Factors guiding therapist decision making in the rehabilitation of physical function after severely disabling stroke – an ethnographic study

Abstract Purpose Whilst strong evidence supports rehabilitation to improve outcomes post-stroke, there is limited evidence to guide rehabilitation in the most severely disabled group. In an era of evidence-based practice, the aim of the study was to understand what factors guide physiotherapists (PTs) and occupational therapists (OTs) to select particular interventions in the rehabilitation of physical function after severely disabling stroke. Material and methods An ethnographic study was undertaken over an 18-month period involving five London, UK stroke services. Seventy-nine primary participants (30 PTs, 22 OTs, and 27 stroke survivors) were recruited to the study. Over 400 h of observation, 52 semi-structured interviews were conducted. Study data were analysed through thematic analysis. Results Key factors guiding therapist decision making were clinical expertise, professional role, stroke survivors’ clinical presentation, therapist perspectives about stroke recovery, and clinical guidelines. Research evidence, stroke survivors’ treatment preferences, organisational type, and pathway design were less influential factors. Therapy practice did not always address the physical needs of severely disabled stroke survivors. Conclusions Multiple factors guided therapist decision making after severely disabling stroke. Alternative ways of therapist working should be considered to address the physical needs of severely disabled stroke survivors more fully. Implications for rehabilitation Multiple factors guide therapist decision making after severely disabling stroke, some of which result in the use of interventions that do not fully address stroke survivors’ clinical needs. Therapists should critically reflect upon their personal beliefs and attitudes about severely disabling stroke to reduce potential sources of bias on decision making. Therapists should consider the timing and intensity of therapy delivery as well as their treatment approach to optimise outcomes after severely disabling stroke.


Introduction
Stroke is a clinical syndrome defined as "a neurological deficit attributed to an acute focal injury of the central nervous system by a vascular cause" [1].The Global Burden of Disease study reported that stroke was the second most common cause of death and disability globally [2].In recent years, medical technologies designed to limit the extent of the stroke event itself, such as thrombolysis and mechanical thrombectomy, have become part of routine clinical practice in high income countries [3][4][5][6].However, rehabilitation is considered to be the mainstay of treatment post-stroke in all health systems [7].
Rehabilitation can be defined as a process designed to optimise function and reduce disability in individuals in interaction with their environment [8,9].Due to the high prevalence of sensorimotor impairments post-stroke [10], rehabilitation of physical function comprises a large component of stroke rehabilitation programmes delivered by healthcare professionals, such as physiotherapists (PTs) and occupational therapists (OTs) [11].Physical function refers to the ability to perform various bodily activities, ranging from basic movements to complex activities [12,13].There is good evidence supporting the use of rehabilitation interventions to improve key aspects of physical function, such as motor function, activities of daily living (ADLs), and balance, in the general stroke population [14][15][16].However, there is a paucity of evidence demonstrating the effectiveness of interventions in the rehabilitation of physical function specifically for survivors of severely disabling stroke [17].
Severely disabling stroke can be conceptualised as a stroke that significantly impacts upon an individual's functioning and causes significant residual disability, which could be any combination of impairments, activity limitations, and participation restrictions [17][18][19].Survivors of severely disabling stroke may present with significant motor, communicative, and/or cognitive impairments, require assistance for ADLs and be unable to walk.Severely disabling stroke may be classified using scores on standardised outcome measures.However, there is discrepancy in the literature regarding what scores should be used.Consequently, it is estimated that between 9.6% and 33.8% of people admitted to hospital with stroke have sustained a stroke initially classified as severe [20][21][22][23][24].Despite this uncertainty, understanding how rehabilitation interventions address the needs of severely disabled stroke survivors is important because this group experience high rates of mortality and post-stroke complications [25][26][27][28][29][30][31], longer hospital length of stay [32,33], less complete recovery of ADLs [23,[34][35][36], increased likelihood of institutionalisation [37,38], reduced quality of life [39], and high levels of caregiver burden [39].
Although organised stroke unit (SU) care can reduce mortality and the likelihood of institutionalisation after severely disabling stroke [40], it is not clear what particular interventions contribute to these outcomes.In addition, the lack of evidence demonstrating the effectiveness of interventions in the rehabilitation of physical function in this group presents clinical challenges.A recent survey of therapist practice highlighted the most frequently used rehabilitation interventions were whole-body positioning, sitting balance practice, and carer training [19].However, it is not clear if current therapy practice is effective at improving outcomes in this group [17].In the current era of evidence-based practice (EBP), treatment decisions should be underpinned by the best available research evidence and be guided by patient preferences [41,42].However, in the absence of research evidence and the increased likelihood of communication and cognitive impairments after severely disabling stroke [43][44][45], it is not clear how healthcare professionals decide which interventions to use to optimise physical function for survivors of severely disabling stroke.
As the selection of rehabilitation interventions may have direct implications on outcomes post-stroke, understanding therapists' clinical decision making more fully may provide greater insight into how therapy addresses the physical needs of severely disabled stroke survivors.Clinical decision making can be defined as the process of making a choice between options, such as assessment procedures, diagnoses, and clinical treatments [46,47].Whilst this definition may seem simplistic, clinical decision making represents a complex phenomenon that involves multiple factors and individuals; occurs in dynamic, time-pressured contexts; and is often characterised by high levels of uncertainty [46][47][48][49].As clinical decision making is a fundamental aspect of autonomous practice for healthcare professionals [50], understanding the factors that positively influence or negatively impact upon therapist clinical decision making is important.
Previous research has identified several factors that may influence therapist decision making after stroke.These factors include compliance with organisational practice, resource availability, and family or social support [51][52][53][54][55].However, these factors were identified in studies that focused on clinical decision making regarding broader aspects of therapy practice, such as general treatment approaches, suitability for ongoing stroke rehabilitation, and discharge planning.These studies did not focus on decision making regarding more specific aspects of therapy practice, such as the selection of different interventions, nor focus on decision making after severely disabling stroke.Therefore, the aim of this study was to explore the factors that guide PT and OT decision making in the rehabilitation of physical function after severely disabling stroke.

