A multi-level location-allocation problem in human milk banking network expansion

Neonatal infections are the leading cause for neonatal deaths in developing countries, resulting in more than 1.5 million infant fatalities annually. The most effective prevention against neonatal infections is exclusive breastfeeding. However, reasons such as maternal death during birth, maternal illnesses such as HIV and TB-meningitis, and lack of rooming-in facilities prevent the infant to be breastfed by his/her mother. One way to mitigate these infections is by supplying pasteurized donor-expressed breastmilk. In this paper, we consider the network expansion of the donated breastmilk distribution supply chain in South Africa. As with the distribution of most public sector and humanitarian relief goods and services, the transportation of donated breastmilk is hampered by the inherent uncertainty in the environment, and by the fact that in addition to efficient usage of resources, distribution should be made in an equitable manner. We incorporate uncertainty into our models by means of multiple scenarios, which are determined based on different assumptions about population size, HIV prevalence, and status of public health in the country, income, and education. We consider various equity-based objectives and propose rounding-based heuristics to solve these. We focus on two delivery schemes; one which uses out-and-back transportation, and one that makes multiple stops on the delivery route. Using computational experiments, we analyze the trade-offs between the objectives as well as the effects of various public health policies, network expansion budget, and assumptions on supply and demand. We also describe the teaching materials resulting from this paper, which include a case study, a supply/demand estimation tool, and an interactive decision support tool.