Research Ethics

Overview

Thus far in the course we have focused on core principles in medical ethics, principles such as autonomy, beneficence, informed consent and truth-telling. In this module we examine these principles within the context of medical research. On the face of it, randomized clinical trials, by their very nature, violate these core principles of medical ethics.

In the readings for this module, Maurie Markman examines several specific difficulties that arise with randomized clinical trials. The central argument for randomized clinical trials, as Markman notes, is that without them, modern medicine would be impossible.

[I]n the absence of such studies it is not possible to be certain that a new drug or clinical intervention is actual beneficial to patients with a particular disease or condition, compared either to a “standard” (accepted or approved) therapeutic strategy or to no treatment at all (an untreated control population). (p. 741)

The core argument against randomized clinical trails is that they violate the physician’s primary duty, which is to care for the patient.

The major argument against the performance of randomized clinical trials is that the individual’s physician’s principal ethical responsibility is to the individual patient that he or she is treating, and not to future patients who may benefit from the potentially important information gained through a well-designed and well-conducted randomized trial. (p. 741)

In the next reading, Hellman and Hellman describe the moral dilemma surrounding randomized clinical trials as a conflict of role—the practicing physician’s role is to treat his patient; however, the physician-scientist’s role is to expand scientific knowledge. Without a solution to this dilemma, they issue a call for discussion of alternative methods to test the effectiveness of new drugs and new procedures.

In the next article Benjamin Freedman takes issue with views presented by Markman and Hellman and Hellman. His core question is this: how exactly have participants in (well designed and well conducted) randomized clinical trails been harmed? Freedman contends that there is no conflict between patient’s rights and the social interests furthered by clinical trials.

In the article “Clinical Trials: Are They Ethical?” (included in this module) Eugene Passamani notes that randomized clinical trials, like any other human activity, can be performed in an ethical manner and an unethical manner. His article aims to elucidate the conditions under which randomized clinical trials are ethical. Passamani agrees with Freedman that clinical trials must begin with clinical equipoise. However, Passamani adds several other constraints: the participants must give informed consent, the study must be stopped if accumulating data destroy clinical equipoise, and the clinical trial must be designed to answer a significant open question.

In “AZT Trials and Tribulations” Crouch and Arras question the morality of the AIDS Clinical Trial Group (ACTG) 076 and by extension other studies that make use of vulnerable populations who are unlikely to benefit from the research. The particular study in question sought to determine whether a shorter than normal course of AZT would be effective in reducing the transmission of the HIV virus from pregnant woman to their fetus/newborn. The subjects for this study were impoverished African woman who pregnant and HIV positive. The discussion question for this module will focus on this study, so I will leave you to assess the morality of the study.

In the final reading for this module (included below) Danstan Bagenda and Philoppa Musoke-Mudido defend the ACTG 076 trial. Here is the question: Is the ACTG 076 trial morally justified?

Objectives

By the end of this unit, I will be able to:

· describe the basic features of randomized clinical trials

· assess the pros and cons of randomized clinical trials

Key Terms

· Randomized Clinical Trials

· Control Group

· Experimental Group

· Clinical Equipoise

· Informed Consent

Review Questions

[1] What is a randomized clinical trial?

[2] Why are randomized clinical trails an essential component to modern medicine?

[3] What is the basic argument against conducting randomized clinical trials?

[4] What specific examples of the ethical difficulties with randomized clinical trials does Markman discuss?

[5] How do Hellman and Hellman describe the moral dilemma surrounding randomized clinical trials?

[6] What final conclusion do Hellman and Hellman reach about the morality of randomized clinical trials?

[7] How does Freedman criticize the views presented by Markman on the one hand and Hellman and Hellman on the other?

[8] Freedman claims that clinical trials are ethical if they begin with clinical equipoise. Discuss.

[9] Passamani lists three features of clinical trials: equipoise, stopping the trail if equipoise is disrupted and informed consent. Discuss. Do you think these features are necessary and/or sufficient to justify particular clinical trials?

[10] Carefully explain the ACTG (076) protocol.

[11] What reasons do Crouch and Arras adduce for their conclusion that the clinical trail was morally unjustified?

[12] What reasons do Danstan Bagenda and Philoppa Musoke-Mudido provide in support of the protocol?

