Occupational Therapy intervention for residents with stroke living in UK Care Homes trial design summary
The OTCH study was a Phase III pragmatic parallel-group cluster RCT with an economic evaluation. The evaluation assessed the clinical efficacy and economic impact of providing an OT service for stroke survivors living in care homes. The primary research objective was to assess the influence of a 3-month course of OT, according to whether or not it helped participants maintain levels of independence in ADL compared with usual care. Secondary outcome measures assessed the influence of the OT intervention on participants’ mobility, mood and HRQoL.
More care homes were recruited than originally planned because of a larger than expected number of small clusters. The number of eligible residents within each care home with a history of stroke was lower than expected.11,15–17 All baseline characteristics were similar across randomisation arms in regards to age, sex, ethnicity and comorbidities. A large proportion of the participant population had significant physical disabilities, substantial limitations on activity (see and ) and increased dependence for personal ADL. Moreover, the assistance of a consultee to consent on the participant’s behalf was needed in 61% of cases,68 indicative of reduced autonomy. There was a high prevalence of cognitive and language impairment: 57% of participants scored below 15 points on the Sheffield screening test, indicative of significant language impairment, and 70% of participants scored in the range signifying cognitive impairment on the MMSE. The participant profile overall is suggestive of significant frailty.
Principal findings
Primary outcome at 3 months
The principal findings are neutral. The 3-month course of individualised OT for care home residents living with stroke, involving patient-centred goal-setting staff training, provision of facilitatory equipment and environmental adaptation, showed no benefit on participants’ capacity to engage in personal ADL at each of the trial end points. Findings observed in the pilot phase were not replicated.48 OT provision did not help maintain independence in personal ADL compared with usual care. The number of participants with a poor outcome (intervention 54% vs. control 52%), moderate outcome (intervention 30% vs. control 34%) and good outcome (intervention 15% vs. control 14%) at the primary end point was similar between treatment arms.
Exploratory subgroup analyses focused on the primary outcome did not reveal any interactions of interest. The forest plot in indicates that participants’ age, severity on the BI at baseline, type of care home, proxy data and severity of cognitive impairment had no discernible influence on the primary outcome at 3 months. However, there is a suggestion from this exploratory analysis that the OT intervention may be effective at maintaining functional activity for residents that have less severe limitations on activity. Referring stroke survivors on an individual basis as opposed to a routine basis may still be of benefit.
Additional sensitivity analyses which included a complete case analysis, tested the potential ceiling effect, imputed missing data and removed clusters with fewer than three residents, did not change the neutral findings. A process evaluation that examined the fidelity of the OTCH intervention is presented in more depth elsewhere.88 The process evaluation summary presented here indicates that the intervention was implemented as intended.
Primary outcome assessed at 6 and 12 months
At the 6- and 12-month follow-up stages, the BI data showed no significant differences between groups. The results from the composite BI analyses, employing the change in BI score from baseline at 6 and 12 months, were similar between groups (see Table 26). The data did not support the hypothesis that a 3-month course of OT for care home residents living with stroke-related disabilities will help maintain levels of function activity in self-care tasks over the longer term compared with usual care.
Secondary outcomes
Analyses assessing mobility, mood, and HRQoL showed no benefit of the OT intervention at 3 months (see Table 24). Similarly, the secondary outcome measures at the 6- and 12-month follow-up showed no evidence of benefit of the intervention (see Table 25). Mean mobility scores decreased over the course of the trial in both randomisation arms (see Table 22). The mean mood scores on the GDS-15 remained similar for both treatment arms across all end points (see Table 23). The mean mood scores are indicative of mild depression.85
Further exploratory analyses
These additional exploratory analyses considered the potential dose effect of OT on depression scores, and the relationship between baseline RMI score and survival. No evidence of an association between OT dose and GDS-15 score was observed. However, there was strong evidence that mobility is associated with survival. Participants with a mobility score of 2 or below at baseline were 1.57 times as likely to die than those with greater mobility over the course of the 12-month trial duration (hazard ratio 1.57, 95% CI 1.15 to 2.15).
Economic evaluation discussion
Based on the mean incremental cost per QALY gain, it is unlikely that the OTCH intervention is cost-effective when compared with usual treatment. Although outcomes were virtually equivalent in both arms, costs were higher in the intervention arm and the intervention did not lead to a reduction in health resource use from other sources.
We acknowledge some limitations in this economic analysis. We requested information on equipment on the assumption that this would have been purchased by the NHS or personal social services. However, it became apparent that some of this equipment may have been purchased by the care home (although who purchased it was not systematically recorded), and that this could be considered outside of the NHS and personal social services perspective. In addition, apart from the training of care home staff, we did not consider all wider social costs, although we considered that, as all participants were in residential care, the burden on community care workers, friends and family would be likely to be less than if the participants lived in their own homes.
