The importance of the family planning service environment and community-level factors on contraceptive use has long been studied. Few studies, however, have been able to link individual and health facility data from surveys that are nationally representative, concurrently fielded, and geographically linked. Data from Performance Monitoring and Accountability 2020 address these limitations. To assess the relative influences of the service delivery environment and community, household, and individual factors on a woman’s likelihood of using a modern contraceptive in five geographically and culturally diverse sub-Saharan African countries. Nationally representative, cross-sectional data from PMA2020 were linked at the household and service delivery level. Country-specific and pooled multilevel multinomial logistic models, comparing non-users, short- and long-acting method users were used. The variables elected for inclusion in our multivariate analyses were guided by the conceptual framework to profile the different levels of influences on individual use of modern contraception. Average marginal effects were calculated to improve interpretability. We find that the effect of contextual factors varies widely but that being visited by a health worker who spoke about family planning in the past 12 months was consistently and positively associated with individual use of short-acting and long-acting contraception. Characteristics of the nearest health facility did not generally exercise their own independent influences on a woman’s use of contraception, except in the case of Burkina Faso, where the average distance between individuals and the nearest family planning provider was significantly greater than other countries. Inclusion of country fixed effects in the pooled models and the relevance of covariates at different levels in the country-specific models demonstrate that there is significant variation across countries in how community, individual, and service delivery environment factors influence contraceptive use and method choice. Context must be taken into account when designing family planning programs.

Estimated use of emergency contraception (EC) remains low, and one reason is measurement challenges. The study aims to compare EC use estimates using five approaches. Data come from Performance Monitoring and Accountability 2020 surveys from 10 countries, representative sample surveys of women aged 15 to 49 years. We explore EC use employing the five definitions and calculate absolute differences between a reference definition (percentage of women currently using EC as the most effective method) and each of the subsequent four, including the most inclusive (percentage of women having used EC in the past year). Across the 17 geographies, estimated use varies greatly by definition and EC use employing the most inclusive definition is statistically significantly higher than the reference estimate. Impact of using various definitions is most pronounced among unmarried sexually active women. The conventional definition of EC use likely underestimates the magnitude of EC use, which has unique programmatic implications.

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