Occupational therapy for care home residents with stroke

Joanna C Fletcher‐Smith; Marion F Walker; Christine S Cobley; Esther MJ Steultjens; Catherine M Sackley

doi:10.1002/14651858.CD010116.pub2

. 2013 Jun 5;2013(6):CD010116. doi: 10.1002/14651858.CD010116.pub2

Occupational therapy for care home residents with stroke

Joanna C Fletcher‐Smith ^1,^✉, Marion F Walker ¹, Christine S Cobley ¹, Esther MJ Steultjens ², Catherine M Sackley ³

PMCID: PMC6464854 PMID: 23740541

Abstract

Background

Stroke is a worldwide problem and is a leading cause of adult disability, resulting in dependency in activities of daily living (ADL) for around half of stroke survivors. It is estimated that up to 25% of all care home residents in the USA and in the UK have had a stroke. Stroke survivors who reside in care homes are likely to be more physically and cognitively impaired and therefore more dependent than those able to remain in their own home. Overall, 75% of care home residents are classified as severely disabled, and those with stroke are likely to have high levels of immobility, incontinence and confusion, as well as additional co‐morbidities. It is not known whether this clinically complex population could benefit from occupational therapy in the same way as community‐dwelling stroke survivors. The care home population with stroke differs from the general stroke population living at home, and a review was needed to examine the benefits of occupational therapy provided to this specific group. This review therefore focused on occupational therapy interventions for ADL for stroke survivors residing in care homes.

Objectives

To measure the effects of occupational therapy interventions (provided directly by an occupational therapist or under the supervision of an occupational therapist) targeted at improving, restoring and maintaining independence in ADL among stroke survivors residing in long‐term institutional care, termed collectively as 'care homes'. As a secondary objective, we aimed to evaluate occupational therapy interventions for reducing complications such as depression and low mood.

Search methods

We searched the Cochrane Stroke Group Trials Register (August 2012), the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, September 2012), MEDLINE (1948 to September 2012), EMBASE (1980 to September 2012), CINAHL (1982 to September 2012) and 10 additional bibliographic databases and six trials registers. We also handsearched seven journals, checked reference lists and obtained further information from individual trialists.

Selection criteria

Randomised controlled trials investigating the impact of an occupational therapy intervention for care home residents with stroke versus standard care.

Data collection and analysis

The lead review author performed all searches. Two review authors then independently assessed all titles and abstracts of studies and selected trials for inclusion, with a third review author resolving any discrepancies. The same two review authors independently extracted data from all included published sources to ensure reliability. Primary outcomes were performance in ADL at the end of scheduled follow‐up and death or a poor outcome. Secondary outcomes aimed to reflect the domains targeted by an occupational therapy intervention.

Main results

We included in the review one study involving 118 participants. We found one ongoing study that also met the inclusion criteria for the review, but the data were not yet available.

Authors' conclusions

We found insufficient evidence to support or refute the efficacy of occupational therapy interventions for improving, restoring or maintaining independence in ADL for stroke survivors residing in care homes. The effectiveness of occupational therapy for the population of stroke survivors residing in care homes remains unclear, and further research in this area is warranted.

Keywords: Adult, Humans, Nursing Homes, Stroke Rehabilitation, Activities of Daily Living, Disabled Persons, Disabled Persons/rehabilitation, Occupational Therapy, Occupational Therapy/methods, Randomized Controlled Trials as Topic

Occupational therapy for care home residents with stroke

Stroke is the leading cause of disability worldwide and is prevalent in the care home population. Whilst care home residents with stroke are likely to be more impaired and dependent than those with stroke residing in their own homes, they are less likely to receive ongoing stroke specialist rehabilitation such as occupational therapy. Occupational therapy aims to help people achieve their maximum level of independence in everyday activities. Evidence can be found to support the benefits of occupational therapy for community‐dwelling stroke survivors. However, the care home population with stroke differs from the community‐dwelling population. For example, they are more likely to have high levels of immobility, incontinence and confusion, along with other co‐morbidities. This review of one trial including 118 participants found that evidence is currently insufficient to conclusively state the benefits of occupational therapy for care home residents with stroke. Additional randomised controlled trials that test occupational therapy interventions for care home residents with stroke are needed. One such trial is currently ongoing.

Summary of findings

Summary of findings for the main comparison.

Occupational therapy compared with standard care for care home residents with stroke
Patient or population: care home residents who have had a stroke Settings: care homes (nursing and residential homes) Intervention: occupational therapy Comparison: standard care
Outcomes	Illustrative comparative risks (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Standard care	Occupational therapy
Function in ADL at the end of scheduled follow‐up (Barthel ADL Index score) Barthel ADL Index score 6‐month follow‐up Analysis 1.1	The mean Barthel across the control group was 8	The mean Barthel in the intervention group was 2 points higher (mean Barthel of 10) (95% CI ‐0.11 to 0.90)		118 (1 study)	⊕⊕⊝⊝ low	Clustering design effect was accounted for
Global poor outcome Death or a drop in Barthel ADL score 6‐month follow‐up Analysis 1.2	Medium‐risk population		OR 0.34 (0.11 to 1.01)	118 (1 study)	⊕⊕⊝⊝ low	Clustering design effect was accounted for
	759 per 1000	516 per 1000 (258 to 759)	OR 0.34 (0.11 to 1.01)	118 (1 study)	⊕⊕⊝⊝ low	Clustering design effect was accounted for
Function in ADL at the end of intervention (Barthel ADL Index score) Barthel ADL Index score 3‐month follow‐up Analysis 1.3	The mean Barthel across the control group was 8	The mean Barthel score in the intervention group was 3 points higher (mean Barthel of 11) (95% CI ‐0.03 to 0.99)		118 (1 study)	⊕⊕⊝⊝ low	Clustering design effect was accounted for
Death Number of deaths from any cause at 6‐month follow‐up Analysis 1.4	Medium‐risk population		OR (0.09 to 0.98)	118 (1 study)	⊕⊕⊝⊝ low	Clustering design effect was accounted for
	242 per 1000	485 per 1000 (28 to 239)	OR (0.09 to 0.98)	118 (1 study)	⊕⊕⊝⊝ low	Clustering design effect was accounted for
Quality of life	See comment	See comment	Not estimable	0	See comment	Included study did not measure 'quality of life' as an outcome
Mobility Rivermead Mobility Index (RMI) score 6‐month follow‐up Analysis 1.5	The mean RMI score across the control group was 4.5	The mean RMI score in the intervention group was 0.5 higher (mean RMI of 5) (95% CI ‐0.36 to 0.64)		118 (1 study)	⊕⊕⊝⊝ low	Clustering design effect was accounted for
Mood	See comment	See comment	Not estimable	0	See comment	Included study did not measure 'mood' as an outcome
Global cognition	See comment	See comment	Not estimable	0	See comment	Included study did not measure 'global cognition' as an outcome
Adverse events	See comment	See comment	Not estimable	0	See comment	Adverse events were not reported in the included study
Satisfaction with care	See comment	See comment	Not estimable	0	See comment	Included study did not measure 'satisfaction with care' as an outcome
Health economic outcomes	See comment	See comment	Not estimable	0	See comment	Included study did not measure 'health economic outcomes' as an outcome
CI: confidence interval; OR: odds ratio
GRADE Working Group grades of evidence: High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Open in a new tab

