Determining a Bayesian predictive power stopping rule for futility in a non-inferiority trial with binary outcomes