School of Public Health - International Health
A Bayesian Network Learning Approach to Health System Forecasting and Visualization in Mozambique
Mozambique’s cadre of health workers and other human resources for health (HRH) are instrumental to the country’s participation in the United Nations Sustainable Development Goals (SDG) 2030 Agenda. The Instituto Nacional de Saúde (INS) is responsible for implementation of a surrogate strategy, the National Evaluation Program (NEP), a way to assess progress and HRH infrastructure through comprehensive data collection and generation of evidence based policy recommendations. However, pressure from the World Health Organization (WHO) to rapidly scale up current HRH infrastructure may prove to be detrimental, and recent studies highlight how unsustainable it is. More specifically, external funding appears to be a critical limiting constraint to where and how Mozambique’s HRH should be placed. This study aims to address this issue through the operationalization of Mozambique’s HRH theory of change via a Bayesian structural model. This model will be fitted to household survey data, health system human resources data, and spatially mapped data. From here, optimization algorithms will identify specific locations to implement HRH scale up that works within financial costs to maximize decreases in mortality. Such a tool may aid in INS’s decision making, as well as assist with simulating alternate world scenarios to compare different policy and program recommendations for the NEP. The long-term implications of this model could be far reaching, as it has the potential to uniquely inform resource allocation and health system forecasting in any country, given available data.
PI Mentor: Timothy Roberton