Analysis 1.1 — Comparison 1 Occupational therapy versus standard care, Outcome 1 Function in ADL at the end of scheduled follow‐up (Barthel ADL Index score).

Analysis 1.2 — Comparison 1 Occupational therapy versus standard care, Outcome 2 Global poor outcome (death or a drop in ADL score) at the end of scheduled follow‐up (6 months).

Analysis 1.3 — Comparison 1 Occupational therapy versus standard care, Outcome 3 Function in ADL at the end of intervention (Barthel ADL Index score).

Analysis 1.4 — Comparison 1 Occupational therapy versus standard care, Outcome 4 Death at the end of scheduled follow‐up.

Analysis 1.5 — Comparison 1 Occupational therapy versus standard care, Outcome 5 Mobility (Rivermead Mobility Index score) at the end of scheduled follow‐up.

Background

This review aimed to evaluate occupational therapy interventions directed at reducing dependency in activities of daily living (ADL) for people with stroke residing in care homes.

Description of the condition

Stroke is a worldwide problem. In the UK, stroke is the third main cause of death (National Audit Office 2010) and the largest cause of adult disability (National Audit Office 2005). Every year in England alone, approximately 110,000 people have a stroke (National Audit Office 2010). In the USA, around 795,000 people experience a new or recurrent stroke each year, ranking stroke as the fourth biggest cause of death (American Heart Association 2011). In Australia, stroke is the second biggest killer after heart disease (Australian Bureau of statistics 2008), with an estimated 60,000 new and recurrent strokes each year (National Stroke Foundation 2011a). Worldwide, approximately 15.3 million strokes occur annually, with stroke accounting for around 10% of all deaths (Johnston 2009). Globally, an estimated 30 million people are living with stroke, most of whom have residual disabilities (World Stroke Organization 2011). This makes stroke a leading cause of adult disability worldwide (World Stroke Organization 2011) and a major contributor to the global burden of disease (Warlow 2008).

The main encumbrance of stroke is survival with disability, dementia, depression, epilepsy, falls and other such complications (Rothwell 2005). Around 80% of stroke survivors have motor impairment (Langhorne 2009). This may result in residual mobility problems (Jorgensen 1995) and loss of capability in ADL that restricts survivors' ability to resume their previous lifestyle. Around half of those who survive a stroke are left dependent on others for help in everyday activities (National Audit Office 2010), and this may persist for the rest of their lives. Three‐quarters of those affected by stroke are older than 65 years of age (National Audit Office 2010). For some stroke survivors, it is possible to return home from hospital with informal support from family or with organised care provided by health and social services. However, in the UK, approximately 10% to 11% of stroke survivors are admitted directly from acute care into a care home setting after stroke (National Audit Office 2010; RCP 2011). The proportion of people with stroke discharged to a care home for the first time after stroke has declined from 13% in 2006 (RCP 2011). The percentage of people newly admitted to care homes is similar between the three nations involved in the National Sentinel Stroke Clinical Audit of the Royal College of Physicians (RCP) , with 10% in England, 10% in Wales and 9% in Northern Ireland (RCP 2011). Data for Scotland are not included in the RCP National Sentinel Stroke Clinical Audit. Scotland has been collecting national data about stroke care since 2002 and since 2005 has published this information in the Scottish Stroke Care Audit report (Scottish Stroke Care Audit Team 2011). However, this report does not contain data on the discharge destinations of stroke survivors.

It is estimated that around 20% to 25% of all care home residents in the USA (Quilliam 2001) and in the UK (National Audit Office 2005) have had a stroke, and stroke is reported to be the second most common cause of disability after dementia in a UK nursing home population (Martin 1998). A US study by the National Heart, Lung and Blood Institute in Boston reported that 26% of ischaemic stroke survivors from the Framingham cohort study, involving 5209 participants from the population of Massachusetts, were institutionalized in a care home with nursing care provided at six months after stroke (Kelly‐Hayes 2003). The Australian National Stroke Audit (National Stroke Foundation 2011b) reported that of 3548 people audited, 338 (11%) were discharged to a care home after stroke. Stroke survivors who reside in care homes are more physically and cognitively impaired and therefore more dependent than those living in their own homes (Bowman 2004). A cross‐sectional study using a US population‐based data set of 53,829 care home residents with stroke described these stroke survivors as a clinically complex population of frail elderly people with a high prevalence of co‐morbid conditions (Quilliam 2001).

