The implementation of evidence-based rehabilitation services for stroke survivors living in the community: the results of a Delphi consensus process

Abstract

Background:

Recovery from stroke requires the provision of specialist rehabilitative care, yet there is a lack of research evidence on optimal methods of delivery and inequitable service provision across countries. Following consensus on Early Supported Discharge, our aim was to define core components of evidence-based community stroke services.

Design:

We used a modified Delphi approach with a purposive sample of 26 UK-based expert panellists (10 academics, 15 stroke service leads or commissioners, one stroke survivor). Statements based on research literature and policy documents were generated by an independent, mixed academic and service improvement team of nine. In three rounds of consultation panellists indicated their level of agreement with statements. Free text comments were analysed thematically.

Results:

Consensus of opinion (>70% agreement) was obtained on 76 of 80 statements. Panellists agreed that stroke specialist care tailored to clinical need should be provided following discharge from hospital. Flexible care pathways should be commissioned and provided through strategic and collaborative leadership across health and social care. Teams need to deliver stroke specialist rehabilitation, handing over responsibility to non-specialist services when patient rehabilitation goals have been met. Lack of consensus on whether to provide services exclusively for stroke patients and the need of support for carers highlights areas for further research.

Conclusions:

Consensus on the provision of evidence-based community stroke rehabilitation has been reached. Commissioning services with a clearly defined remit, which can deliver tailored care to individual stroke patients, is a challenge that needs further exploration.

Introduction

Stroke is a leading cause of disability worldwide. Its greatest impact on the health of individuals and the cost to society is from the long term consequences faced when stroke survivors leave hospital. ¹ There is increasing emphasis on the delivery of healthcare in the patient’s home, particularly as health services face up to the challenges of delivering better care while simultaneously reducing costs. In stroke, optimal recovery requires the provision of specialist rehabilitation, and cumulative research evidence has reported the benefits of a variety of interventions. ^{2 –4} As a consequence, National Clinical Guidelines from various countries recommend a wide range of treatment options that clinicians and commissioners need to consider. ^{5 –8} However, there has been a lack of systematic evaluation of optimal methods of delivery of such interventions, particularly in the community, and as a consequence service provision for stroke patients once they leave hospital is patchy and often poorly organised. ^9,10

Over the last decade there has been increasing focus on Early Supported Discharge services: stroke specialist multidisciplinary teams that facilitate the transfer of patients from hospital to home (contributing to a reduction in hospital bed days) and provide rehabilitation and support in the early stages of being at home. ^{11 –13} Since Early Supported Discharge is beneficial for a selected group of stroke survivors only (mild to moderate disability), ^{11 –13} guidance is lacking on how to implement evidence-based services that provide care for the wider stroke population and that operate over a longer term.

The aim of this study was to establish the core components of evidence-based community stroke services by using a modified Delphi consensus approach and building on a recently published Early Supported Discharge consensus document. ¹³ Since a lack of robust research evidence concerning the delivery of community-based stroke services was identified, this study was informed by learnings from practice through collaboration with the NHS Stroke Improvement programme, founded by the Department of Health in England to implement the National Stroke Strategy. ¹⁴

Methods

A core team of nine people, including two academic researchers (authors RF and MW), a community stroke team lead and members of the NHS Stroke Improvement Programme (including authors IG and DJ), agreed the framework and scope for the consensus activity. Core team members conducted a literature review to identify and select research literature documents (systematic reviews or randomized controlled trials) ^{3,15 –23} or clinical guidelines and policy documents ^{5 –8} that were deemed suitable for generating statements about the implementation of community stroke services. Within the agreed framework, statements were generated based on findings from the research literature or recommendations from policy documents. Each statement was discussed at length and final wording agreed by all core team members over three dedicated meetings.

A consensus panel of 26 individuals, separate from the core team and to whom the statements would be sent, was assembled. Purposive selection of panel members by the core team was based on agreed criteria: the individuals’ knowledge of national stroke policy, their perceived ability to respond to statements from a national perspective, their field of expertise or knowledge of research literature and leadership experience within stroke care. Of the 26 individuals invited to participate (contact facilitated by authors MW and DJ), 25 agreed to take part with just one individual suggesting a more appropriate representative. The decision to limit the panellist number to 26 was based on the authors’ experience of pragmatic use of the modified Delphi approach (and the number of statements that had been generated), in addition to the core teams’ knowledge of the combined expertise of the selected panel members. The consensus panel was composed of 10 academic researchers, 15 service leads or commissioners and one stroke survivor. Panellists were selected from a cross section of stroke services and research across England, with 10 panellists working in a hospital setting, eight in the community and eight at a University with a mixed geographical representation: East Midlands (nine), South West/Coast (five), London (four), Yorkshire (three), West Midlands (two), Glasgow (two) and Northumberland (one).

