3125 steps to perfect health: a nonparametric approach to developing the EQ-5D-5L value set

Area of Science:

• Health Economics
• Biostatistics
• Psychometrics

Background:

• Health state utility values are crucial for health technology evaluations and evidence-based resource allocation.
• The accuracy of these values depends on the quality of the valuation set used.
• Existing methods for developing EQ-5D-5L value sets rely on specific functional forms, which may limit accuracy.

Purpose of the Study:

• To introduce a novel, robust method for developing an EQ-5D-5L value set for Canada.
• To generate a value set with improved accuracy applicable to other populations.
• To provide an alternative to current methods that does not require pre-specification of functional forms.

Main Methods:

• Utilized nonparametric local constant regression to estimate the EQ-5D-5L value set for Canada.
• This method avoids the need to specify a functional form at the beginning of the modeling process.
• Applied preference elicitation techniques to a population sample to value health states.

Main Results:

• The nonparametric method significantly improved in-sample fit, reducing average squared prediction error by 94.46% and mean absolute error by 79.37%.
• Out-of-sample performance was superior to nine comparison models in four of five studies.
• The generated value set demonstrated logical consistency and substantial differences from comparator models.

Conclusions:

• Nonparametric regression is a promising tool for estimating EQ-5D-5L valuation sets.
• This method offers a potential standardized approach for future value set development.
• The findings suggest improved accuracy and consistency in health state utility valuation.