Cognitive Interviewing to Improve Infant and Young Child Dietary Assessments in the Nepal Demographic and Health Survey.

Background: The Demographic and Health Survey (DHS) is the primary source of nationally representative nutrition data for many countries and is important in shaping national nutrition policy, advocacy, and research. Without careful testing and adaptation to local contexts, findings are at risk of inaccurately representing respondents' behaviors and key nutritional indicators. Objectives: This study used a cognitive interviewing approach to assess mothers' comprehension, retrieval, judgment, and response to the infant and young child feeding (IYCF) dietary assessment module questions in the Nepal DHS-8 Woman's Questionnaire. Methods: A cross-sectional qualitative study was conducted in December 2021, using semistructured cognitive interviews, among 54 mothers of children 6-23 mo of age across 3 regions of Nepal. Interviews were analyzed to identify issues in participant understanding of questions, as well as cognitive challenges or benefits of a longer introduction to the IYCF dietary assessment module, as compared with the original shorter introduction and challenges or benefits to using closed-ended compared with open-ended questions. Results: Cognitive challenges were identified, including issues with judgment of certain food categories resulting in misclassification, comprehension of certain concepts, and comprehension of certain instructions for the dietary assessment that resulted in the variability of food and liquids recalled by mothers. Conclusions: These findings carry substantial implications for the measurement of the diet quality of infants and young children (IYC) in Nepal, and similar cognitive challenges may exist in other contexts. Although this current study is not able to answer the exact extent to which the accuracy and precision of IYCF indicators are affected by these cognitive challenges, these findings indicate that current measurements may be subject to over or underestimation and there is a need for further research in this area to understand the degree of measurement error related to these challenges.