Objectives: Tertiary Referral Hospital(TRH)in China are most important hospitals that delivery medical service to most population in the country.Because they are accredited by national authorities and has largely determined the quality of medical service and health economics in the country,so it is also the important information sources of making health policies.Reforms of the supply front is an economic strategy that is promoted by our government.A deep information mining into hospitalization expense in TRH can not only help us grasp the overall structure and influence factors of hospitalization expense,but can also improve the management ability of big comprehensive hospitals,improve medical service quality;It is also crucial parameters for making reform of suppy front strategy in public hospitals.The objectives of this study is to conduct information mining on hospitalization expense of one TRH,explore the value of recursive system model in hospitalization expense analysis,to see the influence factors of hospitalization expense and to provide evidence for making reasonable health policies that can lighten the economic burden of the societyand patients.Methods: Collected the medical records of all hospitalization cases in one of tertiary referral hospital(TRH)in Guangxi with the fixed location and time.Establish medical cost database with Excel 2016.Statistic description,single factor analysis,recursive system model combining with path analysis were adopted to analyze the overall hospitalization expense,large scale hospitalization expense and expense of main diseases.Recursive system model and path analysis were completed by SPSS 18.0 and Amos22.0.Results: The average duration for patient admitted in hospital was 8 days,the average hospitalization expense was RMB 6245.01 yuan and the median age was 50.92 years old.The medical cost of patients older than 60 is the highest,accounting for 35.28%.The proportion of female patients was 52.42%,higher than male,48.58%;Hospitalization expense of per capita in male was higher than female.The expence under RMB 10,000 yuan accounted for 73.14% while the higher hospitalization expense of over RMB 50,000 yuan accounted for2.70%;The hospitalization expenditure pattern in descending order: drug fee(38.29%),materials fee(15.1%),treatment fee(14.06%),testing fee(12.64%),surgery fee(7.13%),examination fee(6.58%),other fee(5.3%),nursing fee(0.90%).The top four expenditure pattern accounted for 80.29% of the total hospitalization expenses.Regards to classification of diseases,the proportion of tumor was the largest,accounting for 20.03%,followed by circulation system disease and digestive system disease.The top three diseases in average hospitalization cost were: tumor(¥9335.88),injury and poisoning(¥8221.25),circulation system disease(¥7791.07).The main factors with big influence on hospitalization expense included length of stay,surgery,rescue,age,tumor,etc.Regards to the higher hospitalization expenses,the maininfluence factors include length of stay,rescue,respiratory system disease,surgery,etc.Conclusions: Drug fee ranked the highest cost in the inpatient expenditure,while the technical service value of medical workers is lower.Chronic non-communicable diseases such as tumor,circulatory system diseases account for a large proportion,with a high average hospitalization expense.Large hospitalization expense accounts for a low proportion,but a large expense and the material expense increases evidently.The length of stay has the largest influencing effect on hospitalization expense and shortening the average length of stay will be the key point to control the increase of expense.Patients involving in surgery,rescue,old age and tumor have a high expense in hospitalization.Recursive system model can better reflect the hierarchical structure of main influence factors and has certain advantages compared with multiple-factor regression method.It also can find the direct and indirect influence factors of hospitalization expense,so can be used for effective control of the hospitalization expense. |