Discussion Question

Carefully explain the ACTG (076) protocol. Do you think this research is morally permissible? Why or why not? (Be sure to consider the views of Crouch and Arras as well as the views of Bagenda and Mudido. You may also wish to consider whether affordability of medications in third-world countries impacts clinical equipoise, whether you think pregnant, HIV positive African woman are capable of giving informed consent (note: to say they are not capable of giving informed consent smacks of paternalism), whether the participants (whether in the control group or the experimental group) benefited from the research (for example, by having access to prenatal vitamins, clean water and a doctor’s examination), and whether the research sought to answer a significant “open question” in the sense outlined by Passamani. You may also wish to consider whether it is ever permissible to use impoverished and vulnerable populations in clinical trials.)

(Minimum World Count: 500 Words)

Clinical Trials: Are They Ethical?[footnoteRef:1] [1: From the New England Journal of Medicine, Vol. 324, no. 22 (1991), pp. 1589-1591. Reprinted by permission.]

Eugene Passamani

Biomedical research leads to better understanding of biology and ultimately to improved health. Physicians have for millenniums attempted to understand disease, to use this knowledge to cure or palliate, and to relieve attendant suffering. Improving strategies for prevention and treatment remains an ethical imperative for medicine. Until very recently, progress depended largely on a process of carefully observing groups of patients given a new and promising therapy; outcome was then compared with that previously observed in groups undergoing a standard treatment.

Outcome in a series of case patients as compared with that in nonrandomized controls can be used to assess the treatment of disorders in which therapeutic effects are dramatic and the pathophysiologic features are relatively uncomplicated, such as vitamin deficiency or some infectious diseases. Observational methods are not very useful, however, in the detection of small treatment effects in disorders in which there is substantial variability in expected outcome and imperfect knowledge of complicated pathophysiologic features (many vascular disorders and most cancers, for example). The effect of a treatment cannot easily be extracted from variations in disease severity and the effects of concomitant treatments. Clinical trials have thus become a preferred means of evaluating an ever increasing flow of innovative diagnostic and therapeutrial is a powerful technique because of the efficiency and credibility associated with treatment comparisons involving randomized concurrent controls.

The modern era of randomized trials began in the early 1950s with the evaluation of streptomycin in patients with tuberculosis. Since that time trial techniques and methods have continuously been refined. In addition, the ethical aspects of these experiments in patients have been actively discussed.

In what follows I argue that randomized trials are in fact the most scientifically sound and ethically correct means of evaluating new therapies. There is potential conflict between the roles of physician and physician-scientist, and for this reason society has created mechanisms to ensure that the interests of individual patients are served should they elect to participate in a clinical trial.

Clinical Research

The history of medicine is richly endowed with therapies that were widely used and then shown to be ineffective or frankly toxic. Relatively recent examples of such therapeutic maneuvers include gastric freezing for peptic ulcer disease, radiation therapy for acne, MER-29 (triparanol) for cholesterol reduction, and thalidomide for sedation in pregnant women. The 19th century was even more gruesome, with purging and bloodletting. The reasons for this march of folly are many and include, perhaps most importantly, the lack of complete understanding of human biology and pathophysiology, the use of observational methods coupled with the failure to appreciate substantial variability between patients in their response to illness and to therapy, and the shared desire of physicians and their patients for cure or palliation.

Chance or bias can result in the selection of patients for innovative treatment who are either the least diseased or the most severely affected. Depending on the case mix, a treatment that has no effect can appear to be effective or toxic when historical controls are used. With the improvement in diagnostic accuracy and the understanding of disease that has occurred with the passage of time, today's patients are identified earlier in the natural history of their disease. Recently selected case series therefore often have patients who are less ill and an outcome that is considerably better than that of past case series, even without changes in treatment.

Randomization tends to produce treatment and control groups that are evenly balanced in both known and unrecognized prognostic factors, which permits a more accurate estimate of treatment effect in groups of patients assigned to experimental and standard therapies. A number of independent randomized trials with congruent results are powerful evidence indeed.

