Advancing Methods for Measurement of Complementary Feeding Interventions and Practices at Scale: Outcomes from Two Rounds of National Surveys in Burkina Faso and Kenya (P10-135-19)

Objectives Promotion of improved complementary feeding (CF) practices for children 6-23 m is a priority intervention to prevent stunting and also childhood obesity. However, global household survey programs do not include CF intervention coverage or "unhealthy" diet practices. We aimed to develop and refine indicators and questions for measuring these outcomes in large-scale household surveys. Methods In 2017 and 2018, we carried out nationally-representative household surveys in Burkina Faso (BF) and Kenya (K) that included children 0-59 m and women 10-49 yrs. Over two rounds per country we modified the questionnaire, tools and enumerator training to better capture the intended information. In 2018, we used both prompted and unprompted approaches to ask about specific CF messages received. Results Coverage of any CF counseling among caregivers of 6-23 m olds who received counseling in the specified recall period (within 1 m for 6-11 m olds, within 3 m for 12-23 m olds) remained constant over the two years in both countries (2017: 16% Burkina Faso, 20% Kenya; 2018: 17% Burkina Faso; 18% Kenya). Between years, we changed the structure of questions about the timing of their last counseling visit. The revised 2018 method allowed more flexibility in defining and comparing recall periods by age group (Figure 1). Unprompted questions about CF messages resulted in much lower coverage compared to prompted (Figure 2).The proportion of children achieving minimum dietary diversity increased slightly across years (2017: 16% BF, 40% K; 2018: 20% BF, 43% K). Consumption of unhealthy foods, particularly sugar-sweetened beverages (SSB) increased with age (Figure 3). However, when we excluded "milk tea with sugar" from the SSB definition in Kenya, consumption was only 11% for children 6-59 m. Perceptions around unhealthy foods and SSBs varied by cultural context, making it challenging for enumerators to classify foods into these categories. Conclusions Consideration should be given to recall periods, prompted versus unprompted responses, and culturally appropriate training around dietary data collection to elicit the most accurate results in survey settings. Our findings are generalizable to global and national nutrition surveys programs including the Demographic and Health Survey. Funding Sources Bill & Melinda Gates Foundation. Supporting Tables Images and/or Graphs