Pošalji zapis e-poštom: Reducing bias in open-label trials where blinded outcome assessment is not feasible: strategies from two randomised trials