Design
Ethnography was chosen to explore and understand the factors that guide therapist decision making in the rehabilitation of physical function after severely disabling stroke.Ethnographic research uses a combination of participant observation and interviews to describe the perspectives and meanings constructed by people in their usual settings [56,57].The use of participant observation and interviews allows for comparisons between what people say in interviews and what they do in practice, thereby providing a greater understanding of people's reality [58].As ethnography has been previously used in healthcare research exploring the decision-making processes of healthcare professionals [53,[59][60][61], it was considered an appropriate research approach to achieve the study's aim.

Research team
The primary researcher, MM, conducted the study as part of a PhD in Health Services Research.MM is a PT with over 20 years clinical experience in stroke and neurological rehabilitation.CM, CS, and RFG supervised the research.

Ethical approval
HRA and ethical approval was granted by the National Research Ethics Committee London-Queen Square (REC reference: 17/LO/1243; IRAS project ID: 218301).Local R&D approval was also obtained from each stroke service that participated in the study.

Study setting
Stroke services within the Greater London, UK region were considered for study inclusion (Figure 1).Purposive sampling was used to select stroke services across the London stroke pathway to investigate therapy practice.These services were those that severely disabled stroke survivors were likely to access in London based on published pathway design criteria and data from SSNAP, the National Stroke Register.Consideration was also given to stroke services in various parts of London as well as stroke services that formed part of the same stroke pathway, such as selecting a SU and then selecting a community stroke service that accepts referrals from this SU.Consequently, five stroke servicesone hyperacute stroke unit (HASU), two SUs and two community services -were selected for the study.

Participants and recruitment
The primary researcher approached the therapy manager of each stroke service to identify potential therapist participants, who were any PT or OT working in the stroke service at the time of the study.Following a brief presentation outlining the study, participant information sheets were provided to all potential therapist participants, who were given at least 24 h to decide about study participation.Once consented, therapists identified potential stroke survivor participants based upon outcome measures used within each stroke service to determine stroke severity and disability.In the HASU and SUs, this was any stroke survivor with an initial National Institutes of Health Stroke Scale (NIHSS) score �16.In the community stroke services, this was any stroke survivor with a Barthel Index (BI) score �9 or Modified Rankin Scale (mRS) score �4.Therapist participants approached potential stroke survivor participants, usually in the presence of a family member, regarding study participation.Interested stroke survivors and/or family members were then approached by the primary researcher, who further discussed the study and provided a patient information sheet tailored according to the stroke survivor's level of comprehension.A minimum of 24 h was provided to stroke survivor participants to decide about study participation and obtain consent.For stroke survivors who lacked capacity to provide informed consent, personal consultees (e.g., family member) or nominated consultees (e.g., ward matron) were identified, based upon guidance from the Mental Capacity Act 2005 [62], to enable stroke survivor participation.
In total, 79 study participants were recruited to the study -30 PTs, 22 OTs, and 27 stroke survivors (Supplementary Materials 1 and 2).As observed sessions sometimes involved carers or other staff members, these individuals were invited to participate in the study as secondary study participants.In total, 37 secondary participants were recruited to the study -31 staff members (therapy assistants, nurses, doctors, and speech and language therapists) and six carers.

Data collection
Data were primarily gathered through participant observation of therapists and stroke survivors, as well as in-depth interviews with therapist participants.The primary researcher observed therapy sessions, goal setting sessions, and team meetings.Sessions and meetings were mutually agreed upon by therapist participants and the primary researcher prior to the session or meeting occurring.Observations were documented contemporaneously using a small note pad and were guided by Spradley's dimensions of participant observation (location, time, event, actors, acts, and goals) [63].Throughout the study, the primary researcher maintained a reflective diary to record pertinent reflections from observation sessions and critically reflect upon his role in the research process [64].Interviews were conducted towards the end of the fieldwork at each stroke service.Interviews were conducted by the primary researcher and involved each therapist participant (30 PTs and 22 OTs).Interviews were semi-structured in nature and facilitated by using an interview guide (Supplementary Material 3).Interviews were recorded using a digital voice recorder and lasted between 30 and 60 min.Interviews were primarily transcribed by the primary researcher, with support from a professional transcriber.

Data analysis
Interview and observational data were analysed by the primary researcher using Braun and Clarke's six-phase process of thematic analysis [65].This is a widely used approach in health and social care research to analyse qualitative data to generate broader themes [60,66].Interview transcripts and observation fieldnotes were read several times in order to become familiar with the dataset (phase 1).Interview transcripts were uploaded onto NVivo and observation fieldnotes were transferred to a Word document.Phase 2 involved the generation of codes, or units of meaning, from both data sources.Interview transcripts were coded first using by line-by-line coding.The rationale for coding interview transcripts first was that interview data covered a wider range of concepts -decision making, prognostication, attitudes and beliefs, reflections on specific observations -than observational data, which tended to focus on the immediate therapy sessions.The phases of generating and developing themes (phases 3 and 4) involved condensing codes into categories, or larger groups of codes, and then identifying factors guiding decision making from these categories.The model for evidence-based decision making was referred to during the identification of these factors.Themes were generated based upon grouping similar factors together into therapist related, patient related, evidence related, and organisation related themes.Finally, the sixth phase of thematic analysis involved writing up the report.An example of how codes were combined to form categories, factors and themes is included as supplementary material (Supplementary Material 4).Throughout the data analysis phase, the primary researcher discussed code generation and thematic development with the research team to ensure rigour.