Description of the intervention

Occupational therapy aims to help people reach their maximum level of function and independence in all aspects of daily living (Legg 2006). Occupational therapists achieve this outcome by enabling individuals to accomplish things that will enhance their ability to participate, or by modifying the environment to better support participation in daily life (World Federation of Occupational Therapists 2010). Occupational therapists define 'occupation' as much more than an individual’s chosen career. Occupation refers to every activity that people engage in during everyday life. This review focused on the effectiveness of occupational therapy for increasing independence in ADL. ADL include personal care activities such as washing, dressing, grooming, toileting and feeding, as well as 'extended' ADL leisure activities such as gardening, crafts, reading and other purposeful activities in which people choose to participate. This review focused on occupational therapy interventions for ADL for stroke survivors residing in care homes.

For the purpose of this review, the term 'care home' encompassed homes with and without nursing care and included various public and private institutions, sometimes referred to as 'residential homes', 'nursing homes', 'rest homes', 'old people's homes', or 'long‐term care institutions'. We defined a 'care home' using the definition used in two previous reviews (Forster 2009; Ward 2009) as a setting that provides overnight accommodation and communal living facilities for long‐term care and provides nursing or personal care, or both, for people with illness, disability or dependence. We included care homes from all funding models (private, charitable, not‐for‐profit and government‐owned).

How the intervention might work

Defining 'occupational therapy' can be complex because the role of an occupational therapist is diverse. The interventions provided by an occupational therapist in a care home setting may vary, but the focus of the intended outcome will be to increase, restore or maintain independence in performing ADL (including self‐care tasks and recreational and leisure activities), increase comfort and safety and prevent stroke‐related complications. Possible occupational therapy interventions were defined in a systematic review of occupational therapy for people with stroke (Steultjens 2003) and may include:

the provision of equipment and adaptations to the environment, as well as instruction in the use of assistive devices (Barrett 2001);
individual resident training of daily living skills such as washing and dressing (Walker 1996);
individual resident training of sensory‐motor functions such as grasp and release (Feys 1998; Kwakkel 1999);
individual resident training of cognitive functions such as memory and visual scanning (Carter 1983);
provision of splints to achieve increased range of movement and to reduce contractures in the hand (Langlois 1991); and
education and training of primary caregivers (care home staff) and family in areas such as correct moving and handling procedures.

Evidence suggests that occupational therapy can specifically target the consequences of stroke by aiming to improve independence in ADL and by improving the ergonomics of the environment. A Cochrane systematic review and meta‐analysis of nine trials, involving 1258 participants, of occupational therapy provision to people with stroke in the community, specifically focusing on personal ADL only, showed improved performance and a reduction in the risk of poor outcomes such as death, deterioration or dependency in personal ADL (Legg 2006). For every 100 people who received occupational therapy intervention, 11 (95% confidence interval (CI) 7 to 30) were spared a poor outcome. Although this Cochrane systematic review did not purposely exclude studies with participants who were care home residents, one‐third of the trials included in the review and meta‐analysis did exclude people who were residents in, or were to be discharged to, a residential or nursing home (Legg 2006). Only one of the nine trials included in the review and meta‐analysis (Sackley 2003; Sackley 2004) involved delivering an occupational therapy intervention specifically to care home residents with stroke within a care home setting.

Why it is important to do this review

Three‐quarters of strokes occur in people over 65 years old (National Audit Office 2010), and an increase in stroke in members of this age group of the population is predicted over the coming decade, inevitably leading to a rise in demand for care home placements. Residents of care homes have been reported to have complex healthcare needs, reflecting multiple long‐term conditions with significant disability and frailty (British Geriatrics Society 2011). Adverse consequences of stroke may include high dependency in self‐care tasks, falls, pain, pressure ulcers and emotional distress (Kelly‐Hayes 2003; Langhorne 2000; Sackley 2002). Stroke survivors residing in care homes are likely to be amongst the most disabled, dependent and vulnerable of stroke survivors, yet few care home residents receive rehabilitation (O'Dea 2000; Sackley 2001a).

Despite evidence of the efficacy of occupational therapy in improving independence in personal ADL and preventing deterioration, it has been estimated that as few as 3% of care home residents in the UK have access to occupational therapy services (Barodawala 2001) compared with 93% in the Netherlands (Sprangers 2000). The prevalence of therapy (occupational and physical) in care homes with nursing input across the world was investigated over a decade ago and was reported to be 11% in the USA, 14% in Italy, 23% in Denmark and 30% in Japan, rising to 31% in Iceland (Berg 1997). One plausible reason for this variation in therapy provision may be the variation in the size and facilities of care homes between the different countries. For example, the average care home in the UK contains around 30 beds (Office of Fair Trading 2005), compared with an average of more than 160 beds in the Netherlands (Ribbe 1997). The question of whether we are comparing similar phenomena is critical for international comparisons (US Department of Health and Human Services 1993). If care homes in the Netherlands more closely resemble intermediate care, respite‐oriented facilities or rehabilitation wards than typical UK care homes, it would be unfair to draw such comparisons on occupational therapy provision between the two nations. In 1985 the percentage of older persons (older than 65 years of age) living in care homes with nursing care was comparable between Australia (4.4%), Canada (4.2%), the Netherlands (3%), Norway (4.8%) and the USA (4.6%). However, care homes with nursing care varied in their role and function. The percentage of older people residing in care homes without nursing care was less comparable between countries: for example, 0.9% of older people in the USA reside in care homes without nursing care compared with 9% of older people in the Netherlands residing in this type of care home without nursing input (US Department of Health and Human Services 1993).

