Assessing baseline comparability in randomised trials

摘要

Randomisation is used to allocate participants to groups (eg, anintervention group and a control group) in clinical trials becausethis process tends to generate groups with similar characteristics.1When the groups have similar characteristics at baseline, and onlyone of the groups then receives an intervention, the difference inoutcomes between the groups can therefore be attributed to thatintervention. Occasionally, randomised trials produce groups thatare not comparable at baseline, which can bias estimates of theeffect of the intervention.