Results
Over an 18-month period (December 2017 and June 2019), approximately 400 h of participant observation and 52 semi-structured interviews were conducted.Data collected and analysed generated five themes: professional expertise, therapist perspectives about stroke recovery, attributes of the severely disabled stroke survivor, documentary evidence, and therapy within the wider stroke pathway (Figure 2).Each theme comprised two factors guiding decision making, and these factors will be presented here.Some factors were more influential in guiding therapist decision making and more consistently observed or reported across the different stroke services.The more influential factors, termed major factors, will be prioritised for presentation in this paper.The major factors were clinical expertise, professional role, the prioritisation of functional recovery over disability management, therapists' understanding of recovery timeframes, the stroke survivor's clinical presentation, and clinical guideline recommendations.The less influential factors which were less consistently observed or reported across the different stroke services, termed minor factors, will be presented in a summarised format.The minor factors were the stroke survivor's treatment preferences, research evidence, type of organisation, and stroke pathway design.
Data that could identify individual stroke services and study participants have been anonymised.Pseudonyms have been used to refer to the stroke services and study participants.Identifying features have been modified to prevent unnecessary disclosure.

Professional expertise
The first theme, professional expertise, encompasses two major factors guiding decision making in the rehabilitation of severely disabling stroke: clinical expertise and professional role.Expanding upon the concept of clinical expertise, as identified in the model for evidence-based clinical decisions [42], professional expertise relates to the state of and interplay between a therapist's clinical practice and professional development.

Clinical expertise
Notable differences were observed between less expert and more expert therapists in the selection, skilled execution, and modification of rehabilitation interventions.More expert therapists appeared more decisive in selecting and confident in implementing different interventions.They also demonstrated a greater degree of fluency in the transition from one intervention to the next.One therapy session to demonstrate these findings involved Christine, a Band 7 PT, 1 treating Jacinta, a stroke survivor, for the first time: Robert (physio assistant) arrives at Jacinta's bay with a hoist and waits for Christine.One minute later, Christine arrives at Jacinta's bay, looks at the wall clock at the cubicle and says to Robert that because time is tight, let's do something different today with the Sara steady.Robert says okay and takes the hoist away.Christine enters Jacinta's cubicle and introduces herself to Jacinta -Jacinta looks at Christine and says hi. Robert comes back with the Sara steady and says hi to Jacinta.Christine asks Jacinta if she is up for some physio -Jacinta says yes. Christine says to Jacinta that she knows what the physiotherapists are doing with her, and she knows what her goals are -to get onto a commode and to go to the toilet.Christine suggests trying a different piece of equipment -not the hoist -to practice getting onto the commode, which will involve standing up.Christine asks Jacinta if that is okay -Jacinta says yes, and that standing and getting on the commode is important to her, but she is scared.Christine reassures her and says she and Robert will be with her all the time.Christine then asks Jacinta to bring her legs over the edge of the bed, getting Jacinta to do as much possible.Christine then raises the head of the bed and asks Jacinta to sit up.Robert is behind Jacinta and assists a little, but Jacinta pushes up by herself.Robert congratulates Jacinta as this is the first time that she has actually sat up by herself -Jacinta nods and smiles but says nothing.(Extract from fieldnotes) Grasping the whole situation -the limited remaining therapy time, Jacinta's goals, and the need to progress Jacinta towards her goals -Christine selected the most appropriate intervention at that point in time.The decisiveness in selecting and confidence in implementing a new treatment intervention without having treated Jacinta before is indicative of Christine's clinical expertise in treating stroke survivors.
In interviews, most therapists reported that severely disabled stroke survivors were a complex and challenging group of patients best managed by more expert therapists.More expert therapists were reported to possess more treatment ideas, resulting in the selection of rehabilitation interventions best suited to address stroke survivors' needs.More expert therapists were perceived to be more confident in giving advice about treatment effectiveness, and were more adept at predicting recovery: I particularly remember one of the Band 7s saying to me, 'I don't know if this patient's going to do very well'.And I really wanted the patient to get better, like, 'No, let's give it another week.'But I can see now obviously she just had that experience to know.And it is hard to pinpoint what was that she knew about this patient.And I suppose that it's the all-encompassing severe stroke, where she knew the patient had physical, communication and cognition all significantly impaired.(Susana, Band 5 OT) On the opposite end of the expertise spectrum, less expert therapists were more likely to replicate interventions performed in previous sessions, which sometimes were not as successfully executed compared to more expert therapists.Less expert therapists were slower and often demonstrated more hesitancy before modifying or changing interventions.These findings were exemplified during a therapy session led by Clive, a newly qualified Band 5 PT, retraining standing with Harry, a non-English speaking stroke survivor, for the first time: Clive blocks Harry's right knee with his knees and puts his left hand on Harry's right hip.Erin (physio assistant) is behind Harry.Clive asks Harry to lean forwards and stand up -both therapy staff guide Harry to stand up.Harry's right knee is hyperextended, and his trunk is very flexed.Clive tells Harry to "Find the middle … back tall" -Harry looks at Clive and shakes his head.Harry leans heavily to the right and bends further forwards but Clive or Erin don't say or do anything.After 30 seconds, Clive guides Harry to sit down without saying anything to Harry.Clive says, "It's warm in here" then stands up -Erin remains behind Harry.Clive walks to the paper towel dispenser, takes a paper towel, and wipes his brow.(Extract from fieldnotes) The lack of skilled execution in practising standing -the degree of physical effort experienced by Clive, the incomprehensible verbal commands given to a non-English speaking individual, the lack of correction of Harry's standing posture -may reflect Clive's lack of experience in treating Harry and in stroke survivors generally.Having observed many therapy sessions involving Harry, Clive appeared to be replicating the same interventions delivered in previous physiotherapy sessions without necessarily reflecting upon his skills to perform the intervention.
In interviews, less expert therapists were reported to be less confident about treating patients with profound impairments, such as global aphasia or severe spasticity.Less expert therapists themselves reported high levels of uncertainty in the effectiveness of treatment interventions, determining future prognosis, and knowing when to stop therapy completely.In these situations, it was suggested that more support was required from more expert therapists to guide decision making: I think my lack of experience is always going to be a limiting factor to treat patients, because I always think that if someone's more experienced then they're going to have further ideas, further knowledge about what they could be doing or, you know, 'I wouldn't be handling it this way, I could be doing it other ways'.As well, sometimes when you're not seeing that improvement and you're thinking, 'Oh, is this something that I'm doing or is this just the way that the patient is, or should I be trying something different to optimise their rehab?' and so on.(Sabrina, Band 5 PT)

