OBJECTIVES: To assess the psychometric performance of 6 bolt-ons (vision, breathing, tiredness, sleep, social relationships and self-confidence) in a myasthenia gravis (MG) patient population.

METHODS: MyRealWorld-MG is an observational, longitudinal study among MG patients across 10 countries. Patients entered demographics and disease characteristics, and completed several patient-reported outcome measures (PROMs), including the EQ-5D-5L, PROMIS-Sleep Disturbance, FACIT-Fatigue and PROMIS-Dyspnea.

The following psychometric properties of the EQ-5D-5L+bolt-on(s) were assessed: ceiling, informativity, divergent and convergent validity, known-groups validity, and explanatory power. Responsiveness for bolt-ons tiredness, breathing and sleep was assessed by comparing changes over time with changes in relevant condition-specific measures.

RESULTS: 1510 MG patients completed the EQ-5D-5L+bolt-ons at baseline; follow-up time and frequency of administration varied per PROM (N=477 to 4649). Bolt-ons tiredness, vision and sleep reduced the EQ-5D-5L’s ceiling the most (43%, 30% and 30% relative reduction). These bolt-ons also had the highest absolute informativity (Shannon H’: 2.04, 1.90, 1.87) and relative informativity (Shannon J’: 0.88, 0.82, 0.80) compared to no bolt-on (Shannon H’ and J’: 1.77, 0.76).

Divergent validity of bolt-ons was generally modest as most bolt-ons were correlated with EQ-5D-5L dimensions, especially tiredness with usual activities and self-confidence with anxiety/depression. The bolt-ons showed excellent convergent validity, correlating well with items from other instruments intended to capture similar constructs.

Vision, breathing and tiredness bolt-ons together improved the known group validity of the EQ-5D-5L with a relative efficiency of 1.29.

Moderate-to-strong correlations were observed between changes in the scores of condition-specific measures and changes in bolt-ons (r(tiredness, FACIT-Fatigue)=0.62; r(sleep, PROMIS-Sleep Disturbance)=0.56; r(breathing, PROMIS-Dyspnea)=0.49), demonstrating responsiveness.

Finally, adding tiredness to the EQ-5D-5L improved the explained variance of the EQ-VAS by 5.2%.

CONCLUSIONS: Tiredness, vision and breathing bolt-ons improved the EQ-5D-5L’s capacity to capture health problems associated with MG, indicating their potential for future use in MG and other relevant patient populations.