Predicting Hospital Related Adverse Events: Interactions Between Frailty and Patient Characteristics in Acutely Admitted Older Adults

Background Frailty is an age-related syndrome that increases vulnerability to adverse outcomes. Hospital-related adverse events (AEs) are complications not directly caused by patients’ pre-existing conditions. While age has been widely studied, less is known about the effect of frailty and its interaction with patients’ characteristics on the risk of having at least one adverse event among acutely admitted older adults. Methods Poisson regression models were used to estimate Relative Risk (RRs) and 95% Confidence Intervals (CIs) for the association between frailty and the likelihood of experiencing at least one AE during admission (p < 0.01). Frailty was first modelled as the primary predictor, followed by individual models assessing crude associations for age, gender, ethnicity, Emergency Department (ED) wait time, and In-patient (IP) Length of Stay (LOS). Interaction terms between frailty and each characteristic were tested using Likelihood Ratio Test. Backward stepwise elimination was used to obtain a multivariable model retaining only variables that significantly improved model fit. Results A total of 158,470 hospital admissions with a recorded frailty score were included. Statistically significant interactions were found between frailty and age, ED wait time, and IP LOS (all p < 0.01). Multivariable modelling showed that the interaction between frailty and IP LOS was the only interaction that significantly improved model fit. Conclusion The risk of AEs increased with frailty, and this effect was most strongly influenced by IP LOS. Early frailty identification and targeted interventions for frail patients, especially those with extended admissions, may help reduce the risk of AE and harm. Clinical trial number: Not applicable.