Improving the rehabilitation of older people after emergency hospital admission.

PURPOSE: Older adults are at risk of functional decline during emergency hospital admissions. This review aims to understand which exercise-based interventions are effective in improving function for older adults who experience unplanned admissions.

METHODS: Database searches identified randomised control trials (RCTs) comparing exercise-based interventions with usual hospital care. The primary outcome was functional status measured by activities of daily living (ADL) scores. Secondary outcomes were length of hospital stay (LOS), mortality and readmissions. Sub-group meta-analyses were conducted on interventions delivered in-hospital only compared with interventions provided both in hospital and after discharge.

RESULTS: After reviewing 8365 studies, nine were eligible for inclusion. Seven were included in the meta-analysis. Participants from five countries had a mean age of 79 years (1602 participants). Usual care varied considerably and the interventions showed heterogeneity, with different combinations of strengthening, resistance, high-intensity or mobility exercises. There were limited descriptions of exercise intervention delivery and participant adherence. There is low-quality evidence supporting exercise interventions that have both in-hospital and post-discharge components (3 trials, SMD 0.56 (-0.02, 1.13)). Trials involving only in-hospital interventions were inconclusive for functional gains (5 trials, SMD -0.04 (-0.31, 0.22)).

CONCLUSIONS: Exercise-based rehabilitation for older patients after emergency hospitalisation improves functional ability if the intervention starts in hospital and continues after discharge. No conclusions can be made regarding the effective exercise 'dose' or content.

IMPLICATIONS: Understanding the components of exercise interventions will improve service planning and delivery. Further studies are needed to understand the effective 'dose' and content of exercise for hospitalised older adults.