Professional role
Professional role differences between PTs and OTs resulted in the selection of different rehabilitation interventions.In the observed therapy sessions, PTs predominantly performed transfer practice (e.g., moving from a bed to a chair), balance tasks in sitting and standing, and mobility tasks (e.g., stepping and walking practice).OTs predominantly performed self-care tasks (e.g., washing and dressing), upper limb splinting, and task-specific object use (e.g., using cutlery during feeding): focussed on manual handling methods during transfers, whereas OTs focussed on monitoring safety during self-care tasks.
The practice of multi-disciplinary working in stroke rehabilitation meant that many observed sessions were delivered by both PTs and OTs.In most instances, mutual negotiation before and sometimes during the joint therapy session about the aims of the session occurred, which resulted in the selection of interventions designed to achieve the session's aim.A discussion prior to a therapy session involving Alicia, a Band 6 PT, and Erica, a Band 5 OT, illustrated these findings: Alicia and Erica are standing at the patient notes trolley.Erica says she hasn't seen Robert [stroke survivor] since before the weekend.Alicia says she saw Robert yesterday, they did some standing practice with him, which wasn't very successful because he was pushing a lot.Alicia gets Robert's notes from the trolley and both therapists start reading through the notes.Alicia and Erica discuss plans for the sessionstanding might be too difficult therefore sitting on the edge of the bed, grooming practice might be better.Both therapists agree to get Robert sitting in his chair, take him to the gym and practice grooming tasks in sitting -Erica mentions this is one of Robert's goals that she is working on.(Extract from fieldnotes) The session lead was usually the most expert therapist, although individual interventions were usually led by the professional traditionally responsible or more experienced in delivering that intervention.Some therapists reported in the interviews that joint working between OTs and PTs provided an opportunity to support and learn from each other: So, I suppose some of it is what you're confident with doing.So, I wouldn't necessarily go down an FES [functional electrical stimulation] route on my own.I'd, obviously, I would ask the physios to come in and support me with that.But it might be that I would look at those kinds of tasks.(Nina, Band 5 OT)

Therapist perspectives
The second theme comprises therapist perspectives on the focus of therapy and the timeframes for recovery after severely disabling stroke.These perspectives were major factors guiding decision making and noted across all stroke services.Therapist perspectives were based upon the normative expectation of functional recovery after severely disabling stroke and shaped by a general preference to treat less disabled stroke survivors.

Prioritisation of functional recovery over disability management
Observation and interviews revealed that the primary focus of therapy was to restore as much function and independence as possible within the constraints of severe brain damage.Function was understood by most therapists as the ability to perform specific activities, such as transfers and grooming.Functional recovery, therefore, referred to a recovery in the ability to perform functional tasks.In a functional recovery approach, interventions were delivered in a sequential and progressive manner, commencing with practising more basic functional tasks (e.g., sitting independently, and grooming in bed) and then progressing to more complex functional tasks (e.g., standing, stepping, washing in a bathroom): I certainly think, I mean at the end of the day, it's similar in that you're dealing with patients who have had a stroke and a lot of it is improving function and trying to get them back to their baseline as close as possible.But I think with the severe strokes, you're a lot more mindful that they might not get back as quickly or back to the same level.(Jane, Band 6 OT) Therapists initially adopted a restorative approach when improving function, which focussed on the restoration of impaired body functions (e.g., muscle power) during task performance.If restoration of impaired functions was slow or limited, therapists shifted to a compensatory approach.This approach focussed on the performance of a task as independently as possible, with less concern about the type of movement pattern executed during the task.The transition from a restorative to a compensatory approach varied considerably between therapists and stroke services, and was influenced by therapist preference and workplace philosophy as much as the rate of restoration of impaired body functions: But when you make that total change towards being more adaptive is probably, I don't know if I could put a timeline on it, but when the comorbidities are feeding into it, when you've tried the restorative things again and again and again and again and you're still not getting activity in the hand or, activity in the leg.Then I would work toward, more towards the adaptive things at that point.(Kylie, Band 7 PT) In situations where stroke survivors on SUs made limited or no functional recovery, therapists shifted to disability management, an approach designed to maintain the stroke survivor's current level of functioning and prevent the development of post-stroke complications.Some interventions designed to manage disability, such as positioning and resting splints, were performed alongside therapy designed to improve functional recovery.Although therapists highlighted the importance of disability management in preventing complications, disability management was not perceived as "regular" rehabilitation, which focussed on functional recovery and restoration.Disability management was also viewed as less rewarding than "regular" rehabilitation: I know I said that physios want to fix people and I think we do hold restoration of function in higher regard than we do management of disability.And no one I've ever met aspires to want to go and work in a disability unit, to passively range people and tilt table, like, GCS [Glasgow Coma Scale] 3 patients.And I think there is an element of job satisfaction you want -you want to work with patients because you like building that relationship with patients and therefore all of a sudden you don't have that because someone's GCS is very low.You don't get any reward.(John, Band 7 PT) The preference for therapy to optimise functional recovery rather than manage disability, as highlighted by John, reflected a wider preference amongst therapists to treat motivated stroke survivors able to engage in therapy and demonstrating faster functional recovery.These stroke survivors were usually those less disabled by stroke.Many therapists recognised that, although personal preferences should not influence clinical practice, a preference to treat those less disabled by stroke may influence the provision of therapy to those severely disabled by stroke: I think some people probably don't like treating the more severe stroke.So, you kind of go into one session and you see, or you compare to a patient that you saw two months ago that was the same who didn't progress.So, you, kind of, make that expectation.(Monica, Band 7 OT)

