Adding the Short Physical Performance Battery to a Fall Risk Score for Time-Dependent Prediction of Falls in Older Inpatients: A Retrospective Observational Study

AIM: To evaluate whether incorporating the short physical performance battery (SPPB) assessed at hospital admission into a fall risk assessment score improves time-dependent predictive performance for falls among older inpatients. METHODS: This retrospective cohort study was conducted at a single center. The SPPB was assessed within 2 days of admission. Fine-Gray subdistribution hazard models were used to examine associations with falls during hospitalization, treating discharge as a competing event. Model discrimination was evaluated using time-dependent area under the curve (AUC) at 20, 40, and 60 days. Differences in AUC (ΔAUC) were estimated using bootstrap resampling. RESULTS: A total of 428 older inpatients were included. Within 60 days, 80 patients (18.7%) experienced at least one fall. Lower SPPB scores were significantly associated with increased fall risk (SHR: 0.739, 95% CI: 0.682-0.800). The model incorporating the total SPPB score showed higher AUC values at 20, 40, and 60 days (0.761, 0.769, and 0.776, respectively) compared with the baseline model. The largest improvement in discrimination was observed for the model including balance at 20 days (ΔAUC: 0.124, 95% CI: 0.107-0.146). CONCLUSIONS: Incorporating the SPPB assessed at admission into the fall risk assessment score was associated with improved time-dependent predictive performance for falls during hospitalization. The SPPB may provide incremental value for identifying patients at increased risk of falls and supporting targeted fall prevention strategies.