Charles M. Beasley, Jr and Roy Tamura: What We Know and Do Not Know by Conventional Statistical Standards About Whether a Drug Does or Does Not Cause a Specific Side Effect (Adverse Drug Reaction)
2. Introductory comments
Edward Shorter’s comments
Well, it would be hard to offer critical comments on these introductory thoughts. The difference between adverse effects not related to the drug and adverse drug reactions seems elementary. I would offer only the thought that there are other techniques for establishing the relationship between a drug and putative side effects, such as challenge-dechallenge-rechallenge. This, too, will answer the question without a lot of statistical complexities. We are hung up on RCTs simply because the US Congress appears to have mandated them in the Kefauver-Harris legislation of 1962 (the legislation used the phrase “well controlled” rather than “RCTs”). RCTs are not the word of God.
One further thought: in the world of drug litigation, retrospective data from RCTs must be analyzed rather than prospective data (as in challenge-dechallenge-rechallenge). It would be unfortunate if our discussion of these matters were entirely guided by the exigencies of litigation.
May 9, 2019