In accordance with a modified Delphi process, consensus panel members were asked to indicate their level of agreement with each statement, ^24,25 with the option to make comments via free text as appropriate. Panellists were asked to adopt a national and aspirational perspective when responding to statements, rather than focusing on the current state of local services with which they may work. Consensus of agreement was determined by over 70% (19 panellists or more, P = 0.05) of panellists giving ‘Strongly agree’ or ‘Agree’ responses (combined score); similarly consensus was also reached if over 70% of panellists disagreed and gave ‘Disagree’ or ‘Strongly disagree’ responses. ²⁴

All statements (eleven in total) for which consensus had not been reached in round 1 were circulated again in round 2 and panellists were given the opportunity to review their level of agreement with each statement, in light of the group response. In round 3, six of the statements for which consensus had not been reached in rounds 1 and 2 were slightly reworded, based on comments from panel members and one statement was split into two. ²⁴ Following round 3, final consensus levels for each statement and the median response were calculated.

Free text comments from all rounds of the modified Delphi process were analysed qualitatively. Consensus headings (e.g. Organisation of community stroke services, stroke specialist care beyond hospital, etc.) allowed a top-down structure to be imposed on the data. ²⁶ An iterative form of content analysis involved organising the free text into themes based on frequency and relevance to consensus headings, which were then subject to further review and interpretation.

Results

A consensus document was created listing statements that the consensus panel, as a group, had reached consensus agreement on, statements that the group had reached consensus and disagreed on, and any statements that the group were unable to reach consensus on. Consensus of opinion (>70% agreement) was obtained on 76 of the 80 statements. Clear guidelines are provided by these statements that are presented in Tables 1 to 9 presented online. For brevity’s sake only summaries of each section of the consensus document will be provided here. Specific statements commented on below are those for which consensus was not reached and which were informed by free text from panellists.

Organisation of community stroke services

Panellists agreed that there is a need for an integrated pathway of stroke care following discharge from hospital that relies on strategic leadership and a dedicated network of service representation across health, social care, voluntary and third sectors. Consensus was also reached on the fact that those responsible for procuring and commissioning community stroke services should be active participants, by being involved in the design, monitoring and ongoing support of services (Table 1).

Stroke specialist care beyond hospital

Consensus was not reached on some of the statements relating to what defines a community rehabilitation team as stroke specialist (Table 2). A statement was split into two parts in round 3 of the modified Delphi process, so that issues concerning the treatment of stroke patients could be dealt with separately from the experience and knowledge of staff. Whereas panellists reached consensus on the fact that a stroke specialist team should predominantly treat and have experience and knowledge of people with stroke, whether they should treat stroke patients exclusively, remains undecided. Two panellists argued that the service should be dedicated to the care of stroke patients, “Non-stroke patients take up capacity, and can be treated by non-specialist therapists.” Seven other panellists believed that a neurological rehabilitation team offering stroke care as a speciality, would be equally appropriate, as the skills of staff are transferable. There was concern that it might not be viable to implement a team that just supports stroke patients, even if it was the ideal. “There may also be regional variances according to geography and demography – it isn’t necessarily one size fits all.”

Decision making about pathways of care according to need

Panellists agreed on the need for distinct stroke care pathways based on stroke severity, highlighting the need for measurement of stroke severity before transfer from hospital (Table 3). Consensus was reached on the fact that stroke survivors not eligible for Early Supported Discharge, but that have ongoing rehabilitation needs, should be referred to a community stroke rehabilitation service when transferred from hospital. It was agreed that severe stroke survivors, not eligible for Early Supported Discharge owing to complex stroke-related needs, require rehabilitation in hospital and should only be transferred into the community when they can be supported at their place of residence.

Provision of information

Panellists agreed that an information strategy should be produced, that clearly describes a pathway of provision of information and support to all stroke patients and their carers, following transfer from hospital. Although four panellists felt it would be better to define who the provider of an information and support service for patients and carers should be, four others suggested that an interdisciplinary approach was required (Table 4). It was suggested that a “Key care coordinator might be best to avoid multiple people thinking they are responsible for everything”. One panellist pointed out that “Partnership working with adult social care is critical. Assisting stroke survivors and carers to access fully personalised budgets and proactive use of care packages would support self management.”

Community stroke rehabilitation team: intervention

These statements explored critical issues around the length and intensity of the intervention provided by the community stroke rehabilitation team and panellists’ understanding of Early Supported Discharge, and how this relates to a community stroke team (Table 5). In rounds 1 and 2, an initial statement version referred to durations of intervention of “between 8 weeks for mild stroke survivors and 17 weeks for those with severe and complex deficits” for which consensus was not reached. Seven panellists questioned the need to specify intervention length, particularly as this was perceived to be “contrary to the concept of individual need”. Thus in round 3, the statement was altered, based on panellist feedback, to state that duration of intervention “should be based on clinical need tailored to goals and outcomes” and consensus was reached (Table 5).

