Table 2 Summary of identified biases
Feature of trial setting | Direction of bias | Cause of bias |
|---|---|---|
Transmission intensity (EIR) | Downwards | At higher transmission intensities, there is more frequent reinfection and reduced detectability of recurrent infections |
Relapse rate | Downwards | With a longer time to first relapse, fewer trial participants relapse during follow-up |
Duration of follow-up | Downwards | With a longer duration of follow-up, more trial participants have an observed reinfection by mosquito biting |
Duration of prophylaxis | Upwards | With a longer duration of prophylaxis, fewer trial participants are reinfected by mosquito biting |
Efficacy metric | Downwards | Measured efficacy is lower when based upon incidence rates or proportion at risk than upon proportional hazards |
Assay sensitivity | Downwards | For incidence rates and the proportion at risk, a more sensitive assay detects more recurrent infections |
Intervention action | Upwards | Efficacy based upon proportional hazards is biased upwards for an “all-or-none” intervention, because the “all” group increases the person-time at risk |