OBJECTIVES: Health-related quality of life (HRQoL) data are commonly collected as an outcome in clinical studies. However, unlike other outcomes, these data are rarely analysed using meta-analytic techniques – particularly when drawn from different yet related studies (e.g. when an indirect comparison of treatments is required). Using atopic dermatitis as an example, this research examines the impact of using data directly from studies or analysed as an absolute or relative effect measure in an indirect comparison.

METHODS: Using a pre-existing network of randomised controlled trials for the treatment of atopic dermatitis, available data were extracted on baseline utilities, utilities at 12/16 weeks, or change from baseline depending on availability. Utility data were converted to EQ-5D-3L if not already provided in that format.

RESULTS: The network comprised of 4 studies connected via placebo. The baseline utility value was taken from the placebo group of SOLO 1. Estimated utility values at 12 weeks for all treatments were: 0.753 (best supportive care - all methods), 0.790 vs 0.693 vs 0.876 (abrocitinib 100mg), 0.790 vs 0.683 vs 0.876 (abrocitinib 200mg), and 0.873 vs 0.714 vs 0.888 (dupilumab 300mg) for raw data vs absolute difference NMA vs relative difference NMA.

CONCLUSIONS: Utility values should be estimated based on the available data with a strong preference for meta-analytic techniques when multiple trials or indirect comparisons are required. Best practice suggests a relative treatment approach would be preferred.