Establishing the evidence and basis for drug policy change
Following the official transition from CQ to SP in 1998 the MoH's Division of Malaria Control (DOMC) and research partners maintained a series of surveillance studies on the sensitivity of SP and amodiaquine (AQ) [
9,
10], the recommended first and second-line treatments for uncomplicated malaria respectively. By 2001, the year the National Malaria Strategy was officially launched, concerns were raised about growing evidence of a decline in SP clinical efficacy as measured through the then standard WHO day 14 clinical and parasitological sensitivity test. By mid-June 2001, 6/15 (40%) studies undertaken by the DOMC and 3/6 (50%) by other partners showed that SP clinical and parasitological failure rates by day 14 were in excess of 25% (the WHO suggested change rubric of 25% failure rate is commonly used in the sub-region to inform antimalarial drug policy changes) [
10‐
13]. Conversely, there appeared to be better day 14 cure rates (≥ 75%) across studies where AQ was tested (19/20 studies) [
10,
13]. In June 2001 a meeting was convened by the DOMC and its partners to discuss strategies principally around how better to deliver medicines through the retail sector. However, the meeting also raised the urgent need to assemble the evidence on SP failure rates noting that "...
plans for the introduction of a replacement [were] now urgent..." [
14]. A second meeting was held just four months later to "...
review the national antimalarial drug policy and build a national consensus on malaria treatment..." [
15]; however, it wasn't until the final quarter of 2003, that the status of SP was deemed desperate requiring the formation of a national task force [
16‐
18]; at this time, seven out of nine studies (more than 75%) conducted between 2002 and 2003, including those examining patients through to day 28, showed SP failure rates in excess of 25% [
10,
19]. Whilst it was accepted that a change to a new first line therapy was urgently required, the possibilities for replacements were limited.
The decline in the clinical efficacy of SP in Kenya was happening within the context of an international push towards ACT in countries where monotherapies were failing [
1,
2,
21] at a time when attention was focused on increasing international funding for effective malaria therapies [
21]. The Global Fund to Fight AIDS, Tuberculosis and Malaria (GFATM) was launched in January 2002 as a financing mechanism for commodities and delivery of effective control of HIV/AIDS, TB and malaria. In January 2004, the publication of claims of "medical malpractice" by the GFATM for financing countries requesting drugs that weren't clinically efficacious heightened the debate on rapid deployment of ACT across Africa [
2]. Kenya was included in this criticism as they requested in September 2002 approximately two million US dollars (US$) worth of SP in their successful Round-2 GFATM bid [
22].
The first Kenyan Drug Policy Technical Working Group (DPTWG) meeting was convened on 6
th November 2003 [
16]. The working group comprised the MoH, non-governmental organization (NGO) partners, bilateral donors, representatives of the research community and technical support from the World Health Organization (WHO). Data on SP and AQ sensitivity were reviewed and possible first-line replacements considered. Options considered were: moving AQ from second-line treatment to first-line therapy; combining SP with artesunate; using the recently registered chlorproguanil-dapsone (LAPDAP
®) product; combining AQ with artesunate (AQ-AS); or using artemether-lumefantrine (AL). However, unlike the previous antimalarial drug policy change in 1998, there was a lack of any nationally generated, comparable sensitivity data, with the exception of AQ monotherapy, on the suggested alternatives. This posed a problem for informed choices by the DPTWG. The DPTWG felt that given the current pressure to avoid monotherapy and the short-term gains in combining SP with another partner drug eliminated these choices. Further, the WHO was unhappy about wide scale use of LAPDAP
® without further safety data [
23]. Thus, the choices were between AQ-AS versus AL. The need to assemble more information on AQ sensitivity and undertake carefully conducted open label, day 28 sensitivity studies of AQ-AS and AL were recommended to fill the information void.
