The Affordable Medicines Facility-malaria (AMFm): are remote areas benefiting from the intervention?

Malaria burden remains high in sub-Saharan Africa and among various reasons for the sustained high burden of the disease in the region is the low uptake of the key malaria control interventions, including prompt treatment with recommended anti-malarial medicines for all the population in need, due in part to high cost of drugs [1]. The Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund) hosted a Pilot Phase of the Affordable Medicines Facility-malaria (AMFm) in 2008 to increase uptake of effective anti-malarial medicines [2, 3]. AMFm is a financing mechanism consisting of: (1) price reductions through negotiations with manufacturers of quality-assured artemisinin-based combination therapy (QAACT); (2) Global Fund subsidy to buyers, through a co-payment to participating manufacturers for purchases made by eligible public, private and non-governmental organization importers; and, (3) interventions to support AMFm implementation and promote appropriate anti-malarial use [2, 3]. Under the AMFm mechanism, approved public and private importers or first-line buyers buy ACT from manufacturers at the subsidized price ranging from US$0.005 to US$0.220 for a treatment course, and then distributed them through the standard public and private sector distribution channels [4].

In 2010, the Pilot Phase of the AMFm was launched in eight national level programmes in seven countries in eastern Africa (Kenya, Uganda and the United Republic of Tanzania), western Africa (Ghana, Niger and Nigeria), and southern Africa (Madagascar) [2]. About 155.8 million doses of QAACT financed through AMFm were delivered to participating countries from August 2010 to December 2011 [5]. In addition to the medicines supplied to countries, several supporting interventions were implemented, including population awareness campaigns, setting recommended retail prices and training providers [2] to ensure effective implementation of the intervention.

The Global Fund commissioned an independent evaluation (IE) of AMFm to assess the achievement of its four objectives of reducing QAACT prices and increasing QAACT availability, market share and use at the national level in each pilot [5]. The IE used a pre- and post-test design and documented the implementation process and context in each pilot independently. National outlet survey of outlets stocking anti-malarial medicines was conducted at the baseline (2009/10) and the endline (2011) in each pilot. While assessing the performance of the programme at the national level, the Global Fund was also interested in knowing if the intervention had reached disadvantaged groups, particularly people living in areas considered remote. For this reason, the Global Fund commissioned a remote area study as part of the IE.

Traditionally, areas are classified as urban and rural with the latter often considered remote. However, some urban areas may be more remote than other urban areas, while some rural areas are more remote than other rural areas. Some rural areas that are better connected than isolated urban areas in some countries. To account for this potential misclassification, for in this study remoteness was defined as a function of distance from population settlements to service centres.

The remote area study involved a supplementary sample of remote areas in the AMFm outlet survey at endline. It was carried out in two of the pilot phase countries considered fast moving in implementing the AMFm intervention. Choosing fast-moving countries would allow sufficient implementation intensity and time for the intervention to have plausibly reached remote areas and, therefore, for an assessment to be informative. Fast-moving countries were expected to have received the co-paid medicines, started the intervention and been implementing the supporting interventions for about 12 months before the endline survey. At the time of selecting countries for the study, from all indications, Kenya and Ghana were fulfilling these criteria. This expectation was borne out by the main evaluation results, which found that Ghana and Kenya were among those that achieved several AMFm benchmarks.

In Ghana, the availability of QAACT increased by 52 % points from baseline to endline, meeting largely the benchmark 1 of a 20 % point increase in QAACT availability. Similarly, the country met the benchmark 5 of 10 % point increase in market share of QAACT with an increase of 40 % points of QAACT market share from baseline to endline. The median price of QAACT with AMFm was three times less than the median price of the most popular anti-malarial, which is not a QAACT in tablet form, just missing the benchmark 2 less than three times.

Kenya met, largely, the benchmark 1 with an increase of 35 % points of availability of QAACT from baseline to endline. Market share of QAACT increased significantly by 31 % point, meeting the benchmark 5. Compared to Ghana, Kenya met the benchmark 2 (price) with the median price of QAACT with AMFm equal to the median price of the most popular anti-malarial, which is not a QAACT in tablet form [4, 5].

However, the question remained of whether the AMFm programme benefitted more remote populations. This paper examines the availability, price and market share of QAACT in remote areas compared with non-remote areas in Kenya and Ghana at the endline of the AMFm intervention.

The study compared the availability, price and market share of QAACT in remote areas and non-remote areas, defined as distance and travel time to major service centres, at the endline of the AMFm pilot phase in Ghana and Kenya. Although there had been considerable uncertainty about whether AMFm would reach remote areas, the results presented here confirm that remote areas benefitted from the intervention to a considerable extent.

