The Epidemiology and Control of Malaria in a Low Malaria Transmission Setting along the Thai-Myanmar Border

Abstract

Background: Thailand is considered to be a good candidate for malaria elimination. Currently, malaria transmission is concentrated only in areas along the international borders where additional specific interventions are required. In an area along the Thai-Myanmar border, we have studied the epidemiology of malaria, in terms of the disease distribution and its determinants. In addition, we have developed mathematical models to critically evaluate malaria transmission related to different malaria control strategies. Methods: We conducted a retrospective cohort study using data of a community-cohort of residents living in seven hamlets along the Thai-Myanmar border between 1999 and 2006. Spatial and temporal clusters of P. falciparum and P. vivax cases were identified using scan statistics. Malaria determinants that characterize people (individual-level), place (hamlet-level), and time (temporal change) were examined using the generalized linear mixed model (GLMM). Mathematical models were developed to simulate the P. falciparum transmission in the area; an additional model was constructed to consider P. falciparum gametocyte dynamics related to primaquine (gametocytocidal agent) pharmacokinetics. Results: Malaria incidence in the area along the Thai-Myanmar border was clustered across space and time. The seasonal variation of malaria cases was best explained by temperature rather than rainfall. The inter-annual reduction of malaria incidence was significantly associated with the deployments of active surveillance and artemisinin-mefloquine combination therapy (ACT). The spatial clustering patterns of malaria cases in this small area could not be solely explained by the variation of geographical characteristics. An individual risk of contracting malaria was largely affected by the density of malaria cases occurring within a hamlet in the previous month, relative to the topographical characteristics of individuals' houses. In the transmission modeling, early detection and treatment with ACT and primaquine had a substantial impact on reducing malaria transmission; however, primaquine was most beneficial when the administration was delayed about 8 days following initial treatment. Conclusions: The knowledge of spatial-temporal clustering provides guidance for targeted intervention to be conducted more efficiently. Transmission from nearby human was the most significant factor for continuation of malaria transmission in the area. Reducing the gametocyte carriage rate through a modified treatment regimen should have a potential impact on malaria transmission.