Therapists' understanding of post-stroke recovery timeframes
In interviews, therapists expressed different understandings about the timeframe for recovery after a severely disabling stroke, which influenced the selection and timing of interventions.In line with the normative expectation of post-stroke recovery, many therapists reported that most recovery occurred in the first three months after a severely disabling stroke: So, I think initially in their first three months, we're expecting the most gains.(Karen, Band 6 PT) Um, but really, if anybody's going to improve, it's the three months that is the most important part.(Samantha, Band 6 OT) Accordingly, these therapists felt that most therapy should be provided in the acute inpatient phase, which usually comprised once daily therapy sessions focussing on functional recovery.Therapists also felt that there was limited benefit in referring stroke survivors not demonstrating early functional recovery to community services or specialist inpatient rehabilitation facilities to monitor for further signs of improvement: But objectively, if we set goals and they make no changes, we then need to be realistic.We need to advocate for a patient not to go to a rehab unit where they will not make changes.It's not fair on those patients who are making gains and would benefit from more rehab in a unit.(Megan, Band 7 OT) Similarly, community therapists were less likely to reassess functional tasks or perform task-specific practice for stroke survivors who had not improved functionally in the acute inpatient phase, as it was felt that further recovery was limited or not possible.In these instances, interventions focussed more on disability management: What we don't do, if someone has come out of hospital, and they haven't managed to get sitting balance, is that we don't go into doing a complete sitting balance assessment.So, with the lady we saw, she had a stroke, I think it was four months, and she never managed to get sitting balance and she never progressed beyond hoist.So, the first time we saw her, we did a full assessment and then got her positioned out into the chair.There was no point focussing on sitting balance or exploring standing with the standing hoist.(Leonie, Band 7 PT) A smaller number of therapists believed that functional recovery was possible in the chronic phase post-stroke.These therapists were usually more expert and had direct experience of treating stroke survivors in the chronic phase post-stroke.Therapists recalled instances where stroke survivors demonstrated improvements in function months and sometimes years post-stroke, despite making variable gains in the acute and subacute phase: The rehab centre where I worked in was easily the best rehab centre I've worked in … And we used to see really severe strokes, and we used to see them quite far down the line, where they had lots of secondary complications because they could wait up to a year to come to us.And we used to get a lot of them up walking in the end.But if you don't know, if you haven't maybe seen beyond the HASU, beyond the stroke unit, beyond the outpatients, if you haven't seen their whole spectrum, you may not know how far to push someone or, maybe, when less is more at that certain time.(Darren, Band 6 PT) Therapists who believed in later recovery highlighted the limitations of early decision making regarding future prognosis and advocated for stroke survivors to have access to community therapy or therapy "further down the line."The focus of this latter stage therapy was to monitor for signs of engagement, optimise functional recovery, as well as enable societal participation: But, you know, these patients still may have the potential to improve in the community six months down the line.I think we pin too much decision making sometimes on this acute stage.And I think the decision making for the longer-term needs should be made six, nine months down the line.Not at six weeks post stroke, in an acute hospital, when they've been medically unwell, they've gone through a massive life event where family are stressed, where the transition into the community is very stressful from the hospital.Actually, when things have settled down six months down the line, it would be interesting to see how these patients actually do.(Natalie, Band 8 OT)

Attributes of the severely disabled stroke survivor
The third theme, attributes of the severely disabled stroke survivor, incorporates two patient-related factors guiding decision making in the rehabilitation of severely disabling stroke: the stroke survivor's clinical presentation and their ability to express their treatment preferences and wishes.Whilst a stroke survivor's clinical presentation was a major factor guiding decision making, ascertaining a stroke survivor's treatment preferences was sometimes limited due to the increased likelihood of communication and cognitive impairments.

Stroke survivor's clinical presentation
Noted during the therapy sessions and reported during the interviews, the stroke survivor's clinical presentation was determined following an initial therapy assessment, which focussed on determining the stroke survivor's impairments and activity limitations.Therapists assessed these aspects of function in a logical and sequential process, which guided the formulation of an individualised treatment plan at a level appropriate for the stroke survivor.
So, I would do a full physiotherapy, neurological assessment.And then I would make a problem list.I would make an impairment activity participation, so an ICF type problem list, to identify and separate the impairments, the activities that are impaired, and the participating patient elements.And then I would make a treatment programme to address those things.(Sandy, Band 7 PT) The stroke survivor's ability to actively engage and participate in therapy sessions determined the type of interventions delivered.More passive interventions, such as bed positioning, splinting, and sitting in a recliner chair, were delivered if a stroke survivor was drowsy or not following commands.More active interventions, such as task-specific practice, were delivered if a stroke survivor was alert and able to follow some commands.Within each intervention, therapists made smaller decisions based upon the stroke survivor's immediate response to the intervention.Smaller decisions included aspects such as the number of repetitions and the length of time to perform an intervention.Continual reflection upon these interventions led the therapist to modify or change the intervention.Fatigue, determined by therapists as a stroke survivor becoming visibly tired or demonstrating reduced movement quality, was reported to be a sign to change the intervention to something easier or to stop the session completely.These findings were noted during a therapy session staffed by Kristina, a Band 6 OT, and Sabrina, a Band 5 PT, treating Simon, a stroke survivor: Sabrina is perched behind Simon and supporting his trunk.Kristina is sitting in front of Simon and holds up a washcloth.Kristina asks Simon to reach for the washcloth -Simon's eyes close for a few seconds and then open.Simon slowly reaches towards the washcloth and grasps it -Kristina says, "Brilliant".Kristina says she will practise wiping Simon's forehead -Kristina guides Simon's hand to his forehead to facilitate the movement and does 6 wipes.Kristina asks Simon for the washcloth -Simon closes his eyes and his head lowers.Sabrina says, "Looks like he's falling asleep" -both therapists agree.Kristina asks Simon to look at his fingers and let go of the washcloth -no response from Simon.About 10 seconds later, Kristina suggests to Sabrina that they reposition Simon in bed and apply the boxing glove splint -Sabrina agrees.(Extract from fieldnotes) The stroke survivor's ability to actively engage and participate in therapy also determined the frequency of therapy sessions and how much therapy was provided overall.During observation sessions and reported during interviews, goal attainment and active participation in SU therapy sessions led to continued and often daily therapy provision.Stroke survivors who demonstrated variable or less engagement in SU therapy sessions over time experienced a reduction in the frequency of therapy sessions: So, if somebody who's engaging with the sessions or hugely motivated, or they've got great goals, then we try and see them as much we canobviously as the caseload allows.So, we're trying for at once a day and maybe a session with the assistant for some upper limb exercises.(Lucy, Band 5 PT) If they're unable to engage or distressed by the intervention, then we wouldn't continue to pursue -unless there is something that is in their best interests that we have to do.But, if they're not going to actively engage in rehab on a consistent basis, then I would probably withdraw.(Jasmine, Band 6 OT) Reported in interviews, therapists expected stroke survivors who actively engaged in SU therapy would continue to engage in the community.Conversely, stroke survivors not actively engaging in SU therapy were deemed ineligible for community therapy.However, several community therapists noted that the clinical presentation of some stroke survivors did not follow these expectations: Yeah, I've certainly seen it go both ways, to be honest.I've seen people massively plateau or not do well in hospital just because the environment was so unfunctional to them, and they were just very low in mood and desperate to get home.But when they've gone home, it's been their own home and much more meaningful.They actually want to get out of bed, and I've seen people then massively improve when they were deemed not to be able to or have the potential to.But then I've also seen the opposite where people were doing really well in the ward, progressing daily, or progressing weekly.And then they've gone home and that has massively affected their mood, because they've actually realised their deficits much more than when they were in hospital.And these people almost go backwards.(Sandy, Band 7 PT)