A physician's daily practice includes an array of preventive, diagnostic, and therapeutic maneuvers, some of which have been established by a plausible biologic mechanism and substantial evidence from randomized clinical trials (e.g., the use of beta-blockers, thrombolytic therapy, and aspirin in patients with myocardial infarction).8 It is unlikely that our distant descendants in medicine will discover that we late-20th-century physicians were wrong in these matters. However, new therapeutic maneuvers that have not undergone rigorous assessment may well turn out to be ineffective or toxic. Every therapy adopted by common consent on the basis of observational studies and plausible mechanism, but without the benefit of randomized studies, may be categorized by future physicians as useless or worse. Physicians are aware of the fragility of the evidence supporting many common therapies, and this is why properly performed randomized clinical trials have profound effects on medical practice. The scientific importance of randomized, controlled trials is in safeguarding current and future patients from our therapeutic passions. Most physicians recognize this fact.

Like any human activity, experimentation involving patients can be performed in an unethical and even criminal fashion. Nazi war crimes led to substantial efforts to curb abuse, beginning with the Nuremberg Code and the Helsinki Declaration and culminating in the promulgation of clearly articulated regulations in the United States and elsewhere. There are abuses more subtle than those of the Gestapo and the SS. Involving patients in experiments that are poorly conceived and poorly executed is unethical. Patients who participate in such research may incur risk without the hope of contributing to a body of knowledge that will benefit them or others in the future. The regulations governing human experimentation are very important, as is continuing discussion and debate to improve the scientific and ethical aspects of this effort.

Several general features must be part of properly designed trials. The first is informed consent, which involves explicitly informing a potential participant of the goals of the research, its potential benefits and risks, the alternatives to participating, and the right to withdraw from the trial at any time. Whether informed consent is required in all trials has been debated. I believe that patients must always be aware that they are part of an experiment. Second, a state of clinical equipoise must exist. Clinical equipoise means that on the basis of the available data, a community of competent physicians would be content to have their patients pursue any of the treatment strategies being tested in a randomized trial, since none of them have been clearly established as preferable. The chief purpose of a data-monitoring committee is to stop the trial if the accumulating data destroy the state of clinical equipoise—that is, indicate efficacy or suggest toxicity. Finally, the trial must be designed as a critical test of the therapeutic alternatives being assessed. The question must be clearly articulated, with carefully defined measures of outcome; with realistic estimates of sample size, including probable event rates in the control group and a postulated and plausible reduction in the event rates in the treatment group; with Type I and II errors specified; and with subgroup hypotheses clearly stated if appropriate. The trial must have a good chance of settling an open question.

Ethical Dimensions of Properly Constituted Trials

Experimentation in the clinic by means of randomized, controlled clinical trials has been periodically attacked as violating the covenant between doctor and patient. Critics have charged that physicians engaged in clinical trials sacrifice the interests of the patient they ask to participate to the good of all similarly affected patients in the future. The argument is that physicians have a personal obligation to use their best judgment and recommend the "best" therapy, no matter how tentative or inconclusive the data on which that judgment is based. Physicians must play their hunches. According to this argument, randomized clinical trials may be useful in seeking the truth, but carefully designed, legitimate trials are unethical and perhaps even criminal because they prevent individual physicians from playing their hunches about individual patients. Therefore, it is argued, physicians should not participate in such trials.

It is surely unethical for physicians to engage knowingly in an activity that will result in inferior therapy for their patients. It is also important that the community of physicians be clear in distinguishing between established therapies and those that are promising but unproved. It is this gulf between proved therapies and possibly effective therapies (all the rest) that defines the ethical and unethical uses of randomized clinical trials. Proved therapies involve a consensus of the competent medical community that the data in hand justify using a treatment in a given disorder. It is this consensus that defines an ethical boundary. The physician-investigator who asks a patient to participate in a randomized, controlled trial represents this competent medical community in asserting that the community is unpersuaded by existing data that an innovative treatment is superior to standard therapy. Arguments that a physician who believes that such a treatment might be useful commits an unethical act by randomizing patients are simply wrong. Given the history of promising but discarded therapies, hunches about potential effectiveness are not the ideal currency of the patient-doctor interchange.

Lest readers conclude that modem hunches are more accurate than older ones, I have selected an example from the current cardiovascular literature that reveals the problems inherent in relying on hunches to the exclusion of carefully done experiments.

