Background: The probability of contracting malaria in a given individual is determined not only by the individual's characteristics, but also the ecological factors that characterize the level of human-vector contact in the population. Examination of the relationship between "individual" and "supra-individual" variables over time is important for understanding the local malaria epidemiology. This is essential for planning effective intervention strategies specifically for each location.
Methods: A retrospective cohort study was conducted, which followed a community-cohort of about 3,500 residents in seven hamlets along the Thai-Myanmar border between 1999 and 2006. Potential malaria determinants measured at different levels (temporal variables, individual variables, and hamlet variables) were incorporated into multilevel models to estimate their effects on an individual's risk of malaria attack.
Results: The monthly minimum temperature was significantly associated with the seasonal variation of malaria risk. An individual risk of malaria attack decreased by about 50% during the period that active surveillance was conducted; an additional 15% and 25% reduction of Plasmodium falciparum and Plasmodium vivax incidence, respectively, was observed after the use of artesunate-mefloquine combination therapy (ACT) for treatment of P. falciparum. Male children (age < 16 years old) were at highest risk of both P. falciparum and P. vivax attack. An increase in the hamlet's incidence of P. falciparum and P. vivax by 1 per 100 persons in a previous month resulted in 1.14 and 1.34 times increase in the risk of P. falciparum and P. vivax, respectively, among individuals in a particular hamlet.
Conclusion: In a small area with low malaria transmission intensity, the variation in mosquito abundance is relatively similar across the residential areas; incidence of malaria between hamlets, which reflects the community level of human infectious reservoirs, is an important predictor for the malaria risk among individuals within these hamlets. Therefore, local malaria control strategies should focus on interventions that aim to reduce the gametocyte carriage in the population, such as early detection and treatment programmes and the use of ACT for P. falciparum.