Fifty-four RCTs (26102 participants) were included in the review.
Multifactorial assessment and management interventions (19 trials, 7,099 participants)
Most trials in this group were rated as fair quality and used prospective methods to assess falling. Most trials enrolled patients considered at high risk of falling. The pooled result did not demonstrate superiority of the intervention: RR 0.94 (95% CI, 0.87 to 1.02). This remained the case when only the six trials with "comprehensive" interventions were included. Overall heterogeneity was high (Ι²=61.5%, p <0.001). Meta-regression models did not identify any variable that clearly accounted for the heterogeneity. There were limited data on harms of the intervention across the included trials.
Exercise or physical therapy (18 trials, 3,986 participants)
Overall, trials in this group were rated as fair quality. The pooled result showed a statistically significant reduction in falling: RR 0.87 (95% CI, 0.81 to 0.94). There was no statistically significant heterogeneity. There was no evidence of harms of the intervention across the included trials.
Single clinical treatments
There were nine trials of vitamin D supplementation (14 trials, 5,809 participants). All trials were rated as fair quality but only three assessed falls prospectively. The pooled result showed a statistically significant reduction in falling with vitamin D: RR 0.83 (95% CI, 0.77 to 0.89). There was no statistically significant heterogeneity. There was no evidence of harms of the intervention across the included trials. There were four trials of vision correction (1,437 participants). Overall, trials were rated as fair quality with all trials including high risk populations. There was no reduction in the proportion of fallers and one study identified an increase in fallers in the intervention group. There was one trial of medication assessment and withdrawal (48 participants) which found that intervention was not associated with a reduced fall rate. No harms were reported in this trial.
Home hazard modification (three trials, 2,348 participants).
The pooled result did not demonstrate superiority of the intervention: RR 0.81 (95% CI, 0.63 to 1.04) and significant heterogeneity was noted. There was no evidence of harms of the intervention across the trials.
Clinical education or behavioural counselling (one trial, 310 participants)
The trial was rated "good". There was no reduction in fall risk, no increase in falls or fallers and no additional adverse events reported.