Elaboration and verification of the validity and reliability of a nutrition literacy instrument among people with diabetes

Abstract Background The literacy about nutrition of people with diabetes is important since healthy eating can improve the life quality of these people. Objective The objective of this study was to create an instrument called “Nutrition Literacy among people with Diabetes (NLD)” and evaluate its validity, reliability and interpretability. Method Methodological study with verification of content validity; pre-test; reliability estimation (internal consistency: Cronbach's alpha; reproducibility: kappa-K, Intraclass Correlation Coefficient - ICC); estimation of the validity of concurrent criterion and interpretability made by counting the correct words associations with scores from 0 to 24 (cut ≤ 18) and hypothesis test in sample for infinite population through logistic regression p≤0.05; (OR / CI 95%). Results The content validity was satisfactory. Reliability (alpha = 0.68, K≥0.60 and CCI = 0.68) was good. Regarding the concurrent validity, a correlation between NLD and schooling (rs = 0.88, p = 0.000) was found. From the 212 participants, 75.8% presented NLD> 18. In the hypothesis test, it was found that the higher the schooling, the higher the NLD (1.20 / 1.09-1.31); the NLD was lower among men (0.50 / 0.25-1.01). Conclusion The NLD was considered valid, reliable and easy to interpret, and it could be used in future research or even in health services that provide assistance to people with diabetes.