Vitamin and mineral deficiencies (VMD) result in a broad range of adverse health and functional consequences, and these deficiencies continue to be a public health concern in low- and middle-income countries. Despite the high prevalence and the adverse health outcomes of VMD, there is a lack of information from nationally representative surveys. With funding from the Gates Foundation, researchers from the Institute for Global Nutrition, the Center for Targeted Machine Learning and Causal Inference at the University of California Berkeley and the Micronutrient Forum aim to catalyze and facilitate the development of new predictive models of VMD prevalence as proxies for biomarker-based data on micronutrient deficiencies.
The overarching goal is to generate national and sub-national prevalence estimates for VMD validated against the prevalence of these same VMD determined by biomarker assessment. The VMD prevalence modeling exercise builds on multiple steps: 1) identify a large set of potential proxy measurements from conceptual frameworks and a landscaping of data availability; 2) clean, prepare, and spatially and temporally align data; 3) test if VMD can be accurately predicted from a large set of proxy measures; 4) explore if the proxy measures can reliably predict VMD across a diverse range of countries in Africa, the target region of the project’s pilot phase; 5) identify a parsimonious set of easily or commonly measured proxies that can predict VMD with limited loss of accuracy from the full model; 6) estimate uncertainty around predictions to identify countries or subnational regions that would most benefit from targeted VMD surveys; 7) disseminate the conceptual and analytical process, and, 8) engage with national stakeholders and decision-makers throughout the project and disseminate the results of this work along with geo-coded maps to identify areas at greatest risk of deficiency.
Contact: Sonja Hess, PhD