Clinical Ethics in Anesthesiology. A Case-Based Textbook
4. Research and publication
30. Ethical function of human subjects review boards: a US perspective
Jeffrey H. Silverstein
An investigator submits a project that proposes to use an antibiotic for treatment of an unusual infection that afflicts as small number of patients following colonic interposition following esophagectomy. Although the number of cases of this type of postoperative esophagitis are small, the suffering of the patients with this complication is severe. The proposed antibiotic is approved for clinical treatment of anaerobic infections but is a second line drug and is not frequently used for this purpose in clinical practice. A related drug is one of the most commonly used therapies for treatment and prophylaxis of anaerobic infections, being administered to thousands of patients every day in the US.
The application is extensive and includes a number of case reports of successful treatment of post esophagectomy colonic interposition inflammation and a small animal study that also suggests the treatment will be effective. The side effect profile of the medication is similar to many antibiotics – including gastric distress, possible allergy and the risk of subsequent resistant infection.
The project accrues a small number of patients and is reviewed annually for two subsequent years. Few adverse or unanticipated events are reported. During the third annual evaluation, one of the IRB reviewers notes that there is a specific warning from the Food and Drug Administration (FDA) indicating that a similar drug (not the one in use, but the more common one that is administered regularly) has been shown to cause cancer in a small number of mice when administered for prolonged periods of time. The drug proposed for this project has never been reported to cause cancer in either animals or humans and the related drug has never been reported to cause cancer in humans. In reviewing the literature and documentation, it is clear that this warning was known when the project was initially reviewed, but was not included in the presentation from the investigator, nor was it noted by the IRB reviewers during the initial review or the first two annual reviews of the project.
Was it ethical to approve the study initially? Did the IRB do its job in reviewing this project? What is the nature of the lapse in initial reporting by the investigator? Should the participants be notified of a failure to inform them of this risk of malignancy?
In a previous chapter, Dr. Viens describes the ethical basis for the conduct of research on human subjects. All researcher who have participated in these activities rapidly learn that adherence to the ethical principles outlined are not left up to the individual investigator, but require a formal review process by independent individuals. This review process is the basis of extensive regulatory statues.
What is the purpose of an IRB and how did they evolve?
The history of protecting human subjects in research goes back into antiquity. While it might be thought that societies would proactively define norms and mores regarding how one can experiment on fellow humans, in practice, our regulatory structure has evolved almost exclusively in the wake of scandal and disaster. Indeed, the history of human subjects protections and the initial creation of formalized institutional review boards followed a specific series of incidents that were both unacceptable and well publicized.
Most treatises on human subjects protection begin their discussion with the effort at codifying research ethics undertaken by the Judges for the Nazi Doctors Trial following World War II. In this case, experiments were conducted, some with little legitimate scientific intent, which caused extreme suffering and frequent deaths. Many were specifically designed to understand the limitations of human tolerance to hypothermia, low oxygen tension high altitude situations and immersion in salt water, all problems suffered by combatants on both sides of the conflict. These experiments were intended to assist scientists in developing means to support soldiers and sailors engaged in the war effort. Some were published in reputable scientific journals. Nonetheless, these experiments were widely considered unethical and the war crimes tribunal following the war prosecuted the doctors involved in these experiments. The judges assigned to the trial found that there were essentially no coherent codes of conduct regarding human experimentation for reference. Therefore, in 1947, they elaborated what has become known as the Nuremberg Code (Table 30.1) as a guide document for understanding what should have happened in the wake of tragedy.
On the basis of this construct, many of the doctors and participants in those experiments were convicted. Many of the concepts were elaborated upon in a subsequent document created by the World Health Organization, which is called the Declaration of Helsinki. The Declaration is an ethics document, as opposed to a regulatory or legal document. Its continued evolution has created controversy in the last few years.1
A key moment in the evolution of American research oversight was the seminal publication in 1966 of an article in the New England Journal of Medicine entitled “Chronicle of 22 unethical studies.”2The author Henry Knowles Beecher was one of the most famous anesthesiologists of his day and the anesthesia laboratories at Harvard still bear his name. This article delineated a large series of published studies that Dr. Beecher contended had failed to follow the ethical principles of the day.
The Tuskeegee experiment and the Belmont Report
The first set of human subjects regulations were elaborated by the then nascent National Institutes of Health. However, the true watershed event in US research ethics was the description of the Tuskeegee experiment. Officially titled “ Tuskegee syphilis study or Public Health Service syphilis study,” the Tuskegee experiment, at least in its initial conception, was a scientifically valid observational study of the natural history of syphilis. When this started in 1932, syphilis was a major health problem with no cure and which created a hugely varied pantheon of symptoms. A similar study was under way in Sweden at the time. Unfortunately, when it became known that syphilis could be cured with penicillin in the mid 1940s, the scientists involved did not stop or redesign the study. The subjects were all male African-American prisoners in southern jails who had no idea that they were part of a research study. The national scandal that
Table 30.1. The Nuremberg Code
erupted led to the formation of the National Bioethics Commission and the elaboration of what became known as the Belmont Report.
