Skip to main content
Fig. 2 | BMC Medical Research Methodology

Fig. 2

From: A re-randomisation design for clinical trials

Fig. 2

Analytical results showing the increase in power through a re-randomisation design for different re-randomisation rates. This graph shows the difference in power between an unadjusted analysis (ignoring patient-effects) and an adjusted analysis (accounting for patient-effects) for a re-randomisation design across different re-randomisation rates. The number of total observations is fixed at 200; the number of individual patients is calculated as: total observations/(1 + proportion of patients re-randomised). The ICC is 0.50, the treatment effect is 0.40, with the within-patient standard deviation set to 1

Back to article page