It is not known whether the same benefits of occupational therapy found amongst community‐dwelling stroke survivors (Walker 2004) would be seen in the care home population with stroke who have a high prevalence of immobility, incontinence and confusion (Bowman 2004). Stroke survivors living in care homes (with and without nursing care) are more likely to have co‐morbidities such as dementia (38% of residents), arthritis, cardiovascular disease, respiratory disease, deafness, depression, fractures and blindness (Bebbington 2001). Overall, 75% of care home residents are classified as severely disabled (Office of Fair Trading 2005). Assistance is required with at least one self‐care task in 57% of women and 48% of men in UK care homes (Office of Fair Trading 2005). In addition to differences in disability levels, stroke survivors living in care homes differ from those able to remain in their own homes in that they have a very small personal living space (as little as 10 metres² in area) (Hanson 2003), and much of their day is likely to be spent in homogeneous facilities that provide shared toilets and bathrooms and shared living areas such as a communal lounge and dining area (Help the Aged 2007). Most equipment required to complete daily activities is also likely to be shared with other residents rather than being specific to each resident’s needs. Care home residents are required to live as part of a small community usually with shared daily routines such as mealtimes and time‐tabled activities. In contrast, those with stroke living in their own home are likely to have greater freedom and choice over their daily routine. Clearly the care home population with stroke differs from the stroke population living in their own homes, and a review is needed that examines the benefits of occupational therapy provided to this specific group.

A Cochrane systematic review of rehabilitation for older people in long‐term care concluded that provision of physical rehabilitation interventions to long‐term care residents is worthwhile and safe, reducing disability with few adverse events (Forster 2009). This was a narrative review as a meta‐analysis could not be performed because of the heterogeneity of outcome measures used in the included studies. This review examined physical rehabilitation, defined as 'all interventions which primarily aim to maintain or improve physical function, rather than those relating to personal care or nursing needs'. The authors also excluded interventions that addressed cognitive deficits or mood disorders unless they also aimed to improve the physical state (Forster 2009). No review has examined the efficacy of occupational therapy interventions targeted specifically at improving and maintaining independence in ADL after stroke for those residing in care homes.

The 'My Home Life' document (Help the Aged 2007) states that occupational therapy can improve everyday functioning and quality of life of older people (Sackley 2001a; Sackley 2001b; Sackley 2004), and that the consequences of a lack of occupational therapy services in care homes can lead to unnecessary dependency and high rates of immobility‐related complications (Sackley 2004). Despite evidence in support of the benefits of purposeful and meaningful activity (Ballard 2001; Baum 1995; Kiely 2003) (a key philosophy of occupational therapy), historically the level of physical activity and positive stimulation in care home residents has been low (Challis 2000; College of Occupational Therapists 2007; Help the Aged 2006; Nolan 1995). Recent studies report that care home residents spend as much as 63% of their day on non‐therapeutic activities, such as sitting passively doing nothing and not interacting with others (Cohon‐Mansfield 1992; Huijben‐Schoenmakers 2009; Sackley 2006a). In one pilot observational study involving residents from an 18‐bed local authority residential home in England (Sackley 2006a), the residents were observed to be "busy doing nothing" with residents sitting (with their eyes open or closed) for 97% of observations. It is known that inactivity and immobility are associated with further deterioration of function (Sackley 2008), and members of the care home population with stroke are more likely to experience additional complications as compared with stroke survivors living in their own homes.

Currently in the UK, there is no requirement for care home staff to have stroke‐specific training. A recent review by the Care Quality Commission (CQC 2011) (the independent regulator of healthcare and adult social care services in England) reported concerns around levels of staff knowledge and skill in stroke care. This review reported that whilst local stroke pathways (policies setting out how care should be delivered) are in place across England, only 32% of these specifically cover people who have had a stroke and are residing in care homes (CQC 2011). Similarly, the National Clinical Guideline for Stroke (RCP 2008) includes more than 300 recommendations for improving the care of people who have had a stroke in the UK, but only 16 of these relate to the care given to people longer than six months after stroke (National Audit Office 2010). No UK guidelines pertain to the longer‐term stroke specialist rehabilitation required to meet the complex needs of those with stroke residing in UK care homes. Similarly, such specific guidelines are absent from the Australian Clinical Guidelines for Stroke Management (National Stroke Foundation 2010). A Scientific Statement from the American Heart Association (Miller 2010) reported on the need to educate nurses and other members of the interdisciplinary team about the potential for recovery in later or more chronic phases of stroke care. In addition, the importance of access to relevant health professionals for those living in care homes has been emphasised (RCP, RCN & BGS 2000).

It could be argued that the care home population has the greatest need for ongoing therapy and rehabilitation post‐stroke because they have such high levels of dependency and co‐morbidities and low levels of activity, yet an inequitable level of therapy is currently provided compared with therapy provided to those living at home. It is known that commissioners require evidence to support the effectiveness of longer‐term rehabilitation therapies if they are to commission the provision of such stroke services, and at present, this evidence is lacking. The purpose of this review is to examine available evidence specifically showing the benefits of occupational therapy interventions aimed towards increasing independence in ADL (including both personal and extended ADL) for people with stroke who are residing in care homes.

Objectives

To measure the effects of occupational therapy interventions (provided directly by an occupational therapist or under the supervision of an occupational therapist) targeted at improving, restoring and maintaining independence in ADL among stroke survivors residing in long‐term institutional care termed collectively as 'care homes'. As a secondary objective, we sought to evaluate occupational therapy interventions provided to reduce complications such as depression and low mood.

Methods

Criteria for considering studies for this review

Types of studies

We included all randomised controlled trials (RCTs) or cluster‐RCTs that evaluated occupational therapy interventions with the specific aim of facilitating, restoring or maintaining independent function in any ADL (or that aimed to reduce complications) for stroke survivors (or that included a defined subgroup of stroke survivors) who were permanently residing in a care home with or without nursing care. We included studies that compared interventions provided by a qualified occupational therapist or by an occupational therapy assistant under the direction of a qualified occupational therapist versus standard care (i.e. routine care usually received by residents or no intervention). We also included studies that compared occupational therapy interventions targeting ADL with usual care interventions, and studies that compared different types of occupational therapy interventions with each other.

We excluded quasi‐randomised trials that used, for example, alternate days of the week as the method of randomisation to eliminate the possibility of systematic bias affecting outcomes (Creswell 2009). When trials were described in a way that implied that they were randomised, and when the demographic details of participants in each group were similar, we included the trial and undertook sensitivity analysis in the presence or absence of these data.

We included cross‐over studies, but we planned to include only data from the first phase of cross‐over studies in the meta‐analysis.

