Exploring different models of stroke unit care and outcome : the Stroke Rehabilitation Outcome (SRO) study

Abstract

Introduction: Stroke is a significant cardiovascular event requiring sub-acute rehabilitation, best provided in a stroke unit (SU). These units include dedicated neurological SUs usually catering only for patients with stroke and more generic SUs existing within geriatric rehabilitation units (GRUs). There exists a "grey" group of survivors of stroke whose allocation to one type of rehabilitation facility over another is arbitrary, in that the referring physician had no evidence to suggest advantages of SU versus GRU rehabilitation.Objectives: The aim of this inception cohort study was to provide a direct comparison of quality of life and functional outcome between two commonly applied models of organised multidisciplinary SU rehabilitation for the "grey" group of stroke survivors. Further, it evaluated differences in the intensity of treatment and the environment in which rehabilitation was implemented.Method:All patients presenting to Royal Perth Hospital-Wellington Street Campus acute stroke unit (RPH-WSC ASU) with a diagnosis of recent stroke requiring hospitalisation and subacute rehabilitation were considered for inclusion into the study. Patients were selected based on their age, absence of dementia and their acceptance by incumbent medical staff for rehabilitation transfer at either Royal Perth Hospital-Shenton Park Campus stroke unit (SPC SU) or at a GRU geographically closest to their home (located at either Mercy hospital, Bentley hospital or Swan health campus).Baseline data was collected in order to establish the underlying level of disability and compare groups for comparability, and also to be used as covariates in data analysis.All treatments received were those considered standard for the individual facility, administered as usual by registered health professionals. During the study, periodic behaviour mapping at each of the study facilities was undertaken by a research assistant in order to quantify differences in the rehabilitation environment. In addition, attending therapists at each facility recorded the frequency and duration of their intervention with individual patients involved in the study in a patient diary designed for that purpose. Six and twelve months following their transfer from RPH-WSC ASU, patients attended follow-up outpatient appointments at neutral rooms where objective and subjective assessments were undertaken by an independent assessor (a physiotherapist) who was blinded as to which rehabilitation facility the patient had attended. The primary outcome measure was the MOS 36-Item Short Form Health Survey (SF-36) and secondary outcome measures included the Functional Independence Measure (FIM) and other functional measures.Results: Between July 2004 and June 2007, 354 patients with stroke were age appropriate (60 years of age or older) for recruitment into the study and of these, 94 consented to participate (SPC SU n=22; GRUs n=72). Patients referred to SPC SU were younger, more likely to be male, and have speech abnormality, peripheral vascular disease and diabetes than those referred to GRUs. Otherwise there were no significant differences between groups in any of the characteristics measured at baseline. Rehabilitation data demonstrated a significant difference in both the total allied health professional (AHP) therapy time (p<0.001) and the indirect support time such as telephone calls and meetings with family (p=0.022), with SPC SU therapists utilising more time compared with GRU therapists. There was no significant difference in time spent undertaking administration including writing notes and reports (p=0.957). Data showed significant difference in length of stay (LOS), whereby patients spent a longer time at SPC SU (p=0.036), however there was no significant difference in discharge destination between facilities (p=0.312). Of the 10 unadjusted patient measures in this study, there were significant differences between groups in only two, the Berg balance score and the Chedoke McMaster posture inventory. The differences in both of these secondary outcomes favoured the SPC SU group. In addition there were differences in the SF36 Mental component summary (MCS) and Physical component summary (PCS) scores that approached significance. The difference in the PCS scores also favoured the SPC SU group but for the MCS score it was the GRU group that had more favourable scores.As the study was not randomized, age and gender, which differed between groups at baseline, and Barthel Index score, known to be associated with length of stay in stroke patients, were added to the models as covariates. As data from 6 and 12 month follow-ups was included in the dependent variable, "visit" was added to the models. After these adjustments there were no significant differences between facilities in any quality of life or functional outcomes.Discussion:Overall there was relatively high quality of life, and low anxiety and depression reported and results were not influenced by where rehabilitation took place. Selection criteria excluding dementia and young age may in part explain this, as both have been found to predict worse quality of life outcome in stroke. Significant differences in both where patients were, and what they were doing throughout the day reflected different ethos between facilities in the way rehabilitation was delivered. However, there was no difference in functional outcome despite these environmental differences and the fact that patients experienced more intensive treatment over a more prolonged hospital stay at SPC SU.Conclusion:In most cases, rehabilitation of this "grey" subgroup of the wider population of stroke may be more cost-effective if carried out at GRUs (with higher patient/ staff ratios, less intensive treatment and shorter LOS) rather than the neurological SUs.