The need for randomised controlled trials

SPEAKER: The Global Challenge of Vector-borne Diseases and How to Control Them. This session focuses on the need for randomised controlled trials of new vector control interventions. The aim of this session is to understand the importance of randomised controlled trials of new vector control tools. And the objectives are firstly to understand the development process of new vector control tools, to understand why randomised controlled trials are considered the gold standard study design to evaluate new vector control tools, thirdly, to understand the roles of entomological and epidemiological outcomes in phase 3 vector control trials, and lastly, to know the importance of conducting a sample size calculation. The development of new vector control tools typically follows a stepwise process.

The first step is to come up with a intervention concept and a draft target product profile. Then lab assays are conducted to determine the mode of action of the product. After this stage, the product enters phase 2 testing in semifield and small-scale field trials with entomological outcomes.

If successful, the product can enter phase 3 field trials where the epidemiological effect of the intervention is measured. This is in order to determine the public health value of the product. Subsequently, there may be phase 4 studies or implementation pilots to determine how the products work under operational conditions. There are a number of different study designs available to estimate the efficacy of vector control tools. These studies can be either experimental or observational. Experimental studies such as before and after studies and randomised controlled trials identify participants, intervene with an intervention, and then measure the effects of the intervention.

In contrast, in observational studies such as cross-sectional studies, case control, and cohort studies, participants are identified, their characteristics recorded, including use of vector control interventions, and the investigators look for associations between the intervention and disease. Why is it a good idea to randomise? Randomization is when individuals or clusters are allocated to intervention and control groups in a random fashion, typically using a table of random numbers. If the number of experimental units is sufficiently large, randomization is expected to result in no systematic differences between groups. As a result, randomised controlled trials, provide the least biassed and most robust estimate of an intervention’s efficacy and are considered gold standard study designs for measuring the efficacy of new vector control tools.

One of the main reasons we randomise is because randomization can reduce confounding bias. A confounder is a variable that is associated with both the exposure and the outcome but that does not lie on the causal pathway between the exposure and the outcome. For example, in a study evaluating the effect of the use of long-lasting insecticidal nets on malaria incidence, socioeconomic status can confound the relationship between the exposure and the outcome. Randomization, as long as there are sufficient units, makes it more likely that socioeconomic status will be balanced between the two study arms. In randomised controlled trials, the unit of allocation can either be an individual or a cluster, which I mentioned earlier.

Which of these is chosen depends on the intervention and the study question you wish to answer. For example, environmental management or introduction of genetically-modified mosquitoes can only feasibly occur in large geographic areas while topical repellent can be allocated at an individual level. Studies which also want to evaluate the indirect effect of the intervention are typically allocated at cluster level, usually village or group of villages or a larger area. For example, in some circumstances, long-lasting insecticidal nets can result in mass killing of mosquitoes which can provide indirect protection to those that are not sleeping under bed nets. This slide describes how a cluster randomised controlled trial works. First, houses or clusters of houses are identified.

Then they are randomly allocated to receive either intervention, here in peach colour, or control, depicted here in blue. And then they are followed over time for the outcome of interest. What outcomes should be measured? Epidemiological outcomes are required to determine the public health benefit of new interventions and can be measured using a number of different approaches. Passive case detection is where detection of a case is triggered when study participants attend a health care provider because they are sick while active case detection is where cases are detected by study staff who visit participants at their homes or in the community and screen them for disease.

Alternatively, cross-sectional surveys can be used where study staff test all or a sample of participants at a fixed time point. Entomological outcomes are usually supportive only and secondary to an epidemiological outcome in phase 3 vector controlled trials. However, they are useful to evaluate the claimed and logical effect of a vector control intervention. It is essential that sample size calculations are conducted for both epidemiological and entomological outcomes. A sample size calculation determines the number of participants that need to be recruited to answer the study’s question, assuming a given level of power and level of statistical significance. Too small a sample size and the study may not be sufficiently powered to be able to show an effect.

So this is a really important issue. In summary, new vector control tools typically follow a stage development process. Several different study designs are available to measure the efficacy of a new vector control tool, but randomised controlled trials are considered the gold standard.

Randomization is used to reduce the introduction of bias as long as a sufficient number of units are randomised. Vector control trials are often cluster randomised, for example, at the level of a village or even a larger area. Measurement of epidemiological outcomes is essential to demonstrate the public health benefit of new vector control interventions.