Does delirium prevention reduce risk of in‐patient falls among older adults? A systematic review and trial sequential meta‐analysis
Steven He, Kaye Rolls, Katrina Stott, Rozina Shekhar, Vaulina Vueti, Kelli Flowers, Margaret Moseley, Bernadette Shepherd, Mandana Mayahi‐Neysi, Briony Chasle, Bradley Warner, Danielle Ní Chróinín, Steven A. Frost
Abstract
OBJECTIVES: To determine whether delirium prevention interventions reduce the risk of falls among older hospitalised patients. METHODS: A systematic search of health-care databases was undertaken. Given the frequency of small sample sized trials, a trial sequential meta-analysis was conducted to present estimate summary effects to date. A Bayesian approach was used to estimate the posterior probability of the delirium prevention interventions reducing falls risk by various clinically relevant levels. RESULTS: Five randomised controlled trials were included in our final meta-analysis. There was a 43% reduction in the risk of falls among participants in the delirium prevention intervention arm, compared to the control; however, confidence intervals were wide (RE RR = 0.57, 95% CI 0.32; 1.00, p = 0.05). This result was found to be statistically significant, according to traditional significance levels (z > 1.96) and the more conservative trial sequential analysis monitoring boundaries. The posterior probabilities of the delirium prevention intervention reducing the risk of falls by 10%, 20% and 30% were 0.86, 0.63 and 0.29 respectively. CONCLUSIONS: The results of this systematic review and trial sequential meta-analysis suggest that delirium prevention trials may reduce the risk of in-hospital falls among older patients by 43%. However, despite significant risk reduction found upon meta-analysis, the variation among study populations and intervention components raised questions around its application in clinical practice. Further research is required to investigate what the necessary components of a multifactorial intervention are to reduce both delirium and fall incidence among older adult in-patients.