Fig. 1From: Overestimation of benefit when clinical trials stop early: a simulation studyScatterplot of observed relative risk reduction versus true relative risk reduction for all simulated typical truncated trials where true benefit existed. Figure shows overestimates in purple and underestimates in orange. Overestimates are 8.5 times more numerous than underestimates. Superimposed is a histogram of observed RRR/true RRR to show the range of overestimationBack to article page