Predicting First Time Falls: Validating a Novel Algorithm in Long Term Care

To determine the predictive validity of the 1^stFall algorithm in long-term care (LTC) residents across four Canadian provinces. This retrospective cohort study included all clients admitted to LTC between 2006-2017 with no history of falls in the past 30 days. The outcome was occurrence of a fall and logistic regression analysis was performed to assess predictive validity. A total of 199,997 LTC residents were studied (71% were >80 years old, 66% women, and 17% had severe cognitive impairment). For the total sample, clients in the 2^nd, 3^rd, 4^th and 5^th risk categories had 1.15, 1.58, 2.66, and 3.76 times greater odds of falling than the 1^st category, respectively. Similar trends were observed across provinces. 1^stFall was developed to predict the risk of a first-time fall event in individuals with no history of a recent fall. 1^stFall identified LTC residents at risk of a first-time fall, supporting its use in routine care.