The costs of introducing artemisinin-based combination therapy: evidence from district-wide implementation in rural Tanzania

BACKGROUND: The development of antimalarial drug resistance has led to increasing calls for the introduction of artemisinin-based combination therapy (ACT). However, little evidence is available on the full costs associated with changing national malaria treatment policy. This paper presents findings on the actual drug and non-drug costs associated with deploying ACT in one district in Tanzania, and uses these data to estimate the nationwide costs of implementation in a setting where identification of malaria cases is primarily dependant on clinical diagnosis. METHODS: Detailed data were collected over a three year period on the financial costs of providing ACT in Rufiji District as part of a large scale effectiveness evaluation, including costs of drugs, distribution, training, treatment guidelines and other information, education and communication (IEC) materials and publicity. The district-level costs were scaled up to estimate the costs of nationwide implementation, using four scenarios to extrapolate variable costs. RESULTS: The total district costs of implementing ACT over the three year period were slightly over one million USD, with drug purchases accounting for 72.8% of this total. The composite (best) estimate of nationwide costs for the first three years of ACT implementation was 48.3 million USD (1.29 USD per capita), which varied between 21 and 67.1 million USD in the sensitivity analysis (2003 USD). In all estimates drug costs constituted the majority of total costs. However, non-drug costs such as IEC materials, drug distribution, communication, and health worker training were also substantial, accounting for 31.4% of overall ACT implementation costs in the best estimate scenario. Annual implementation costs are equivalent to 9.5% of Tanzania's recurrent health sector budget, and 28.7% of annual expenditure on medical supplies, implying a 6-fold increase in the national budget for malaria treatment. CONCLUSION: The costs of implementing ACT are substantial. Although drug purchases constituted a majority of total costs, non-drug costs were also considerable. It is clear that substantial external resources will be required to facilitate and sustain effective ACT delivery across Tanzania and other malaria-endemic countries.