What is the best way to choose the block sizes in a randomised clinical trial?
Randomisation for an ongoing trial is stratified by site (2 sites) and gender (m/f) for 4 strata - it uses random permuted blocks (block sizes 2-8) to randomise 1:1. The number of sites is increasing to 5 to increase the rate of recruitment (10 strata). There is expected to be uneven randomisations per strata (some sites will recruit more individuals). What is the best way to select the block sizes to reduce the chance of imbalance at the end of the trial?