There was only one clinical efficacy study undertaken in Kenya on AL available for review pre-publication in January 2004. This study was undertaken at Kilifi, on the Kenyan coast, as part of multi-country Phase III regulatory approval studies for the 6-dose AL recommendation [
24]. The interim results showed that over 95% of children attained adequate clinical and parasitological response by day 28 for both the supervised and unsupervised groups. The data generated from this study were presented to the DOMC during the 3
rd DPTWG who felt that one study was insufficient and insisted on a multi-site comparison of AL with AQ-AS. Deliberations on who would undertake these studies and how they would be funded continued through the 4
th and 5
th DPTWG meetings between January and March 2004 [
13,
19]. These studies were never completed. WHO provided technical support to the DOMC and provided examples of other countries, such as Zanzibar and Zambia, who changed policy using only international and regional evidence without comparative national data [
23]. The WHO also encouraged the DPTWG to consider the operational complexities of carrying out drug policy change and that issues of implementation should be defined early. In January 2004 the DOMC formed sub-groups of the DPTWG to address specific drug and non-drug issues related to the drug policy change, including therapeutic efficacy testing; legal issues; guidelines and formulations; logistics, procurement and supplies; case management; and later Information, Education and Communication (IEC) [
13].
Debate continued over the quality and amount of evidence available on the efficacy of AQ. Most of the day 28 studies undertaken by the DOMC under the auspices of the East African Network for Monitoring Antimalarial Treatment (EANMAT) were not corrected for possible new infections and that these data were at variance with data generated by the Africa Medical Research Foundation (AMREF) at other sites that showed high failure rates for AQ monotherapy [
13]. These discrepancies were never fully resolved and a view persisted that AQ was a valuable long-term combination partner by some members of the DPTWG despite evidence of AQ failure from neighboring countries [
10].
On the 3
rd March 2004 an urgent meeting of the DPTWG was convened in anticipation of the deadlines for the Round 4 GFATM application on the 5
th of April 2004 [
19]. It was clear that a decision had to be made on what ACT was to be announced as the replacement therapy for SP, an estimation of drug needs and costs was required and a strategy for delivery articulated and costed. A decision based on international, regional, sub-regional and country efficacy data and country experiences was taken at this DPTWG meeting in favour of AL, with oral quinine recommended as the second-line treatment. The decision was based largely on issues related to AL being the only co-formulated ACT at the time, doubts that AQ could be withdrawn from the informal sector as a monotherapy while deploying AQ-AS in the formal sector and the assumed rising levels of existing AQ resistance across Kenya. Concerns were, however, raised about the high cost of AL despite subsidized arrangements between WHO and Novartis Pharma AG [
3], and the notion that the global supply of artemisinin was in jeopardy [
19]. Both issues affected both AQ-AS and AL. It is notable that at the time of the final decision to adopt AL there were no comparative efficacy data on AL versus AQ-AS nor any substantiated DOMC generated data on day 28 AQ sensitivity.
On the 5
th April 2004 the "National Symposium on Next Anti-Malaria Treatment Policy in Kenya" was held at Naivasha. The DPTWG sub-committees and other key participants summarized their deliberations before the Minister for Health informed the gathering that, after negotiations with relevant bodies and development partners, Kenya had opted to change policy to the WHO recommended ACT and AL was now the recommended first-line treatment for uncomplicated malaria [
25]. The new policy was reiterated on the 25
th April 2004, Africa Malaria Day, in a speech delivered by the Deputy Director of Medical Services [
26]. From this point began the long process of policy implementation.
The GFATM Round-4 malaria proposal was submitted on the 5
th April 2004 and approved by the GFATM Technical Review Panel (TRP) in July 2004 amounting to over 188 million US$ over 5 years including key components of implementation of the new drug policy and scaling up insecticide treated net coverage. The funds were approved to be disbursed in two phases, the first amounting to over 82 million US$, including approximately 40 million US$ to cover the costs of procuring approximately 11 million treatment doses of AL annually; second, the balance of funds to be accessed following successful implementation of the first phase [
27,
28]. Kenya proposed to phase in the new drug policy, beginning with a two year (2006–2007) introduction of free distribution of AL through the public formal sector (government, NGO and mission health facilities), followed by its use within the private formal sector from 2008 and finally scaling-up to facilitate distribution through the private for-profit retail sector from 2009. The initial three year restriction to the formal sector was proposed as a way to build national confidence in the new drug and to establish a better pharmacovigilance profile before deregulating use for over-the-counter (OTC) in 2009. During this period, it was proposed that AQ be promoted as an alternative to AL in the retail sector [
29].
In summary, the decision to replace SP with AL as the nationally recommended therapy for uncomplicated malaria was a pragmatic choice. It was based on limited comparative scientific data (between AL and competing alternatives such as AQ-AS); legitimate international pressure to abandon as soon as possible drugs that were no longer efficacious; and the opportunity to fund more expensive medicines through an international financing mechanism. The GFATM announced on the 1st July 2004 that Kenya had been successful in its application for US$ 82 million for phase 1 of Round-4. However, the practicalities of a rapid translation of policy into practice were only fully appreciated after the policy was announced.
