심혈관질환 위험 기반 생물학적 연령 산출을 이용한 전체 사망 및 의료비 예측: 건강보험공단 맞춤형 자료 분석 = Estimation of Biological Age Based on Cardiovascular Risk and Its Predictive Validity for All-Cause Mortality and Healthcare Costs: Evidence from National Health Screening Data

다국어 초록 (Multilingual Abstract)

Background: Chronological age (CA) has limitations in predicting individual health risks and healthcare expenditures. Biological age (BA), estimated from health status indicators, has gained attention as an alternative. This study calculated BA based on the predicted risk of cardiovascular disease (CVD) hospitalization using health screening data and defined the difference between BA and CA (BA–CA) as accelerated aging (AA). The study aimed to evaluate whether BA has equal or better predictive power than CA for all-cause mortality and healthcare expenditure, and to quantify medical cost differences according to AA level through simulation analysis. Methods: We analyzed a cohort of approximately 11.7 million Korean adults aged 40 and over who underwent health check-ups between 2012 and 2014, with follow-up until 2023. BA was derived from clinical indicators, lifestyle factors, and environmental exposure using a sex stratified Cox model estimating 10-year CVD hospitalization risk, and AA (BA–CA) was used as the main explanatory variable. Primary outcomes were all-cause mortality and annual total medical cost, and area under the receiver operating characteristic curve (AUROC) was used to compare predictive performance between BA- and CA-based models. Regression models were used to assess the association between AA and medical cost, and cost differences under varying BA levels within the same CA group were illustrated via simulation. Results: Higher BA was significantly associated with increased mortality and medical expenditure. BA-based models showed comparable AUROC values to CA-based models. A non-linear relationship was observed between AA and cost: as AA increased, medical cost rose continuously in women but plateaued or declined in men beyond a certain threshold. Simulation results estimated that among 65-year-old men, those with BA of 70 incurred Korean won 340,000 more in annual medical cost than those with BA of 60 based on the 2012 medical expense usage. Conclusion: BA derived from routine health screenings data accurately predicts individual mortality and healthcare costs, complementing traditional CA-based models. AA is a practical indicator for identifying high-risk populations, evaluating intervention effects, and simulating healthcare budget scenarios.