The Cardiac Arrhythmia Suppression Trial

Sudden death occurs in approximately 300,000 persons in the United States each year and is thus a problem worthy of our best efforts. In the vast majority of cases the mechanism is ventricular fibrillation superimposed on a scarred or ischemic myocardium. It had been observed that the ventricular extrasystoles seen on the ambulatory electrocardiographic recordings of survivors of myocardial infarction were independently and reproducibly associated with an increased incidence of subsequent mortality. It had been established that a variety of antiarrhythmic drugs can suppress ventricular extrasystoles. Accordingly, physicians had the hunch that suppressing ventricular extrasystoles in the survivors of myocardial infarction would reduce the incidence of ventricular fibrillation and sudden death.

The Cardiac Arrhythmia Suppression Trial (CAST) investigators decided to test this hypothesis in a randomized, controlled trial. They sought survivors of myocardial infarction who had frequent extrasystoles on electrocardiographic recordings. The trial design included a run-in period during which one of three active drugs was administered and its effect on extrasystoles noted. Those in whom arrhythmias were suppressed were randomly assigned to active drug or placebo. The trial had to be stopped prematurely because of an unacceptable incidence of sudden death in the treatment group.15 During an average follow-up of 10 months, 56 of 730 patients (7.7 percent) assigned to active drug and 22 of 725 patient (3.0 percent) assigned to placebo died. Clinical equipoise was destroyed by this striking effect. It is quite unlikely that observational (nonrandomized) methods would have detected this presumably toxic effect.

The CAST trial was a major advance in the treatment of patients with coronary disease and ventricular arrhythmia. It clearly revealed that the hunches of many physicians were incorrect. The trial's results are applicable not only to future patients with coronary disease and ventricular arrhythmia but also to the patients who participated in the study. By randomizing, investigators ensured that half the participants received the better therapy—in this case placebo—and, contrary to intuition, most of them ultimately received the better therapy after the trial ended prematurely and drugs were withdrawn.

To summarize, randomized clinical trials are an important element in the spectrum of biomedical research. Not all questions can or should be addressed by this technique; feasibility, cost, and the relative importance of the issues to be addressed are weighed by investigators before they elect to proceed. Properly carried out, with informed consent, clinical equipoise, and a design adequate to answer the question posed, randomized clinical trials protect physicians and their patients from therapies that are ineffective or toxic. Physicians and their patients must be clear about the vast gulf separating promising and proved therapies. The only reliable way to make this distinction in the face of incomplete information about pathophysiology and treatment mechanism is to experiment, and this will increasingly involve randomized trials. The alternative—a retreat to older methods—is unacceptable.

Physicians regularly apply therapies tested in groups of patients to an individual patient. The likelihood of success in an individual patient depends on the degree of certainty evident in the group and the scientific strength of the methods used. We owe patients involved in the assessment of new therapies the best that science and ethics can deliver. Today, for most unproved treatments, that is a properly performed randomized clinical trial.

We're Trying to Help our Sickest People, Not Exploit Them[footnoteRef:2] [2: Reprinted with the permission of the authors from the Washington Post, 1997 September 28, p. C3. Copyright © 1997 by The Washington Post Company. ]

Danstan Bagenda and Philoppa Musoke-Mudido

Every day, like the beat of a drum heard throughout Africa, 1,000 more infants here are infected with HIV, the virus that causes AIDS. At Old Mulago Hospital, we are trying to educate people about AIDS, as well as study new therapies to prevent the disease's rampant spread. Recently, some of these studies have been attacked, with comparisons made to the notorious Tuskegee experiment in which black men in the United States were denied treatment for syphilis. Tuskegee? Is this really what is happening here in our mother-child clinic?

Our country lies in the heart of Africa, along the Great Rift Valley and Lake Victoria. It is one of those hardest hit by the AIDS epidemic. A few years ago, visitors here in the capital were greeted by the macabre sight of empty coffins for sale—piled in pyramids from adult to baby size—along the main road.

These grim reminders have since been removed by city authorities, but the AIDS epidemic is omnipresent. In this city of 1 million, about one out of every six adults is infected with HIV. Hospitals and clinics like ours, which provide free medical care and therefore serve the poorest communities, are stretched beyond their resources.

At the Mulago Hospital, where more than 20,000 women deliver each year, we are trying to find effective therapies to stop transmission of HIV from pregnant women to their babies. About one in five babies becomes infected with HIV during pregnancy and delivery. If he mother breast-feeds her baby, there is an additional 15- to 25-percent chance that the baby will later become infected. There is no available treatment for the disease in Uganda. After careful consideration among researchers from developing and developed countries, the World Health Organization (WHO) recommended in 1994 that the best way to find safe and effective treatment for sufferers in countries in the developing world is to conduct studies in which new treatments, better tailored to the local population, are compared with placebos (inactive pills).

