Ten year follow-up of a randomised controlled trial of care in a stroke rehabilitation unit

Introduction

Decreased mortality and reduced disability are well recognised short term benefits of care in a stroke unit.1 Early organised management improves survival up to five years after stroke.2 Only one study has examined the effects of care in a stroke unit for longer than five years,3 and it showed that treatment in a combined acute and rehabilitation stroke unit in Norway conferred benefit even 10 years after stroke. We aimed to examine whether the benefits of a non-acute stroke rehabilitation unit persist for 10 years after stroke. This study was a continuation of the five year follow-up by Lincoln and colleagues.2

Participants, methods, and results

We identified participants who had been randomly allocated to receive treatment in a non-acute stroke unit or on conventional wards (general medical wards or wards for the elderly) as part of an earlier trial.4 Ten years after that randomisation, we traced them on hospital and general practice databases. We asked survivors to consent to follow-up with a postal questionnaire. Participants needing help to complete the questionnaire were visited by researchers who were blind to original group allocation and to five year results for individuals.

We recorded place of residence. We used the Barthel index to measure independence in personal activities of daily living5: we classified participants as disabled (0-17) or independent (18-20). We obtained age, sex, and date of stroke from previous records. We compared survival for participants in the two groups (stroke unit and conventional ward) over 10 years using Kaplan-Meier survival curves.

In the original study, 176 participants were randomly allocated to receive treatment in a stroke unit and 139 to receive treatment on a conventional ward.4 Improvements in databases meant that more participants were identified at 10 years than at five years,2 with only 15 participants untraced (stroke unit, eight; conventional wards, seven). Eight traced participants (four in each group) refused to give consent for follow-up but were included in the survival analysis.

Survival times of participants in a stroke unit and on conventional wards. Black crosses indicate censured individuals (those for whom no information was available after five year follow-up—see text)

• Download figure
• Open in new tab
• Download powerpoint

At 10 years, 122 (69%) stroke unit participants and 111 (80%) participants who had been in conventional wards were known to have died; 31 stroke unit participants (67% of the 46 survivors) and nine conventional ward participants (43% of the 21 survivors) were known to be disabled (Barthel score < 18); and nine stroke unit participants (20% of survivors) and two conventional ward participants (10% of survivors) were known to be in institutional care.

Relative risks and confidence intervals were calculated by assuming worst case scenarios (that is, untraced participants all dead, non-consenting participants all disabled). At 10 years, the relative risks of death (0.87; 95% confidence interval 0.78 to 0.97), death or disability (0.99; 0.94 to 1.05), and death or institutional care (0.91; 0.83 to 1.00) all tended towards more favourable outcome for participants who had received care in a stroke unit. Survival was significantly greater in the stroke unit group (log rank test, 6.63, P = 0.01) (figure).

Comment

The relative risks of death, death or disability, and death or institutional care all tended towards more favourable outcome for stroke unit patients. Survival was significantly greater in the stroke unit group. Although this study was not designed to detect differences in long term survival, these findings are consistent with previous work showing that the long term benefits of stroke rehabilitation are maintained over a 10 year period.3 The reasons for this are unclear, but one explanation is that long term survival is related to early reduction in disability.