Decoding the Anxiety Burden in Asia and Developing an XGBoost-Based Online Screening Tool Using Korean Data.

Abstract

Anxiety disorders impose a substantial burden across Asia, yet comprehensive regional surveillance remains limited. In addition, the utility of the Health-Related Quality of Life Instrument with 8 Items (HINT-8) for anxiety screening in the Korean population has not been fully clarified. This study integrated data from Global Burden of Disease (GBD) 2021 and Korea National Health and Nutrition Examination Survey (KNHANES) to comprehensively assess the prevalence, incidence, and disability-adjusted life year (DALY) rates of Anxiety disorders across 48 Asian countries and territories from 1990 to 2021. Joinpoint regression, age-period-cohort modelling, decomposition analysis, and health inequality indices were employed to analyse trends and driving factors. In the Korean dataset, eight machine learning algorithms were compared. SHAP analysis was used to interpret the results, and the best-performing model was ultimately developed into an online prediction model. In 2021, Asia accounted for over half of global anxiety cases, and age-standardized rates increased by 14.6%-17.1% since 1990. After 2019, incidence (APC = 9.532%) and DALY rates (APC = 9.157%) rose sharply. Burden patterns varied by sex and age, and population growth was a major driver. In the Korean dataset, XGBoost showed the best performance (AUC = 0.894). An 8-feature parsimonious model performed comparably to the 17-feature model while improving interpretability and practical feasibility. Asia's growing anxiety burden demands prioritised care for women, adolescents, and middle-aged populations and better crisis mental health services. A Korean machine-learning tool shows potential but requires broader Asian validation.