Robust Causal Inference and Policy Effect Evaluation and Optimization Based on Large-Scale Medical Insurance Big Data

Abstract

With the continuous improvement of the medical security system, large-scale medical insurance big data has become a core resource for insights into the operation of medical services and evaluation of policy implementation effects. However, medical insurance data is plagued by issues such as noise interference and sample selection bias. Traditional causal inference methods struggle to accurately reveal the causal relationship between policies and their effects, hindering the scientific optimization of medical insurance policies. Taking large-scale medical insurance big data as the research object, this paper constructs a causal inference framework of “data preprocessing - model selection - robustness verification”, integrating traditional and machine learning methods such as Propensity Score Matching (PSM) and random forest. Combined with a specific case of DRG/DIP payment method reform in a certain city, it conducts effect evaluation from dimensions including medical costs, service utilization, and fund sustainability, and proposes policy optimization strategies such as payment method reform, fund supervision enhancement, and balanced resource allocation based on the evaluation results. The research findings indicate that robust causal inference can effectively improve the accuracy of medical insurance policy effect evaluation, providing reliable data support for the formulation and adjustment of medical insurance policies and contributing to the further improvement of the medical security system.