The majority of patients requiring heart valve replacement in low- to middle-income countries (LMICs) need it for rheumatic heart disease (RHD). While the young age of such patients largely prescribes replacement with mechanical prostheses, reliable anticoagulation management is often unattainable under the prevailing socioeconomic circumstances. Cases of patients with clotted valves presenting for emergency surgery as a consequence of poor adherence to anticoagulation control are frequent. The operative mortality rates of reoperations for thrombosed mechanical valves are several times higher than those for tissue valves, and long-term results are also disappointing. Under-anticoagulation prevails in these regions that has clearly been linked to poor international normalised ratio (INR) monitoring. In industrialised countries, safe anticoagulation is defined as >60%–70% of the time in the therapeutic range (TTR). In LMICs, the TTR has been found to be in the range of twenty to forty percent. In this study, we analysed >20,000 INR test results of 552 consecutive patients receiving a mechanical valve for RHD. Only 27% of these test results were in the therapeutic range, with the vast majority (61%) being sub-therapeutic. Interestingly, the post-operative frequency of INR tests of one every 3–4 weeks in year 1 had dropped to less than 1 per year by year 7. LMICs need to use clinical judgement and assess the probability of insufficient INR monitoring prior to uncritically applying Western guidelines predominantly based on chronological age. The process of identification of high-risk subgroups in terms of non-adherence to anticoagulation control should take into account both the adherence history of >50% of patients with RHD who were in chronic atrial fibrillation prior to surgery as well as geographic and socioeconomic circumstances.