Assessing baseline imbalance in randomised trials: implications for the Cochrane risk of bias tool | Synapse