New study  on  cluster  randomised controlled trials  for  mosquito  release interventions  such as gene drive 

New technologies for suppressing populations of disease-transmitting mosquitoes involve releasing modified male mosquitoes into field environments.

One  example of these technologies is the sterile insect technique (SIT) which releases male mosquitoes that have been sterilised using radiation treatment.  A second one is genetically modified mosquitoes, including gene drive mosquitoes, which are a potential candidate  tool for malaria control. 

Interest in mosquito release interventions is rapidly growing, spurred by recent global rises in cases of mosquito-borne diseases, and stalling progress in malaria control. Recent field trials have shown these technologies to be very effective in reducing dengue incidence, while very few field trials have yet been conducted to test how effective they could be in malaria reduction. 

In a new study titled  “Requirements for designing cluster randomised control trials to detect suppression of malaria vector”,  our  research teams  in  Burkina Faso and the United Kingdom  have  investigated how to  evaluate  the efficacy of an intervention in suppressing malaria vector populations  using  cluster  randomised controlled trials (CRCTs). 

Cluster randomised controlled trials (CRCTs) are the gold standard for assessing intervention efficacy, comparing intervention performance in a series of clusters or communities that receive it against a series of clusters that don’t receive the intervention and serve as controls. If the efficacy of mosquito release interventions could be demonstrated by a CRCT, this would be a major step forward for these new technologies for vector control. 

In  our  study, we used a rich data set of mosquito population densities collected from field sites in western Burkina Faso to develop statistical models to inform the design of CRCTs for suppressing malaria-transmitting mosquito populations. By  considering  the natural variability in mosquito densities over time, and between  different locations, we investigated how to efficiently design trials so that the suppressing effect of the intervention could be detected with minimal sampling effort.  We found that as few as 20-21 clusters  were required  to detect a suppression effect of 50%, while 9-10 clusters  were required  to detect a larger suppression effect of 90%. 

This research is  timely  in contributing to the design of field trials of gene drive releases, which are currently being actively planned. CRCT  designs do not need  a large number of  clusters to detect moderate to high suppression effects acting on  An. gambiae  malaria vector populations.  

A challenge in designing field trials of gene drive releases involves the spatial spread of gene drives, due to mosquito dispersal. This could potentially carry the gene drive from release sites into control clusters, causing “spillover”,  which could compromise the study design.  The next phase of our work will involve  developing mathematical models to predict the spatial spread of gene drives, and how this might affect the performance of CRCTs. This can further inform CRCT design  and help us  select a spatial arrangement of clusters  to  mitigate the influence of spillover  on  the  trial  results.