Research And Realizing Of Cost Prediction For DRGs

China is now facing some problems in medical field,such as the “difficult and expensive in curing of disease” and the tense doctor-patient relationship and the fees of treatment of the same disease in different hospitals or different doctors are not the same.The huge difference in the cost of the disease comes from the absence of a unified path of disease treatment,all the hospitals and even all the doctors have their own treatment of the same disease.In the past decades,China has adopted a variety of medical payment system,but still can not effectively improve the problems faced.In view of the DRGs(Diagnosis Related Groups)is widely and effectively used in the world,this paper is based on the DRGs to develop the cost in medical treatment and predict the costs in medical treatment.Regulating and unifying the path of disease treatment in this way.Then improving the medical problems faced by our country.First of all,this paper uses the medical record data comes from some hospitals of Sichuan province as the data source,and uses the data analysis and data mining knowledge to carry out the necessary data preprocessing which including the data cleaning,the data reduction,the data integration and the data transformation.Making the data meet the requirements of the data analysis and data mining.Improving the accuracy and effectiveness of the research results.Secondly,this paper studies the grouping principle and the method of grouping logic of CN-DRGs to group the data prepared.Studying the CART and CHAID used to group the data,so as to form a contrast.Thirdly,this paper studies the method of formulating the medical treatment cost of the CN-DRGs.And applying the method to formulate the medical expenses.Studying the decision tree algorithm and applying to formulate the medical expenses.The testing data are used to test the rationality and accuracy of the two model.And comparing the effect of the two model in the cost forecast.At the same time,the paper puts forward the implementation scheme of DRGs payment system.Finally,using Structs2 framework technology and Java Web programming to achieve a simple cost forecasting system,which supports the application of the CN-DRGs and the decision tree DRGs to predict the medical costs.In data mining,the method of regression is used to predict the continuous value,and the decision tree is used to predict the discrete values.This paper uses decision tree to predict medical expenses which are continuous value to get the DRGs which can standardize medical expenses and standardize treatment path,so as to improve the medical care problems faced by our country.The results obtained from the decision tree DRGs are better than those predicted by the CN-DRGs.