Insecticide resistance, particularly pyrethroid resistance in Africa, is increasingly recognised as one of the most important and growing barriers to effective malaria control. Despite the potential importance of insecticide resistance, there are still many unanswered questions about the evolution and spread of resistance, current variation in resistance, and the impact of this resistance on malaria control.
MAP-IR is collaboration between the Malaria Atlas Project and the Liverpool School of Tropical Medicine. The project team is:
The main aims of MAP-IR are to 1) to provide the best available estimates of resistance for all locations within a region to fill the gaps in the available field data,
2) use modelled patterns of resistance in time and space to investigate the relationship between these patterns and the coverage of potential drivers of selection for resistance, and 3) investigate the impact of historical trends in insecticide resistance on malaria transmission.
# Collating field data
All of the following data will be made freely available in the MAP online repository unless it is provided to us as confidential data. We will collate, geo-position (if required) and release data from WHO and CDC susceptibility bioassays and from tests for the mechanisms of resistance (e.g. genetic mutations or overexpression). All data must be linked to a field collection of mosquitoes from a specific site(s) and a specific date(s). Finally, data must be linked to an anopheline species, complex or subgroup, and the identification methods used should also be detailed.
# Modelling insecticide resistance
We will assess the volumes of data available and select geographical regions for inclusion based on the data volume. The cartographic approach used by the Malaria Atlas Project will be used to generate data that models patterns of resistance in time and space. Building on our earlier spatial modelling work, we have recently developed a model that (i) accounts for the non-random space-time distribution of survey data, varying survey sample size, season, and identification method; (ii) incorporates temporally dynamic spatial covariates to improve predictive accuracy and allow investigation of empirical associations; and (iii) provides robust metrics of uncertainty around both predictions and parameter estimates. This model will be tailored to measures of insecticide resistance. We will produce a sequence of annual maps, and maps describing, for each 5×5 km pixel, the magnitude and direction of estimated change between specified time periods together with 95% credible intervals. Regions with strong evidence of a decline, an increase, or no change in resistance will therefore be distinguished from those where data are simply insufficient to support assessment. The model outputs will include robust measures of uncertainty that can be propagated through to other analyses, and will be shared on a free and open access basis.
# The potential drivers of selection
The potential drivers of selection that will be considered are the use of insecticide-treated bednets, coverage of indoor residual sprays, larvicides and agricultural insecticides. Increases in the frequency of genes coding for minor resistance mechanisms may be selected for by the same factors that drive increases in operationally significant resistance in the population, even if the mechanisms coded for do not, themselves, produce a large-fold resistance, therefore, these mechanisms will also be included in the analysis.
# The impact of insecticide resistance
Once historical patterns of resistance have been defined, we will investigate whether they account, in part, for the unexplained variation in the transmission of Plasmodium falciparum malaria in Africa as characterised by the Malaria Atlas Project.