During the development of the approach, the focus was on creating a pragmatic and feasible approach to identify users' research needs within the timeframe and limited resources available. Users were broadly defined to include persons with stroke, their caregivers and health personnel all of whom could be affected by the research. 16,17 The research group consisted of the authors of this article. As several authors encourage describing the approach used 33,34 and the use of checklists to ensure the quality and transparency of the approach, 35,36 the steps in the TracStroke priority-setting approach (Figure 1) will be described in this section.

| Establishing the user group
The stakeholders in this process were broadly defined to include patients with stroke, their caregivers and various health personnel.
Persons with stroke and caregivers were important as they experience the outcomes of a stroke, and their experiences are important when developing research that might meet their needs.
Health personnel in both the hospital and PHC, and caseworkers in the municipalities were included since they communicate, cooperate and coordinate their activities to make sure patients receive appropriate care after discharge from the hospital. Unlike the JLA PSP, where the steering groups work closely with the researchers, 23 the user group had an advisory role and was kept informed and consulted throughout the process. The user group's main tasks were to give feedback and advise about the scope and the priority-setting process, and promote and participate in the online surveys for gathering and prioritizing research needs.
As the priority-setting process entailed meetings, members were recruited from neurologic and allied health units at the university hospital, the municipalities around the hospital and the local branches of the three organizations for persons with stroke. The invitation included information about the scope and purpose of the project, and that their involvement entailed an advisory role. They were also informed that meetings would take place at the hospital, and that refreshments would be provided and travel expenses refunded.
At the first and only meeting with the user group, the scope and priority-setting approach were presented and discussed. The group was encouraged to discuss the different concepts and terms used to describe the scope and the approach, their relevance, and other aspects that triggered engagement. A consensus was reached on the scope, the members' involvement and responsibilities in relation to gathering and prioritizing research needs.

| Gathering research needs
To identify the research needs of patients, caregivers, health personnel and caseworkers, the research group developed an online survey. 23 The questions were phrased differently for persons with stroke and their caregivers, and health personnel and caseworkers to reflect on their specific experiences in the transition ( Table 1). The user group piloted the survey and was asked to comment on the questions, their wording and user-friendliness. There were no limitations on how many inputs each respondent could provide.
Participants were recruited using convenience/snowballing sampling to reach as many as possible within the timeframe. The survey was promoted by the user group and the hospital through

| Analysing research needs
Before the analysis, the data set was managed in Microsoft Excel to organize and group research needs and to ensure that each respondent group's (patients, caregivers, health personnel in hospital, health personnel in PHC, caseworkers) research needs would be apparent as recommended by the JLA. 23 The submitted research needs were organized under the survey question categories: information, collaboration and other aspects of discharge.
As JLA recommends categorizing and grouping similar research needs, 23 the qualitative analysis of the submitted research needs was inspired by thematic analysis. 37  and a pragmatic process that makes the list to be checked against the evidence base shorter and more manageable ( Table 2).

| Checking research needs
After the analysis, to prevent duplicate research, the list of research needs was checked against systematic reviews to verify whether there were true knowledge gaps. 23 Questions that had been addressed and answered by systematic reviews were removed from the list, while relevant research recommendations from the reviews were added to the list. 23  The abstracts of the systematic reviews were read to check whether the research needs categories were mentioned as outcomes or categories for the review.
The analysis and checking of research needs were performed by the first author and checked and discussed by the research group.
The user group received a PowerPoint presentation, with an audio explanation of the analysis in an email, and was encouraged to ask questions if anything was unclear and to give feedback.

| Prioritizing research needs
Due to the COVID-19 restrictions in Norway, which resulted in a national lockdown, the prioritization was carried out in two steps and did not include workshops as recommended in JLA (Figure 1). Before

| Ethical considerations
In Norway, needs-led research does not require approval by the Regional Committees for Medical and Health Research. The study was discussed with the Norwegian Centre for Research Data and, since the survey did not gather personal information and contained informed consent information, it did not require their approval either.
The user group received and signed a written informed consent at the first meeting. The survey contained informed consent information in accordance with the guidelines set out by the Norwegian Centre for Research Data. The first page of the survey contained information outlining the purpose of the study and its anonymous nature and asked participants to refrain from writing confidential information. Participants were informed that their participation was voluntary and that by answering the survey they gave their consent.

| RESULTS
The TracStroke priority-setting process lasted from September 2019 until April 2020. How to communicate with patients with aphasia.