Types of participants

We included studies that recruited people with a clinical diagnosis of stroke regardless of age, sex, gender, time since stroke onset or ethnic group, and those with multiple diagnoses, as long as they permanently resided in a care home. We excluded trials of mixed causes in which the percentage of participants with stroke was less than 50%.

We defined stroke as a focal neurological deficit caused by cerebrovascular disease (confirmation of the clinical diagnosis using imaging was not compulsory).

Within the European Union, different definitions of long‐term care coexist (European Commission 2008). Definitions used by the member states vary in identifying the care recipient and in defining the services provided (European Commission 2008). In this review we used the term 'care home' to include various public and private institutions caring for the dependent elderly, such as 'residential homes', 'nursing homes', 'rest homes', 'old people's homes' and 'long‐term care institutions'. We defined a care home, using the definition used in two previous Cochrane reviews (Forster 2009; Ward 2009), as providing:

communal living facilities for long‐term care;
overnight accommodation;
nursing or personal care;
for people with illness, disability or dependence.

We included care homes from all funding models (private, charitable, not‐for‐profit and government owned).

Types of interventions

We included all occupational therapy and therapy‐based interventions (delivered on an individual or group basis) provided directly by a qualified occupational therapist, or by an occupational therapy assistant under the direction of a qualified occupational therapist, that aimed to increase or maintain occupational performance and independence, and to improve function in ADL ('personal' ADL or 'extended' ADL, or both).
We defined standard care as the routine care that residents usually received whilst residing in a care home.
We included only trials that provided occupational therapy as part of a multidisciplinary team intervention, for which the occupational therapy component of the intervention could be clearly identified and extracted from the results.

Types of outcome measures

We aimed to record outcomes that were likely to reflect the domains targeted by occupational therapy interventions.

Primary outcomes

Performance in ADL at the end of scheduled follow‐up (e.g. Barthel ADL Index score (Mahoney 1965), Nottingham extended ADL Index score (Nouri 1987), Edmans ADL Index score (Edmans 1997)). When both personal ADL outcomes and extended ADL outcomes were available, we used personal ADL outcome data.
Death or a poor outcome. We defined poor outcome as deterioration in ability to perform ADL (a drop in ADL score).

Secondary outcomes

Performance in ADL at the end of intervention (e.g. Barthel ADL Index score, Nottingham extended ADL Index score, Edmans ADL Index score). When both personal ADL outcomes and extended ADL outcomes were available, we used personal ADL outcome data.
Death (the number of deaths from any cause).
Global quality of life (e.g. EuroQoL EQ‐5D score (EuroQol Group 1990)).
Mobility (e.g. Rivermead Mobility Index score (Collen 1991)).
Mood (e.g. Geriatric Depression Scale score (Yesavage 1982)).
Global cognition (e.g. attention, memory, perceptual skills, problem solving) (Mini Mental Status Examination (MMSE) score (Folstein 1975)).
Admission to hospital or other higher‐dependency institution.
Adverse events (e.g. falls, new pressure sores, new contractures).
Satisfaction with care (Satisfaction with Stroke Care questionnaire SASC‐19 (Boter 2003)).
Health economic outcomes (e.g. EuroQol EQ‐5D (EuroQol Group 1990)).

Search methods for identification of studies

See the 'Specialized register' section in the Cochrane Stroke Group module.

We performed both electronic searches and handsearches. We included trials in all languages and where possible arranged translation of articles published in languages other than English. If translation was not feasible, we included possibly relevant trials in the Characteristics of studies awaiting classification table.

Electronic searches

The primary search resource was the Cochrane Stroke Group Trials Register, which was searched for us in August 2012. In addition, we searched the following bibliographic databases:

Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, September 2012) (Appendix 1):
MEDLINE (1948 to September 2012) (Appendix 2);
EMBASE (1980 to September 2012) (Appendix 3);
Cumulative Index to Nursing and Allied Health Literature (CINAHL) (1982 to September 2012) (Appendix 4);
Allied and Complementary Medicine Database (AMED) (1985 to September 2012) (Appendix 5);
Occupational therapy database of systematic reviews and randomised controlled trials (OTseeker) (1980 to September 2012) (Appendix 6);
PsycINFO (1967 to September 2012) (Appendix 7);
Physiotherapy Evidence Database (PEDro) (1952 to September 2012) (Appendix 8);
Applied Social Index and Abstracts (ASSIA) (1987 to September 2012) (Appendix 9);
NHS Economic Evaluation Database (NHS EED) (1991 to September 2012) (Appendix 10);
Education Resources Information Center (ERIC) (1966 to September 2012) (Appendix 11);
Center for International Rehabilitation Research Information and Exchange (CIRRIE) (1990 to September 2012) (Appendix 12);
Web of Science (all years up to October 2012) (Appendix 13);
ProQuest Dissertations & Theses A&I (http://search.proquest.com).

We also searched the following registers of ongoing and completed trials (September 2012):

Current Controlled Trials (www.controlled‐trials.com);
Clinical Trials (www.ClinicalTrials.gov);
EU Clinical Trials Register (www.clinicaltrialsregister.eu);
Stroke Trials Registry (www.strokecenter.org/trials/);
WHO International Clinical Trials Registry Platform (www.who.int/ictrp/en/);
Australian New Zealand Clinical Trials Registry (www.anzctr.org.au/).

We developed the MEDLINE search strategy with the help of the Cochrane Stroke Group Trials Search Co‐ordinator and adapted it for the other databases.

Searching other resources

In an effort to identify additional published, unpublished and ongoing trials, we performed the following additional searches.

Reference searching

We used the Science Citation Index Cited Reference Search for forward tracking of important papers. We inspected the reference lists of identified articles that we obtained in full text to look for evidence of additional studies.

Personal contact

We contacted authors of relevant studies to inquire about other sources of relevant information.

