Taking malaria transmission out of the bottle: implications of mosquito dispersal for vector-control interventions.

Most malaria transmission models assume enclosed systems of people, parasites, and vectors in which neither emigration nor immigration of mosquitoes is considered. This simplification has facilitated insightful analyses but has substantial limitations for evaluating control measures in the field. Here we show that mosquito dispersal can confound conventional approaches to evaluating malaria vector-control interventions, and explore this association with a model of two villages between which mosquito subpopulation exchange occurs. Exchange of mosquitoes between such subpopulations can readily explain the discrepancy between experimental efficacy measurements for insecticide-treated nets and their much lower apparent effectiveness when in use. Our results indicate that current approaches to assessing malaria interventions that confer community-level protection may be severely compromised by mosquito dispersal in many endemic settings. The true effectiveness of many vector-control methods may be much greater than previously appreciated and the application of such interventions should be consolidated into larger contiguous spatial units so that more effective local suppression of malaria can be achieved. Spatially explicit modelling formats that consider vector dispersal as a determinant of malaria transmission and control are needed urgently for rational planning and evaluation of efforts to roll back malaria.