Regression tree modeling of hospitalization expenses and DRG grouping for unilateral inguinal hernia in children
10.3760/cma.j.cn111325-20201008-02006
- VernacularTitle:儿童单侧腹股沟疝住院费用的回归树模型构建及DRG分组探索
- Author:
Mei ZHAO
1
;
Xinyu WANG
;
Hui XU
;
Guoshuang FENG
Author Information
1. 安徽省儿童医院病案室,合肥 230051
- Keywords:
Diagnosis-related groups;
Hernia, Inguinal;
Children;
Hospitalization expenses
- From:
Chinese Journal of Hospital Administration
2021;37(8):658-661
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore the variation of hospitalization expenses and its influencing factors under the same main diagnostic disease(children′s unilateral inguinal hernia), and group them based on regression tree model, so as to provide reference for diagnosis-related groups(DRG) grouping of the disease.Methods:The data of the first page of medical records with unilateral inguinal hernia(ICD-10 coding K40.3 and K40.9)as the main diagnosis discharged from a children′s special hospital from 2016 to 2018 were extracted. With hospitalization expenses as the outcome variable and gender, age, operation type, length of hospital stay and complications as the predictive variables, a regression tree model was constructed to classify the children into subgroups.Results:The regression tree model showed that the variables that had a significant impact on the total hospitalization costs were the length of hospital stay, operation type and complications. Age and gender had no significant impact on the hospitalization expenses. According to these three significant variables, the children were divided into five subgroups and the median hospitalization expenses were 1 190.8, 6 387.2, 7 037.7, 11 413.7, and 18 499.1, respectively. Except that the coefficient of variation of hospitalization expenses in group 1 was greater than 1, the coefficient of variation of the other four groups were≤0.5.Conclusions:Children with unilateral inguinal hernia can be divided into groups according to the length of hospital stay, operation type and complications, and the grouping results are relatively reasonable.