Risk-adjustment methods for all-payer comparative performance reporting in Vermont.

Area of Science:

• Health Services Research
• Health Economics
• Health Policy

Background:

• The shift towards value-based healthcare necessitates comprehensive, all-payer approaches to measure utilization, cost, and quality.
• Disparities among Medicare, Medicaid, and commercial payer populations present challenges for uniform risk-adjustment methods.

Purpose of the Study:

• To evaluate the effectiveness of different risk-adjustment models in comparing primary care practice performance across diverse payer groups.
• To assess the impact of increasing model complexity on the accuracy of risk-adjusted expenditure and utilization outcomes.

Main Methods:

• The study applied varying levels of risk adjustment, from basic age/gender to a full model including demographic, payer, and health status factors.
• Risk adjustment was performed on patient-centered medical home populations in Vermont (283,153 adults, 78,162 pediatric patients).
• Expenditure and utilization outcomes for 2014 were analyzed using comparative practice profiles.

Main Results:

• A full risk-adjustment model explained 29.3% of expenditure variation in adults, significantly higher than age/gender (2.8%) or basic CRG models.
• The full model demonstrated the lowest variation (coefficient of variation 11.7) for adult practices compared to less adjusted models.
• Similar improvements in reducing variation were observed for pediatric populations.

Conclusions:

• More comprehensive risk-adjustment models are effective for accurate, multi-payer population comparisons of healthcare metrics.
• These advanced models are essential for practices and for implementing incentive-based payment systems focused on whole-population outcomes.
• Vermont's Blueprint for Health program utilizes these profiles to drive quality improvement initiatives.