Death rates were similar between study arms. We tested the robustness of our estimates by performing multiple imputation for missing data and also running a sensitivity analysis excluding participants who died during follow-up (not reported); our cost-effectiveness conclusions were consistent.103
Despite these limitations, the cost-effectiveness results were conclusive. At £20,000 per QALY, this probability was 29.2% and at a threshold of £30,000 the probability was 39.2%. Based on the results of this RCT, OT was not a cost-effective routine intervention for stroke survivors living in care homes.
Strengths and weaknesses
This was the largest cluster randomised trial conducted in care homes to date. The potential pitfalls of the cluster design associated with the provision of informed consent, identification of the unit of inference and methods of stratification were addressed in the trial protocol.71,104 The trial was sufficiently powered to detect a clinically significant change in the BI measure following a 3-month course of individualised OT,82 involving task-related ADL practice, provision of adaptive equipment, adaptations to the individual’s environment and caregiver training. The unadjusted ICC for BI scores was 0.36 at baseline; however, for the change in scores from baseline to 3 months, allowing for the effect of treatment, TACs and type of care home, it decreased to 0.09. The mixed modelling approach has been shown to be a reasonable method of analysis when some cluster sizes are small, allowing for appropriate inferences to be made in relation to fixed effects.105 There is additional evidence that even with small cluster sizes, mixed modelling is a better approach than an analysis that ignores clustering.106 Furthermore, results of the sensitivity analysis showed that the conclusions were robust when small clusters were excluded. Results from the complete case analysis, where BI scores were not imputed for participants who died before the primary end point revealed similar neutral results to the primary analysis.
The OT administered to participants was similar to a standard NHS intervention,88 shown to be of benefit to stroke survivors living in their own homes.34 Compliance over the course of the trial was good, resulting in high completion rates among survivors for all assessments at each end point. On average, participants in the intervention arm were visited on five occasions with visits lasting a median of 30 minutes. Retention of care home participation was good throughout the study, with 204 (89%) homes providing data at the final 12-month end point. The main reason for care home withdrawal was the death of all participating residents. A total of 313 out of 1042 (30%) participants died during the trial. Overall, we regard the findings to be robust and conclusive.
We acknowledge several potential limitations. The percentage of care home residents affected by stroke was less than expected, which resulted in a larger number of small clusters than was originally expected. Previous research suggested the incidence of stroke in care home residents to be approximately 20–40%.11,15–17 However, this trial observed incidence of stroke in care home residents to be approximately 16% (1556 out of 9840; see Table 6). This figure concurs with results from a recent census of care homes in the UK.107 In addition, GP surgeries’ response rate to trial correspondence requesting from confirmation of participants’ stroke was low despite multiple attempts.
There were a high proportion of participants with cognitive impairment, and GDS scores were indicative of moderate and severe depression. If these participants had been excluded during recruitment, it would have reduced the external validity of the trial. However, we acknowledge that these participants may have had potentially limited engagement in therapy, as demonstrated by the observed distribution of therapy content (see Table 20). Approximately half of all therapy hours were spent on communication with participants, carers and consultees, as opposed to ADL training (see Table 20). Therapy time attributed to this category was spent on providing information and guidance to residents, staff or relatives on personal ADL, initiating referrals to other agencies and ordering equipment. In the pilot phase the duration of the intervention was also 3 months.48 The median number of visits per resident per month was 2.7 (interquartile range 1–4.2), and the median duration of therapy time per resident per month was 4.5 hours (interquartile range 2–6.9 hours).48 This represents a significant difference in the amount of therapy received by residents in the two trials. Residents received more therapy in the pilot phase. In the main trial, the median number of visits was 5 (interquartile range 3–7 visits) over 3 months and the median duration of therapy per resident was 145 minutes (interquartile range 85–255 minutes). The mean duration of therapy listed in Table 20 is 208 minutes (SD 208 minutes; minimum 10 minutes, maximum 1380 minutes). The difference in the amount of therapy administered between the pilot trial and the main trial indicates that the high levels of disability among participants may have prevented engagement in personal ADL training and reduced therapy time as a result.
