Research On The Diagnosis Related Group And Performance Evaluation Of Ophthalmology Department Of A Third Grade Hospital In Liaoning Province

Abjective: To study the grouping process of the CHS-DRG grouping scheme,explore the regional subdivision group scheme of DRGs and how to optimize it,use DRGs as a reference for evaluating local medical conditions,and clarify that DRGs are the local medical performance system.The changes that have come will provide reference and ideas in the local health system.Methods: Relying on the national version of the CHS-DRG grouping scheme,divide the home page of medical records by ADRG group;using statistical basis such as CV value and RIV value,further subdivision groups were selected from the ADRG group.The secondary diagnosis of the subdivided target group is merged and screened,the secondary diagnosis is modeled by multiple linear regression models,determine the weight of secondary diagnosis through P value and standardized regression coefficient,and the CCI of each patient is calculated;Make a single-factor regression model for various data on the front page of the medical record and the total hospitalization cost to determine the factors that affect the total hospitalization cost;The independent variable is the factor that affects the total hospitalization cost,and the dependent variable is the total hospitalization cost to establish a multiple linear regression model and new type The data mining model compares the results to select the final subdivision group model variables;the subdivision group model uses CHAID,E-CHAID,and CART three models,the total hospitalization cost is the output variable,and the CCI value is the first input variable,The other selected types are the secondary input variables to establish a model,the optimal solution of the subdivision group is obtained by clipping the leaf nodes of the selected model..According to the ADRG grouping in the CHS-DRG grouping scheme,the comprehensive medical performance evaluation of the ophthalmology department of the hospital is carried out.The evaluation indicators adopt existing DRGs’ own indicators.Evaluate the consumption of medical resources in each department.Results: In the results of CHS-DRG classification,the MDCC of the hospital lacked CR1(malignant ocular tumors and borderline tumors)group,and the other 15 groups except for CT1(non-surgical treatment of anterior chamber hemorrhage and ocular trauma)group had CV during hospital stay The value is greater than 0.8,and the others are less than 0.8,indicates that the homogeneity within the group is very strong,but the RIV value is 33.25%,indicating that the heterogeneity between the groups is weak;comprehensively consider the ADRG Indices,the CB3(crystal surgery)group is selected as the subdivision object.A total of 9 types of weights in the secondary diagnosis are included in the CCI calculation.There are a total of 8independent variables in the univariate analysis results into the next two models.After comparing the traditional multiple linear regression model,ANN model,and SVM model,finally selected 4 types of variables(one-time medical material cost for surgery,surgical grade,length of stay,CCI)into the decision tree model;3 decision tree models finally Select the results of the E-CHAID model and cut the decision tree.The final subdivision group plan selects CCI and surgical disposable medical material fees as the grouping basis.The CB3 is subdivided into 4 groups,and the CV and RIV tests pass the same group.Homogeneity and heterogeneity have increased;the comprehensive performance evaluation results show that in two years,Department 1ranked first,and Department 3 ranked last,but Department 2 has made the most progress although it ranks lower.Conclusion: The CHS-DRG grouping scheme has a good matching degree in the ophthalmology department of the hospital;CCI value is effective as a basis for ADRG subdivision;the results of traditional multiple linear regression model and new data mining model are similar and different,and practical applications need to be considered;decision tree model In the process of grouping DRGs,it needs to be tailored,otherwise it will cause too many groupings and it is not easy to practically apply;DRGs grouping scheme can be used as an effective tool for medical performance evaluation,comprehensively evaluating medical resource consumption and output.