Analyses And Applications Of Medicare Insurance Data

To avoid the occurence of disastrous household-expenditure for the medical treatments of urban and rural residents, six ministries of the Central Government of China issued "The Guidance on the Major Illness Medical Insurance for Residents both in Rural and Urban Areas" (short for the Guidance below) in August 2012. Anhui province started to implement the policy in 2013. In order to further improve and promote the policy, we need to understand exactly the implementation situation and guarantee level. Bidirectional referral plays an important role in the control of medical expenses and the use of medical resources and has a guidance and reference significance in medical system reform and the implementation of the new policy in Anhui province.The existing research works focused more on the interpretation and analysis of medical insurance policies, while less on insurance data. We cooperated with Anhui Medical Insurance Research Institution and used the seven years" claims data of urban and rural residents in a city of Anhui province to do some researches on the major illness insurance and referral conditions.Making a study on the Guidance and understand the condition and guarantee level of the major illness insurance in the pilot city.We apply data analysis to the claim data of seven years with 277756 items in the view of the annual hospitalization and hospital level to count length of stay and hospital costs and find that the actual compensation ratio of the major illness insurance in the pilot city is merely 15.59%, which is far lower than the 50% required by the Guidance. Therefore, we design lots of schemes related to the three key factors, the starting line, reimbursement ratio and directories respectively. From the analysis of these schemes, we find that the actual compensation ratio can be raised to 35.69% with the reduce of the starting line and increase of the amount of fund raised. We get the measurement indexs and calculation formulations of the guarantee level of the major illness insurance and propose a new compensation scheme that balanced all the measurement indexs.Exploring the relationship between the referral social network and length of stay and hospital costs. We extracted the referral rules and designed referral data filter algorithm, thus got the interesting referral data and built a referral social network. By choosing a proper community detection algorithm, finding the referral characters and community distribution of the referral social network, we can present a resource sharing relationship among hospitals. Combining with the SIS model, we improved a random walk algorithm for community detection. The result of the improved algorithm is quite better than other algorithms with the measure of modularity. As for the relationship, we explored the key factors of the characteristics of the network affecting the length of stay and hospital costs using linear regression analysis and decision tree classification methods. In addition, for referral sequences, we use Apriori algorithm to mine frequent sequences of referral hospitals and find some practical referral patterns.The research on the guarantee level of the major illness insurance, which has been approved by experts, has a positive effect on improving and promoting the policy. Research of referral shows that the sound referrals not only can balance the use of medical resources but also reduce the expenditure of medical insurance. The obtained research results of this thesis have been applied to a city in Anhui province for health insurance policy-making, and achieved certain results.