To compare the clinical effectiveness of rehabilitation programmes with medical care.

Index Medicus/MEDLINE and Nursing and Allied Health Abstracts were searched from 1974 to 1994 using the following terms: 'rehabilitation'; 'physical medicine'; 'activities of daily living' (ADL) and 'functional ability'. Bibliographies of meta-analyses, reviews and retrieved articles were also searched.

Study designs of evaluations included in the review

Randomised controlled trials (RCTs) and non-equivalent control group designs. In some trials, subjects acted as their own controls, whereby comparison treatment was received at a later date. Treatment duration ranged from 14 to 84.8 days. The trial had to prospectively compare at least one treatment group (i.e. inpatient rehabilitation programme) with an untreated control group (i.e. standard medical care), and report the results such that a quantitative analysis could be undertaken.

Specific interventions included in the review

Multidisciplinary inpatient physical rehabilitation programmes versus standard medical care for physically disabling disorders.

Participants included in the review

Patients with a physically disabling diagnosis. The included studies assessed stroke and geriatric patients. The average age of study subjects was 74.9 (SD=2.97).

Outcomes assessed in the review

Survival, functional ability and discharge location were assessed.

How were decisions on the relevance of primary studies made?

The authors do not state how the papers were selected for the review, or how many of the authors performed the selection.

The authors do not report a method for assessing validity. It is stated that studies were coded according to patient characteristics study design, statistical techniques and study retrieval characteristics. However, the coding system is not explained.

Information on the patients' characteristics, study details, outcomes and retrieval characteristics were abstracted independently by three raters.

How were the studies combined?

The Mantel-Haenszel procedure was used to test survival and discharge location (dichotomous data), and combined probability tests (the Stouffer and the Fisher combined test) were used to test functional ability (continuous data).

How were differences between studies investigated?

Tests for heterogeneity of the studies were calculated.

Eleven studies (2,183 patients) were identified, 9 were RCTs and 2 were non-equivalent control group designs. Three studies (378 patients) evaluated rehabilitation for geriatric patients, the other 8 (1,805 patients) evaluated rehabilitation for stroke patients. Not all studies are included in the meta-analyses.

Survival:There was a significant difference in the odds of survival at discharge in favour of the rehabilitation group compared to the control group (P<0.01). The overall odds ratio was 1.46 (99% CI: 1.13, 1.78). There was no significant difference between groups at follow-up (calculated as the odds of survival between discharge and follow-up). Tests for heterogeneity indicated a constant treatment effect at both time periods.

Discharge location:significantly more patients from rehabilitation programmes return to their home at discharge (P<0.001) and remain there during follow-up (P<0.001). The odds ratio was 2.08% (99% CI: 1.81, 2.35) at discharge and 1.87 (99% CI: 1.57, 2.17) at follow-up. Tests for heterogeneity indicated a constant treatment effect at both time periods.

Functional ability:rehabilitation patients had significantly higher functional ability at discharge as rated by both the Stouffer (P<0.01) and Fisher (P<0.001) combined test. The results at follow-up were not significant. Tests for homogeneity were not significant.

To estimate the effect of publication bias, the 'fail-safe N' was calculated, this is the number of unpublished studies confirming the null hypothesis required to offset the results obtained. For survival at discharge, the number of studies required is 13, and, for functional ability at discharge, the number of studies required is 60.

Rehabilitation services are effective in improving short-term survival, functional ability and the most independent discharge location. Rehabilitation patients are more likely to remain in their home during the first year after hospital stay, however, the lack of long-term benefits of short-term rehabilitation may suggest that therapy should be extended to home or subacute care settings rather than being discontinued at discharge.

Not enough detail is given about the specific interventions in the primary studies, or the techniques for assessing the validity of studies. The trials included in each of the various analyses are not listed. It is not clear why the publication bias test was not reported for discharge location, as well as survival and functional ability at discharge. The literature search was restricted to English language articles, and no handsearching is reported.

Evans RL, Connis RT, Haselkorn JK. Evaluating rehabilitation medicine: effects on survival, function, and home care. Home Health Care Services Quarterly 1997;16(3):35-53.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.