Comparing EQ-5D-5L, PROPr, SF-6D and TTO utilities in patients with chronic skin diseases

Abstract

Objectives We aim to compare the measurement properties of three indirect (EQ-5D-5L, PROPr, SF-6D) and one direct (time trade-off, TTO) utility assessment methods in patients with chronic skin diseases. Methods 120 patients with physician-diagnosed chronic skin diseases (psoriasis 39%, atopic dermatitis 27%, acne 19%) completed a cross-sectional survey. Respondents completed the EQ-5D-5L, PROMIS-29+2 and SF-36v1 questionnaires and a 10-year TTO task for own current health. Utilities were computed using the US value sets. Ceiling, convergent and knowngroup validity were compared across the utilities derived with these four methods. Known-groups were defined based on general, physical and mental health. The agreement between utilities was assessed using intraclass correlation coefficients (ICC). Results Mean utilities for the EQ-5D-5L, PROPr, SF-6D and TTO were 0.79, 0.47, 0.76 and 0.89. In corresponding order, the ceiling was 28%, 0%, 2% and 65%. The SF-6D showed excellent agreement with the EQ-5D-5L (ICC = 0.770). PROPr demonstrated poor agreement with the EQ-5D-5L (ICC = 0.381) and fair with SF-6D utilities (ICC = 0.445). TTO utilities showed poor agreement with indirectly assessed utilities (ICC = 0.058–0.242). The EQ-5D-5L better discriminated between known groups of general and physical health, while the SF-6D and PROPr outperformed the EQ-5D-5L for mental health problems. Conclusion There is a great variability in utilities across the four methods in patients with chronic skin conditions. The EQ5D-5L, despite its higher ceiling, appears to be the most efficient in discriminating between patient groups for physical health aspects. Our findings inform the choice of instrument for quality-adjusted life year calculations in cost-utility analyses.