Personalized Precision in Healthcare Payment Systems: Insights From a Real-World Cancer Patients Cohort Study

Area of Science:

• Health Economics
• Oncology
• Medical Informatics

Background:

• Rising global medical expenses for cancer patients necessitate improved cost assessment.
• Current diagnosis-related group (DRG) payment systems inadequately reflect individual patient cost burdens.
• Gastric cancer (GC) surgery patients face significant and variable hospitalization costs.

Purpose of the Study:

• To develop and validate a prediction model for identifying gastric cancer surgery patients at high risk of elevated hospitalization costs.
• To enhance the accuracy of cost prediction beyond existing DRG-based systems.
• To support individualized and precise medical cost management.

Main Methods:

• A real-world cohort study of 3981 GC surgery inpatients in Beijing, China.
• Definition of 'reasonably high cost' as exceeding the 75th percentile of hospitalization expenses.
• Logistic regression analysis (univariate and multivariate) with Akaike information criterion for predictor selection.

Main Results:

• A final prediction model incorporating 13 predictors was developed.
• The model demonstrated good discrimination with an area under the receiver operating characteristics curve (AUC) of 0.718 (95% CI: 0.700-0.736).
• Cross-validation yielded a comparable AUC of 0.708 (95% CI: 0.690-0.727), and good calibration was observed.

Conclusions:

• The developed prediction tool for high hospitalization costs in GC surgery patients shows robust performance.
• Integrating this tool with DRG payments can lead to more personalized and precise cost control.
• This approach aims to ensure rational, efficient, and equitable healthcare services.