Constrained optimization: evaluating possible packages of community health interventions with competing resource requirements in Galmudug, Somalia

TL;DR

This paper uses constrained optimization to design an efficient community health program in Somalia that maximizes lives saved within limited resources.

Contribution

The study introduces a new optimization model for selecting maternal and neonatal health interventions under resource constraints.

Findings

An optimized package of interventions could avert 15% of projected maternal and neonatal deaths in Galmudug, Somalia.

Sensitivity analyses showed how changes in resources affect the optimal intervention combinations and lives saved.

The model provides a practical tool for health practitioners to evaluate community health programs under competing resource demands.

Abstract

Investment in community health worker (CHW) programs has allowed health systems to reach previously underserved rural and remote populations. As a result, CHWs are often burdened with responsibilities to deliver large packages of services, at times without sufficient human, financial, or health resources. To design a community-level program that saves maternal and newborn lives while operating within resource limitations, we used constrained optimization (a mathematical process for finding the solution to a stated objective while accounting for listed requirements) to construct a model for select villages in Galmudug State, Somalia. After establishing the resource requirements for delivering 25 evidence-based maternal and neonatal interventions, we used the Lives Saved Tool and optimization techniques to determine the package of care that leads to the most projected lives saved. With a cadre of 1450 female health workers and a budget of $435 000 for maternal and neonatal health commodities and programming over 1 year, we calculated that the optimized set of interventions for Galmudug could avert 15% of the 4132 projected maternal and neonatal deaths in 2024. We also conducted sensitivity analyses to show how the optimal combination of interventions and the number of lives saved change as the resource levels change. The model provides practitioners with a new tool and accompanying approach to evaluate possible packages of community health interventions with competing resource requirements.

Figures