Log in with your email address username.


[Comment] ISIS and the emergence of large, simple trials

By the 1970s randomised trials had become quite common but almost all were small, certainly too small to give reliable answers to many important questions. At that time, sample size was rarely determined in relation to the ability to detect a clinically important difference; such considerations began to appear in trial reports in the 1960s.1 A review of 132 cancer trials showed that the median sample size was less than 50 participants per treatment group and only two of the reports discussed statistical power.