IS IT REALLY BAD? A Proposal for the Toxicity-Testing of Drugs By M. A. SCHNEIDERMAN* Presented December 4, 1962, New York City ABSTRACT Clinical research differs from laboratory research in three major areas: size of the sample, variability in the experimental animal and possible experimental pro- cedures. Within the limitations imposed by these differences, the clinician must strive to conduct a good ethical experiment. In the past, many ingenious attempts have been made to carry out acceptable experimentation. Most of the problems have not yet been solved, leaving the clinician still at loose ends about many of the things he would like to do. Some logically tight systems are available to him, and through very carefully pre-planned trials he may sometimes be able to answer the question, "Is it really bad ?" A few of these systems are described. T•E REVORTS during the spring and summer of 1962 concerning thalidomide and the recent headlines about Preludin (1-3) and other drugs have made us more toxicity conscious than ever before. The charges and counter charges have led the statisticians to examine their ability to plan experiments that will yield adequate safety information. We do not come out too well, but here is what I see available to us, now. Sometimes the newspaper accounts I saw reported things that looked like nonsense, or nearly nonsense. There were remarks like, "It has not been proved that Drug X caused the illnesses ascribed to it" or "It has been proved that Drug Y is thoroughly safe" or "There is no evidence that Drug P is responsible for deformities." These are nonsense for several reasons. First, I think we can agree that "proof" outside of deductive mathematics is almost impossible of achievement. Honest men can, and do, disagree as to what has been "proved." Second, these are not problems of "proof" with which we are faced. These are problems of decision and action. To talk of proof here is to drag red herrings across the trail. To talk of "re- sponsibility" seenas to me to be using language to cloud the issue. The New England •ournal of Medicine speaks of the "benign hesitancy" of vitally interested parties "to accept any causal relation (4)." New materials for entering the market can be looked at from the opposite poles. They can be considered safe until evidence accumulates to show * National Cancer Institute, Bethesda 14, Md. 227

228 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS that they are unsafe (innocent until proven guilty). Or, they can be con- sidered unsafe until evidence accumulates that they are innocuous (guilty until proven innocent). Neither of these approaches is wholly sensible, because both neglect important elements. The first approach might represent the inherent bias of the drug producer. The second is contained in the sound advice from the American Medical Association that pregnant women take only "essential" drugs during the first trimester of pregnancy. It is also rather well summed up in these remarks of Blackader (5): "[Speaking of the many new drugs] a few may have some slight value the majority, unfortunately, have little or none, and some appear to be distinctly harmful. Of the exact action of these drugs for good or evil we have at present little knowledge, except the statements, which are invariably much prejudiced, of the commercial houses introducing them. "Many of these so-called new drugs, introduced even by reputable houses, h.ave been shown by analysis to be merely a mixture of old and well-known drugs .... It appears to me, therefore, to be of the utmost importance that physicians should accept all unproved state- ments very cautiously and should still rely almost entirely on the standard official preparations of which the exact composition and phys- iological action is well known." This was not testimony given before the Kefauver Committee. It was not argument given to support a more vigorous Food and Drug Administra- tion. It was contained in a paper published fifty years ago, in the Canadian Medical ztssociation •7ournal. Why is neither of these approaches satisfactory? Something is being neglected. This something is the value judgment that must be made and weighed. What are the gains to be achieved--at what cost? Sometimes the issue is clear. Suppression of morning sickness is a minor gain. (I say this as an unbiased, never-pregnant male.) Phocomelia, even in a few cases, is a dreadful cost. The issue of whether it has been proved that thalidomide causes or is responsible for phocomelia is not the dominant ele- ment here. The high cost associated with a high probability associated with a small gain are enough to decide actions--whether an air-tight case has been made or not. In the Preludin problem we have a similar high cost, and small gain, but this time these are tied together by what today looks like a low probability. It is no wonder that people do not agree on what to do with this one. Aside from being eligible for membership in the American Medical As- sociation, how else does the physician differ from laboratory scientists? Primarily, in his daily practice he has to make decisions and take action in the face of (what he knows to be) incomplete evidence. We do not. We can delay. We can hesitate. We can back-track. We can go back to the laboratory to get more data. Most of the time he cannot. Fortunately,