OBJECTIVES: Cost-effectiveness models typically use reported aggregate utility values from scientific literature. Ideally, these models should incorporate country-specific utility values. We introduce a novel copula-based simulation method to estimate mean EQ-5D utility values for any value set using reported mean values and the original value set. We demonstrate the usefulness of this approach by comparing estimated and observed mean values for arthritis and diabetes patient groups in the Multi Instrument Comparison study, using UK crosswalk values as input to estimate German and Spanish mean utility values.

METHODS: We generated EQ-5D data using a copula approach to reflect a natural distribution of responses modeled on self-reported health states in a mixed patient population. The modeled data included EQ-5D-5L questionnaire and EQ-VAS responses, maintaining correlations between dimensions and cumulative densities. The resulting simulated dataset was then re-weighted to produce mean values matching the target value in the target value set. Finally, mean values for alternative value sets were estimated for using the simulated data and fitted weights.

RESULTS: In our MIC data analysis, the observed mean utility values for arthritis patients were 0.626 (UK crosswalk), 0.733 (German 5L), and 0.697 (Spanish 5L). Our simulation estimated values of 0.742 (German) and 0.693 (Spanish). For diabetes patients, observed values were 0.695 (UK crosswalk), 0.793 (German 5L), and 0.754 (Spanish 5L), with estimated values of 0.799 (German 5L) and 0.754 (Spanish 5L). Across all disease groups, the mean absolute error of estimation was approximately 0.005.

CONCLUSIONS: Our copula-based simulation method provides reliable and precise estimates of EQ-5D utility values for different country-specific value sets using reported aggregate means from one value set. This approach can be further refined using empirical EQ-5D data from similar patient populations.