Linking Data from Demographic and Agricultural Surveys to Examine the Drivers of Stunting and Wasting in Nigeria: Lessons Learned


PDF document icon wp-20-241.pdf — PDF document, 1,351 kB (1,383,624 bytes)

Author(s): Emily H. Weaver, Siân Curtis, John Spencer, Gustavo Angeles

Year: 2020

Linking Data from Demographic and Agricultural Surveys to Examine the Drivers of Stunting and Wasting in Nigeria: Lessons Learned Abstract:

Stunting and wasting are still global issues, with an estimated 149 million children under five with stunted growth and 49 million children under five suffering from wasting worldwide. Wasting and stunting can have severe health effects on children and are therefore a major health concern for most low-middle income countries where stunting and wasting rates are highest (UNICEF/WHO/World Bank Group, 2019). Both stunting and wasting share underlying risk factors that derive from several different levels of influence. Existing studies focus on demographic and health indicators, such as those that are available in the Demographic and Health Surveys (DHS). However, additional influences on these outcomes are also agricultural and community-level indicators that are not included in conventional demographic and health surveys. Studies are needed to trial the linking of these data and to provide lessons learned for others seeking to do the same.

The increased availability of data from multiple sources in low- and middle-income countries in recent years, combined with advances in data science, have stimulated an increased interest in using existing data in innovative ways to bring new insights to population, health, and nutrition problems. MEASURE Evaluation was contracted to do just that—to conduct an analysis of publicly available secondary data using innovative linking methods to better understand a broader range of drivers of wasting and stunting, particularly in contexts with stagnant or increasing wasting levels and decreasing stunting trends. The study links data from the Nigeria DHS (NDHS) with a Living Standards Measurement Survey Integrated Survey on Agriculture (LSMS-ISA) that contains agricultural and community information. This study also sought to use machine learning to identify additional or unique patterns of indicators that influence stunting and wasting. Neither of these two methods are prevalent in current research; therefore, these analyses also serve as proof of concept for these two approaches and provide lessons learned for future research.