Fall risk prediction in older adults at the emergency department: where the guidelines do not fit

Falls represent 10% of emergency department (ED) visits in older patients. Identification of those at risk for future falls is important to allow for preventive interventions. The aim of the study was to investigate the accuracy of a new algorithm, based on an adaptation of the World Falls Guidelines (WFG) for falls prevention and management, to identify patients at high risk of recurrent falls in a cohort of older patients with fall-related visits to the ED.

The FALL-ER registry is a prospective, observational, multipurpose cohort including consecutive, community-dwelling patients age ≥65 years, attending the ED of five Spanish hospitals after a fall during 52 randomly selected days between 2014 and 2015. Variables necessary to operationalise the algorithm or a proxy when necessary were recorded. The primary outcome was a new fall within 6 months after the index visit. Survival and logistic regression analyses were conducted.

The cohort included 1241 patients (median age 80 years (IQR 73-85), 69.1% female). The algorithm allowed the classification of 1039 patients (88.7%) as high risk, 39 (3.3%) as intermediate risk and 93 (7.9%) as low risk of future falls. Overall, there were 134 patients (11.4%) who met the outcome and experienced a new fall within 6 months after the index ED visit. The cumulative probability of suffering a new fall was 13% (95% CI 10.8% to 15.1%), 17% (95% CI 4.8% to 29.1%) and 8.5% (95% CI 2.4% to 14.6%) in the high-risk, intermediate-risk and low-risk groups, respectively, without significant differences between groups (log-rank=0.422). Being classified as high risk had a sensitivity of 90% (95% CI 85% to 95%) and a specificity of 11% (95% CI 10% to 13%) for new falls.

A new fall-risk screening tool, based on the WFG algorithm, had poor discriminatory capacity in our ED cohort to predict new falls within 6 months of the index fall.