Abstract
PURPOSE: This analysis evaluated the psychometric properties of the EQ-5D-3L among patients with Mycobacterium avium complex lung disease (MAC-LD). METHODS: Data from the Phase III CONVERT trial were analyzed. Study measures including the EQ-5D-3L, St. George's Respiratory Questionnaire (SGRQ), the 6-minute walk test (6MWT), and percent predicted forced expiratory volume in 1 second (ppFEV(1)) were collected at baseline, month 3 (M3), month 4 (M4; only 6MWT), and month 6 (M6). Item characteristics and validity were evaluated at baseline. Test-retest reliability was assessed using intraclass correlation coefficients (ICC) and a weighted kappa statistic among a subgroup of stable patients. Validity was tested by evaluating correlations of the EQ-5D-3L index/visual analogue scale (VAS) scores with SGRQ and 6MWT and comparing mean index/VAS scores across known groups defined using 6MWT and ppFEV(1). Responsiveness of the EQ-5D-3L was assessed using 6MWT, SGRQ, ppFEV(1), and culture conversion as anchors. RESULTS: The index score was subject to ceiling effects, with 32.6% of patients reporting perfect health at baseline. ICCs for the index (0.80) and VAS (0.85) scores and weighted kappas for the domains (0.5-0.72) indicated adequate test-retest reliability. Correlations between the index/VAS scores and related domains of the SGRQ and 6MWT were as hypothesized (0.31-0.62), and the mean index/VAS scores were significantly different between the 6MWT and ppFEV(1) known groups (p<0.05), supporting the validity of the EQ-5D-3L. No evidence was found supporting the responsiveness of the EQ-5D-3L to changes in any of the anchors. CONCLUSION: EQ-5D-3L exhibited evidence of validity and reliability but poor responsiveness to clinically meaningful changes in patients with MAC-LD.