Bridging Claims‐Based and Clinical Frailty Assessment: Translation of the
mFI
‐v10 to the Clinical Frailty Scale
Chien‐Chou Su, Yu‐Tai Lo, Yi‐Ching Yang, Yu‐Huai Yu, Wei‐Chun Cheng, Wen‐Ping Lin, Deng‐Chi Yang ABSTRACT
Background
While the Clinical Frailty Scale (CFS) is intuitive for clinical use, its reliance on clinician interviews limits its feasibility for large‐scale population monitoring. The claims‐based multimorbidity frailty index‐version 10 (mFI‐v10) is highly scalable but has not been directly linked to the CFS standard. This study aimed to develop a model to translate mFI‐v10 scores into clinically interpretable CFS categories.
Methods
In this study, 1038 individuals aged ≥ 65 years were recruited from a tertiary medical center between January 2020 and December 2021. They underwent assessments including the CFS, cognitive function, mood, physical health, and quality of life. The mFI‐v10 score was calculated from electronic medical records. Logistic regression with 5‐fold cross‐validation was used to develop a prediction model for moderate‐to‐severe frailty, adjusting for age and sex. Model performance was evaluated using accuracy, sensitivity, specificity, precision, the F1‐score, and the area under the receiver operating characteristic curve (AUC‐ROC).
Results
Among the 1038 participants, 331 (32%) had moderate‐to‐severe frailty. The mean age of the individuals with moderate‐to‐severe frailty was 83.31 years (SD, 7.32), and 40% were male. The mFI‐v10 showed moderate positive correlations with the CFS ( ρ = 0.42) and Charlson Comorbidity Index ( ρ = 0.57), and a negative correlation with ADL functional independence ( ρ = −0.46). The prediction model incorporating mFI‐v10 score, age, and sex demonstrated good discriminative ability with a mean AUC‐ROC of 76.5% (range: 70.7%–82.6%).
Conclusions
The mFI‐v10 is a valid, scalable surrogate for clinical frailty assessment. This translation framework allows efficient population‐level case‐finding, prioritizing high‐risk older adults for targeted clinical intervention.