OBJECTIVES: Recall period is important when selecting appropriate patient-reported outcomes (PROs) for generating health utilities. People living with migraine experience variation in their daily experience, which may not be well captured using generic quality-of-life measures such as the EQ-5D with its 1-day recall period. In this post hoc analysis of the DELIVER (NCT04418765) migraine-prevention trial, we compared utilities estimated from 3 PRO measures—EQ-5D-5L, Migraine-Specific Quality of Life Questionnaire (MSQ; 4-week recall), and 6-item Headache Impact Test (HIT-6; 4-week recall) to assess the extent to which these measures capture migraine impact. The relationship between utilities and monthly migraine day (MMD) reduction was also examined with the 3 PRO measures.

METHODS: MSQ and HIT-6 were mapped to EQ-5D-3L using previously published algorithms. Utilities were estimated for each study visit (up to 24 weeks) at which the HIT-6, MSQ, and/or EQ-5D-5L was completed. A mixed linear model comparing number of MMDs and utility score was estimated for each instrument. The base-case model estimated the relationship between utilities and MMDs independent of treatment group.

RESULTS: Analysis of MSQ and HIT-6 showed a similar relationship with changes in MMDs (average utility decrement per MMD [95% CI]: MSQ, 0.0189 [0.0180, 0.0198], p<0.001; HIT-6, 0.0188 [0.0181, 0.0195], p<0.001). Relative to MSQ and HIT-6, EQ-5D-5L was less responsive to MMD changes (0.0053 [0.0048, 0.0059], p<0.001). When pooling or stratifying patients (n=337) by migraine type, a linear correlation between migraine presence during the day of EQ-5D-5L completion, migraine severity, and scores obtained using EQ-5D-5L was observed: presence and severity impacted the utility estimate (values ranged from 0.62–0.92).

CONCLUSIONS: Measures with a short recall period may not fully capture migraine’s impact on a patient’s life and functioning. Thus, this work highlights the importance of investigating different approaches that may be used to generate health utility data.