Comparative models on low multiplier DRG classification for advanced lung cancer.
Mingming Yu1,2, Yanxi Zhang3,4
1School of Economics and Management, Shanghai Technical Institute of Electronics & Information, Shanghai, China.
Frontiers in Public Health
|September 29, 2025
Summary
Machine learning models accurately predict low multiplier Diagnosis Related Groups (DRGs) for advanced lung cancer. Random Forest excelled, with cost and length of stay being key predictors over demographics.
Area of Science:
- Oncology
- Health Informatics
- Machine Learning
Background:
- Advanced lung cancer patient care involves complex Diagnosis Related Groups (DRGs) and resource allocation.
- Accurate prediction of low multiplier DRGs is crucial for financial management and quality assessment in healthcare.
Purpose of the Study:
- To compare the predictive performance of various machine learning models for low multiplier DRGs in advanced lung cancer.
- To identify the optimal machine learning algorithm and key factors influencing these predictions.
Main Methods:
- Four machine learning algorithms were employed: logistic regression, hybrid naive Bayes, Support Vector Machine (SVM), and Random Forest.
- Model performance was assessed using metrics like AUC, accuracy, and precision.
- Key contributing features were identified to understand prediction drivers.
Main Results:
- The Random Forest algorithm demonstrated superior performance, achieving the highest AUC, accuracy, and precision.
- Cost-related features and length of hospital stay were more significant predictors than demographic factors (e.g., gender, age).
Conclusions:
- DRG classification for advanced lung cancer patients is reasonably structured, aiding subgroup analysis.
- Predictive models suggest potential upcoding and medication underuse, necessitating monitoring of examination fees and medication costs.
- Larger datasets enhance model stability; algorithm choice depends on analytical goals (Random Forest for precision, Logistic Regression/SVM for recall).
