Fig. 3From: Adaptive designs in critical care trials: a simulation studyPerformance of the designs under the null scenario. Panels a, b and c show the probability of making an early correct decision in percentage, the probability of making an early incorrect decision in percentage and the expected sample size using various designs with different adaptive methods (red: Haybittle-Peto, brown: Hwang-Shih-DeCani, green: O’Brien-Fleming, blue: Posterior probability approach, purple: Predictive probability approach) across different numbers of interim analyses conducted under the null scenario for the ADRENAL trial, respectively. Panels d, e and f show the same metrics for the NICE-SUGAR trialBack to article page