A nationally representative sample of all outlets that could sell or provide antimalarials to a consumer was taken through a census approach in 38 clusters across two malaria‐endemic strata, urban and rural, in Madagascar. A cluster sampling approach was used because there were no reliable lists of all outlets stocking antimalarials. Clusters were communes, with an average of 10,000 to 15,000 inhabitants. Clusters were selected with probability proportional to size (PPS)—a sampling technique in which the probability that a particular commune is selected is proportional to its population size. Oversampling of public health facilities, pharmacies and drug stores (dépôt de médicament) was conducted in administrative districts surrounding the selected clusters. The sample size was powered to detect a change of 20% percentage points in availability of ACTs over time.
Outlet inclusion criteria for this study included outlets which stocked an antimalarial at the time of survey or in the previous three months. An outlet is defined as any point of sale or provision of commodities for individuals. Outlets included in the survey are as follows: 1) public health facilities (government hospitals and health centres); 2) pharmacies (pharmacies licensed by the National Drug Authority and Pharmacists’ Council) 3) drug stores (depot de medicament) (licensed by the National Drug Authority and Pharmacists’ Council); 4) private health facilities (private clinics, private practices, NGO health centres and dispensaries); 5) grocery stores (épicerie [small groceries], épi bars [small groceries with bars] and épi gargotes [small groceries with food stalls]);6) Community health workers (Agent de Santé à Base Communautaire and Agent de Vente à Base Communautaire); and 7) other outlets (gargotes [food stalls with sit‐down eating areas], bars, and other outlets).