Controversies and achievements in clinical trials

Abstract

In attaching the label “keynote” to my remarks, the Program Committee took a grave risk. Some works of music, like Beethoven’s First Symphony, start in the wrong key, and have to be wrenched back as soon as possible. Others start in the right key and end in the wrong one. Others have no key at all. All I can hope to do is to sound a moderately confident note which I hope will recur from time to time during the conference, no doubt interspersed with some surprising modulations. It is certainly a great privilege to be asked to open this meeting. The Society for Clinical Trials has made a remarkable impact during its first few years of existence, through its meetings and through its journal, Controlled Clinical Trials. This particular branch of scientific enquiry, the clinical trial, has existed in something like its present form for about 40 years, and yet it is only in the last 5 years or so that it has started to engender the traditional signs of a scientific discipline-a society, a journal, monographs, training courses, and the like. The Society, which is in the forefront of these developments, took the wise step of recognizing at the outset the collaborative nature of the clinical trial, encompassing the clinician, the statistician, the data manager-and of course the patient. We shall be discussing all these aspects during the meeting and it would be pointless for me to try to cover the whole field now. I happen to be a statistician, and I want to concentrate on some statistical aspects of trials, but in doing so of course I in no way imply that the other aspects play a subordinate part. There is, though, one sense in which statistics does occupy a key role in the concept of the clinical trial. Quite apart from all the technical apparatus randomization, stratification, significance tests, models for covariates, and the like-anyone participating in a clinical trial is committing himself or herself to a statistical investigation. That is, he or she believes that it is possible, by