The Research Of Medical Insurance Cost Control Based On Transformation Model And Mixed-effects Model

Social medical insurance is vital to every one of us. As a very important part of the social security system, It is the "stabilizer" of social economic development and the "buffer" of reform. Our country has been basically achieved "wide coverage" and solved the problem of " high cost of the medical treatment ". However, in recent years, the rapid growth in medical costs, makes some patient shard to bear the burdens. Even some health insurers have appeared the fiscal deficit. In addition, some studies have found that the growth of medical expense was faster than the GDP growth rate in some places. If we don't control it in time, it will make the level of our social security down.Now, there are five kinds of basic medical insurance payment:fee for service, global budget, capitation, service unit, diseases related to groups. In general, all kinds of ways are built around medical insurance cost control and medical service quality. They have different focus, leading to different advantages and disadvantages. Fee for service is widest-use in our country. However, there are many drawbacks about this payment. Since hospitals' income is proportional to the services provided and the high asymmetry of medical market information between doctor and patient, it is likely to lead to medical service encouraging patients to do unnecessary tests, using imported rare medicinal herbs, repeating outpatient service, and so on. It's a waste of medical treatment insurance fund and even caused conspiratorial swindling by patients and hospitals. Because of the good effect of medical expenses control, global budget has become the first selection of payment in many places.The implementation key of global budget is estimating the total budget reasonably and scientifically for every hospital. The current way of accounting total budget is based on the respective hospitals total health care costs, the per capita cost and the total number of patients. It's unstable and leading to low prediction accuracy from a view of statistics. At present, most of the researches on distribution and control of health care and costs are qualitative, and little research do a quantitative analysis through mathematical models. So in this paper, aimed at the skew-ness of the medical data, we propose the transformation model and mixed-effects model to estimate the total budget reasonably and identify the hospitals with unreasonable costs scientifically. Our model is an expand of Lin's model(Lin, et all,2015) and has two advantages:1. The transformation model is estimated by nonparametric way which doesn't request the pre-specified model.2. Our model estimating is based on the data of other hospitals with the same level in the same time. It's more scientific than the way before. We apply our model to the data of urban workers hospitalization medical expenses in 2011, Sichuan province. We get the following conclusions:1. Compared to Lin's way (Lin, et all,2015), our model can fit the data better which make the total budget estimating more reasonably.2.Through the model, we preliminarily identified 23 hospitals with unreasonable costs. It's important to the study of health care cost control.