Effectiveness of exercise intervention on physical and health outcomes in patients admitted to an acute medical ward: A systematic review and meta-analysis.

The evidence for both positive findings was rated low certainty, and the four walking trials differed widely in their results (I² = 90%), so how large the benefit is remains unclear. Only 13 trials with 1273 participants, of fair to high quality. Trials lost quality points for not blinding therapists, assessors or participants, for not reporting exercise dose and for not monitoring what the control group did. The trials used different exercises, doses and intensities, and intensity was self-chosen and unmeasured in several, so the review cannot say which exercise prescription works. Data on sit-to-stand, tandem stance and readmissions could not be pooled because the original authors did not publish or share the numbers, leaving some outcomes answered by single small trials. Participants were almost all older adults. Only two trials included anyone under 60 and the youngest average age was 51, so this says little about younger medical inpatients. People admitted with stroke, and rehabilitation and surgical wards, were excluded. No search of unpublished or grey literature was done, so studies that were never published in a journal may have been missed. The authors suggest the lack of effect on health outcomes may reflect other differences between patients, such as how unwell they were, or discharge happening before any effect of exercise could appear.

The authors declared no conflicts of interest and received no funding for the research, authorship or publication of the article.

Do not drop balance work from a ward exercise programme on the strength of this review. Balance was measured in only one trial, using the timed up-and-go test, and that one trial found no difference between exercise and usual care. That is a gap in the evidence, not proof that balance training does not work.