Casey J. Allen, William M. Hannay, Clark R. Murray, Richard J. Straker, Mena M. Hanna, Jonathan P. Meizoso, Juliet J. Ray, Alan S. Livingstone, Carl I. Schulman, Nicholas Namias, Kenneth G. Proctor

As the population ages, mortality from falls will soon exceed that from all other forms of injury. Tremendous resources are focused on this problem, but how these patients die is unclear. To fill this gap, we tested the hypothesis that falls among the elderly are related to patient, rather than to injury factors when compared with falls among younger adults. From January 2002 to December 2012, 7,293 fall admissions were reviewed. Data are reported as mean ± SD if normally distributed or median (interquartile range) if not. In 2002 to 2007, 25% of all falls were in elderly patients (≥65 years), but in 2008 to 2012, this proportion increased to 30% (p < 0.001). When comparing adult (n = 5,216) with elderly (n = 2,077) admissions, characteristics were as follow: Injury Severity Score (ISS) of 8 (4-13) versus 9 (5-17), length of stay (in days) of 3 (1-7) versus 6 (2-11), and mortality of 3.8% versus 13.7% (all p < 0.001). After controlling for variables associated with mortality using multiple logistic regression, elderly age was the strongest independent predictor of mortality (odds ratio, 8.18; confidence interval, 4.88-13.71). When comparing adult (n = 198) with elderly (n = 285) fatalities, ground-level falls occurred in 31% versus 91%, ISS was 27 (25-41) versus 25 (16-36), and length of stay (in days) was 2 (0-6) versus 4 (1-11) (all p < 0.001). Death occurred directly from fall in 82% versus 63%, from complications in 10% versus 20%, and from a fatal event preceding the fall in 8% vs. 17% (all p < 0.001). The proportion of fall admissions in the elderly is growing in this trauma system. Elderly age is the strongest independent predictor of mortality following a fall. In those who die, death is less likely a direct effect of the fall. Epidemiologic study, level III.