Defining a Dichotomous Indicator for Population-Level Assessment of Dietary Diversity Among Pregnant Adolescent Girls and Women: A Secondary Analysis of Quantitative 24-h Recalls from Rural Settings in Bangladesh, Burkina Faso, India, and Nepal
Eric O. Verger, Sabrina Eymard‐Duvernay, D. Bahya-Batinda, Giles Hanley‐Cook, Alemayehu Argaw, Elodie Becquey, Loty Diop, Aulo Gelli, Helen Harris–Fry, Shivani Kachwaha, Sunny S. Kim, Phuong Hong Nguyen, Naomi Saville, Lan Mai Tran, Rock R. Zagré, Edwige Landais, Mathilde Savy, Yves Martin‐Prével, Carl Lachat
Abstract
Background: The Minimum Dietary Diversity for Women of Reproductive Age (MDD-W) indicator was validated as a proxy of micronutrient adequacy among nonpregnant women in low- and middle-income countries (LMICs). At that time, indeed, there was insufficient data to validate the indicator among pregnant women, who face higher micronutrient requirements. Objective: This study aimed to validate a minimum food group consumption threshold, out of the 10 food groups used to construct MDD-W, to be used as a population-level indicator of higher micronutrient adequacy among pregnant women aged 15-49 y in LMICs. Methods: = 4909). We computed the 10-food group Women's Dietary Diversity Score (WDDS-10) and calculated the mean probability of adequacy (MPA) of 11 micronutrients. Linear regression models were fitted to assess the associations between WDDS-10 and MPA. Sensitivity, specificity, and proportion of individuals correctly classified were used to assess the performance of MDD-W in predicting an MPA of >0.60. Results: In the pooled sample, median values (interquartile range) of WDDS-10 and MPA were 3 (1) and 0.20 (0.34), respectively, whereas the proportion of pregnant women with an MPA of >0.60 was 9.6%. The WDDS-10 was significantly positively associated with MPA in each survey. Although the acceptable food group consumption threshold varied between 4 and 6 food groups across surveys, the threshold of 5 showed the highest performance in the pooled sample with good sensitivity (62%), very good specificity (81%), and percentage of correctly classified individuals (79%). Conclusions: The WDDS-10 is a good predictor of dietary micronutrient adequacy among pregnant women aged 15-49 y in LMICs. Moreover, the threshold of 5 or more food groups for the MDD-W indicator may be extended to all women of reproductive age, regardless of their physiologic status.