Dietary Diversity Predicts the Adequacy of Micronutrient Intake in Pregnant Adolescent Girls and Women in Bangladesh, but Use of the 5-Group Cutoff Poorly Identifies Individuals with Inadequate Intake

Background: The Minimum Dietary Diversity for Women (MDD-W) indicator based on a 10-food group women dietary diversity score (WDDS-10) has been validated to assess dietary quality in nonpregnant women. Little is known about its applicability in pregnant women, and specifically pregnant adolescent girls with higher nutrient requirements. Objectives: This study aimed to 1) compare the adequacy of micronutrient intakes between pregnant adolescent girls and women, 2) examine the performance of WDDS-10 in predicting the mean probability of adequacy (MPA) of 11 micronutrients, and 3) assess how well the MDD-W cutoff of 5 groups performed in pregnant adolescent girls and women. Methods: We used data from a 2015 household survey in Bangladesh (n = 600). Nutrient intakes were estimated with a multiple-pass 24-h recall and WDDS-10 was assessed through the use of a list-based method. Multiple linear regression models adjusted for geographical clustering assessed the association between WDDS-10 and MPA. Sensitivity and specificity analysis assessed the accuracy of MDD-W in correctly classifying individuals into high (MPA >0.6) or low MPA. Results: Dietary intakes of pregnant adolescent girls and women were similar in energy intake, WDDS-10 (5.1 +/- 1.4), MPA (0.40 +/- 0.12), and micronutrient intakes. Probabilities of adequacy were similar to 0.30 for riboflavin, vitamin B-12, calcium, and zinc; 0.12-0.15 for folate; 0.16-0.19 for vitamin A; and extremely low for iron at 0.01. The WDDS-10 was significantly associated with MPA in both groups and predicted MPA equally well at population level (SD of residuals 0.11 for both). Use of the 5-food groups cutoff for MDD-W to classify individuals' diets into MPA >0.6, however, resulted in a low correct classification (similar to 40%). A cutoff of 6 food groups markedly improved correct classification. Conclusions: The WDDS-10 predicted MPA equally well for pregnant adolescent girls and women at population level. The MDD-W indicator performed poorly in classifying individuals with MPA >0.6.