The Bias in Trials

John Sutter: Much of the bias in these trials is not so much in how the data is analyzed and how it's spawned, but how it's designed. The study with 60% dropout, you know, when you get into that sensitivity analysis, boy, your power to show anything is getting to near zero, especially with this enormous non-interiority margin," he says. "This is the medicine that I am the least interested in practicing"