Similarly, nine panellists expressed concern about a reference to 45 minutes of therapy in an initial version of a statement circulated in rounds 1 and 2: “the community stroke rehabilitation team should offer 45 minutes of therapy to each patient, during each visit, unless there are reasons for this to be inappropriate”. Panellists argued that it was practically difficult to set time limits on rehabilitation duration and that “there is not strong evidence for 45 minutes as the baseline.” Consensus was reached in round 3 on the amended statement that intensity of therapy should be “based on clinical need tailored to goals and outcomes” (Table 5).

It was apparent from eight panellist responses that they recognised a clear distinction between an Early Supported Discharge and a community stroke rehabilitation team, with Early Supported Discharge being associated with a higher intensity of intervention. “Early Supported Discharge should be separate from a community stroke team or the team becomes blurred and will not meet the different needs of patients”. Five other panellists suggested a more flexible approach motivated by a concern that conceptualising these services as separate entities would compromise a fully integrated pathway of care. “I run an Early Supported Discharge/Community Neurorehabilitation team which is completely integrated – the patient would see no distinction between the two ‘models’ apart from intensity of treatment which reduces naturally in line with patient need/goals.”

Although it was agreed that carers of stroke survivors should receive individualised training (Table 5), consensus was not reached as to whether support (including provision of information) should be provided by stroke specialist rehabilitation services (Table 4). Seven panellists pointed out the benefits of other sources of support and that “other providers should be identified and trained accordingly”.

Community stroke rehabilitation team: model of team

Although panellists reached consensus on guideline whole time equivalent figures for staff members of a community stroke rehabilitation team (Table 6), five panellists argued that some of the figures quoted were too low, particularly with regard to rehabilitation assistants. “Rehab assistants are extremely useful flexible team members and should feature more.” This also relates to the lack of consensus around the staffing level of a social worker, despite agreement that this discipline should feature in the team. Lack of a clear distinction between Early Supported Discharge and community stroke teams for some panel members was also apparent in comments around the model of team. “If Early Supported Discharge were offered as part of the team’s work, the complement of staff would need to rise significantly from the suggested whole time equivalent above in order to manage the intensity required by Early Supported Discharge patients.” One panellist suggested that “community stroke rehabilitation teams should be staffed appropriately post careful demand and capacity calculations.”

Community stroke rehabilitation team: access and transfer of care

The reference to the need for the identification of clear goals for rehabilitation in order for stroke survivors to access services (Table 7) stimulated a number of comments from panellists. Nine panellists expressed concern that patients not able to participate in goal setting may be deprived of stroke specialist rehabilitation. It was felt that “Goals may not be particularly clear initially due to psychological, communication or cognitive factors.” Also that advice on seating, transfers and feeding may not relate to clear goals but to adjustment to disability. Four panellists recognised the importance of goal setting as a means to determine access and discharge from the service and the need to avoid advocating “unlimited and unfettered access” in order to assist with service planning. Two panellists drew attention to the type of intervention the team offers and the desire to offer both rehabilitation and longer-term disability management, “The service I run is not purely ‘rehab’ focussed. We offer disability management and ongoing advice and support.”

Community stroke rehabilitation team: performance indicators

Statements in this section cover outcome measures that should be used by teams to measure their effectiveness (Table 8). Panellists agreed that teams should standardised outcome measures, including measures of degree of dependency (e.g. Barthel score), general health, mood evaluation, activities of daily living and goal attainment. It was also agreed that teams should monitor carer health status, quality of life and mood.

From healthcare to reintegration

Comments from seven panellists reflected a reluctance of community stroke services to hand over responsibility to other services, although panellists agreed with a need for the planned withdrawal of stroke specialist rehabilitation (Table 9). “I would argue that stroke specialist rehabilitation should also be about participation and re-integration to community life and not just impairment orientated”. It was highlighted that the General Practitioner or community doctor might not at present recognise their stroke-related responsibilities and should have support from stroke consultants and the opportunity for joint training in stroke.

Discussion

Consensus has been reached between a mixed panel of academic researchers, service leads, commissioners and a stroke survivor on the provision of services for stroke survivors living in the community. This consensus document identifies core elements of evidence-based community stroke services and provides guidelines for implementation.

