Analysis of influencing factors and equity in hospitalization expense reimbursement for mobile populations based on random forest model: a cross-sectional study from China.

Background: Continuously improving the accessibility of hospitalization expense reimbursement and reducing the medical expense burden on the migrant population are crucial objectives of China's health insurance system reform. Existing research lacks comprehensive analyses of the current status of hospitalization expense reimbursement for the migrant population, and insufficiently addresses the factors influencing reimbursement and equity. The study aims to identify the key factors influencing the hospitalization expense reimbursement for China's migrant population and to further analyze the equity of this reimbursement. Methods: Data were obtained from the 2018 China Migrants Dynamic Survey. After data cleaning, a sample of 3,186 individuals who incurred hospitalization expenses was selected for analysis. First, the current status of hospitalization expense reimbursement (occurrence, location, method, and amount) was analyzed using percentages and chi-square tests. Secondly, the random forest algorithm was applied to evaluate the importance of the factors influencing hospitalization expense reimbursement. Third, the regression analysis was used to quantify the key factors. Finally, the concentration index was utilized to assess the equity of hospitalization expense reimbursement for the migrant population and the contribution of key factors to this equity. Results: Regarding reimbursement rates, 69.83% of the migrant population chose to reimburse hospitalization expenses, while 30.17% still did not. In terms of reimbursement location, 55.69% reimbursed hospitalization expenses at their place of household registration, and 44.31% at their place of inflow. Regarding reimbursement method, 88.36% chose the Basic Medical Insurance System for Urban and Rural Residents, while 11.64% used the Basic Medical Insurance for Urban Employees. The mean of total hospitalization expenses for the migrant population was 3,058.7 (USD), with health insurance reimbursing 1,213.4 (USD) and individuals paying 1,845.3 (USD) out-of-pocket. The health insurance reimbursement ratio was 39.67%, and the out-of-pocket share was 60.33%. The results of random forest analysis identified the key factors affecting whether the reimbursement occurred as: education, health, age, income, and local insurance enrollment. Key factors affecting the level of reimbursement were: health status, insurance type, total medical expenditure, illness status, and mobility scope. Equity analysis revealed pro-rich inequity (favoring high-income groups) in both the probability and level of hospitalization expense reimbursement. Factors contributing to hospitalization cost reimbursement probability inequity, listed in descending order of impact are education (42.3%), income (34.1%), health (12.4%), age (8.2%), and enrollment location (3.0%); factors contributing to the level of hospitalization reimbursement inequity, listed in descending order of impact are health (58.12%), mobility range (21.74%), total healthcare expenditures (9.35 %), type of healthcare coverage (9.28%), and illness (1.51%). Conclusion: There is still much room for improvement in the reimbursement rate of hospitalization expenses for the migrant population. Future efforts to strengthen protection should: (1) further improve the coordination level of medical insurance to narrow the treatment differences between different regions; (2) encourage migrant populations to enroll locally (in the inflow area) and participate in the Basic Medical Insurance for Urban Employees to increase reimbursement levels; and (3) simplify reimbursement policies, optimize information dissemination channels, and enhance the policy comprehensibility and acceptance to narrow accessibility gaps.