Collective Medical Decisions
Dr. Yi Junjian
Associate Professor
Department of Economics
Chinese University of Hong Kong
We analyze a model of collective medical decisions between patients and physicians, building on insights from the classical collective model for household decisions (Browning & Chiappori, 1998; Cherchye et al., 2015). We identify the model based on the assumption that the claimed expense is a “private good” for physicians, as it is paid by public insurers instead of patients. The model has three contributions. First, based on the treatment bundle, we use a machine-learning algorithm to predict the medical treatment effect on health that directly enters patient utility function, which is unobserved. Second, we apply a revealed-preference-based moment-inequality method to allow flexible unobserved heterogeneity in both patients’ and physicians’ utilities, including two types of unobserved errors—expectational errors and structural errors—in these utility functions. Third, we develop a novel and efficient moment selection method to deal with the high-dimensional treatment bundle. To estimate the model, we use a large public health insurance claim dataset in China. Estimation results show that both patients and physicians significantly contribute to medical decision-making, but their relative importance differs by disease type. Patients (physicians) play much more important roles in medical decisions for chronic (acute) diseases.