Hospitalization Expenses Of Single Diseases With Multiple Factor Analysis And Model Selection

Objective: To provide methodology reference for the analysis of the hospitalizationexpense and related research, and for the administrative departments to make reasonablerelated policies, the factors that influence the hospitalization expenses of patients withdifferent characters were analyzed by multiple linear regression method and path analysismethod, and then their results will be compared with each other ones, based on theexpense structure data of senile cataract patients, cholecystolithiasis with cholecystitispatients, leiomyoma of uterus patients, cesarean section patients collected from an AAAcomprehensive hospital in Xinjiang province. Methods: The hospitalization expensestructure information of all the eligible patients with the above mentioned four diseasesfrom2008to2010were collected, given a statistic description and analyzed univariatelyin this paper. And deployed SPSS16.0and AMOS17.0perform statistic description,single factor, multiple linear regression and path analysis. Results: In general, medicineaccounts for the largest proportion among all expenses, while nursing and bed charges afairly low proportion; in addition to leiomyoma of uterus, the expense growth rate of theother3disease decreased; both multiple linear regression and path analysis results showedthat the length of hospital stay, the proportion of prescription charge and discharge yearwere the major factors that influenced hospitalization expenses; the results of path analysisrevealed that the preoperation days in hospital, the proportion of prescription charge andthe condition of additional diagnosis are the main factors that caused the hospitalizationdays extended; several adapter indices showed that the path analysis model are better thanthe results of multiple linear regression analysis. Conclusion: The path analysis can foundout the relationships among the factors, as well as their impact on the hospitalization costs.The findings of path analysis are generally consistent with those of the more commonlyused method—multiple linear regression, but with more abundant information. Path analysis should be applied to deal with similar problems in future research. Scientificstatistical findings will provide more effective and comprehensive information forconcerned administrative departments to make more reasonable policies.