When viewing these figures it is also important to acknowledge differences in eligibility criteria between the two trials. The pilot trial included 118 residents with a moderate to severe BI score that was between 4 and 15 out of 20. The Phase III definitive trial reported here used a pragmatic approach and included residents with a BI score between 0 and 20. As a consequence the mean baseline BI scores differed between the two trials. In the pilot trial, the mean baseline BI score in the intervention arm was 10.1 (SD 5.7) and 9.5 (SD 5.2) in the control arm. The mean baseline BI scores in both arms in the pilot trial are classed as moderate. The mean baseline scores in the main trial were in the severe range across both arms. Table 16 displays the baseline BI scores. Overall, 48% of participants in the intervention arm had a BI score between 0 and 4, indicating very severe limitations when engaging in personal ADL. It is also important to reflect on the conclusions from the pilot trial.48 After receiving a higher dose of therapy, participants in the intervention arm showed a tendency for improvement between baseline and the primary outcome time point at 3 months for the Barthel composite measures only (see Table 21 and Table 25 for similar results from this trial). However, participants in the intervention arm deteriorated in a similar way to the control group at all subsequent follow-up time points.48 Therefore, despite the difference in the amount of therapy administered between the two trials potentially being viewed as a limitation, the main trial has provided a realistic portrayal of the high level of impairment exhibited by care home residents living with stroke-related disabilities throughout England and Wales (see Table 16 and and ). A realistic conclusion is that the majority of participants may have lacked the capacity to directly engage with structured, repetitious activity-based therapy.
Safety
The OT provided to residents was not an experimental intervention. Therapy given to residents was similar to what stroke survivors living in the community would receive from the NHS. Overall, tolerance of the OT provided to participants was very good, resulting in no adverse events occurring that were attributable to the intervention. There was a higher incidence of falls among participants in the intervention arm. It is a possibility that, by facilitating an increase in functional activity, it may have led to an increased risk of falling. However, according to recently published figures,108–110 the average fall rate of residents in long-term institutional care varies from 1.45 to 2.50 per annum. This suggests the quarterly fall rate of 0.18 recorded in the intervention arm is below the previously observed parameters.
Generalisability
The large sample size, the geographical distribution of different types of care home, the involvement of a large number of qualified therapists and the inclusion of all stroke survivors, regardless of cognitive and communication impairments, increase the potential for generalisability of the observed results across all UK care homes. The baseline participant characteristics, including age, level of comorbidity, cognitive impairment and incidence of depression observed in the present study, are similar to rates recently reported in another large cluster RCT carried out in UK care homes.111
Research in context
The neutral results observed in the OTCH trial concur with results observed in other RCTs conducted in a care home setting. These trials assessed the benefit of activity-based interventions at reducing levels of depression, and incidence of falls.111,112 Collectively, these neutral findings suggest that we may be expecting too much from this predominantly frail population with a high prevalence of dementia and depression and with low autonomy to respond to activity-based therapies (see and in relation to levels of severe disability). It seems more appropriate to seek alternative care approaches to ameliorate disability for this very inactive patient group. The established OT intervention has good evidence of efficacy when administered to patients in their own homes.34 This suggests that the barriers to success are more attributable to the care home environment and the care home population. Furthermore, the suitability and application of a patient-centred goal-setting approach in this population may require further scrutiny. Currently, patient-centred care does not have a universally accepted definition or accepted methods of how it can be measured for adherence.113
A changing role of care homes is acknowledged in a report from the Centre for Policy and Aging.16 Residents are being admitted to care homes with a higher level of dependency and more complex care needs than ever before.26 In the census of UK Bupa (The British United Provident Association Limited) care homes in 2012, 87% of residents were classed as having high-support needs and total dependency.16 Providing care for residents with high-support needs requires careful consideration of the care home environment. The therapists delivering the OTCH intervention in the participating care homes reported that the use of adaptive equipment and environmental modifications (e.g. the installation of bed levers, grab rails or raised toilet seats) was highly variable.
Future work
In order to advance the level of care currently provided to care home residents, future research should identify applicable criteria to promote an enabling environment where possible. The emphasis in this population is suggested to be about providing care for residents with high support needs for a short period towards the end of life. Further research is needed to demonstrate how the care home environment can be suitably modified to tackle these needs. A key factor in promoting an enabling environment is minimising health-related complications caused by inactivity (such as urinary incontinence and pressure sores). In addition, further work is needed in measuring health-related complications in this population in order to identify problems early.
Conclusions
The neutral results from this Phase III cluster randomised trial are deemed robust and conclusive. The trial was geared towards evaluating whether or not there is sufficient evidence to support an improved NHS provision of OT for care home residents with stroke-related disabilities. The clinical and health economic evidence presented here does not support commissioning a routine OT service in this population of care home residents. There may be benefit for residents with less severe limitations on functional activity; however, these residents were in the minority of the sampled population of stroke survivors. It appears justified to suggest that referrals of care home residents with stroke-related disabilities for OT may be of benefit on an individual basis if left in the hands of the health professional initiating the referral. OT as a service for stroke survivors still able to live in their own home has good evidence of success.34,37 However, the benefit of this style of therapy as a routine service was not evident for stroke survivors resident in UK care homes.