The Belmont Report elaborated three basic principles: autonomy, beneficence, and justice, which remain the ethical underpinnings of American human research protections.3 This was soon followed by federal regulations that created a standard for evaluation, review and consenting for humans participating in research subjects. The role of Institutional Review Boards (IRBs) was carefully described “The IRB shall be sufficiently qualified through the experience and expertise of its members and the diversity of the members, including consideration of race, gender, and cultural backgrounds and sensitivity to such issues as community attitudes, to promote respect for its advice and counsel in safeguarding the rights and welfare of human subjects…the IRB shall be able to ascertain the acceptability of proposed research in terms of institutional commitments and regulations, applicable law, and standards of professional conduct…The IRB shall therefore include persons knowledgeable in these areas.”4
Of note, there are two separate, related, but distinct sets of regulations in the USs. Most research undertaken at medical schools with funding from the NIH are governed by the Department of Health and Human Services Protection of Human Subjects found at 45 CFR 46 (CFR = Code of Federal Regulations). However, there is a separate, similar, but not identical set of regulations that govern the Food and Drug Administration. Anyone proposing to submit a drug for FDA approval must follow those regulations.5 In many cases, both sets of regulations apply simultaneously.
Even with this set of regulations in place, IRBs remained somewhat sleepy backwaters of compliance in major medical schools. Following another series of scandals, federal authorities removed the authorization to use federal monies for research from a number of major educational institutions in the late 1980s and early 1990s. This resulted in a major overhaul and significant augmentation of the human subjects protections programs in all major schools and the rapid evolution of independent institutional review boards. Since that time, there has been an appreciation that human subjects protections is a profession, which has evolved it’s own certification process for individuals (CIP – certified IRB professional) and an accreditation process for programs (AAHRPP – Association for the Accreditation of Human Research Protection Programs).
IRBs have frequently been referred to as ethics boards and indeed the charge includes the assessment of the ethics of a proposed study. However, in practice modern IRBs are responsible for assuring that investigators are in compliance with the regulations regarding human subjects research. Thus, the content of informed consent documents, which is delineated in US federal regulations and compliance with specific protections afforded to special groups of potential participants are required activities that IRBs must document. This compliance function frequently appears to take precedence over the pure ethics of a clinical trial. Whether this type of review is more or less beneficial to the goal of protecting human subjects is an interesting ethical question in its own right, but one beyond the scope of this chapter.
In the index case presented above, both the IRB and the investigator apparently failed to properly identify a specific risk. This failure could be described as a failure of the principle of beneficence, and nonmaleficenc (the principle to do no harm). It can also be seen as having severely limited the autonomy of the participants, whom, in the absence of a potentially critical piece of information, might not have been able to make an informed decision regarding their participation. Appropriate actions following the identification of such risks should include the notification of participants.
Determination of risk, risk/benefit ratio, and risk management
One of the principal obligations of IRBs is the determination of the level of risk posed to a participant and assuring that the risks are both acceptable and minimized by appropriate research design risk. This idea is frequently couched in the phrase risk/benefit ratio, which makes this determination sound like a calculation with a definable answer. Risk involves the expected value and likelihood of one or more future events. Unfortunately, the concept of risk is not defined in federal regulations and remains a particularly difficult concept to apply in the setting of human experimentation. Bioethicists, such a Ezekiel Emmanuel from the National Institutes of Health have argued for an actuarial definition of risk, based on the statistical likelihood of a an untoward occurrence. In this construct, human research is, for the most part, much safer than many daily human activities such a driving a car or participating in a contact sport such as football. On a practical basis, IRBs find themselves considering risk in a relation to a relative level of tolerance or acceptability. So, while the risk of driving a car is well understood, it is also generally acceptable in that very few individuals would consider not using an automobile based on the risk of death or injury, while a much lower absolute likelihood of injury or death in a biomedical research is not so well accepted.
The index case provides an interesting example. The cancer warning for the drug was established based on animal experiments of a related drug. There was no indication that the related drug actually did produce cancer in any humans and the actual drug in use was never tested for chronic use in animals, so the actual drug proposed for the study had never been reported to cause cancer. The warning was based on very little information and could arguably be dismissed from a scientific basis; however, failing to note the warning was clearly an error on the part of both the investigator and the review board. Any risk of cancer, no matter how small or theoretical, might be legitimately considered to be significant.
What is the IRB’s role in scientific review?
A frequent complaint leveled against current IRBs is the tendency to review the scientific design of a research proposal. Investigators argue, with some justification that scientific review occurs at other venues. Particularly for major grants, such as those funded by the National Institutes of Health, extensive review by panels of experts have critiqued the scientific content of the proposal. One could legitimately argue that IRBs, even with some direct expertise, are not as well equipped and provide little added value to previous professional scientific review. Even in the absence of this highly sophisticated professional review process, many institutions have internal scientific review bodies that can provide high quality review of the science underlying a proposal. These groups should be at least as good at evaluating the current standard of care and determining whether an experiment is justified and well designed.
