Abstract
BACKGROUND: Satisfaction with healthcare security is a critical indicator of the effectiveness of health systems. Social equity and trust and the financial burden of healthcare are key socioeconomic factors that can significantly influence residents' perceptions of healthcare security. This study aims to investigate the impact of social equity and trust and medical burden on satisfaction with healthcare security and to analyze their potential interaction mechanisms. METHODS: Using data from 7,052 participants in the 2021 China General Social Survey, this study employed machine learning methods, including neural networks (NN), random forests (RF), and logistic regression (LR), to predict and classify satisfaction with healthcare security. Additionally, causal inference techniques were applied to identify the key determinants and estimate their effects on satisfaction levels, thereby uncovering the underlying causal mechanisms. RESULTS: The predictive performance of the three machine learning methods was similar (p < 0.001). In the original models, the AUCs for LR, NN, and RF were 0.549, 0.563, and 0.534, respectively. After including factors related to social equity and trust, the AUCs for LR, NN, and RF improved to 0.633, 0.638, and 0.611, respectively. Among the three ML models, medical expenses and social equity and trust were identified as the most influential factors. Further causal analysis confirmed that higher levels of social equity and trust increased satisfaction with healthcare security, while a heavier medical burden reduced it. The analysis also revealed significant marginal effects, suggesting that the impact of social equity and trust varied across different levels. CONCLUSION: This study highlights the complex relationship between social equity and trust, medical burden, and satisfaction with healthcare security, offering theoretical support for understanding residents' perceptions of healthcare security in various social contexts.