Optimal distribution of medical backpacks and health surveillance assistants in Malawi

Abstract

Despite recent progress, Malawi continues to perform poorly on key health indicators such as child mortality and life expectancy. These problems are exacerbated by a severe lack of access to health care. Health Surveillance Assistants (HSAs) help bridge this gap by providing community-level access to basic health care services. However, the success of these HSAs is limited by a lack of supplies and long distances between HSAs and patients. To address this issue, we used large-scale weighted p-median and capacitated facility location problems to create a scalable, three-tiered plan for optimal allocation of HSAs, HSA designated medical backpacks, and backpack resupply centers. Our analysis uses real data on the location and characteristics of hospitals, health centers, and the general population. In addition to offering specific recommendations for HSA, backpack, and resupply center locations, it provides general insights into the scope of the proposed HSA backpack program scale-up. In particular, it demonstrates the importance of local health centers to the resupply network. The proposed assignments are robust to changes in the underlying population structure, and could significantly improve access to medical supplies for both HSAs and patients.

Improving access to care and community health in Haiti with optimized community health worker placement

“…The parameter values in each of the four scenarios corresponding to Calculations were performed using the Gurobi commercial solver version 8.1 [46] and the lpsolveAPI R package [47,48], which retrieves the exact solution of the optimization problem. Although heuristic methods exist, it was chosen, as in Kunkel et al [18], to use an exact solver with a controlled tolerance (called "Gurobi gap" in [18]). Additionally, the optimization program was run for each department independently (with Artibonite, Centre and Reste-Ouest each sub-divided into two), in order to reduce computational burden, but also because the operational decision-making is done at the department level.…”

“…In particular, in Haiti, Parker et al used linear programming methods, not to study CHW geographical allocation, but to optimize how much time and resources CHWs should dedicate to different tasks in order to maximize health benefit [9]. In Malawi, Kunkel et al implemented p-median models to optimize CHW geographical allocation as well as the location of their commodity resupply in nearby health facilities [18]. Compared to p-median models, covering-based models [17] are well suited for universal coverage objectives, as they include a maximum distance constraint and hence exclude the possibility that a small number of individuals would need to travel a long distance in the optimal solution [16].…”

Optimizing the location-allocation problem of pharmacy warehouses: A case study in Gaziantep

“…Determining three location for healthcare facility in China Kunkel et al [26] P-median and capacitated facility location models Distribution of health resources in Malawi Guerriero et al [27] P-median, P-center, set covering models and mathematical formulation for network reorganization problem Public hospital network reorganization in Italy…”

Optimizing community healthcare coverage in remote Liberia

“…Cocking et al (2006) use a modified max covering model to locate health facilities in Burkina Faso. As noted in Kunkel et al (2014), much of the prior work on community healthcare networks focus on a single tier of facilities; however, Kunkel et al (2014) consider a multiple tier network of health surveillance assistances and medical backpacks in Malawi. The authors formulate both a decomposed model that designs two tiers separately and an integrated model in which the tiers are determined simultaneously.…”