Fig. 2From: Advantages of Bayesian monitoring methods in deciding whether and when to stop a clinical trial: an example of a neonatal cooling trialProbabilities of treatment benefit (log RR) on predischarge mortality for three experimental cooling groups. Negative values favor the experimental group. Deeper cooling (β 1; left panel), longer cooling (β 2; middle panel), and both (β 1 + β 2 + β 3; right panel) are compared with standard cooling (33.5 °C for 72 h). Top (bottom) panel shows the neutral (enthusiastic) prior and corresponding posterior probabilitiesBack to article page