Commentary

Stroke is a major cause of long term disability. Much research has been done to determine the best way of delivering care after a stroke. The studies by Anderson et al and Mayo et al add to our understanding of the effectiveness of programmes designed to facilitate early discharge and deliver rehabilitation in the home. In both studies, this was achieved by coordinated, multidisciplinary care provided in the patients' homes. Both studies were done in urban settings.

The studies were well designed, with random allocation of participants and outcomes assessed by people who were unaware of the treatment patients had received. Similar to previous research, these studies showed that early discharge schemes can be used to reduce length of hospital stay after a stroke.¹ An accompanying paper by Anderson et al included details of costs and resource use,² and indicated savings. Other limited evidence suggests that early discharge is less expensive,¹ but generalisation of these data is difficult as the cost of providing resources varies across settings.

The study by Anderson et al found no improvement in outcomes at 6 months for early discharge. In this study, 66% of control patients received inpatient rehabilitation in a stroke unit. Stroke units are more effective than care in general wards.³ Hospital stays for the control group were on average 15 days longer than for the early discharge group, suggesting that some patients in this group spent a substantial period of time receiving rehabilitation.

The study by Mayo et al showed some improvement in the physical health of the intervention group at 3 months. Only 12.5% of the control group received care in a rehabilitation unit, although a stroke care team was available in 4 of the 5 centres. Time for inpatient rehabilitation was less than in the Anderson et al study. Duration of acute stay was on average only 3 days longer for those not transferred to rehabilitation (8 days including rehabilitation unit care).

The studies by Anderson et al and Mayo et al were large enough to detect some important differences in patient condition caused by the early discharge programme, but they were too small to conclude that statistically non-significant results equate to no difference in modest, but clinically important, effects. For example, the findings of Anderson et al are consistent with a true reduction in the rate of falls of 19% or an increase of 10%. Neither study reported trends for increased adverse outcomes, which suggests that it is unlikely that patients were harmed by early discharge.

The results of both studies suggest that early discharge programmes that involve coordinated, multidisciplinary rehabilitation are feasible and at least as effective as routine discharge and follow up care. Future studies with larger samples and pooled analysis of existing data will be needed to compare the effects of the early discharge programme with usual care on substantive outcomes such as falls, dependency, and hospital readmission.

The patients studied are not representative of all patients with stroke. The study by Anderson et al included only 22% of the population of patients with stroke. Both studies required patients to be relatively independent. The study by Mayo et al required that patients had a caregiver, and the study by Anderson et al required that the home environment be suitable for home based rehabilitation. These findings may not therefore be applied to stroke patients with greater degrees of disability or other variation in characteristics. Neither study allows full consideration of the effects on family caregivers. Although weak evidence from Anderson et al suggests that caregivers may experience additional mental strain, which is consistent with findings from studies of other hospital at home schemes,⁴ the data were based on a small subset of the total study population.

Clinicians who are considering implementation of an early stroke discharge programme need to examine carefully their own situation. A reduction in length of stay can be achieved without apparent harm for relatively independent patients with stroke. Weak evidence suggests that early discharge is better and less expensive than usual non-stroke unit care, and the effects on caregivers need further investigation. There is, however, no evidence that early discharge is better or less expensive than stroke units. Providing some form of coordinated, multidisciplinary rehabilitation after stroke improves patient outcome. Further research is needed to determine how best to deliver this care.