Placebos Under Attack
The fact that researchers and research subjects could differ in their understandings of equipoise was unlikely to lead to controversy in comparisons between two active treatments, one old and one new. But comparisons between a potentially active drug and an inert placebo were far more capable of sparking an uproar among patients with a life-threatening disease. The use of placebos in the Phase II AZT trial was one of the first such cases to be criticized by AIDS treatment activists. In blunt terms, in order to be successful the study required that a sufficient number of patients die: only by pointing to deaths in the placebo group could researchers establish that those receiving the active treatment did comparatively better. Furthermore, to avoid introducing confounding variables into the study, the protocol forbade participants to receive other medication during the study. All this made a certain sense from the standpoint of experimental design, but it was difficult to justify to those people occupying the dual social roles of "patient" and "research subject"—people who began with the assumption that the purpose of medicine was to help them. Researchers insisted that clinical trials should not be confused with treatment—that being a research subject is not the same as being a patient. But this was a difficult distinction to put across in the best of circumstances, and it did not resonate with people with AIDS who were fighting to stay alive. In essence, the same practices and procedures that gave biomedicine its credibility as a science were threatening the credibility of medicine as a healing profession.
Mathilde Krim, a New York cancer researcher who had become the co-chair of the American Foundation for AIDS Research (AmFAR), argued at a New York demonstration in summer 1986 that "the double-blind clinical trial on AZT is an insult to morality." But defenders of placebo-controlled trials characterized them as the quickest route to the truth and pointed to their track record in weeding out ineffective drugs that practicing physicians had believed in. Without the science of clinical trials in general, and without double-blind, placebo-controlled trials in particular, physicians were left with nothing but anecdotes and hunches. Douglas Richman, the AZT researcher at the University of California at San Diego, argued in 1988: "In the field of antiviral therapy alone, numerous anecdotal claims were made for the benefits of corticosteroids for chronic hepatitis B, of iododeoxyuridine for herpes simplex encephalitis, and cytosine arabinoside for disseminated herpes zoster. These clinical observations made by concerned physicians were proved to be erroneous in randomized, double-blind, placebo-controlled studies. In fact, the study drug in each case did more harm than the placebo."
The opposite error—erroneously rejecting an effective therapy—was also possible in the absence of placebo-controlled trials. Indeed, said Richman, if there had been no double-blind, placebo-controlled trials, AZT probably would have been discarded. Since AZT showed no impact on the rate of opportunistic infections for the first six weeks of the study and no impact on survival for an even longer period, it would have been easy to conclude from an uncontrolled study that AZT was toxic and ineffective. In response, some, such as Krim, argued that placebo controls weren't the only option for a controlled study. Data obtained from treatment groups could be compared with the medical records of matched cohorts of other AIDS patients who had been followed in the past in studies of the natural history of AIDS (a method called "historical controls"). Or patients in treatment groups could be compared against their own medical records from the weeks prior to their entry into the study. Similar methods had been employed successfully in research with cancer drugs.
Beyond the questions about whether double-blind, placebo-controlled trials were ethical, there began to emerge, in response to the Phase II AZT study, a growing concern about whether such trials were in practical terms possible . The essence of a double-blind trial is that neither the subject nor the investigator knows whether the subject is receiving the drug or the placebo. But how can such information be
disguised in the case of a relatively toxic drug that produces symptoms like nausea and headaches? And how do researchers anticipate the actions of patients understandably anxious about the possibility that they were squandering their remaining days swallowing sugar pills? Even before the trial had ended, rumors began to trickle in from various quarters: some patients were seeking to lessen their risk of getting the placebo by pooling their pills with other research subjects. In Miami, patients had learned to open up the capsules and taste the contents to distinguish the bitter-tasting AZT from the sweet-tasting placebo. Dr. David Barry, the director of research at Burroughs Wellcome, complaining implausibly that never before in the company's history had any research subject ever opened up a capsule in a placebo-controlled trial, quickly instructed his chemists to make the placebo as bitter as AZT. But patients in both Miami and San Francisco were then reported to be bringing their pills in to local chemists for analysis.
Presumably such practices were not invented by AIDS patients. But the prevalence of AIDS within relatively well-defined communities, and the growing sophistication of the emergent treatment underground, made it likely that strategies for "beating the system" diffused more rapidly and more extensively among AIDS patients than among, say, research subjects in cancer trials. (Ironically, such behavior also risked extending the length of the trial, by increasing the time required to show a statistically significant difference between the AZT group and the placebo group—an example of the clash between the individual and social good that makes such trials so vexing.)
Researchers insisted that their own monitoring methods revealed little abuse: blood tests identified few patients in the placebo arms of studies who had obtained the active drug. Still, reports of "noncompliance" raised serious questions about just how "objective" the much-vaunted double-blind trials really were. Those seeing only the tidy graphs and reading only the crisp prose in the New England Journal of Medicine might conceive of such trials as the essence of scientific rigor and, hence, the most solid basis for forming clinical and regulatory judgments. Those observing the conduct of a trial "from the inside" might conclude that knowledge was resting on something rather less solid than bedrock, and they might wonder why the research establishment chose to fetishize this mechanism for establishing biomedical truth.