Introduction: Life-long adherence to gluten-free diet (GFD) and assessment of adherence to GFD is essential for celiac disease (CeD) patients. In view of lack of celiac-expert dietitians and lack of a cultural-specific adherence tool, we developed and validated a tool for assessing adherence to GFD “Celiac Disease: Compliance Assessment Test (CD-CAT)”.
Methods: Phase-I: Development, content validation and assessment of reliability of tool. Phase-II: Validation of tool against standard dietary evaluation (SDE) (gold standard), IgA-anti-Ttg and gluten immunogenic peptides in urine (Urine-GIP). Overall, 310 CeD patients (Phase-I, n=100; Phase-II, n=210) were recruited. Relevant tests of factor analysis were conducted to identify the subscale structure. The sensitivity and specificity of the final tool was assessed through receiver operating characteristic (ROC) curve analysis in comparison with gold standard.
Results: In phase-I, of initial 90-point questionnaire, 84-items (CD-CAT.v1) were retained after content validation and pilot testing. Upon administering CD-CAT.v1 on 100 CeD patients, a comprehensive 35-items tool (CD-CAT.v2, α=0.86) was obtained after removing items with low test-retest reliability and inter-item correlation. In Phase-II, upon administering CD-CAT.v2 on 210 CeD 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), and high construct validity with IgA-anti-tTG (p=0.0081) and urine-GIP (p<0.001). Factor analysis identified three relevant sub-scales with a cumulative-variance of 62%. The expert panel reviewed the factors identified and by consensus renamed them based on their clinical features: 1) Intentional gluten intake and psychosocial assessment, 2) Availability of flour mill and label reading and 3) Eating out and traveling.
The ROC analysis obtained a cut-off score of <19 for good adherence to GFD with an area under curve of 0.905 with SDE, high sensitivity (80%) and specificity (91%). Calculated positive and negative predictive values were 77% (64.5 – 86.5%) and 91.9% (86.4 – 95.8%), respectively. To define the level of adherence, the results were compared with the scores of SDE and the cut-off score of <19 was further classified as: 13 = excellent,14 - 18 = very good, 19 – 28 = average and >28 = poor adherence to GFD.
Conclusion: CD-CAT is a new tool for monitoring of dietary adherence to GFD with high sensitivity and specificity. CD-CAT is rapid and can be easily administered by both physicians and dietitians. It can be used alone or in combination with serological and biochemical parameters to assess the adherence to GFD in patients with CeD.

Ashish Agarwal

Appearances