Antimalarial drug resistence and artemisinin based combination therapy : a bio-economic model for elucidating policy choices
Antimalarial drug resistance is a major cause of the increasing burden due to P. falciparum malaria. Artemisinin-based combination therapies (ACTs) are now recognised to be the ideal choice for the first-line treatment of uncomplicated malaria, in order to achieve two beneficial outcomes: improvement of treatment efficacy and delay in the development of drug resistance. However uncertainties remain about the current and future benefits, risks and costs of ACTs and in particular how these outcomes are affected by differences in malaria epidemiology, health care settings, human behaviour and implementation strategies. This thesis seeks to address these uncertainties by creating a comprehensive, dynamic, bio- economic model of malaria transmission and the spread of drug resistance, which incorporates vector factors, human immunity, human behaviour, drug characteristics and costs. Central to the model is a biological model, developed in collaboration with a mathematician, which outputs the proportion of drug resistant infections and the incidence of new and recrudescent infections. Parasite biomass is also tracked in order for human "infectiousness" to be measured and fed-back into the model. Sub-models are used to calculate severe malaria, deaths, costs and cost-effectiveness. Data were obtained to develop and populate the model. This included a community drug usage survey in Cambodia, which was undertaken in order to document the adherence and coverage rates to ACT following the implementation of locally blister-packaged ACT. Coverage was found to be extremely low, and the use of artemisinin derivatives on their own was widespread. However, both of these outcomes were improved by interventions to increase coverage, particularly village malaria volunteers. Application of the model in a low transmission setting suggests that with a 10-year time-frame, switching from monotherapy to an ACT is very cost-effective and results in overall cost savings in a range of scenarios. High coverage rates with an ACT are required to delay the spread of drug resistance if resistance has already arisen to one of the partner drugs. Running the model with data from Cambodia suggests that even in settings with low coverage, the change will be cost-effective and significant benefits are gained from the implementation of the specific delivery interventions. Strategies for optimising the implementation of ACTs are discussed in light of the findings.