From Code to Care and Digital Detection of Frailty: A Scoping Review of electronic Frailty Index (eFI) Applications Using Electronic Health Records

The electronic Frailty Index (eFI) is a cumulative deficit model that quantifies frailty by summing health-related deficits recorded in a patient’s electronic health record. This scoping review evaluated the emerging applications of eFI models, providing an overview of the current landscape. A comprehensive search of three databases identified peer-reviewed studies published between August 2020 and July 2025. Study characteristics, including purpose, sample characteristics, eFI model details, cut-off thresholds and outcome measures were extracted. A total of 80 studies were included. Most were conducted in the UK (n = 29, 36.2%), followed by the US (n = 23, 28.7%) and Australia (n = 6, 7.5%). The most common purpose was risk stratification (n = 40, 50%), followed by model development and validation (n = 26, 32.5%). The predominant model included 36 deficits (n = 35, 46.1%), followed by 54 deficits (n = 10, 13.2%) and 31 deficits (n = 8, 10.5%). By type, the Clegg-based eFI (36 deficits) was the most frequently used (42.5%), while the Pajewski-based eFI (45-59 deficits) accounted for 21.3% and the Veterans Affairs Frailty Index (VA-FI) or adaptations represented 7.5%. Among cut-off schemes, the four-level classification (Fit: 0-0.12, Mild: >0.12-0.24, Moderate: >0.24-0.36, Severe: >0.36) was most common (24 studies, 30%), followed by the simplified three-level scheme (14 studies, 18%). Mortality was the most frequently evaluated outcome (65%), followed by hospital-related utilization such as admissions, readmissions, length of stay and emergency visits (48%). These findings highlight the widespread adoption of eFI models for risk stratification and validation, with notable variability in deficit selection and model adaptation across settings.