Stroke survivor's treatment preferences
It was recognised by therapists that involving severely disabled stroke survivors in decision making was challenging due to communication and cognitive impairments.Most therapists reported that stroke survivors with aphasia limited the therapist's ability to inquire about stroke survivors' preferences and for stroke survivors to express their preferences.Some therapists highlighted that stroke survivors with anosognosia may not understand what therapy was required to address their problems.In these situations, the therapist almost exclusively made intervention selection.
And especially with the severe stroke patients … the fact that a lot of them can't communicate or they might not have the cognitive ability to make a goal or have the insight into their deficits and so on.So, I guess decisions can be led more by therapists.(Kane, Band 5 PT) Therapists highlighted strategies to modify their communication for stroke survivors presenting with communication and cognitive impairments in order to understand their preferences or involve them in decision making, as well as utilising the expertise of speech and language therapists.However, alternative sources of information to understand the stroke survivor's preferences, such as family members, were inconsistently used in inpatient settings and more frequently used in community settings.Family members were more likely to provide information about the stroke survivor's premorbid function rather than participate in intervention selection.
I think it's quite easy as a therapist when you're in somebody's home and they're telling you what they want for the patient, it can be quite easy to be swayed by what the family want.But I think it's just about trying to keep the patient at the centre and thinking, actually, what does patient want?And it doesn't really matter what the family want, it's about the patient and it's trying to work cohesively with both the patient and the family to make sure that it's the best outcome for the patient based on my assessment and my clinical reasoning as a therapist.(Rita, Band 6 OT)

Documentary evidence
The fourth theme, documentary evidence, relates to the influence of clinical guideline recommendations and research evidence on decision making in the rehabilitation of severely disabling stroke.Across all stroke services, clinical guideline recommendations were more influential in guiding therapy practice than research evidence.

Clinical guideline recommendations
Every inpatient therapist and most community therapists referred to guidance from the UK's National Clinical Guideline for Stroke recommending the daily provision of at least 45 min of each appropriate therapy [3].This recommendation guided the timing, length, and frequency of therapy sessions, which shaped the nature and content of therapy sessions.
So, we're very much indoctrinated into the daily 45-minute SSNAP target which is the fundamental reason why we're here and why we're so well staffed as a therapy service.The way it affects us is we have targets.So, all patients should receive an assessment from a therapist, whether it's speech and language or physio within 24 hours and all therapists, all specialities that should be seeing a patient should have assessed them within 72 hours.(Martin, Band 7 PT) In order to adhere to the clinical guideline recommendation of at least 45 min of daily therapy, SU therapists used a timetable system to plan their day and allocate therapy sessions to stroke survivors.
I think our team is very good in that we do our timetable every day … it makes sure that everyone gets seen regularly.So, aiming for, if we can, six times a week.It also means that the national guidelines tend to be met.(Beth, Band 6 PT) Most therapy sessions were listed as 45 or 60 min on timetables.The longer, 60-min session length incorporated time to prepare the stroke survivor for the session, such as hoisting them into a wheelchair, taking a stroke survivor to the gym, and completing documentation of the session after its completion.In some instances, therapists had 15-or 30-min gaps in between longer therapy sessions, which resulted in therapists selecting interventions requiring less time to deliver, such as re-positioning stroke survivors or applying hand splints.Therefore, timetabling guided the type of interventions therapists were able to deliver in the allocated session length.
In the interviews, most therapists felt that a singular, 45-min therapy session was not always appropriate for survivors of severely disabling stroke.SU therapists suggested shorter and more regular sessions may be preferable due to the stroke survivor's reduced ability to physically and cognitively engage for an extended period: But realistically, with a lot of these patients, they don't have the attention span for it.And a lot of them suffer from fatigue, which means that they can't tolerate those kinds of sessions.So, for patients, I often feel that short bits throughout the day would be more appropriate.(Claire, Band 6 OT) However, therapists highlighted that a more regular frequency was not always feasible due to difficulties in coordinating multiple therapy staff and the availability of specialist seating.Community therapists reported that fatigue was less prevalent once stroke survivors were residing in the community.Community therapists suggested that session length should be determined by the session's purpose and the time required to perform interventions: Once you see someone in their own home, that guides you with what you need to do.You can see how their environment influences your session.So, if they need to be able to do stairs and it take 10, 15 minutes, then that's what it takes.If they need to get to their bathroom and it takes just as long, then so be it.So, I'm guided by what they need to achieve and however long it takes to achieve it.For some clients, it can be 60 minutes or an hour and a half.But then, it might be once a week that is required.(Kelis Band 6 OT)

