Five RCTs (2,287 participants) were included in the review. The mean van Tulder quality score was 11.8 out of 19 points. Length of follow up ranged from 30 days to 18 months.
There were no significant differences between intervention and control groups for mortality (five RCTs; I2=0%), institutionalisation (three RCTs; I2=63%), quality of life (one RCT), cognition (one RCT) or readmission (five RCTs; I2=42%).
One RCT reported a significant improvement for functional outcomes in the intervention group compared with the control group (SMD 0.41, 95% CI 0.21 to 0.61), but it was unclear whether the difference was clinically meaningful.
There were no significant differences for sub-group analysis by intervention type. A sensitivity analysis, excluding one trial which only scored 7 points on the van Tulder scale, did not significantly alter the direction of estimates (data not presented).
There was no evidence of publication bias (data not presented).