Fig. 4

Simulation results across different re-randomisation proportions We compared two methods of analysis: (a) an unadjusted analysis (ignoring patient effects); and (b) an adjusted analysis (accounting for patient effects using a mixed-effects model). The ICC was set to 0.50 for all scenarios. The estimated treatment effect was unbiased for both methods of analysis. Standard errors for the estimated type I error rate and power are 0.3 % and 0.6 % respectively (assuming true values of 5 % and 80 %)