Handsearches

We handsearched the following journals that were not already included in the handsearching carried out by The Cochrane Collaboration and were not included in the Cochrane Central Register of Controlled Trials (CENTRAL):

American Journal of Occupational Therapy (1997 to November 2012);
Australian Journal of Occupational Therapy (1980 to November 2012);
British Journal of Occupational Therapy (1980 to November 2012);
Canadian Journal of Occupational Therapy (1996 to November 2012);
Clinical Rehabilitation (January 2012 to November 2012);
Occupational Therapy International (2009 to November 2012);
Scandinavian Journal of Occupational Therapy (1997 to November 2012).

Data collection and analysis

Selection of studies

Two review authors (JCF‐S, CSC) independently assessed all titles and abstracts of the records identified by searches of the electronic databases and excluded all studies that clearly did not refer to an RCT or a cluster‐RCT of an occupational therapy intervention for care home residents. We obtained the full text of the remaining studies, and the same two review authors independently assessed each study to determine whether it met the pre‐defined review selection criteria. These two review authors resolved any disagreements by discussion and, if necessary, in consultation with a third review author (MFW) until they reached a consensus. The review authors were not blinded to the names of study authors, institutions or journals of publication. We report excluded studies and the reasons for exclusion in the Characteristics of excluded studies table.

Data extraction and management

Data extraction

Two review authors (JCF‐S, CSC) independently extracted data from all included published sources to ensure reliability. When necessary, we contacted study authors for clarification. These two review authors discussed any disagreements with the third review author and documented the decisions. We extracted data presented only in graphs and figures whenever possible. We contacted study authors to request missing information or clarification.

Management

We used Review Manager 5.2 (RevMan 2011) to prepare and maintain the review, to perform meta‐analyses of the data and to present the results graphically. The extracted data were independently entered using the Review Manager software and included full citation details of the study, numbers and characteristics of participants (inclusion and exclusion criteria), descriptions of interventions, outcome measures, intention‐to‐treat analyses, withdrawals and losses to follow‐up.

Forms

We extracted data onto standard simple forms that assisted us when we examined the methodological quality of identified studies.

Scale‐derived data

We planned to include continuous data from rating scales only if the measuring instrument was (1) a self‐report, or (2) completed by an independent rater or relative (not the therapist).

Endpoint versus change data

We planned to use primarily endpoint data and to use change data only if the former data were not available.

Skewed data

Continuous data on clinical and social outcomes often are not normally distributed. To avoid applying parametric tests to non‐parametric data, we aimed to apply the following standards to all data before inclusion.

Standard deviations and means were reported in the article or could be obtained from the authors.
When a scale started from the finite number zero, the standard deviation, when multiplied by two, was less than the mean (as otherwise, the mean was unlikely to be an appropriate measure of the centre of the distribution) (Altman 1996).

Endpoint scores on scales often have a finite start and endpoint, and these rules can be applied. When continuous data are presented on a scale that includes a possibility of negative values (such as change data), it is difficult to tell whether or note data are skewed. Skewed data pose less of a problem in looking at means if the sample size is large.

Common measure

To facilitate comparison between trials, we planned to convert variables that could be reported in different metrics, such as days in hospital (mean days per year, per week or per month) to a common metric (e.g. mean days per month).

Conversion of continuous to binary

When possible, we planned to convert outcome measures to dichotomous data. This can be done by identifying cut‐off points on rating scales and dividing participants accordingly into 'clinically improved' or 'not clinically improved'. When necessary, we contacted study authors to ask for information.

Direction of graphs

When possible, we intended to enter data in such a way that the area to the left of the line of no effect indicated a favourable outcome for occupational therapy intervention.

'Summary of findings' table

We anticipated including the following outcomes in 'Table 1:

function (also referred to as 'occupational performance') in ADL (personal ADL and/or extended ADL). (When both personal ADL and extended ADL outcomes data were available, we used personal ADL outcomes data.);
global poor outcome;
death;
quality of life;
mobility;
mood;
global cognition;
adverse events;
satisfaction with care;
health economic outcomes.

Assessment of risk of bias in included studies

JCF‐S and CSC worked independently to assess risk of bias in accordance with the Cochrane Collaboration's tool for assessing quality and risk of bias (Higgins 2011). This tool addresses evaluation of the following specific components for each trial. We identified the following factors as potentially important for sensitivity analyses but did not use them as exclusion criteria.

Method of generation of the randomisation sequence.
Method of concealment to treatment allocation (it was considered adequate if the assignment could not be foreseen).
Blinding of outcomes assessors, participants and clinicians.
Completeness of outcomes data (including attrition and exclusions from analysis).
Presence of an 'intent‐to‐treat' analysis.
Selective reporting.
Other biases (concerns about other biases not addressed in the other domains of the tool).

We then categorised the trials as follows:

low risk of bias;
high risk of bias;
unclear ‐ uncertain risk of bias.

We did not include in the meta‐analysis trials with a high risk of bias (defined as at least three out of five components categorised as 'high risk'). If the two review authors (JCF‐S, CSC) disagreed, the final decision was made by consensus with the involvement of a third review author (MFW). When inadequate details of the trial were provided, we contacted the study authors to request further information.

Measures of treatment effect

Dichotomous outcomes

For dichotomous outcomes (i.e. death, drop in Barthel ADL Index score), we planned to express the intervention effect as an odds ratio (OR) with 95% confidence interval (CI).

Continuous data

For continuous outcomes (i.e. personal ADL (PADL) score, Quality of Life (QoL), depression score), our intention was to present the mean difference (MD) with corresponding 95% CI.

When studies assessed the same outcome but measured it in different ways (e.g. different questionnaires used to measure performance in PADL), we intended to present the data as standardised mean difference (SMD) with corresponding 95% CI.

Unit of analysis issues

Cluster trials

Analysis and pooling of clustered data can pose problems, as authors often fail to account for intra‐class correlation in clustered studies, leading to a 'unit of analysis' error (Divine 1992), whereby P values are low, CIs unduly narrow and statistical significance overestimated. When clustering was not accounted for in primary studies, we planned to present data in a table, in which a (*) symbol would be used to indicate the presence of a probable unit of analysis error. When clustering had been incorporated into the analysis of primary studies, we planned to present the data as if from a non‐cluster randomised study, while adjusting for the clustering effect.

