Abstract
OBJECTIVES: Aging anxiety is not only a health issue but also a stress response to the structural risk of insufficient medical resources. This study aims to reveal the impact of aging anxiety on individual healthcare utilization and the complex psychosocial mechanisms behind it. METHODS: Based on large-scale data from the 2021 Chinese General Social Survey (CGSS), the study employs a Double Machine Learning (DML) method to build a causal inference model. The random forest algorithm is used to estimate the marginal effect of aging anxiety on healthcare utilization. The robustness checks and placebo tests are conducted to further verify the model's stability and validity. Finally, heterogeneity analysis explored the differential impact of independent variables across groups by age, education, household health status and kid number. RESULTS: Aging anxiety has a significant positive effect on healthcare utilization (β = 0.110, t = 4.895). It mediates through multiple pathways including healthcare accessibility anxiety (β = 0.344, t = 16.904), affordability anxiety (β = 0.384, t = 19.845), physical deterioration (β = 0.160, t = 7.286), psychological pessimism (β = 0.175, t = 7.819), sleep disorder (β = 0.104, t = 6.124), and self-efficacy loss (β = 0.160, t = 5.595). Heterogeneity analysis shows significant differences in this effect across groups with different socio-demographic characteristics and health statuses, reflecting variations in medical demand and anxiety responses among populations. CONCLUSION: To alleviate anxiety related to medical resource shortage and promote healthy aging, a multidimensional response system is needed. This includes improving medical insurance, advancing primary healthcare management, enhancing health literacy, and building family-community support networks. Policy design should emphasize the synergy between psychosocial factors and institutional frameworks, providing theoretical and empirical support for equitable, inclusive healthcare utilization and sustainable health development.