Chinese Medicine Hospital Of Patients With Diabetes Hospitalization Costs Of Factor Analysis And Forecasting

Objective:The study deeply analyse the composition and factors of total TCM Hospital expenses of DM, the decision tree technology and multiple linear regression equations to predicte different kinds of TCM Hospital expenses of DM.Methods:The study base on national medical quality monitoring center. Using the hospital medical record of Diabetic patients discharged from hospital in 2008, to describe the basic situation of the patients and the stucture of hospital fee. Using a chi-square test, multiple linear regression, stepwise regression to analyse impact factors of total TCM Hospital expenses of DM and using regression and decision tree technology to classify different kinds of DM patients, and to predict the charge of different categories in hospital, finally to judge the prediction effect of this two model with the 4 aspect Average Squared Error(ASE), Loss Charts, Lift Charts, Diagnostic Charts under the SAS8.2's Enterprise Miner model.Results:Drug costs account for hospitalized patients with diabetes about 50% of the total cost, the main part of TCM Hospital expenses of DM. Regression reveals the important factors in the total cost of hospitalization followed by length of stay, hospital grade, age, treatment, areas, the same result gets from one-way ANOVA, Decision tree analysis. Decision tree devides the patients in provincial hospital into 22 categories. Each category is given the prediction interval of cost. Patients in Municipal hospitals, in county hospitals are also established models by Decision tree to classify patients and predict the cost of various types of patients off the assembly line.Conclusion:Currently, TCM Hospital expenses of DM is unreasonable. Length of stay has a big impact on the cost of hospital,we should be further explored to lower drug costs, shorten the average length of stay and improve the Hospital Information System, full use the Decision tree to predicte different types diabetic patients' cost of hospitalization.