Newton Opiyo, Gavin Yamey, Paul Garner

Malaria causes ill health and death in Africa. Treating illness promptly with artemisinin-based combination therapy (ACT) is likely to cure people and avoid the disease progressing to more severe forms and death. In many countries, ACT use remains low. Part of the problem is that most people seek treatment from the retail sector where ACTs are expensive; this expense is a barrier to their use.The Global Fund and other international organisations are subsidising the cost of ACTs for private retail providers to improve access to ACTs. The subsidy was initially organised through a stand-alone initiative, called the Affordable Medicines Facility-malaria (AMFm), but has since been integrated into the Global Fund core grant management and financial processes. To assess the effect of programmes that include ACT price subsidies for private retailers on ACT use, availability, price and market share. We searched the Cochrane Central Register of Controlled Trials (CENTRAL; 2015, Issue 1, The Cochrane Library, including the Cochrane Effective Practice and Organisation of Care (EPOC) Group Specialised Register); MEDLINE (OvidSP), EMBASE (OvidSP), CINAHL (EbscoHost), EconLit (ProQuest), Global Health (OvidSP), Regional Indexes (Global Health Library, WHO), LILACS (Global Health Library, WHO), Science Citation Index and Social Sciences Citation Index (ISI Web of Science) and Health Management (ProQuest). All databases were searched February 2015, except for Health Management which was searched November 2013, without any date, language or publication status restrictions. We also searched the International Clinical Trials Registry Platform (ICTRP; WHO), ClinicalTrials.gov (NIH) and various grey literature sources. We also conducted a cited reference search for all included studies in ISI Web of Knowledge, checked references of identified articles and contacted authors to identify additional studies. Randomised trials, non-randomised trials, controlled before-after studies and interrupted-time-series studies that compared the effects of ACT price subsidies for private retailers to no subsidies or alternative ACT financing mechanisms were eligible for inclusion. Two authors independently screened and selected studies for inclusion. Two review authors independently extracted data, assessed study risk of bias and confidence in effect estimates (certainty of evidence) using Grading of Recommendations, Assessment, Development and Evaluation (GRADE). We included four trials (two cluster-randomised trials reported in three articles and two non-randomised cluster trials). Three trials assessed retail sector ACT subsidies combined with supportive interventions (retail outlet provider training, community awareness and mass media campaigns). One trial assessed vouchers provided to households to purchase subsidised ACTs. Price subsidies ranged from 80% to 95%. One trial enrolled children under five years of age; the other three trials studied people of all age groups. The studies were done in rural districts in East Africa (Kenya, Uganda and Tanzania).In this East Africa setting, these ACT subsidy programmes increased the percentage of children under five years of age receiving ACTs on the day, or following day, of fever onset by 25 percentage points (95% confidence interval (CI) 14.1 to 35.9 percentage points; 1 study, high certainty evidence). This suggests that in practice, among febrile children under five years of age with an ACT usage rate of 5% without a subsidy, subsidy programmes would increase usage by between 19% and 41% over a one year period.The ACT subsidy programmes increased the percentage of retail outlets stocking ACTs for children under five years of age by 31.9 percentage points (95% CI 26.3 to 37.5 percentage points; 1 study, high certainty evidence). Effects on ACT stocking for patients of any age is unknown because the certainty of evidence was very low.The ACT subsidy programmes decreased the median cost of ACTs for children under five years of age by US$ 0.84 (median cost per ACT course without subsidy: US$ 1.08 versus with subsidy: US$ 0.24; 1 study, high certainty evidence).The ACT subsidy programmes increased the market share of ACTs for children under five years of age by between 23.6 and 63.0 percentage points (1 study, high certainty evidence).The ACT subsidy programmes decreased the use of older antimalarial drugs (such as amodiaquine and sulphadoxine-pyrimethamine) among children under five years of age by 10.4 percentage points (95% CI 3.9 to 16.9 percentage points; 1 study, high certainty evidence).None of the three studies of ACT subsidies reported the number of patients treated who had confirmed malaria.Vouchers increased the likelihood that an illness is treated with an ACT by 16 to 23 percentage points; however, vouchers were associated with a high rate of over-treatment of malaria (only 56% of patients taking ACTs from the drug shop tested positive for malaria under the 92% subsidy; 1 study, high certainty evidence). Programmes that include substantive subsidies for private sector retailers combined with training of providers and social marketing improved use and availability of ACTs for children under five years of age with suspected malaria in research studies from three countries in East Africa. These programmes also reduced prices of ACTs, improved market share of ACTs and reduced the use of older antimalarial drugs among febrile children under five years of age. The research evaluates drug delivery but does not assess whether the patients had confirmed (parasite-diagnosed) malaria. None of the included studies assessed patient outcomes; it is therefore not known whether the effects seen in the studies would translate to an impact on health.