Using data from a meta-analysis including 606,715 women, we found that the predictive value of a fall the past year for future hip fractures in women significantly decreases with age, resulting in a diminishing population attributable risk with increasing age.

Purpose: In a recent meta-analysis of 40 cohorts, we demonstrated that a fall history in the past year was associated with an increased risk of hip fracture. An interaction between fall history and age was observed in women, with lower hazard ratios (HR) for older women. This study aimed to determine the population-attributable risk (PAR) for hip fracture due to increased fall risk in women of different ages.

Methods: Fall history associated attributable risk (AR, %) for hip fracture was calculated [100·(1-1/relative risk (HR))] for women per age stratum, using previously calculated HRs. PAR (%) of hip fractures in the female population (≥ 50 years) that could be prevented if the fall history-mediated risk increase could be eliminated was calculated as 100·P(exp)·(HR-1)/[1 + P(exp)·(HR-1)] where P(exp) was the exposed proportion of the population (i.e. the proportion with past falls).

Results: A total of 606,715 women included from 40 cohorts, with fall risk documented in the past year (fall history yes/no or 2 or more falls) and prospective information about hip fracture and death, were analysed. The proportion of fallers increased progressively with age from 24.6% at age 50-54 years to 45.5% at age 90-94 years. In contrast, the AR due to falls decreased, from 54.8% at age 50-54 years to 8.3% at age 90-94 years, and the PAR diminished with age, from 23.9% in women 50-54 years old to 3.9% in women 90-94 years old.

Conclusions: As falls become more common with age, their predictive value for future hip fractures in women significantly decreases. This suggests that the effectiveness of fall prevention strategies in reducing hip fractures is lower in older women, who are at higher risk for serious falls and hip fractures.