Development and evaluation of a Nutrition Transition-Food Frequency Questionnaire for adolescents in South India

Background There are no validated Food Frequency Questionnaires (FFQ) for adolescents in India, a population that may be undergoing nutrition transition – the rapid changes in dietary patterns from globalization. Objective To develop and evaluate the reproducibility and validity of a Nutrition Transition-FFQ (NT-FFQ) among adolescents ages 14–18 years in South India. Methods A 125-item semi-quantitative NT-FFQ was developed to include traditional and global foods. The reference period of the NT-FFQ was one month. Foods were condensed to 21 food groups for food-based analysis. The NT-FFQ was administered at home by interviewers in a sample of 198 school-going adolescents (14–18 y) at baseline (NT-FFQ1) and after two months (NT-FFQ2). A subsample of 97 adolescents also completed three interviewer-administered 24-hour dietary recalls during the two months between NT-FFQ1 and NT-FFQ2. Reproducibility of the NT-FFQ was assessed by comparing intakes between NT-FFQ1 and NT-FFQ2. Validity of the NT-FFQ was assessed by comparing the intake of food groups from the NT-FFQ2 to the average intakes from three 24-hour diet recalls. The estimates of reproducibility and validity were calculated using Spearman correlations, Intraclass Correlation Coefficients (ICC), Bland-Altman analysis, and the misclassification of food group intake using tertiles. Results Reproducibility: The Spearman correlations for foods ranged from 0.12 (cooked lentil) to 0.80 (red meat) (table 1). For food groups, Spearman correlations ranged from 0.33 (pulses and nuts) to 0.80 (red meat) and ICC from 0.05 (fruits) to 1.00 (pulses and nuts and tea or coffee). On average, 60% were classified in the same tertiles and 8% were classified in opposite tertiles, indicating good agreement. The classification of food groups in the same tertiles (concordance) was high for the global-healthy food group comprising oats, cereal, and multigrain biscuits (90%) and for the lean meats group (75%). Discordance was high (>10%) only for fruit juices (12%) and snacks (11%). Validity of NT-FFQ: The Spearman correlations ranged from 0.14 (buns) to 0.79 (chocolate milk powder) for individual foods and from 0.09 (sugar and jaggery) to 0.71 (tea or coffee) for food groups. The ICC for food groups ranged from 0.41 to 1.00. The classification of food group intake in the same and opposite tertile were 48% and 8%, respectively, suggesting good agreement. Concordance was high for the global-healthy food group (87%) and red meat food group (75%). Discordance was high only for the fried snacks (14%) and fried traditional foods (14%). The Bland-Altman plots showed acceptable agreement for food groups between the two administrations of the NT-FFQ and between NT-FFQ2 and the 24-hour diet recalls. Conclusion The results show that the NT-FFQ has good reproducibility and acceptable validity for foods and food groups and can be used to assess nutrition transition among adolescents in South India.