Improving anticoagulation in sub‐Saharan Africa: What are the challenges and how can we overcome them?

Patients in sub‐Saharan Africa generally have poor anticoagulation control. We review the potential reasons for this poor control, as well as the potential solutions. Challenges include the affordability and centralisation of anticoagulation care, problems with access to medicines and international normalised ratio monitoring, the lack of locally validated standardized dosing protocols, and low levels of anticoagulation knowledge among healthcare workers and patients. Increasing numbers of patients will need anticoagulation in the future because of the increasing burden of noncommunicable disease in the region. We propose that locally developed “warfarin care bundles” which address multiple anticoagulation challenges in combination may be the most appropriate solution in this setting currently.


| INTRODUCTION
Anticoagulation is used to treat and prevent venous thrombosis and to prevent intracardiac thrombosis due to some structural heart diseases or dysrhythmias. Depending on the indication, treatment may had mean TTRs of 55%, 1 46% 2 and 58%. 3 In an atrial fibrillation (AF) registry mean TTR in nine African countries was 33% (vs 62% in 19 Western European countries) 4 and in a rheumatic heart disease registry predominantly conducted in African countries PTR was 28%. 5 A number of smaller observational studies from SSA are presented in Table 1; in all but one of these PTR or TTR ranged between 28% and 52%.
In this review, we explore potential reasons for this poor anticoagulation control in SSA and explore strategies to overcome these challenges.
T A B L E 1 Observational studies from sub-Saharan Africa which measured time in therapeutic range (TTR)

| METHODS
For this narrative (nonsystematic) review, we searched Pubmed and Africa-Wide Information through EBSCOhost, using combinations of index terms (eg, "anticoagulants", "Africa South of the Sahara") and free text (eg, "anticoagulation", "warfarin", individual country names).

| CHANGING LIFE EXPECTANCY AND DISEASE BURDEN
SSA is undergoing an epidemiological transition: with increasing life expectancy the increasing burden of noncommunicable disease is colliding with the pre-existing burden of infectious diseases. In the context of anticoagulation, this transition is evident in the increasing prevalence of (nonvalvular) AF, adding to the large number of people requiring anticoagulation for valvular heart disease, which in SSA is still mostly caused by rheumatic heart disease. 23  A recent systematic review suggested that AF prevalence in SSA may be higher than previously thought, 33  Rheumatic heart disease prevalence among school children in SSA is 1.5-3.0%. [42][43][44][45][46][47] In rheumatic heart disease registries, only two-thirds of patients with an indication for oral anticoagulation received it. 5,37,48 Venous thromboembolism (VTE) epidemiology in SSA has not been well described, 49 but HIV infection is a well-established risk factor, associated with a 1.5-fold increased hazard (95% confidence interval 1.1 to 2.0). 50 As the HIV pandemic epicentre, HIV-associated F I G U R E 1 Example observational studies from sub-Saharan Africa which describe resource limitations VTE is commonly seen in SSA. In all the studies in  71 and are often beyond the means of patients and public health services. In South Africa, which has statutory private sector medicine ceiling prices, 1 month's supply of dabigatran, apixaban or rivaroxaban costs the equivalent of approximately 60 hours' minimum wage. 72 To our knowledge, no DOAC has yet been shown to be costeffective in any SSA country's public health sector. 73 A weak medicines regulatory environment in much of SSA means that substandard and falsified medicines may be found on the market. 74 We are not aware of any SSA data on the quality of warfarin on the market, but in a medicines quality assessment across 10 West and Central African countries, there was fortunately no evidence of poor quality acenocoumarol. 75 Dedicated anticoagulation clinics using standardized approaches may achieve better anticoagulation control than routine models of care, where anticoagulation patients are seen as part of the general patient mix. [76][77][78][79][80][81] Such anticoagulation clinics are, however, not common in SSA; anticoagulation is often managed in outpatient cardiology, cardio-thoracic surgery and haemato-oncology clinics 20 and in a Kenyan service the mean age was 43 years. 20 Younger patients often show reduced adherence compared to older patients. 114 Two possible reasons for this are that they are economically active and therefore may be unable to attend follow-up appointments, and that they may have reproductive wishes and expectations and therefore intentionally reduce their intake of a potentially teratogenic medicine. 115,116 Four studies reporting patients' self-reported adherence to anticoagulants in SSA are summarised in Table 4. Notably, from these, fewer than half of patients considered themselves highly adherent to warfarin. One study suggested that warfarin nonavailability may contribute to poor adherence. 64 In an analysis of the "care cascade" of rheumatic heart disease patients in Uganda, retention in care was the stage with the highest patient drop-out. 48 We are not aware of any studies reporting SSA patients' attitudes and beliefs about anticoagulation. A few studies (Table 5) 129 In SSA such a warfarin care bundle must include both process-centred and patient-centred activities, the exact combination of which will be specific to each setting and will depend on cost-effectiveness to guide rational allocation of limited resources. In making process changes, it will be important to leverage off existing successful systems, such as HIV treatment programmes, 130  have to make to the clinic. 87 However, the high cost of test strips is problematic, and relying on donations and subsidies is not sustainable.
Localised dose initiation and dose adjustment algorithms must consider the comorbidities and potential drug interactions that are prevalent in SSA, such as HIV, tuberculosis, antiretrovirals, antituberculosis therapy, co-trimoxazole and herbal/traditional medicines. These algorithms must be easy to implement, for example be paper-based, and should recommend small, percentage-based dose adjustments. 103 One example of an effective anticoagulation programme combining multiple interventions has come from Rwanda. 133 In this programme, specialist noncommunicable disease nurses deliver postoperative care to valvular heart disease patients in decentralised clinics. Standard dosing algorithms are used, while nurses are supervised and supported by cardiologists, using mobile communications.
Adherence support, as well as financial support, is offered to patients.
While this small study did not report the effect of this programme on TTR, low mortality was described and there were no bleeding or thrombotic complications. 133 DOACs may be a solution to some anticoagulation challenges in SSA: while they are still prohibitively expensive to most African patients and healthcare systems, they will in the future come off patent and generics may be more affordable.