Analysis On Hospitalization Expenses And Its Influencing Factors Of Patients With Gastric Cancer In A 3A Hospital Of Gan Su

Objective The rising medical cost has become a common problem for all countries in the world.Although the problems of “Difficulty and High Cost of Getting Medical Service” have alleviated because of the continuous reform of the health system,the allocation of health resources and hospitalization costs still remain unreasonable.The data of hospitalization costs of gastric cancer patients in a 3A hospital of Gansu province from 2007 to 2016 is analyzed to describe the basic characteristics,the internal composition and the potential influencing factors using the grey relational grade,degree of structure variation,linear regression and BP neural network model.The results and conclusion could be used to optimize the allocation of health resources,control the growth of hospitalization costs,and reduce the burden of diseases.Methods The data of 13,398 gastric cancer patients in a 3A hospital of Gansu province from January 1,2007 to December 31,2016 were included.The grey relational grade and degree of structure variation were calculated.Wilcoxon rank test and multiple linear regression was used for univariate analysis and multivariate analysis respectively.The results from multiple linear regression were compared with that of BP neural network model.Statistical analyses were performed using Excel 2010 and SPSS version 20.0.Results Among the 13,398 patients,there are more male than female with a male to female ratio of 3.29:1.98.81 % of the patients are married,and there are few patients who were single,divorced,and widowed.The patients belonging to the new rural cooperative medical system and the urban employee medical insurance account for 47.79% and 23.51%,respectively.The length of hospitalization is between 1 and 97 days with the median of 6 days.The number of patients diagnosed with gastric cancer increases year by year,and the total costs of hospitalization continue to be at a high level from 2007 to 2016.Drug costs account for 63.17% of the total hospitalization costs,followed by examination costs(10.54%)and laboratory costs(10.14%).The grey relational grade shows that drug costs are highest associated with hospitalization costs,followed by examination costs and laboratory costs.The degree of structure variation shows that material fees,laboratory costs,examination costs,and drug fees are dominant in all hospitalization costs(66.24%).The univariate linear regression shows that gender,age,marital status,medical insurance type,year of admission,length of stay,first admission or not,surgery,whether more than one medical diagnosis and whether the patient received blood transfusion are associated with hospitalization costs.Multivariate linear regression and BP neural network model show that the length of stay,surgery,and the year of admission are the main influencing factors.The BP neural network model is superior to multivariate linear regression analysis in analyzing the influencing factors of hospitalization costs.Conclusions The hospitalization costs in our study are very high,especially for farmers.The internal composition of hospitalization costs is unreasonable.The proportion of drug costs is very high which cover up the work of medical workers.Length of stay,surgery,year of admission,gender,and medical insurance categories are the main factors influencing hospitalization costs of gastric cancer patients.The key of lowing the total cost of hospitalization is to reduce the hospitalization days and the proportion of drug costs.BP neural network model is more suitable for the analysis and prediction of influencing factors of hospitalization costs compared with multiple linear regression analysis.