Patient and caregiver needs
Do the patients understand the information they are given?
How do patients experience the information given by health personnel?
How do the patients perceive the information given to them?
Collaboration How can the hospital and the community collaborate to make each other better?
How do health personnel cooperate across levels of care and how can the collaboration be improved?
Is there any collaboration and how can it be improved?
How to prioritize collaboration in a hectic work situation.

| The user group
The user group consisted of 12 members (Table 3). The three regional stroke organizations were invited to appoint a member each. The hospital appointed two nurses, one neurologist, one physiotherapist and one consultant. Two physiotherapists, one occupational therapist and a physician represented the municipalities.

| Participation in the online survey
The survey gathered research needs from 122 patients, caregivers, health personnel and caseworkers through the online survey ( F I G U R E 2 Flow-chart of the TracStroke process.
'Common Understanding' were submitted by health personnel, and patients and caregivers submitted 81% of the needs relating to 'Disabilities after stroke'.

| Checking the research needs
The systematic search identified 26 relevant systematic reviews.
The 62 research needs were checked against the systematic reviews. Twelve of the reviews answered seven research needs, six gave rise to nineteen questions, while eight did neither ( Figure 2

| The prioritaztion of the top 10 list
In the interim prioritization, the research group produced a shortlist of 19 questions that were represented by at least three of the six respondent groups (patients, caregivers, health personnel in hospital, health personnel in PHC, caseworkers, systematic reviews).
In the final prioritization, 19 users (

| Strengths and limitations
In the process of identifying research needs, we have followed the JLA, but have made some pragmatic adjustments to the approach.
The TracStroke approach did not include a JLA consultant nor a steering group, and these roles were filled by the research group. By assigning these responsibilities to the research group, we could save both time and costs, although this affected the level of involvement compared to a JLA PSP. Involving the user group in the analysis was not planned for as this requires specific competencies. 23 We could have involved the users in the interim prioritization through an online survey. This could have changed the outcome list for prioritization.
However, the interim prioritization ensured that the research questions for final prioritization were broadly supported by the different respondent groups, and it also saved us time.
We aimed for and, in part, achieved a broad composition of participants in the user group. Unfortunately, caregivers and caseworkers from the municipalities were not represented and more effort should have been made to include these groups. We believe that the user group felt ownership of the project since it promoted and distributed the survey through different media and reached the different stakeholder groups. The research needs gathered through the survey added new perspectives to the research group. Hopefully, this participation will contribute to research being more relevant and to it meeting the needs of those who will benefit from and implement the research. 20 Although the survey resulted in broad participation from different user groups, we did not gather any demographic information. Online surveys are unlikely to reach the voices of those in low socioeconomic groups, those with the greatest unmet needs, or those with low health and/or online literacy. 53 These voices tend to be less represented in research. Face to face meetings in the communities could be more suited to reach these groups. 54 The prioritization, although group discussions and agreement were lacking, included more and broader user participation than the user group (Table 4). In addition, the online survey may have been a positive feature as the participants could prioritize their own opinions, without being influenced by those with strong voices. 23 Online prioritization could also have been used for the interim prioritization, which could have affected the final list for prioritization. Following the pandemic, JLA has described how workshops can be adapted to an online setting. 55 An online conference was not considered feasible to hold due to the strain on the healthcare personnel during the pandemic and a lack of facilities, such as laptops, computers and space to participate.
One strength of the TracStroke process is its qualitative approach, which includes input not written as questions. This