Bridging the gap: leveraging data science to equip domain experts with the tools to address challenges in maternal, newborn, and child health

The United Nations Sustainable Development Goals (SDGs) advocate for reducing preventable Maternal, Newborn, and Child Health (MNCH) deaths and complications. However, many low- and middle-income countries remain disproportionately affected by high rates of poor MNCH outcomes. Progress towards the 2030 sustainable development targets for MNCH remains stagnated and uneven within and across countries, particularly in sub-Saharan Africa. The current scenario is exacerbated by a multitude of factors, including the COVID-19 pandemic’s impact on essential services and food access, as well as conflict, economic shocks, and climate change.

Traditional approaches to improve MNCH outcomes have been bifurcated. On one side, domain experts lean heavily on expert-driven analyses, often bypassing the advantages of data-driven methodologies such as machine learning. Conversely, computing researchers often employ complex models without integrating essential domain knowledge, leading to solutions that might not be pragmatically applicable or insightful to the community. In addition, low- and middle-income countries are often either data-scarce or with data that is not readily structured, curated, or digitized in an easily consumable way for data visualization and analytics, necessitating non-traditional approaches, data-driven analyses, and insight generation. In this perspective, we provide a framework and examples that bridge the divide by detailing our collaborative efforts between domain experts and machine learning researchers. This synergy aims to extract actionable insights, leveraging the strengths of both spheres. Our data-driven techniques are showcased through the following five applications: (1) Understanding the limitation of MNCH data via automated quality assessment; (2) Leveraging data sources that are available in silos for more informed insight extraction and decision-making; (3) Identifying heterogeneous effects of MNCH interventions for broader understanding of the impact of interventions; (4) Tracking temporal data distribution changes in MNCH trends; and (5) Improving the interpretability of “black box” machine learning models for MNCH domain experts. Our case studies emphasize the impactful outcomes possible through interdisciplinary collaboration. We advocate for this joint collaborative research approach, believing it can accelerate the extraction of actionable insights at scale. Ultimately, this will catalyse data-driven interventions and contribute towards achieving SDG targets related to MNCH.