Delays in translating policy into practice
Despite several consultative meetings of the DPTWG and other WHO advisory meetings, an announcement by the Minister for Health and a successful application to the GFATM, there remained several concerns about the implementation of the recommendation to adopt AL as first line therapy. These revolved around long-term predictable financing, procurement and supply, ensuring adequate access across all service providers, the concerns of pharmaceutical manufacturers, and regulatory issues with regard to widening access to AL.
Financing, procurement and supply challenges
Concerns were raised regarding the long-term financial sustainability of the new policy, especially in light of the recent MoH's experience with the
Haemophilus influenzae type B (Hib) vaccine. Hib was introduced in the country in 2001 [
30] with an initial funding commitment from the Global Alliance for Vaccines and Immunization. It was initially agreed that at the end of the five-year funding cycle in 2006, the MoH would incorporate Hib in the Expanded Programme for Immunization. The agreement was based on optimistic estimates that the cost of the vaccine would come down substantially [
30,
31]. However, this did not happen and the MoH found itself saddled with an expensive intervention for which it did not have money. In light of this, the MoH viewed AL as an expensive first-line antimalarial drug replacement, costing at least ten times more than SP at the proposed subsidized prices, which could not be financed directly and solely by the MoH. In the absence of external funding, the cost of AL alone would absorb the entire MoH budget for rural drug supply [
32‐
35].
In June 2004, the Director of Medical Services (DMS) requested an economic evaluation of
not changing policy. In September 2004, the DMS called for an urgent meeting to discuss the AL drug policy decision where it was resolved that an official assurance should be sought from the GFATM that the new policy would not be jeopardized over the next five years because of a failure after two years to meet all the milestones set for implementation of non-drug policy related funding requests [
32]. A few weeks later, on 30
th September 2004, a meeting was called by the GFATM to offer countries technical assistance on re-programming committed funds from Round-2 funding for monotherapies. The Kenyan DMS asked for an unequivocal assurance on continued funding, however the GFATM maintained that it was unable to provide this assurance given the vagaries of its own donors [
36].
The economic analysis became an ever increasing concern for the MoH and was discussed at length during subsequent meetings of the DPTWG during the delay in signing the GFATM Round 4 agreement [
37,
38]. Some bilateral donors such as the UK Department for International Development (DFID) tried to allay the fears of financial sustainability by pledging funds to kick-start the purchase of AL on one hand and trying to lobby long term financial support for the policy from other partners on the other. The economic analysis was never undertaken and decisions to continue with AL were taken without this analysis.
Another challenge facing the DOMC and MoH was navigating the complex financial and procurement arrangements for AL [
39‐
42]. The main challenge centered on how to manage the financial flows to make sure funds were availed in time for orders to be placed and processed. Theoretically, funds would flow from the GFATM, to the principal recipient (Ministry of Finance), then to the MoH (sub-recipient). The MoH would, after consultation with the national procurement consortium established to manage the tendering and ordering of commodities purchased with GFATM funds, place an order with WHO to forward the order to the supplier (Novartis Pharma AG in Switzerland). As per GFATM rules a Local Fund Agent (LFA), KPMG Kenya, should verify disbursement requests and financial reports on behalf of the GFATM. Quarterly release of monies already approved under Rounds-2 and 4 for malaria were contingent on the LFA signing off on how well the funds disbursed during the previous quarter were used; this, in turn, was used as a proxy for meeting performance targets. Although the GFATM TRP approved the Round-4 proposal in July 2004, concerns were raised by the GFATM about meeting broad performance targets, local fund management and procurement related to funds allocated for HIV/AIDs, TB and malaria during Round-2. This delayed considerably the final signing of the Round 4 agreement which happened 9 months after the TRP approval [
43].
