Common scale valuations across different preference-based measures: estimation using rank data.

Area of Science:

• Health Economics
• Psychometrics
• Econometrics

Background:

• Preference-based measures (PBMs) estimate quality-adjusted life years (QALYs) but yield varying utility values due to different methodologies.
• Existing studies compare pairs of PBMs using patient data, lacking a unified approach for multiple PBMs on a common scale.

Purpose of the Study:

• To develop a novel econometric model for comparing various PBMs on a common scale.
• To address limitations of standard rank-ordered logit models in analyzing health state preference data.

Main Methods:

• Analysis of general public survey data (n=501) where respondents ranked health states from 6 PBMs.
• Development of a mixed logit model to overcome independence of irrelevant alternatives and repeated observations issues.

Main Results:

• Substantial differences in estimated parameters (mean difference 0.07) and health state value orderings between the standard and new models.
• The proposed model yields more consistent results for health state values compared to the standard model.
• Limitations include a small sample size and limited health state coverage.

Conclusions:

• The developed flexible econometric model appropriately handles rank data features.
• Standard models are inappropriate for this data, producing inconsistent results.
• The new approach enables PBM comparison on a common scale.