Study On The Diagnosis Related Groups Of Hospitalization Expenses For Patients With Nervous System Disease

ObjectiveAnalysis of hospitalization expenses in patients with neurological diseases.To explore the influence factors of hospitalization expenses for patients with nervous system disease.Build of DRGs in nervous system diseases.And according to the DRGs,to set reasonable standard of the hospitalization cost payment,evaluation of medical resource consumption situation.Provide the basis for medical cost control,provide a reference for the implementation of the DRGs prospective payment system in China.MethodsCollected 20827 patients with diseases of the nervous system of hospital medical record home page data between 2013 and 2016 in three branch of a hospital in Henan province,mainly including sex,age,marital status,occupation,times of hospitalization,admission form,length of stay,inpatient department,disease diagnosis,discharge status,case classification,operation condition,hospitalization expenses,etc.using nonparametric test to single factor analysis of factors affecting hospitalization expenses,using BP neural network multiple factors analysis of influencing factors,then axis of hospitalization expenses for classification,using SPSS Modeler software CHAID algorithm of decision tree,build case mix model,and with the CV,ROC curve and non-parametric test method to evaluate the case mix model;According to the case mix scheme,to set up the hospitalization expenses payment standard,calculation the case mix index.ResultThe 20827 patients with nervous system disease,men(53.09%)more than women(46.91%).The results on the analysis of the composition of hospitalization expenses,show that the proportion of medicine expenses(42.09%),examination expenses(28.54%)is higher.The results on the single factor analysis of hospitalization expenses,show that sex(Z=-3.770,P<0.001),age(?2=70.625,P<0.001),marital status(?2=12.855,P<0.001),occupation(?2=140.741,P<0.001),admission form(Z=-5.954,P<0.001),length of stay(?2=10460.407,P<0.001),other diagnosis(Z=-23.541,P<0.001),discharge status(?2=128.171,P<0.001),case classification(?2=380.725,P<0.001),whether operation(Z=-31.823,P<0.001)have an effect on hospitalization expenses.Artificial neural network analysis results show that the factors affecting hospitalization expense according to the importance in sorting is: case classification(0.3746),whether operation(0.2599),other diagnosis(0.1046),discharge status(0.0632),age(0.0605),occupation(0.0518),admission form(0.0396),marital status(0.0288),sex(0.0170).According to the analysis of the artificial neural network,five important factors: whether operation,case classification,other diagnosis and admission form by decision tree model into the predictive variables,the hospital total cost as the objective variables,form 9 cases group,and the evaluation results(CV maximum of 9.69%,AUC>0.5,P<0.05)of model show that the case mix model effect is good.According to case mix make standard cost and cost limit,calculate the weights of each combination and calculate the case mix index.ConclusionAccompanied by surgery,other disease,more severe illness,emergency admissions hospital patients with nervous system disease,the higher cost of hospitalization.Through the combination of artificial neural network and CHAID algorithm to build diagnosis related group scheme is reasonable.