Evaluating Frailty Index Integrity: Insights from an International Network Study

Abstract Deficit-accumulation Frailty Indices (FIs) have rapidly been integrated into health systems worldwide to quantify the state of reduced physiologic reserve to recover from a health insult in the aging population. Common data models (CDMs) have further enabled international, interinstitutional applications of FIs. However, the validity of such applications is unknown. We conducted an international network study comparing estimates of frailty from two electronic health record (EHR)-based FIs: United States (US)-based Veterans Affairs FI (VAFI) and United Kingdom (UK)-based electronic FI (eFI) across 5 research and clinical databases. US: All of Us (n=159,721), IQVIA Pharmetrics+ (n=5,099,557); UK: IMRD-THIN (n=3,036,003), IMRD-EMIS (n=832,455), and BioBank (n=207,202). In US databases, VAFI identified higher proportions of frailty ([VAFI; eFI] All of US : 10.3%; 2.5%, Pharmetrics+: 9.6%; 2%) while in UK databases, the eFI identified higher proportions (IMRD-THIN: