Forty-nine studies were included. Study quality ranged from very low to high.
Physical exercise (25 studies):
There was no statistically significant reduction in the risk of falling in high-risk participants following untargeted exercise (six low-quality studies, n=642), targeted exercise (two high-quality studies, n=659) and combination exercise versus home untargeted exercise (two moderate-quality studies, n=138) programmes. There was evidence of substantial statistical heterogeneity for combination versus home untargeted exercise (I2=67.5%).
For the general elderly population there was a statistically significant reduction in risk of falling and of fall-related injuries following untargeted exercise programmes (RR 0.78, 95% CI 0.66 to 0.91; 12 moderate quality studies, n=2,484 and RR 0.44, 95% CI 0.27 to 0.72; two studies) and targeted exercise programmes (RR 0.81, 95% CI 0.67 to 0.98; three low-quality studies, n=566 and RR 0.67, 95% CI 0.51 to 0.89; three studies).
There was no statistically significant reduction in the risk of fall after a short (<6 months) untargeted, targeted or combination exercise programme or a long (≥six month) targeted exercise programme. However, there was a statistically significant reduction in risk of falling following a long untargeted exercise programme (RR 0.76, 95% CI 0.64 to 0.91; eight moderate quality studies, n=896) or long combination versus untargeted exercise programmes (RR 0.73, 95% CI 0.57 to 0.95; one high-quality study, n=70). There was a statistically significant reduction in risk of fall-related injuries following a long targeted exercise programme (RR 0.68, 95% CI 0.51 to 0.90; two studies) and following a long combination exercise programme (RR 0.31, 95% CI 0.13 to 0.74; one study).
Environmental modifications (four studies):
There was a statistically significant reduction in the risk of falls following environmental modification in high-risk populations (RR 0.66, 95% CI 0.54 to 0.81; three high-quality studies, n=374) and populations with both high- and low-risk participants (RR 0.85, 95% CI 0.75 to 0.97; three high-quality studies, n=1,163). However, there was no statistically significant reduction in falls in low-risk populations only (one high-quality study, n=324).
Vitamin D supplements and calcium supplements (three studies):
There was no statistically significant reduction in risk of falls in both men and women who received supplementation with vitamin D and calcium (one moderate quality study, n=445). However, there was a statistically significant reduction in the risk of falls in women who received supplementation with vitamin D and calcium (RR 0.83, 95% CI 0.73 to 0.95; two moderate quality studies, n=2,121), but no statistically significant reduction in their risk of fall-related fractures (two studies). There was evidence of substantial statistically significant heterogeneity (I2=57.7%) for fall-related fractures.
Multifactorial interventions (17 studies):
There were some discrepancies between data reported in the text and in supplementary tables (data reported here were from the main text). There was no statistically significant reduction in risk of falls (six very low-quality studies, n=2,857) and in risk of fall-related injuries in the general elderly population (two low-quality studies, n=1,656). However, there was a statistically significant reduction in fall-related injuries in high-risk elderly populations (RR 0.86, 95% CI 0.75 to 0.98; 10 low-quality studies, n=2,610), but not in fall-related injuries (five moderate quality studies, n=1,554). There was evidence of substantial statistically significant heterogeneity for risk of falls for the general elderly population (I2=61.1%) and the high-risk population (52.3%) individually and combined (I2=52.6%). There was evidence of substantial statistical heterogeneity for the risk of fall-related injuries for the general elderly population (I2=85.4%).
There was no statistically significant reduction in falls following vision assessment and referral (two moderate-quality studies, n=892), cataract surgery in women (two moderate-quality studies, n=545), men and women who received vitamin D supplementation (three high-quality studies, n=752), women only who received vitamin D supplementation (one moderate-quality study, n=137) and in women who received HRT (one moderate-quality study, n=373).
Results for medication withdrawal, gait-stabilising devices and hip protectors were reported in one study each and the results were reported in the review.