BACKGROUND: Frailty is a multidimensional geriatric syndrome recognized as a critical public health challenge in 771 million aging population worldwide. Although electronic frailty index (eFI) is successfully adopted for frailty screening in developed countries, such a tool is still absent in China. Furthermore, for facilitate early illness prevention, China offers annual physical examinations for the elderly which offers a potential opportunity for the early detection of frailty. This study aimed to develop a new eFI algorithm leveraging routinely collected healthcare data and validated it within both the development and an independent external cohort. METHODS: Individuals aged 65 or older from the development and external validation cohort were enrolled in this study. Data were extracted from the annual physical examinations and medical records. Based on the cumulative deficit model, a tailored eFI calculation algorithm was developed. The eFI's validity was assessed through correlation with the established FRAIL scale, and its predictive utility for hospitalization and mortality was prospectively evaluated. RESULTS: A set of 30 variables across 13 functional domains was selected to calculate the eFI. It demonstrated a strong correlation with the FRAIL scale (P <
0.001). In the development cohort, individuals categorized as prefrail and frail had higher (62% and 137% respectively) risk of hospitalization compared to the robust group. Regarding all-cause mortality, the risk was also higher (59% and 117% respectively) for prefrail and frail participants. Similar associations were observed in the external validation cohort. CONCLUSION: Utilizing standardized healthcare records, this study successfully developed and validated an eFI algorithm that can offer a reliable and scalable tool for early frailty screening in China and populations with similar preventive physical examination data.