Women who enroll in our studies undergo intensive education and individual counseling. They are given a comprehensive consent form, written in the local language, which they are encouraged to take home and discuss with their families. It describes the potential risks of participating in the study and their chances of receiving a placebo. Only when they and their counselors are satisfied that all questions have been answered are they asked to sign the form. Our careful attention to these measures has consistently met the standards of national and international ethical review committees.

Results from a clinical trial in the United States and France, known as the ACTG 076 protocol, showed as long ago as 1994 that if a mother takes zidovudine (AZT) daily from the middle of her pregnancy until delivery, receives intravenous AZT during delivery, gives her infant oral AZT for the first six weeks of life and does not breast-feed, the transmission of HIV from mother to child can be reduced by two-thirds. The ACTG 076 protocol immediately became the recommended therapy in the United States. But it is not possible to simply transplant this protocol to Uganda for three main reasons: At a cost of between $800 and $1,000 per person, it is far too expensive; it requires treatment to begin in the middle of a pregnancy; and it means mothers must abstain from breast-feeding.

Some critics in the United States have asserted that we should compare new therapies with the ACTG 076 protocol rather than with a placebo. But in Uganda, the government health expenditure is $3 per person per year, and the average citizen makes less than $1 per day. We think it is unethical to impose expensive treatment protocols that could never be used here. The situations are not parallel. In America, for instance, antibiotics are often over-prescribed; but here in Uganda we have difficulty even obtaining many needed antibiotics—to treat common complaints like ear infections. It is also naive to assume that what works for Americans will work for the rest of the world. Differences in nutrition, economics, societal norms and culture, and the frequency of tropical diseases make such extrapolations dangerously ethnocentric and wrong.

Many pregnant women here never show up for prenatal care and, of those who do, 70 percent make their first visit after the 30th week of pregnancy—too late for the U.S. treatment protocol. Should we make a study available only to the minority of women who come early for care and tell the others, sorry, you came too late? We need to find treatments that will reach the most women possible—ones that can be given late in pregnancy or during labor.

There is also a huge gap between the United States and Uganda in breast-feeding practices. Should we apply the ACTG 076 protocol and tell women in the clinic not to breast-feed and instead give their babies infant formula? Access to clean water is a formidable challenge here, and we still remember the shocking epidemics of infant diarrhea and mortality in the early 1970s, when multinational companies shamelessly marketed formula in Africa. Despite the known risks of transmitting HIV through breast milk, the Ugandan Ministry of Health, UNICEF and WHO still encourage African women to breast-feed as the nutritional benefits outweigh the risks of HIV transmission.

There are other factors we need to take into account. Every day, we treat both mothers and infants for malaria and iron deficiency. Both diseases contribute to anemia, which is also a major side effect of AZT. We are worried that AZT will exacerbate anemia in women and infants here. If we are to find out whether the new treatments are safe, the best way is to compare them with a placebo. How could we evaluate the safety of a new treatment if we compared it with the treatment used in America—one that has its own side effects? Could we really tell Ugandans that we had evaluated a new therapy for side effects using the best possible methods?

The AIDS epidemic has touched all our lives. Each of the 90 staff members in the mother-child health clinic has lost a family member, a loved one or a close friend. There is no dividing line between patients with HIV and those of us who care for them. A few years ago, we all chipped in money when a staff member needed to pay for the burial of a loved one, but recently we realized that we were all giving and receiving the same.

The ethical issues in our studies are complicated, but they have been given careful thought by the local community, ethicists, physicians and activists. Those who can speak with credibility for AIDS patients in Africa are those who live among and know the people here or have some basic cross-cultural sensitivity. We are suspicious of those who claim to speak for our people, yet have never worked with them. Callous accusations may help sell newspapers and journals, but they demean the people here and the horrible tragedy that we live daily.

In the next several months, we expect to see results from our study and others like it in Ivory Coast, South Africa, Tanzania and Thailand. We hope they will help bring appropriate and safe therapies to the people of the developing world. That hope is the driving force that brings us back to our work in the clinic after each and of the all-to frequent burials.