Optimization of diagnosis-related groups for patients with acute appendicitis using a machine learning model

Area of Science:

• Health economics
• Medical informatics
• Healthcare management

Background:

• Diagnosis-Related Groups Prospective Payment System (DRG-PPS) is a global standard.
• Localizing DRG grouping and pricing is crucial for healthcare systems.
• Acute appendicitis (AA) cost factors require local analysis.

Purpose of the Study:

• To identify key factors influencing hospitalization costs for AA patients.
• To develop a DRG grouping model for AA based on local data.
• To inform individualized hospitalization cost reimbursement strategies.

Main Methods:

• Stratified random sampling of hospitals in Hefei, China (2017-2019).
• Analysis of 4,066 AA patient records using single-factor and multiple linear regression.
• Application of a Classification and Regression Tree (CART) model with E-CHAID for DRG grouping.

Main Results:

• Length of stay, marital status, surgery, and hospital level significantly impacted AA hospitalization costs (p<0.05).
• A CART model classified AA inpatients into 10 DRG groups using age, surgery type, and LOS.
• Standardized disease costs varied from 3,047 CNY to 15,569 CNY.

Conclusions:

• AA hospitalization costs are demonstrably linked to clinical and demographic factors.
• The developed decision tree model offers a framework for DRG classification.
• Findings support the implementation of precise, individualized reimbursement policies.