A tree-structured multiobjective optimization framework for constructing diagnosis-related groups.

Area of Science:

• Health Services Research
• Health Economics
• Medical Informatics

Background:

• Diagnosis-related groups (DRGs) are crucial for medical insurance payment standards.
• Current DRG construction methods struggle with grouping rule violations and multiobjective optimization imbalances.
• These limitations hinder accurate reflection of clinical complexities in payment standards.

Purpose of the Study:

• To propose a novel multiconstraint multiobjective optimization model and a tree-structured framework for DRG construction.
• To address challenges in existing DRG methods, including rule adherence and objective balancing.
• To develop a DRG system that accurately reflects clinical complexity and supports robust payment standards.

Main Methods:

• Developed a mathematical model defining intragroup homogeneity and intergroup heterogeneity objectives under grouping constraints.
• Utilized nonnegative adaptive LASSO regression for precise clinical complexity quantification.
• Integrated tree structures with multiobjective optimization algorithms to generate Pareto-optimal DRG sets.

Main Results:

• The proposed framework successfully satisfies predefined grouping constraints.
• The method demonstrates effectiveness in accurately reflecting clinical complexity within DRGs.
• Empirical results validate the framework's ability to generate efficient and compliant DRG sets.

Conclusions:

• The novel framework offers a paradigm for constructing DRGs that adhere to rules and optimize objectives.
• This approach provides decision-makers with efficient DRG sets for improved medical insurance payment standards.
• The study advances DRG methodology for better alignment with clinical realities and economic principles.