Reconciling the complexity of the ordering and procurement procedures delayed the final submission of the AL request to WHO. The request was finally made in June 2005 and approved by WHO within two days. In addition, the delays in flow of Global Funds continued to postpone the timing of programmatic activities such as health worker training on the new drug, planned around the drug's arrival in the country. To circumvent the complex process of paying for the drugs once the order had been placed, a decision was made by the MoH to have the GFATM directly pay Novartis Pharma AG for AL orders but required the approval of the Ministry of Finance (MoF). The MoH wrote to the MoF in September 2005 and approval for this arrangement was obtained in December 2005, six months after an order had been initially placed with the WHO. Following the signed GFATM Round-4 agreement and a means to disburse the funds to WHO, the original order was processed in February 2006 and the first AL supplies arrived in Kenya three months later. The process was reduced in subsequent orders to only three months from placing an order receipt, financial disbursement, to arrival of supplies in country.
National and international pharmaceutical interests
On 8
th April 2004, three days after the AL policy change was first announced, a regional meeting on AQ-AS was launched for approximately 13 francophone countries (which were using the product) with Kenya as the host. The publicity surrounding this meeting overshadowed Kenya's own policy change announcement. This caused confusion, especially in the media, as to what the new first-line policy was. Between April 8
th and 13
th 2004, media reports appeared to endorse AQ-AS as the first line drug for malaria [
44‐
46]. These reports were denied by the then head of the DOMC [
47], but confusion continued in the media where reports in the print media supported the use of competing artemisinin products, such as artesunate-mefloquine [
48] and even a two-day regimen of dihydroartemisinin-piperaquine [
49].
The special arrangements between WHO, Novartis Pharma AG and GFATM posed a problem under the notion of single sourcing of drugs, contrary to the public sector procurement rules that insisted upon competitive bidding for government contracts unless a case could be made for an exception [
50]. Even though WHO was the broker for AL supply, the tender process, paradoxically still had to be adhered to. This was highlighted in a claim by the Pharmaceutical Society of Kenya (PSK) in the national press accusing the WHO of a "monopoly" [
51,
52]. Tender documents, therefore, had to be prepared by the national GFATM procurement consortium in January 2005, with the final drafts sent to the Permanent Secretary, MoH, by February 14
th 2005 and placed in the local newspapers (February 18
th 2005) for the supply of malaria drugs [
53‐
55]. This insistence on adhering to the letter of the law even when there was a clear national need for an exception, further delayed the procurement of AL by the MoH.
The PSK made several representations through letters and position papers to the MoH, presentations to media houses, and organizing workshops questioning the choice of AL as first-line between August 2004 and April 2005 [
35,
56]. The DOMC explained its position in a rejoinder to the PSK in September 2004 [
32,
57] stating the technical reasoning behind the selection of AL. This did not change the position of PSK and its members, who have maintained their strident opposition to the new policy to date. There are large stocks of artemisinin monotherapies, SP and AQ in the Kenyan market [
58] and concerns among local manufacturers and importers of pharmaceuticals of losing a market share to a transnational company.
Regulatory issues regarding widening access to AL
AL remains a prescription-only-medicine (POM) in line with all artemisinin monotherapies and other combinations [
34]. For it to be availed in the non-premium, private retail sector, where a substantial number of Kenyans seek treatment for malaria fevers [
59], requires mechanisms to deregulate its POM status. The policy implementation plan acknowledges the need to deregulate AL for OTC use and develop some evidence based operational research to maximize the best use of this drug as an OTC medicine [
60]. However, theoretically the POM status would only be changed following the assembly of adequate Phase IV safety data. The Pharmacy and Poisons Board (PPB) will launch a pharmacovigilance system that will entail passive reporting of adverse reactions to all medicines, including AL and other antimalarial drugs, in the second quarter of 2007 (Mohamed A, personal communication).
The POM status of AL also affected its effective deployment in the formal health sector. According to the Pharmacy and Poisons Act, not all cadres of health workers can prescribe or dispense POM drugs; only medical doctors, dentists, clinical officers and graduate pharmacists (in emergencies only) are allowed to prescribe. Graduate pharmacists can dispense all prescription drugs, but pharmaceutical technologists can only dispense Part II poisons (the so-called pharmacy only drugs); Part I poisons can only be dispensed under supervision of a graduate pharmacist [
61]. This obviously posed a problem for lower cadre health workers, such as nurses, who manage the majority of malaria prescriptions and dispensing in most government and mission health facilities. The Act invests many powers in the Minister of Health who can make exceptions to the rules and subsidiary legislation for better service delivery. Normally, the MoH would issue a legal notice through the Kenya Gazette to cover such exceptions, but this has not happened to date.