7.1 Randomised Control Trials
Randomised Control Trials (RCT’s) are often regarded as the gold standard of evaluation, primarily because a clear link between cause and effect can be drawn. In summary, the design requires that participants in the study are randomly allocated to either experience the intervention or be allocated to a control group for the same duration. After the set period of the intervention, the outcomes of interest are measured in both groups and any effects seen in the intervention group relative to the control group are attributed to the intervention content. That is, any difference between the two groups is assumed to be caused by the intervention rather than other factors. This is because with a large enough sample any ‘other factors’ that may impact the outcome will be balanced across the two groups through random assignment to condition.
RCT’s are a powerful evaluation tool, however they are not always feasible to conduct. An ethical concern that may be raised is that the control group would not be receiving the intervention and thus may be left ‘at risk’ for the duration of the trial. However, to overcome this, it is suggested to use a wait-list control group who would receive the same intervention as the active group immediately after the study period and outcomes have been measured. An alternative for some intervention activity may be to have an ‘active control’ group who receive a different version of intervention activity. For example, if updating an intervention the active control group could receive the existing intervention whilst the trial group test the new intervention.
Furthermore, RCT’s can be costly and very time consuming, which would likely be disproportionate to smaller intervention activities. Additionally, running an RCT outside of controlled laboratory conditions, especially in a complex social system, can lead to great losses in validity. However, through partnership working, and with the correct resources in place, it is entirely feasible to run an RCT to evaluate KFRS fire safety interventions. There are also variations on the RCT design which may make the design easier to implement for larger scale intervention activity. One such example is the cluster RCT (cRCT), which randomly allocates groups of people to the trial vs. control conditions rather than individuals.
Example: cRCT with linked smoke alarm technology
Proposal to use a cRCT to evaluate home fire safety visits in high-rise buildings, in combination with new linked smoke alarm technology. In collaboration with a smoke alarm supplier, a new linked system of smoke alarms will be fitted in a selection of high rise housing association buildings in Kent. The new technology tracks activations, location of activation (e.g. which room), duration of activation, and potential severity of incident. Using this data, an algorithm predicts the likelihood of a serious fire incident occurring which would allow FRS’s to more accurately target home fire safety intervention activity. With this system installed, KFRS are also able to randomise the residents within buildings (clusters) to receive different intervention messaging to see whether different behaviour change techniques reduce activations and ultimately real incidents.