THE PROGRESS OF EXPERIMENT: SCIENCE AND THERAPEUTIC REFORM IN THE UNITED STATES, 1900-1990

This book explores the origins of our contemporary system of drug regulation and the modern clinical trial. Marks illustrates the symbiotic relationship between the history of modern drug regulation and the history of therapeutic reform. Accompanying this history of public policy is a detailed account of changing experimental ideals and practices. Marks traces the history of therapeutic experimentation, from the "collective investigations" of the past century to the controlled clinical trial that emerged after 1950 as the paradigm of scientific experimentation. The result is the first general history of clinical research in the United States, a book that examines therapeutic experiments in a wide range of diseases, from syphilis and pneumonia to heart disease and diabetes.

What kinds of tests should be performed on a drug before its release? What kinds of professional skills are most appropriate for performing such tests, and how safe must a drug be before it is marketed? Harry Marks, a medical historian at Johns Hopkins University, shows that none of these questions are new; issues of authority, risk, and benefit have been raised since early in the 20th century, when the new "wonder drugs" gave hope that diseases would succumb to the weight of laboratory science.

The primary focus of the book is on the post-World War II era, when randomized clinical trials became the gold standard of clinical drug research. Historians and clinicians have tended to regard the introduction of powerful new statistical methods as part of the march of science, but Marks shows that this progress was not achieved without resistance. Much of the resistance stemmed from the requirement that large-scale drug tests be conducted according to standardized experimental protocols, a requirement that many physicians saw as compromising the doctor-patient relationship.

The point of the new experimental designs was to diminish bias (both "profligate" administration of drugs by physicians and exaggerated claims by drug manufacturers), but many physicians also resented the push for statistical controls, insofar as it diminished their discretion in deciding how to treat their patients. The Veterans Administration's 1947 experiments to test the therapeutic value of streptomycin against advanced tuberculosis, for example, ran up against the concern that the untreated controls would be deprived of the benefits of the drug. Many Veterans Administration physicians were not eager to treat their patients as experimental subjects, and the decision was eventually made to abandon the use of untreated controls. Marks does not mention it, but the timing of the decision (September 1947) suggests that ethical questions raised by the Nuremberg medical trials may have contributed to the decision.

The introduction of double-blind experiments shortly thereafter led to new kinds of ethical quandaries. Marks concentrates on two studies that ran into difficulties, for rather different reasons. The ambitious Diet-Heart Study, designed to assess the effects of fats and meats on arteriosclerosis, came to a premature end when it was realized that the people under study might alter their diets as new evidence of health effects came to be publicized. The University Group Diabetes Program ran into a different kind of problem when it turned out that patients receiving one of the study drugs (tolbutamide) were dying of heart disease at higher rates than patients in a control group. Those in charge of the study had to wrestle with an entirely new kind of ethical question in the context of clinical trials -- namely, how much evidence of the harm of a drug is required to halt a study.

The book is heavily documented, almost overly so. Footnotes occupy about half the text, giving it something of the flavor of an annotated bibliography or a doctoral thesis. There are also points where one wishes the topic could have been brought up to date. There are many ongoing controversies over the value of "evidence-based" medicine, fraud, and statistical standardization that deserve historical scrutiny. There is some discussion -- though more would have been better -- about efforts by consumer activists (largely inspired by the AIDS epidemic) to "fast-track" the process of drug approval, as well as recent congressional mandates to include women and minorities in clinical trials.

Marks's book is slow going at first, but the chapters on the struggles to introduce randomized clinical trials are impressive. The book is sure to change how we regard the origins of clinical trials; it also asks us to broaden our understanding of experimental ethics and to question the political effects of different kinds of disciplinary authority. Marks shows once again that there is no single voice for science, but also that statistical designs can never substitute for political engagement when it comes to the kinds of therapies we decide to bar or allow. There is no better history of the complex weave of science, politics, and ethics in the design of therapeutic research protocols.

Reviewed by Robert N. Proctor, Ph.D.
Copyright © 1998 Massachusetts Medical Society. All rights reserved. The New England Journal of Medicine is a registered trademark of the MMS.