Sex Differences in the Development of a Multidimensional Fall Risk Index for Older Adults

Introduction: Multiple factors predict falls among older adults, yet most screening tools are unidimensional and ignore sex differences despite higher fall occurrence among females. Understanding sex differences in the risks and occurrence of falls is critical to inform clinical decision-making. Using extant data from a population-based study, we developed and validated separate, fall indexes by sex based on the relationship between multiple, intrinsic risk factors for falls. Methods: Two thousand, eight hundred fifty-seven individuals (57% females) comprised the test set, and 1428 individuals comprised the validation set. Cox proportional hazards regression determined the weights for each significant risk factor in predicting incident falls, which subsequently comprised each fall index. Results: Traditional, fall risk factors were largely confirmed for females and males, except that underweight BMI and use of fall-risk inducing drugs (FRIDs) discriminated future fallers from non-fallers among females only. The multidimensional index predicted fall risk such that for every one-point increase on the index, there was a 2.5-fold increased hazard of experiencing a future fall among females (HR = 2.47, 95% CI [2.08-2.93], p <.001), and a 77% increase among males (HR = 1.77, 95% CI [1.56-2.02], p <.001). Classification accuracy (percentage of correct predictions out of all predictions) was 69% for females at one-year post-baseline and 67% for males, increasing to 74% at five years and 79% at ten years post-baseline for both females and males. Conclusions: Significant sex differences in fall risk factors were identified. Sex-specific measures may improve clinical screenings and guide tailored interventions.