Comparing diagnosis-related group systems to identify design improvements

Area of Science:

• Health Services Research
• Health Economics
• Information Management

Background:

• Casemix systems are crucial for healthcare funding and resource allocation.
• International variations exist in casemix system design and performance.
• The United Kingdom's Healthcare Resource Groups (HRG) required evaluation against international benchmarks.

Purpose of the Study:

• To compare the performance of the UK's HRG versions 3.1 and 3.5 with international casemix systems.
• To identify specific areas for improvement in the UK's HRG system design.
• To leverage insights from the US All Patient Diagnosis Related Groups (AP-DRG) and Australian Refined Diagnosis Related Groups (AR-DRG).

Main Methods:

• Utilized a large dataset of over 12 million inpatient and day case discharge records from English NHS acute hospitals (2001-2002).
• Applied four casemix classification systems: HRG v3.1, HRG v3.5, AP-DRG, and AR-DRG.
• Assessed statistical performance using the Reduction in Variance (RIV) statistic.

Main Results:

• The Australian Refined Diagnosis Related Groups (AR-DRG) demonstrated the best overall Reduction in Variance (RIV).
• AR-DRG benefited from a larger number of defined groups compared to other systems.
• Analysis of individual chapters within each system revealed varying strengths, with each system outperforming others in specific areas.

Conclusions:

• The comparison successfully identified actionable changes to enhance the performance of the UK's Healthcare Resource Groups (HRG).
• Future revisions of HRG should prioritize modifications within chapters showing the greatest potential for RIV improvement.
• International benchmarking provides valuable insights for optimizing national casemix classification systems.