The Heterogeneous Risk and Dynamics of Out-of-pocket Healthcare Expenditures Using Multivariate Generalized Random Forest

Out-of-pocket healthcare expenditure uncertainty poses a financial risk to US households. We propose a new model for healthcare expenditure dynamics where parameters vary by household characteristics and develop the multivariate generalized random forest to extract this parameter-heterogeneity in a data-driven way. Individuals in the Health and Retirement Study display extensive heterogeneity in both healthcare expenditure dynamics and risk.  The poorest households have the largest risk of using all financial resources on healthcare implying a negative social gradient in the risk of medical bankruptcy. Higher income, wealth, and a privately provided healthcare insurance associate with massive reductions in bankruptcy risk for the high-risk group. 

• Register here