Research evidence
Research evidence was a rarely reported factor contributing to decision making in the selection of different rehabilitation interventions.In interviews, two PTs suggested that there was limited evidence in severely disabling stroke, although there was uncertainty in their responses and no examples of research evidence were presented.Don't quote me on it, but I think, evidence-wise, maintaining range by passive stretching doesn't have a huge evidence base.So, yeah, I think there is quite limited evidence in severe stroke, from my understanding anyway, as to what is the most effective option.(Gareth, Band 6 PT) I think the evidence out there for severe strokes isn't quite there, which I guess is why you're conducting this study.But I'm not really sure.(Hugo, Band 5 PT) One HASU PT specifically referred to findings from a randomised controlled trial investigating early mobilisation post-stroke and its impact upon clinical practice after severely disabling stroke.In these instances, therapists interpreted research evidence as studies evaluating the effectiveness of different rehabilitation interventions.
During therapy sessions, therapists did not explicitly discuss research findings before or during therapy sessions amongst themselves, with stroke survivors, or with their carers.Therefore, the extent to which research guided therapy practice was determined by noting the types of interventions delivered during therapy sessions and referring to the research evidence base.Many interventions delivered during observed therapy sessions had no strong research evidence supporting their use in the rehabilitation of severely disabling stroke.Examples of these interventions include positioning, passive stretches, and sensory stimulation.However, some interventions supported by research evidence, such as task-specific practice and electrical stimulation, were noted during therapy sessions.

Therapy within the stroke pathway
The fifth theme, therapy within the stroke pathway, refers to the influence of organisational factors and pathway design on therapist decision making in the rehabilitation of severely disabling stroke.Individual rehabilitation interventions were selected by therapists as part of a wider package of therapy to address specific clinical problems.As therapists worked closely with one another in teams across a range of interconnected stroke services, the context in which therapy was delivered shaped therapy practice.

Type of organisation
The type of stroke service in which a therapist worked guided the selection of rehabilitation interventions.For example, the role of a HASU is to assess, medically manage, and triage stroke survivors into the different rehabilitation services.Consequently, following assessment, HASU therapists commenced with functional tasks (e.g., sitting balance practice, task-specific upper limb practice) and early mobilisation (e.g., seating, standing), before patients were transferred to a SU within 72 h post-stroke and once medically stable.
Demand and capacity of the stroke service, related to the number and dependency of stroke survivors and available staffing resources, also guided therapist decision making.Short-term SU staff shortages resulted in the prioritisation of stroke survivors newly admitted, making faster improvements in functional recovery, or those that could be seen with one therapist.Consequently, survivors of severely disabling stroke were frequently prioritised out when staffing levels were reduced.Prolonged SU staff shortages, due to extended annual or sick leave period, often resulted in a reduced therapy frequency for survivors of severely disabling stroke compared to less disabled stroke survivors.
I think when you're down on staff, and you've got so many patients to see, you have to focus on the priorities -getting people better and then discharging them out of hospital.The severe ones who are slower and less engaged or not making the quick wins, they're the ones who you prioritise out.It's a hard decision but that's the way it sometimes is.(Claire, Band 6 OT)

Stroke pathway design
In London, stroke survivors traverse through several types of stroke services in line with the London stroke model.However, local pathway variation was seen across London, resulting in differential access to stroke services dependent upon a patient's place of residence.This guided the therapeutic approach delivered by therapists working in various parts of London.For example, most stroke survivors from Site B were discharged to the community stroke service whereas most stroke survivors from Site C were referred for further inpatient rehabilitation.Accordingly, Site B therapists performed more discharge planning tasks and shifted to a compensatory approach much earlier than Site C therapists.
The time limited nature of community stroke services, determined by local commissioning arrangements, meant that stroke survivors were discharged from the service once their allotted time with the service had finished or, less commonly, once all goals were achieved.Community therapists reported a paucity of longer term rehabilitation services for survivors of severely disabling stroke.The consequence of limited ongoing rehabilitation services, which usually had long waiting lists and provided substantially less therapy, meant that some interventions were handed over to the stroke survivor's carers to continue, whereas some interventions were discontinued.Accordingly, many community therapists expressed frustration at not being able to provide ongoing therapy when clinically required as well as frustration about the lack of ongoing therapy for severely disabled stroke survivors.
I suppose the difficultly with the severe strokes are the ones that are changing when we come to the end of our length of stay.They're really tricky to discharge because morally you feel that they need more input.But realistically the resources aren't there to give them more input.And sometimes, I go back to management, because I really struggle with discharging someone that is continuing to make change.(Adrianna, Band 7 PT)

