Table 4 Challenges faced in modelling insecticide resistance in space and time

(a) Scarcity of data/few observations

 Scarcity and heterogeneous distribution of IR data in some regions

 In another context, few observations resulted in higher precision errors while using Bayesian geostatistical models to model the distribution of vectors in Mali

(b) Challenges in the estimation of various predictor variables

 There were challenges in the estimation of various predictor variables such as quantities of insecticides used in agriculture and where they are used

(c) Causality among the drivers of insecticide resistance

 Establishing causality among the drivers of insecticide resistance in malaria vector populations could be explored further because the variables interact. In addition, causal variables used to develop various models may not have been exhaustive, hence use of additional potential insecticide resistance drivers may result in more robust models

(d) Lack of standardization in the diagnostic tools

 Estimating insecticide resistance in un-sampled locations is hampered by a lack of standardization in the diagnostic tools used and by a lack of data in several regions for both resistance phenotypes and genotypes