Development and Validation of a Tool for Assessing Adherence to Gluten-Free Diet in Patients With Celiac Disease

Abstract

INTRODUCTION: Life-long adherence to gluten-free diet (GFD) and its assessment is essential for patients with celiac disease (CeD). We have developed and validated a tool for assessing adherence to GFD which can be used by both physicians and dietitians. METHODS: Phase 1: Development, content validation, and assessment of reliability of tool. Phase 2: Validation of tool against standard dietary evaluation (SDE) (gold standard), immunoglobulin A - anti-tissue transglutaminase antibodies (IgA anti-tTG Ab), and gluten immunogenic peptides in urine. Overall, 380 biopsy-confirmed patients with CeD (derivation cohort: n = 100 [phase 1], n = 210 [phase 2] and independent validation cohort, n = 70) were recruited. RESULTS: Of an initial 90-point questionnaire, 84 items (Celiac Disease: Compliance Assessment Test [CD-CAT.v1]) were retained after content validation and pilot testing. In phase 1, upon administering CD-CAT.v1 on 100 patients, a comprehensive 35-item tool (CD-CAT.v2; α = 0.86) was obtained after removing items with low test-retest reliability and item-rest correlation values. In phase 2, upon administering CD-CAT.v2 on 210 patients, 22 items were removed having low correlation values (R < 0.4) with SDE. Finally, a 13-item tool (CD-CAT.v3; α = 0.84) was obtained with high criterion validity with SDE (r = 0.806, P < 0.001), moderate convergent validity with celiac disease adherence test (r = 0.602, P = 0.007), and moderate to weak correlation with urine gluten immunogenic peptides (r = 0.46, P = 0.001) and IgA anti-tTG Ab (r = 0.39, P = 0.008), respectively. The final 13-item tool also strongly correlated with SDE (r = 0.78, P < 0.001) in an independent validation cohort of 70 patients with CeD. Principal component analysis identified 3 relevant subscales with a cumulative variance of 62%. The sensitivity and specificity of CD-CAT.v3 were 80% and 91%, respectively, with an area under curve of 0.905 with SDE. The obtained cutoff score of <19 from the receiver operating characteristic curve was further categorized as 13 = excellent, 14–18 = very good, 19–28 = average, and >28 = poor adherence to GFD. DISCUSSION: CD-CAT is a new and rapid tool for monitoring dietary adherence to GFD with high sensitivity and specificity, which can be administered by both physicians and dietitians.