We had planned to follow the statistical recommendation used in a previous Cochrane review (Xia 2002): binary data presented in a report should be divided by a 'design effect'. This is calculated using the mean number of participants per cluster (m) and the intra‐class correlation coefficient (ICC) [Design effect = 1 + (m ‐ 1) *ICC] (Donner 2002). If the ICC was not reported, it was assumed to be 0.1 (Ukoumunne 1999).

If cluster studies have been appropriately analysed with ICCs and relevant data documented in the report taken into account, synthesis with other studies is possible using the generic inverse variance technique.

Cross‐over trials

A concern of cross‐over trials was the possibility of carry‐over effect. This occurs if an effect of the treatment in the first phase is carried over to the second phase. As a consequence, on entry into the second phase, participants can differ systematically from their initial state. Also, cross‐over trials are not considered appropriate if the condition of interest is unstable (Elbourne 2002). These effects are likely in stroke; therefore we intended to use only data from the first phase of cross‐over studies.

Studies with multiple treatment groups

When a study involved more than two treatment groups, if relevant, we presented the additional treatment group in comparisons. When the additional treatment groups were not relevant, we did not reproduce the data.

Dealing with missing data

We sought to obtain relevant missing data from the primary investigators. We evaluated important numerical data such as numbers of people screened, numbers of participants randomly assigned, losses to follow‐up and withdrawals. For any outcome, when more than 50% of the data was unaccounted for, we did not reproduce the data or use it within the analyses. If more than 50% of participants in one treatment group of a study were lost, but the total loss was less than 50%, we marked such data with an asterisk (*) to indicate that the result may be prone to bias. We also investigated attrition rates. When attrition for a binary outcome was between 0 and 50% and data had not been clearly described, we present the data on a 'once‐randomised‐always‐analyse' basis (intention‐to‐treat analysis). We assumed that participants leaving a study early had the same rates of negative outcome as those who completed the study, with the exception of the outcome of death. We planned to undertake a sensitivity analysis to test how prone the primary outcomes were to change when 'completed' data were compared with the intention‐to‐treat analysis. When attrition for a continuous outcome was between 0 and 50% and completer‐only data were reported, we reproduced these.

Assessment of heterogeneity

Clinical heterogeneity

Initially we planned to consider all included studies (without seeing comparison data) to judge clinical heterogeneity. We planned to look for clearly outlying situations or participant groups not predicted to arise. If such outlying situations or participant groups arose, all review authors would discuss these.

Methodological heterogeneity

Again, we planned to initially consider all included studies without seeing comparison data, to judge methodological heterogeneity. We would inspect all studies for clearly outlying methods not predicted to arise. When such methodological outliers arose, all review authors would fully discuss these until we reached consensus.

Statistical heterogeneity

Visual inspection

We planned to visually inspect the graphs to investigate the possibility of statistical heterogeneity.

Employing the I² statistic

We planned to investigate heterogeneity between studies by considering the I² method alongside the X² P value. We identified an I² estimate greater than or equal to 50% accompanied by a statistically significant X² statistic as evidence of substantial levels of heterogeneity (Higgins 2011). If we found substantial levels of heterogeneity in the primary outcome, we intended to explore reasons for heterogeneity (subgroup analysis and investigation of heterogeneity).

Assessment of reporting biases

When funnel plots were appropriate and possible, we tested for funnel plot asymmetry.

Data synthesis

The random‐effects method incorporates an assumption that the different studies are estimating different, yet related, intervention effects. The random‐effects model takes into account differences between studies even if there is no statistically significant heterogeneity. However, a disadvantage of the random‐effects model is that it puts added weight onto small studies, which often are the most biased ones. Depending on the direction of effect, these studies can inflate or deflate the effect size. Therefore, we planned to use a fixed‐effect model and to carry out sensitivity analysis to determine whether there were differences when a random‐effects model was employed.

Subgroup analysis and investigation of heterogeneity

If data were available, we performed subgroup analyses for type of intervention, intensity (dose) and duration of treatment intervention, as well as timing of occupational therapy after stroke (acute: less than six weeks; subacute: six weeks to six months; and chronic: more than six months).

We anticipated carrying out standard tests of statistical heterogeneity and exploring sources of heterogeneity.

Sensitivity analysis

We also planned to carry out sensitivity analyses to determine the effects of omitting trials with a high risk of bias. We intended to base the sensitivity analyses on the method of randomisation, presence of an intention‐to‐treat analysis and blinding of final assessment.

Results

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of studies awaiting classification; Characteristics of ongoing studies.

Results of the search

The search strategies identified a total of 1929 records. We removed duplicates, resulting in 1436 records for initial screening. Two review authors (JCF‐S and CSC) independently screened all 1436 titles and abstracts (when available) for potentially relevant studies. A third review author (MW) screened 14 for which a discrepancy was noted. We obtained copies of 12 articles in full. Among these 12 articles, three studies had produced multiple articles; therefore we discarded three articles to an 'additional study information' pile. The remaining nine articles represented potential trials for inclusion in the review (Braun 2012; Brittle 2009; Corr 1995; Egan 2007; Frandin 2009; OTCH 2012; RICH‐T; Sackley 2006b; Tsaih 2012), of which one was included (Sackley 2006b) and one was an ongoing trial (OTCH 2012). See Figure 1 for the study flow diagram. All included, ongoing and excluded trials were published in English; therefore no translation was required. However, we requested and obtained further details from two study authors to aid our judgement on eligibility for inclusion in the review.

Included studies

One included trial (Sackley 2006b) was conducted in 2001 and included 118 participants from 12 care homes in Oxfordshire, UK. This pilot study was a cluster‐randomised controlled trial with the care home as the unit of randomisation (to avoid the chance of contamination that would be likely to occur if residents were randomly assigned individually). The purpose of the study was to evaluate an occupational therapy intervention to improve self‐care independence for residents with stroke‐related disability. Further details of the study can be found in the Characteristics of included studies table. A further ongoing study (OTCH 2012) appeared to meet the inclusion criteria. However, as no data are yet available for this trial, we would not include it in a meta‐analysis, and we will re‐consider using it in future updates of this review. Further details of this study can be found in the Characteristics of ongoing studies table.