Although the availability of QAACT was lower in remote areas than in non-remote areas, the medicines were widely available in remote areas in Ghana and Kenya at the endline. The availability of QAACT was particularly high in public health facilities; however, substantial private for profit outlets and most medicines were QAACT with the AMFm logo in both countries. Overall, the findings suggest that the AMFm programme was instrumental in making QAACT more available in remote areas in these countries. Yadav and colleagues reported similar finding in Tanzania where AMFm led to a large increase in the availability of QAACT with the AMFm logo in the country, with variation based on remoteness [10]. In Kenya and Ghana, reliable distribution systems, particularly in the private sector, could have played a major role in making the medicines available in remote areas. Ghana and Kenya are among the countries that met the AMFm benchmark indicator of a 20 % point increase in QAACT availability from the baseline to the endline [4].

The median price of QAACT (adult and children’s formulations) in remote areas and non-remote areas was similar in Ghana. On the other hand in Kenya, the median price of QAACT was US$0.46 in non-remote areas and almost zero in remote areas. The near-zero median price in remote areas in Kenya is driven by the substantial number of public health facilities where the drug is mostly free. In contrast, patients paid for QAACT in public health facilities in Ghana [5]. However, there was no difference in the median price of QAACT pediatric formulation in the public health facilities and private for-profit outlets between remote areas and non-remote areas in both countries. These findings are very comforting as for Ghana and Kenya children under-five are the most vulnerable to malaria infection.

At the inception of the AMFm, there were concerns the price of QAACT with AMFm logo will be higher than the recommended retail price in the private sector in remote areas due to potential challenges and additional cost for distribution, and limited competition. However, in Ghana and Kenya, the median retail price for QAACT with AMFm logo in private for profit outlets were in line with the recommended retail price (US$0.94) in Ghana and in Kenya (US$0.46). Despite the challenges and additional cost related to distribution of the medicines in remote areas, private sector provider mark-up was not excessive. These findings are in line with those by Yadav and colleagues in Tanzania mainland where they found no difference in price of QAACT with AMFm logo between remote areas and non-remote areas [9]. In Ghana and Kenya, several factors could have contributed to widespread compliance to the recommended retail price for co-paid QAACT; significant participation by the private sector in the AMFm implementation process in both countries is likely to have been a contributing factor.

In Ghana, although several anti-malarial medicines were on the market, for all sectors combined there was no difference between remote areas and non-remote areas in the market share of QAACT. However the market dominance of QAACT in the private for-profit sector, especially in remote areas, suggests that the increased availability and reduced price may have increased community awareness of QAACT, resulting in its higher level of (relative) sales. In contrast, in Kenya in the private for-profit outlets in remote areas, the market share was dominated by non-artemisinin therapy. The lower market share of QAACTs in the private for-profit outlets in remote areas is probably due to the higher price. Nevertheless, as in Ghana, Kenya achieved the AMFm market share benchmark of a 10 % point increase in market share of QAACT [4]. A possible explanation for the relatively low uptake of QAACT in RAs compared with nRAs in Kenya could be a lack of community awareness of the AMFm programme in remote areas.

There are a number of limitations that should be considered when using the findings. Remoteness was defined as distance and travel time to service centres. The assumptions underlying the travel times assigned to different land surfaces may vary by country-specific context and may therefore, overestimate or underestimate remoteness. Secondly, the outlet survey relied on self-reports of sales volume and prices with potential recall or reporting bias. The authors tried to minimize the recall bias by asking for reported sales volumes only for the week preceding the survey; however, recall may still have been imperfect [11]. Thirdly, the additional remote area outlet survey was conducted in the low malaria transmission season, and non-remote area data collected during a higher transmission season in both countries, therefore, some of the differences between remote areas and non-remote areas might be due to differences in data collection dates. Fourthly, the comparison of remote areas versus non-remote areas was only undertaken for Ghana and Kenya, which were the best-performing countries overall in AMFm; their experience in remote areas may have been different in countries with poorer AMFm outcomes. Finally, the lack of weights for Ghana may have created bias in the comparison between remote areas and non-remote areas in that country, as most of the larger outlets most likely to have QAACT were probably in non-remote areas.

Despite the challenges in geographical access posed by remote areas and the relatively short intervention period, the results show that the AMFm intervention has been able to reach remote areas in Ghana and Kenya. Although this cross-sectional evidence does not permit strong inference about changes over time, it is likely that the availability of QAACT increased in remote areas, as the majority of QAACT were those with the AMFm logo, not available at IE baseline. However, even though co-paid medicines was available in remote areas in the private for profit sector, availability remained lower, price higher, and (in Kenya) market share was lower than in non-remote areas.

The health care delivery system could learn from this success and the challenges in the private sector, and could adapt a similar approach to adequately cover remote and hard-to-reach populations with other interventions. However, further examination of the factors accounting for the deep penetration of the AMFm programme into remote areas is needed to inform actions to improve the health sector commodities distribution systems.