DOI link for Multi-Arm Trials
Multi-Arm Trials book
Leaping ahead a couple of millennia from Daniel, we ﬁnd “the ﬁrst deliberately planned controlled experiment ever undertaken on human subjects” (Stuart and Guthrie, 1953), a six-arm trial with two patients per arm. The study is a tremendous improvement over the biblical trial. It was painstakingly planned, including eﬀorts to eliminate bias (except that two of the worst got sea water), and was reported in suﬃcient detail to judge the quality. Despite the pitifully small sample size, the correct conclusion that citrus prevented scurvy
his treatments a cure. We, having to live with modest treatment eﬀects and high variability, need to consider the problems in conducting multi-arm trials.