Randomisation
Randomisation was performed by a trial pharmacist using an online randomisation sequence generator and opaque envelopes. Participants were stratified into two arms based on their haemoglobin at time of screening, < 105 g.l-1 or ≥ 105 g.l-1. Study participants, clinicians, researchers, and statistician were masked to treatment allocation until after all analyses were complete.
After data review at trial completion, 14 ineligible participants were discovered to have been mistakenly enrolled and randomised. These participants were excluded from the analysis (Figure 1); this is not expected to cause bias. (26)