Our approach is novel in that we recognised that current research evidence did not address specific issues relating to the implementation of services. We included statements based on national guidelines, recommendations from our NHS Stroke Improvement Programme colleagues and the research evidence base. One weakness is that this necessitated some interpretation and statement composition that, if alternatively worded, may have achieved different results. Our mixed consensus panel of academics, service providers, commissioners and stroke survivor were selected on their ability to respond to statements from a national perspective, using leadership experience and knowledge of research and stroke policy. We acknowledge limitations of the study resulting from the inclusion of just one stroke survivor and no carers, however, feel these are partly compensated for by the chosen stroke survivor’s knowledge and experience of stroke and national role in stroke service development. Although we are confident that the consensus panel represented a group of experts in the topic area, we may have achieved different results had we chosen an alternative panel composition. We also acknowledge this work had a UK focus, though we would argue that findings can inform stroke care on an international basis.

One topic that we did not directly explore in this consensus was funding of services. In England, the funding of community services and rehabilitation lies outside of the Payment by Results scheme, ²⁷ resulting in procurers of services developing a variety of solutions to fund services. Our aim was to identify core components of community stroke services that could be implemented regardless of geography, and as such we avoided a level of detail for which local issues might complicate matters. It is acknowledged that financial arrangements are a major determinant of whether services are commissioned, highlighting that those responsible for procuring services should be active participants in the design, monitoring and ongoing support of services.

Lack of consensus on four of the statements highlights remaining areas of uncertainty and the need for further research. A key issue that has been raised is whether a community stroke rehabilitation team should, or should not, treat stroke patients only; this topic is discussed further below and warrants further investigation. Although it was agreed that carers, in addition to stroke survivors, should be supported, it remains unclear whether this needs to be by a specialist stroke rehabilitation team or by voluntary sector staff with stroke specialist training. More research is required to understand the needs of carers of stroke survivors and the most effective ways to meet them. ²⁸

Remits of community stroke and Early Supported Discharge teams have been clarified, reinforcing conclusions from an earlier Early Supported Discharge consensus. ¹³ Early Supported Discharge is a distinct, high intensity intervention, specifically suited for mild to moderate stroke patients, that can be offered as part of a community stroke service or by a separate dedicated team. An area of debate that statements impact on, but perhaps do not resolve, concerns whether Early Supported Discharge and community rehabilitation can or should be provided to stroke patients by more generalist rehabilitation services. One could argue that providing services exclusively for stroke patients creates a conflict, both conceptually and practically, between providing clearly defined services focused on a particular clinical speciality (in this case stroke) and ensuring that patients with similar problems can also access the care they need. Panellists unanimously agreed that a community stroke rehabilitation team is a multidisciplinary group of specialists, who have skills and competencies acquired through stroke-specific training and who have experience of treating and managing stroke survivors. While this does not preclude services providing rehabilitation to patients with other long term conditions, it does emphasise the fact that staff treating stroke survivors need to be stroke specialist, which may be difficult to maintain in a more generalist service. What this consensus document does provide are core, evidence-based components that need to be met by commissioned services.

A key finding was that the intensity and length of intervention delivered by teams should be based on clinical need, tailored to goals and outcomes rather than fixed time periods. This highlights a need for community stroke services to agree and adhere to a standard definition of stroke specialist rehabilitation, ^2,4 with the use of goal setting and monitoring as a means to govern access and discharge from the service. This should not prevent access to services by stroke patients with communication or cognitive problems, for which goals can be established through support from therapists and carers. There is a need for more research to investigate optimal levels of intensity of rehabilitation therapy, particularly with regards to delivering care in practice and from a multidisciplinary team perspective. It is also important that community stroke teams strive to empower people with stroke and their families to move towards reintegration into the community, and hence towards non-specialist services once their rehabilitation goals are met; this will be especially relevant in the current climate of financial austerity.

Furthermore, it suggests that flexibility needs to be built into current service specifications, while at the same time being clear about the service’s overall remit and key performance indicators. There is an insoluble conflict between the clinical imperative to meet the unique clinical needs of the individual stroke patient that are uncontrolled and unpredictable, and the commissioning and management desire to have clear service parameters that increases control and predictability. But this tension is a common feature of complex systems of healthcare and is not unique to this situation. ²⁹ It might be better if service specifications were geared toward a whole systems approach ³⁰ identifying what a care pathway of services should deliver, rather than detailing specific team characteristics. We have established the core principles to guide this approach.

It is clear that stroke specialist care must be provided once stroke survivors leave hospital, and that flexibility around the range of options and services offered is required depending on severity of stroke. This requires close collaborative working across different service providers, between different health disciplines and social care and a procurement system that promotes this. National audits of stroke care in the community are needed to galvanise this process and should be informed by this document.

Clinical messages

• The intensity and length of intervention delivered by stroke specialist teams should be based on clinical needs tailored to goals and outcomes.

• An organised and flexible pathway of stroke-specialist rehabilitative care needs to be provided for stroke survivors once they leave hospital, requiring a procurement system that spans health and social care.