Assessment of the methodologic quality of the included studies.

*Ref is the reference number listed at the end of this paper. †Some data reported in a prior paper by Chapuy et al. in 1992 on the same study population [reference 26]. ‡Some data reported in a prior paper by Sawka et al. in 2005 [reference 23] or the primary references. λIn the hip protector study by Meyer et al. [reference 44], the primary author relayed through e-mail correspondence that hip protector compliance was not measured, but a “worst-case” estimate of 34% for hip protector use in the treatment group was reported.