Background
Systematic, long-term, and spatially representative monitoring of insecticide resistance in mosquito populations is urgently needed to quantify its impact on malaria transmission, and to combat failing interventions when resistance emerges. Resistance assays on wild-caught adult mosquitoes (known as adult-capture) offer an alternative to the current protocols, and can be done cheaply, in a shorter time frame, and in the absence of an insectary. However, quantitative assessments of the performance of these assays relative to the gold standard, which involves rearing larvae in an insectary, are lacking.
Methodology/Principal findings
We developed a discrete-time deterministic mosquito lifecycle model to simulate insecticide resistance assays from adult-captured mosquito collection in a heterogeneous environment compared to the gold standard larval capture methods, and to quantify possible biases in the results. We incorporated non-lethal effects of insecticide exposure that have been demonstrated in laboratory experiments, spatial structure, and the impact of multiple exposure to insecticides and natural ageing on mosquito death rates during the assay. Using output from this model, we compared the results of these assays to true resistance as measured by the presence of the resistance allele. In simulated samples of 100 test mosquitoes, reflecting WHO-recommended sample sizes, we found that compared to adult-captured assays (MSE = 0.0059), larval-captured assays were a better measure of true resistance (MSE = 0.0018). Using a correction model, we were able to improve the accuracy of the adult-captured assay results (MSE = 0.0038). Bias in the adult-capture assays was dependent on the level of insecticide resistance rather than coverage of bed nets or spatial structure.
Conclusions/Significance
Using adult-captured mosquitoes for resistance assays has logistical advantages over the standard larval-capture collection, and may be a more accurate sample of the mosquito population. These results show that adult-captured assays can be improved using a simple mathematical approach and used to inform resistance monitoring programs.