Statistical Cost Allocation in Health Care Management Using Finite Mixture Models

Abstract

The profit loss statement of hospitals contains two properties of profit and policy because hospitals have the characteristics of public goods. Therefore, we need to distinguish and allocate costs between health care policy and profit using segment accounting. However, cost allocation is a difficult problem and always becomes arbitrary in health care management. This paper considers cost allocation in health care management based on statistical view, specifically using finite mixture models. Since cost allocation using conventional finite mixture models is interpreted as either profit or health care policy, this paper proposes a new finite mixture model with the state of both profit and health care policy. Monte Carlo experiments show that the parameter estimates of a new finite mixture model is are unbiased but the convergence speed is quite slow.