Abstract
Insecticide resistance in malaria vector has increased in Africa. Implications for the effectiveness of the most widely implemented vector control (VC) techniques are questioned. To manage resistance, two VC strategies combining two different insecticides were evaluated during a randomised controlled trial in Benin. These strategies did not show a better efficacy than the baseline intervention recommended by the National Malaria Control Program (insecticidal nets alone). The aim of this thesis was to identify operational, biological and environmental factors that could explain this failure. We modelled the risk of exposure to malaria vectors according to environmental factors to (i) identify the determinants of the vectors' presence and density and to (ii) analyse the impact of VC strategies. Our results showed high spatial-temporal heterogeneities of vector populations at the village scale. The vector populations were characterized by different ecological and behavioural niches which could have impacted the effectiveness of VC. We highlighted the importance of An. funestus in malaria transmission during the dry season when people are less likely to use nets because of hot temperatures and low mosquito biting nuisances. We also observed changes in vectors biting behaviour following the implementation of VC. We therefore provide suggestions to improve actual VC tools, to define characteristics of new tools and to better select tools to implement in the field according to the entomological facies.