Questioning the Scientific Method: Ethical Issues in Clinical Trials

lbonnell's picture


Questioning the Scientific Method: Ethical Issues in Clinical Trials



            To many scientists and drugmakers, the form of a scientific experiment is fairly systematic. Drugs under development undergo a series of phases before applying for licensing and use methods such as blind experiments, controls, and placebos to establish scientific validity. While these methods are effective tools that allow researchers to determine the efficacy of a treatment, the use of these methods should not go unquestioned. In some cases the standard scientific method may generate insurmountable ethical issues that suggest an alternative method is needed.

A recent New York Times article by Amy Harmon can be used as a case study of a scientific experiment with major ethical implications. This article highlights the ethical dilemmas that phase trials and randomized control trials can create. Randomized control trials are frequently used to test the efficacy of drugs, by randomly dividing subjects into an experimental group and control groups. The control group may receive treatment that is comparable to the experimental group’s treatment or may be given a placebo. In this article, the control group was given an existing chemotherapy drug and the experimental group was given PLX4032, a new drug developed to treat a lethal skin cancer, melanoma.

In Phase 1 of the trial, a small group of subjects were given the drug to test for side effects and determine proper dosage. The company that manufactures the drug, Roche, reported that nearly all 32 patients saw tumors shrink but most patients saw their tumors grow back within a year. Nevertheless, the drug was considered a major breakthrough in cancer research with potential to extend the life expectancy of patients with advanced melanoma. At the conclusion of the Phase 1, trial Roche was in the position to move in one of two directions with their research. Roche could continue with Phase 2 trials by testing PLX4032 in a larger group of people or alternatively, Roche could seek accelerated F.D.A. approval that would allow the drug to be sold based on early promise and simultaneously continue with Phase 2 trials.

Roche’s choice of action would have a number of ethical and scientific implications. Continuing with Phase 2 trials would allow Roche to further establish the scientific validity of their drug and to determine if there are any unforeseen side effects before releasing it to the market. On the other hand, seeking advanced approval from the F.D.A. would grant patients, who were already begging for the drug, access to PLX4032 and the potential for treatment. Harmon suggests that if this direction was chosen it would be difficult to run a Phase 2 randomized control trial because patients would not want to be placed in the control group.

This background provides logic for Roche’s decision to continue with Phase 2 randomized controlled trial and not seek F.D.A. accelerated approval. With this choice, Roche made the decision to establish scientific validity in favor of helping a small group of people that could potentially benefit from PLX4032. Establishing scientific validity allows future patients to benefit from PLX4032 but, at the same time, leaves a small group of patients from the drugs’ beginnings in the dark.

The New York Times article describes one such patient, who was placed in the control group of the Phase 2 trial. The article notes how the man wished to be in the experimental group after seeing his cousin in the experimental group’s improvement. However, this man was not allow to switch groups because the researcher informed him that switching would impinge the scientific validity of the experiment. Scientific validity is considered a requirement for ethical clinical research. A guideline published in the Journal of the American Medical Association (JAMA), based on national and international codes of research, includes scientific validity among the seven requirements of ethical clinical research. According to one code of research used in the study, the Council for International Organizations of Medical Sciences (CIOMOS), “Scientifically unsound research on human subjects is ipso facto unethical in that it may expose subjects to risks or inconveniences to no purpose.” In the case of PLX4032, for the research to be purposeful and ethical, researchers needed to follow the method of a randomization control trial regardless of the patient’s interests.

Yet, the same guidelines also include respect for the enrolled subject as a requirement for ethical research. In the article, respect for enrolled subjects refers to the subjects right withdraw from the study and have their privacy protected. This definition can plausibly extend to include the subject’s right to switch from the control group from the experimental group.  Subjects should have autonomy and the right to choose their treatment plan. Ideally, scientific validity should not come at the expense of personal liberty. To ensure personal liberty, researchers should have chosen an experimental method that was based on both scientific validity and ethics.

The randomized control trial method also falls short of the favorable risk-benefit ratio requirement published in JAMA. This requirement necessitates that “risks must be minimized, potential benefits must be enhanced, and the potential benefits to the individuals and society must outweigh the risks.” In the case of PLX4032, being placed in the control group posed risks with no benefits for the individual. The control group was given the chemotherapy drug, which was less effective than PLX4032.  The New York Times article describes how, “The standard chemotherapy used in melanoma, dacarbazine, slowed tumor growth in 15% of patients for an average of two months. By contrast, PLX4032 had halted tumor growth in 81% of patients for an average of eight.” Based on this data, the risks associated with the control group and experimental group are clearly not equal. There isn’t a favorable risk-benefit ratio for individuals in the control group.

  At the same time, there is promise of knowledge and benefits for society as a whole. The control group would help establish that PLX4032 is a safe and efficient treatment. This in turn, could help many more people with advanced melanoma in the long run. In this example, there is a divide between the benefits of individuals and the benefits of society. Perhaps historical events can help inform us of whether our priority should be treating the individual or advancing society. Past injustices, including experiments by the Third Reich during the Holocaust and the Tuskegee Syphilis Trial in the U.S., show that aiming to further society without respect for the individual can be dangerous territory. These experiments took advantage of vulnerable populations in order to further their understanding of the human body and medicine. While these examples are very different in scope and severity from the case of PLX4032, they leave the impression that we should have the highest respect individuals in a clinical trial.

If a randomizated control trial is a flawed method for PLX4032, researchers should investigate alternative methods. While there are no clear alternatives, it is possible to imagine some feasible options. In general, a worthwhile alternative method would involve giving subjects autonomy over their treatment and benefits with minimal risks. One solution would be to simply not have a control group.  Maybe in cases like PLX4032 where Phase 1 is successful, the rest of the phases in development can exist with only an experimental group. Or maybe, physicians could compile records of  outcomes of various drugs and use this information as a control. Another alternative is to let patients choose whether they would like to be placed in the control group or experimental group based on unbiased information provided by the researchers.

An alternative method without major ethical implications would be beneficial in the example of PLX4032. But at the same time it is essential to move away from traditional methods of scientific research with caution. The New York Times article describes an incidence twenty years ago when “bone marrow transplants appeared to be so effective that breast cancer patients demanded their immediate approval, only to learn through a controlled trial that the transplants were less effective than chemotherapy and in some cases causes death.” At the same time, drug development has changed in the last 20 years, as the also article points out, and way we test drugs must also change in response. While randomization control trials and other tools such as placebos and blind experiments are useful for researchers they do not preclude the existence of more ethical alternatives. 




Emmanuel, E., Wendler, D., & Grady, C., What Makes Clinical Research Ethical? JAMA. 2000; 283:2701-27.

Harmon, A., New Drugs Stir Debate on Rules of Clinical Trials. The New York Times. Sept 18 2010.

The Mechanisms on a Controlled Trial. Accessed Oct 6 2010.


Paul Grobstein's picture

clinicial trials and "scientific rigor"

This paper raises some very interesting issues about the relationship between "scientific rigor" and "ethical behavior" that I would like to see further pursued.  Is it possible that the concern for "scientific rigor" can be relaxed in many cases where one is looking for large effects rather than small ones?  Is it possible that the "gold standard" of clinical research actually relates more to the needs to drug companies and liability issues than it does to effective health care? 

Post new comment

The content of this field is kept private and will not be shown publicly.