Abstract
BACKGROUND: Studies have shown that supply chain costs are a significant proportion of total programme costs. Nevertheless, the costs of delivering specific products are poorly understood and ballpark estimates are often used to inadequately plan for the budgetary implications of supply chain expenses. The purpose of this research was to estimate the country level costs of the public sector supply chain for artemisinin-based combination therapy (ACT) and rapid diagnostic tests (RDTs) from the central to the peripheral levels in Benin and Kenya. METHODS: A micro-costing approach was used and primary data on the various cost components of the supply chain was collected at the central, intermediate, and facility levels between September and November 2013. Information sources included central warehouse databases, health facility records, transport schedules, and expenditure reports. Data from document reviews and semi-structured interviews were used to identify cost inputs and estimate actual costs. Sampling was purposive to isolate key variables of interest. Survey guides were developed and administered electronically. Data were extracted into Microsoft Excel, and the supply chain cost per unit of ACT and RDT distributed by function and level of system was calculated. RESULTS: In Benin, supply chain costs added USD 0.2011 to the initial acquisition cost of ACT and USD 0.3375 to RDTs (normalized to USD 1). In Kenya, they added USD 0.2443 to the acquisition cost of ACT and USD 0.1895 to RDTs (normalized to USD 1). Total supply chain costs accounted for more than 30% of the initial acquisition cost of the products in some cases and these costs were highly sensitive to product volumes. The major cost drivers were found to be labour, transport, and utilities with health facilities carrying the majority of the cost per unit of product. CONCLUSIONS: Accurate cost estimates are needed to ensure adequate resources are available for supply chain activities. Product volumes should be considered when costing supply chain functions rather than dollar value. Further work is needed to develop extrapolative costing models that can be applied at country level without extensive micro-costing exercises. This will allow other countries to generate more accurate estimates in the future.