Study findings
This ethnographic exploration of therapy practice in five London stroke services has demonstrated that multiple factors guide decision making in the rehabilitation of physical function after severely disabling stroke.More emphasis was placed on the therapist's clinical expertise and the stroke survivor's clinical presentation, whereas less emphasis was placed upon research evidence and the stroke survivor's treatment preferences.This may suggest that the current model of EBP, which incorporates research evidence, clinical expertise, and patient preferences to guide decision making [41,42], does not fully account for all the factors that guide clinical decision making after severely disabling stroke.These additional factors include therapist perspectives about stroke recovery and professional role.The interdependent nature of specific interventions within a larger package of therapy meant that therapists often described their practice in terms of treatment approaches -restoration versus compensation, functional recovery versus disability management -rather than specific interventions.As this larger package of therapy was provided by therapists working in teams across a range of different stroke services, organisational function and the stroke pathway design also guided therapist decision making.
Clinical expertise was one of the most influential factors guiding therapist decision.Previous research has identified differences in the practice of novice and expert clinicians [67,68].For example, expert clinical practice is characterised by faster skill performance, superior short-and long-term memory, possession of a wider and context-specific knowledge based, and advanced reflective practice skills [68].In the current study, the practice of more expert therapists not only resulted in the faster, more confident, and more skilled execution of rehabilitation interventions, but also guided the clinical practice of less expert therapists.This reliance on more expert therapists may stem from the uncertainty of clinical practice in the rehabilitation of severely disabling stroke and the perception that more expert therapists know what to do in these situations.This reliance may also stem from the paucity of research evidence supporting the use of particular interventions in the rehabilitation of severely disabling stroke.In addition, therapists frequently selected interventions according to assessment findings and modified based upon the stroke survivor's response to the intervention.This "reflection-in-action" [69], or reflection upon clinical data derived from assessment tests and observations, may provide more real and immediate evidence about an intervention's effectiveness than data from a research trial conducted in a different healthcare setting.Higgs and Turpin support this view [70], who posited that reliance on research evidence as the sole form of evidence in treatment prescription fails to acknowledge the complexity of individuals' responses to illness and their proposed treatments.They proposed that reflection in action is a necessary requirement for healthcare professionals to practice in an evidence-based manner.
Another limitation of incorporating research evidence into clinical decision making is its relevance to clinical practice.Previous research has reported that, although therapists recognise the importance of EBP within clinical practice, the limited applicability of some research evidence to clinical practice may limit its utilisation with clinical decision making [71,72].Consequently, more research evaluating the effectiveness of research interventions in severely disabling stroke may not automatically lead to its inclusion in clinical decision making.Therapists have also reported other barriers to incorporating research evidence into clinical decision making, such as perceived lack of time and research skills to review the available literature [71,72].As such, in addition to undertaking clinically meaningful and relevant research, there is a need to address these barriers to EBP to increase utilisation of research evidence into clinical decision making.
Therapist perspectives about recovery after severely disabling stroke was a novel factor guiding clinical decision making.Findings from the study suggested that a therapist's reported beliefs were at times inconsistent with the literature.For example, many therapists believed that most recovery after severely disabling stroke occurred within the first three months, resulting in a greater concentration of therapy to optimise functional recovery in the acute and sub-acute phases post-stroke.However, several observational studies report a much slower and longer recovery period after severely disabling stroke [23,[34][35][36]73,74].These findings may support the provision of longer-term therapy in the chronic phases post-stroke, even if early functional recovery is limited.However, in the current study, severely disabled stroke survivors were less likely to access longer-term therapy if they did not demonstrate early functional recovery.As well, post-stroke recovery occurs through a combination of spontaneous and learningdependent processes involving restitution (repair of the lesioned area with its function assumed by adjacent cortical and subcortical structures) and substitution (behavioural compensation by the stroke survivor) [75][76][77].Cortical reorganisation is less likely to occur in larger strokes due to less adjacent cortical structures assuming the function of the damaged area.Therefore, compensatory strategies are likely to play a greater role in functional recovery after severely disabling stroke than after less disabling stroke.However, compensatory strategies were less commonly observed to facilitate functional recovery.Additionally, the wider preference amongst therapists to treat stroke survivors able and motivated to engage in therapy demonstrating faster function recovery as well as less favourable attitudes towards severe disability may have indirectly influenced the relative de-prioritisation of severely disabled stroke survivors during periods of staff shortage compared to those less disabled by stroke.As a result of these therapist perspectives, therapy did not consistently address the physical needs of severely disabled stroke survivors.Consequently, therapists need to be more critical of their beliefs and attitudes, as well as any potential bias that these beliefs and attitudes may cause, on therapist decision making.Greater consideration needs to be given to the timing of therapy as well as the treatment approach used by therapists to optimise outcomes after severely disabling stroke.

Strengths and limitations
There are several strengths of this study.This was an extensive ethnographic exploration of therapist practice in severely disabling stroke conducted over a prolonged period.The primary researcher's in-depth knowledge of the London stroke model resulted in the selection of stroke services that accurately reflected the transition of severely disabled stroke survivors across the stroke pathway.The primary researcher's clinical background enabled a quick immersion in each stroke service and an awareness of the more hidden and unspoken aspects of therapy practice, such as clinical reasoning and inter-professional relationships.There was adherence to key aspects of the ethnographic tradition, such as prolonged fieldwork, the achievement of data saturation, and triangulation of fieldwork findings by combining data obtained from semi-structured interviews and participant observation [56,57].There are also several limitations of this study.The context specific nature of ethnographic research may have generated different findings if conducted by different researchers or in different settings.Therefore, study findings may not be transferable to other health care settings.As well, the primary researcher conducted part of the study in his clinical workplace.This may have resulted in bias in the analysis and interpretation of findings.However, the use of a critically reflexive approach and regular debriefing with the research team throughout the study may have mitigated this bias.

Conclusions
An ethnographic exploration of therapy practice in five London stroke services revealed several factors guiding therapist decision making in the rehabilitation of physical function after severely disabling stroke.Whilst therapists aimed to tailor therapy to stroke survivors' clinical presentation, some factors guiding decision making resulted in the use of interventions that did not fully address the physical needs of severely disabled stroke survivors.This mismatch between therapy practice and clinical need may contribute to the poor outcomes experienced by this cohort of the stroke population.Alternative ways of therapist working and models of care should be considered to address the needs of severely disabled stroke survivors more fully.

Notes
1. Qualified NHS allied health professionals are graded from Band 5 (newly qualified or junior therapist) to Band 8 (clinical specialist or clinical lead therapist).

Figure 2 .
Figure 2. Thematic overview and factors guiding therapist decision making (bold indicates major factor).