Inventory policies for public pharmaceutical distribution in Zambia: improving availability and geographic access equity for essential medicines

Gallien, J, Leung, N-H Z and Yadav, P (2017) Inventory policies for public pharmaceutical distribution in Zambia: improving availability and geographic access equity for essential medicines. Working Paper. Social Sciences Research Network.

Abstract

Problem definition: The inventory control problem associated with the distribution of essential medicines in Zambia involves seasonality and uncertainty in both demand and lead-times, heterogeneous delivery locations, and lost sales. Beyond the traditional performance metrics of system service level and inventory level, the equity of distribution across delivery locations must be considered.

Academic/practical relevance: Stockouts of life-saving medicines are currently widespread in sub- Saharan Africa including Zambia. The companion study Leung et al. (2017) suggests that inventory management is an important opportunity, but ignores the challenge of distributing limited inventory to multiple locations. The existing literature on inventory management does not consider all aspects of this problem.

Methodology: A simulation model generating inventory dynamics of anti-malaria products in 212 (17%) of Zambia's health facilities is used to compare the current and various proposed inventory policies. It leverages extensive data including digital photographs of facility stock reports and its stockout prediction accuracy is validated against independent field measurements.

Results: Inventory control constitutes an important opportunity to improve patient access to medicines in Zambia. The policies proposed in USAID DELIVER (2011b) and Watson et al. (2014) perform well for only two of the three metrics of system service level, facility inventory and distribution equity, and are sensitive to their base stock parameter. Our proposed optimization-based policy outperforms these policies, and seems robust to policy parameters and demand forecasts accuracy.

Managerial implications: The proposed optimization-based policy seems a desirable alternative to current inventory management practices and other proposed enhancements for Zambia's public pharmaceutical distribution system, and may outperform the so-called min/max rule in other settings. In distribution systems with heterogeneous delivery locations, widespread proportional inventory rationing rules may lead to substantial service level discrepancies between facilities. The performance metrics of average service level and distribution equity may be frequently at odds, prompting non-trivial design trade-offs and considerations.