Analysis Of Influencing Factors On Hospitalization Expenses In Injury Inpatients In Shenzhen During 1994 To 2003

Objective To provide counter measures for government to control medical expenses in a rational and effective way by analyzing the main factors and its mean impact value influencing hospitalization expenses of injury inpatients in Shenzhen from 1994 to 2003.Methods A retrospective study to collect medical records of injury inpatients from 39 hospitals in Shenzhen was conducted. Statistical graph and chart were used to describe the data characters, nonparametric test and linear correlation analysis were used for single factor analysis. BP neural network in one hidden layer model to fit function between influence factors and hospitalization charge with SAS8.2 Enterprise Miner. The results from BP neural network were compared with multiple linear regressions.Results The hospitalization expenses of injury inpatients in Shenzhen from 1994 to 2003 presented ascending trend, with 6684±18418 yuan per capita after discount adjusted. The treatment expense accounted major proportion with 40.63%, followed with examination expense(16.56%), operation expense(16.56%), nursing care expense(6.83%) andother cost(20.32%). By BP network analysis, the most important factors influencing hospitalization expenses of injury inpatients became "payment modality" and "injury cause", whose mean impact value showed 0.134 and 0.120.Conclusions A few recommendations were purposed to control hospitalization cost as follow: Hospitalization expenses of single disease need to be made during the reformation of medical payment system. Pay extreme attention to traffic accident injury and improve medical treatment to avoid complication. Neural network could be used to analyze the influent factors of expenses and could assimilate more complicated relationships(main effects and interactions) between inputs and outputs, better than using the traditional multiple linear regressions.