Validation of NINA-DISH food frequency questionnaire with multiple 24-hour dietary recalls among pregnant women in Pune, India

Adequate dietary intake is critical to prevent adverse pregnancy outcomes. India has a high burden of maternal and child morbidity and mortality, but there is a lack of adequate tools to assess dietary intake. We validate a food frequency questionnaire (FFQ), New-Interactive Nutrition Assistant-Diets in India Study of Health (NINA-DISH), among pregnant women living with and without HIV in Pune, India. Women were selected from a cohort study investigating immune responses to HIV and latent TB during pregnancy. The FFQ was administered during the third trimester and validated against multiple 24-hour diet recalls (24-HDRs) collected in second and third trimesters. Data for analysis was available from 58 women out of 70 enrolled into this sub-study, after excluding those with incomplete data or implausible energy intake. The median (Q1, Q3) age of study participants was 23 (20, 25) years. Median (Q1, Q3) daily energy intakes were 2522 (1912, 2858) kcal and 2551 (2034, 3337) kcal by 24-HDR and FFQ, respectively with FFQ overestimating nutrient intake. Pearson correlations between log-transformed estimates from FFQ and 24-HDR for energy, protein, carbohydrate, fat, iron and zinc were 0.47, 0.48, 0.45, 0.33, 0.4, and 0.54, respectively. Energy-adjusted and deattenuated correlations ranged from 0.41 (saturated fat) to 0.73 (sodium). The highest misclassification into extreme tertiles was observed for fat (22%), saturated fat (21%), and sodium (21%). Bias existed at higher intake levels as observed by Bland-Altman plots. In conclusion, NINA-DISH is a valid and feasible tool for estimating dietary intakes among urban pregnant women in western India.