Predictive validity of the Hospital Frailty Risk Score on health service utilization among older adults in New Brunswick

Abstract

The Hospital Frailty Risk Score (HFRS) is a claims-based frailty algorithm designed to identify older adults at risk of adverse outcomes, but its performance in the Canadian healthcare context remains understudied. This study evaluated the predictive validity of the HFRS for adverse health service outcomes among adults aged 65 years and over in New Brunswick using linked administrative data from 2014-2018. Three-year intervals were constructed with a two-year exposure period to calculate HFRS and a one-year follow-up to assess outcomes. Adjusted logistic regression models examined associations between frailty risk (low, medium, high) and cumulative acute length of stay, alternate level of care (ALC) stays, and mortality. Among 57,730 observations, a clear dose–response relationship was observed, with progressively higher odds of prolonged hospitalization, extended ALC stays, and mortality across frailty categories. This study demonstrates the predictive validity of the HFRS for longer-term hospital utilization and ALC outcomes, supporting its use for health system planning and resource allocation.