Taxonomy of Deep Data Knowledge Integration and Modeling For Precision Global Health
Growth faltering and impaired neurocognitive development affect millions of children globally. Abundant data about pathways and risk factors that impact birth, growth, and neurodevelopmental outcomes are dispersed across multiple previous studies, publications, and principal investigators. The Healthy Birth, Growth, and Development knowledge integration initiative was established to provide data-driven solutions to key questions by creating a large, integrated knowledge base for improved statistical power and multidisciplinary deep data analysis and modeling. The multidisciplinary team included > 120 experts in data curation, statistical analysis, modeling, visualization, public health, medicine, and knowledge translation. Experts were organized into 11 surge teams to focus initial modeling and methodology development including data management and visualization, auxology, fetal growth, and ontology. Subsequently, teams were reorganized to focus deeper modeling and promote cross-pollination of expertise about life cycle and neurocognitive outcomes, policy, and knowledge translation. Modeling and deep data analysis provided new data-driven insights about drivers of within- and between-child variability of growth, neurodevelopment, and recovery. Unique challenges included data sharing barriers, limited previous models, maximizing collaborative learning between diverse experts, multifactorial complexity of somatic growth and neurocognitive development, limited biomarker data, and heterogeneity of term definitions and geographic and sociodemographic settings. The diversity of factors affecting growth and development necessitated a customized, precision global health approach, defined as the delivery of the right interventions at the right dosage (“what”) for the right child (“who”) at right time (“when”) using the right distribution channels (“how”), to promote the right response and avoid adverse events.