Comparison of Preference-Based Utilities of the 15D, EQ-5D and SF-6D in Patients with HIV/AIDS

Objective: This article compares preference-based utilities from the multiattribute utility instrument 15D with those derived from the EQ-5D and the Short Form 36 (SF-6D) in patients with HIV/AIDS. In particular, we wanted to examine if the finer descriptive system of the 15D would result in better discriminative capacity or responsiveness. Methods: In a prospective observational study of 60 Norwegian patients with HIV/AIDS from two hospitals, the authors compared scores, assessed associations with disease staging systems, and assessed test-retest reliability and responsiveness of the instruments. Results: On average, the 15D gave higher utility scores than the other two measures, the mean utility scores were: 15D - 0.86, SF-6D - 0.73, and EQ-5D Index - 0.77. Test-retest reliability was acceptable for all measures, with intraclass correlation coefficients between 0.78 and 0.94. The correlation between scores of the 3 scales was substantial (π = 0.74-0.80). There was no major difference in responsiveness between the measures. Conclusions: The different measures gave different utility values in this sample of patients with HIV/AIDS, although many of the measurement properties were similar. There was no evidence for better discriminative capacity or responsiveness for the 15D, than for the two other multiattribute measures.