Excluded studies

We excluded seven studies after considering the full articles. We excluded studies in which participants had a mixed cause for residence in a care home and in which stroke accounted for fewer than 50% of participants; and those in which the participants were not care home residents. We also excluded studies if the intervention was not delivered by an occupational therapist, and those that included occupational therapy as part of a multidisciplinary team intervention but where the occupational therapy component of the intervention could not be clearly identified and extracted from the results. Excluded studies are listed in the Characteristics of excluded studies.

Risk of bias in included studies

Two review authors (JCF‐S and CSC) rated the methodological quality of the study independently using the bias criteria in the risk of bias table. See Risk of bias in included studies. We present our judgements about each risk of bias item for the included study in Figure 2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Sackley 2006b used a clearly concealed randomisation procedure, allocating participants by care home (cluster randomised) to receive or not receive an occupational therapy intervention. Randomisation was carried out independently by a statistician, with care homes grouped into three strata: type of home (residential, nursing or both), funding source (private or local authority) and setting (urban or rural). Computer‐generated random numbers were used to randomly allocate care homes to one of the two groups (occupational therapy intervention or standard care control group). Group allocation was revealed only to the treating therapist and not to the outcomes assessor.

Blinding

The outcome assessor was blinded as to the group assignment of participants. Because of the nature of the intervention, allocation concealment from participants, treating therapist or care home staff involved in the study was not possible.

Incomplete outcome data

Participants in Sackley 2006b were reported to be treated on an 'intention‐to‐treat' basis. All 'missing' data during the course of the study were related to death of participants, which is to be expected in a frail elderly population.

Selective reporting

The risk of selective reporting bias is unclear. The Sackley 2006b study team could not supply the review authors with a copy of the original study protocol. The article reported all outcomes that it stated would be provided. However, it was not possible to ensure that the original intention had been to report on these specific outcomes and no additional outcomes.

Other potential sources of bias

Risk of bias is possible when a cluster design is used. However, the Sackley 2006b study justified the use of a cluster‐randomised trial because of the possibility of contamination if individual participants within each care home were randomly assigned. In a care home setting, equipment is often shared, and staff work with a number of residents. Therefore, the intervention provided by the occupational therapist could have easily affected the control participants unwittingly had a cluster‐randomised design not been used.

Effects of interventions

See: Table 1

We included only one study in this review; therefore a meta‐analysis was not possible. Data were available for the outcomes: function (occupational performance) in ADL at the end of scheduled follow‐up, global poor outcome (death or a drop in ADL score) at the end of scheduled follow‐up, function in ADL at the end of intervention, and mobility. The data for outcomes related to function in ADL and mobility were reported in the study article as mean (SD) values, and data related to global poor outcome (death or a drop in ADL score) were reported as total N and number of participants who had clinically deteriorated in each treatment group.

Primary outcomes

Performance (function) in ADL at the end of scheduled follow‐up

One included trial (Sackley 2006b) recorded the Barthel ADL Index score; this was used in the analysis as the measure of performance in ADL at the end of scheduled follow‐up. As the included trial was a cluster‐randomised trial, we used an intra‐cluster correlation coefficient of 0.1 to calculate average cluster size to take account of the design effect.

Average cluster size in the trial was calculated by dividing the total number of participants by the total number of care home clusters, (63 + 55)/(6 + 6) = 9.83.

The design effect for the trial as a whole is therefore 1 = (m ‐ 1)*ICC = 1 + (9.83 ‐ 1) x 0.1 = 1.883.

This results in an effective sample size in the occupational therapy intervention group of 63/1.883 = 33 and an effective sample size in the control group of 55/1.883 = 29. The design effect was applied to the outcomes data for performance (function) in ADL at the end of scheduled follow‐up.

The SMD using a fixed‐effect model was 0.39 (95% CI ‐0.11 to 0.90; P = 0.13) (Analysis 1.1).

Trials were insufficient to allow firm conclusions to be drawn.

Death or a poor outcome (drop in ADL score) at the end of scheduled follow‐up

At six months, Sackley 2006b reported a reduction in the number of care home residents who died or deteriorated in their ability to perform ADL among participants who received occupational therapy intervention (32/63, 51%) compared with the control group, which received standard care (42/55, 76%) (OR 0.32, 95% CI 0.14 to 0.71; P = 0.005).

However, applying the design effect (1 + (9.83 ‐ 1) 0.1 = 1.883) to the number of residents (participants) who died or deteriorated in their ability to perform ADL (global poor outcome) produces the following results: 17/33 (51%) in the intervention group compared with 22/29 (76%) in the control group (OR 0.34, 95% CI 0.11 to 1.01; P = 0.05) (Analysis 1.2).

Trials were insufficient to allow firm conclusions to be drawn.

Secondary outcomes

Performance (function) in ADL at the end of intervention

Sackley 2006b reported performance in ADL at the end of the three‐month intervention period. When the design effects were applied to the published outcome data, the SMD using a fixed‐effect model was 0.48 (95% CI ‐0.03 to 0.99; P = 0.06) (Analysis 1.3).

Trials were insufficient to allow firm conclusions to be drawn.

Death at the end of scheduled follow‐up

Data were available from Sackley 2006b for the outcome of death at end of scheduled follow‐up (six months). Applying the design effect (1.883) to the reported number of deaths in the intervention group (10/63, 16%) compared with the control group (20/55, 36%) at six months produces the following adjusted results: 5/33 (15%) in the intervention group compared with 11/29 (38%) in the control group (OR 0.29, 95% CI 0.09 to 0.98; P = 0.05) (Analysis 1.4).

Trials were insufficient to allow firm conclusions to be drawn.

Global quality of life