High-quality scientific review should be extremely helpful to an IRB reviewing a project. A human subjects’ reviewer should be able to have confidence that a project reviewed and funded by an auspicious national body represents quality science. One might argue that a project prepared by a major pharmaceutical manufacturer who employs skilled personnel with extensive experience in research design is equally well designed. It is beneficial for these projects to have previous focused scientific review prior to submission to the IRB.
IRBs are supposed to minimize risk by ensuring appropriate research design. This is the standard justification used by most IRBs to question a study design. A number of circumstances arise. For major projects that have been reviewed, it may be that the contingencies of human subjects protections were not considered during the design or review of the project. For projects reviewed by the National Institutes of Health, this changed a few years ago, so that all projects including human subjects must now have a section describing human subjects protections. This change in the scientific review process has been a major improvement in scientific review that should minimize discrepancies seen in the past.
Even with this process, there may remain legitimate differences in the evaluation of a design. A poignant example has arisen in the latest rendition of the Declaration of Helsinki, which seeks to minimize the use of placebo controls. The supporters of this document contend that the use of placebos in rarely justified. For example, if a new analgesic was to be tested, one could argue that we don’t need one that is better than nothing, but rather one that is as good as or better than current therapy. Therefore, one should not design a trial that compares such a drug to placebo, but rather to an active standard of care control drug. Interestingly, the US FDA has rejected the current Declaration of Helsinki over this issue (among others) and continues to support and sometimes require placebo-controlled studies prior to approval of a new drug. IRBs frequently have a difficult time reconciling these different positions.
Furthermore, much research that is reviewed by IRBs has not undergone such extensive, if any peer review and might have been written by individuals with little experience is writing research proposals. These projects probably benefit from the expert reviews provided by an IRB. Much current science could be argued to be repetitive with little to be offered in terms of significant scientific progress. While major funding agencies require a project to be innovative and significant, the tremendous expansion of the medical literature has resulted in large numbers of scientific reports that do not have a major impact. Pharmaceutical companies are particularly interested in maintaining drugs on patent and are frequently accused of testing new articles with limited advantages over existing medications. For the most part, IRBs have been reluctant to take issue with well-designed research which promises minor scientific or medical value. Most IRB members would find that role unappealing and difficult to fulfill.
A difficult problem arises when poorly designed research poses minimal or no risk, but is unlikely to answer the proposed questions, due, for example, to a lack of proper controls or the inclusion of too few patients to achieve statistical significance. Some IRBs would argue that poor science represents an unacceptable risk, even if the primary “risk” involved is wasting the time of a participant. Others would argue that, in the absence of any definable risk of injury, for example, in an innocuous anonymous survey, the IRB is not justified in suggesting alterations to a project. Because the primary model ensconced in the regulations is based on significant clinical trials, it can be argued that applying these principles to minimal risk types of projects is either inappropriate or a waste of resources.6 Although an ongoing problem, there is little appearance that this issue will be resolved at a federal level any time soon.
In the index case, there was no concern regarding the design. Had the information regarding the risk of the test substance been discussed at the initial review, the project most likely would still have been approved, although the information regarding this risk would have been discussed and included in the information provided to potential participants. In the age of the internet, where information is easily accessible, it is reasonable to question whether it is the responsibility of IRB reviewers to “go beyond” what the researcher supplies in order to do a thorough review. In the index case, this did not happen during the initial review, but did occur during subsequent review. It is hard to argue that IRBs should not be looking for this additional information, but it is also difficult to set standards for what should be reviewed.
New issues facing IRBs
IRB review is a continually evolving process. In recent years, issues of privacy and confidentiality have become major focuses of human subjects protections under the guise of the Health Insurance Portability and Accountability Act of 1996. Initially designed to avoid loss of insurance based on violations of privacy, the rules regarding privacy and confidentiality have proven daunting to review and enforce. Consent documents have expanded greatly to accommodate required language. Investigators have legitimately complained that the requirements are excessive and impair research activities. Balancing these concerns are the very real problems associated with identity loss and the concerns that private health information could be used to discriminate against individuals. IRBs have also become embroiled in the desire to minimize conflicts of interest in research. All of this activity has created an increasing burden for investigators, IRB members and human subjects protection program staff members. Fortunately, there seems to be a movement to simplify and streamline some of these review processes. One can only hope that we will achieve a balance of appropriate review and adequate protections which can be accomplished in an efficient and professional manner.
• Regulation of human subjects research has evolved in the wake of scandals involving mistreatment of human research subjects.
• Modern human subjects protections had their roots in the war crimes trials of German physicians who experimented on prisoners during World War II.
• Important early declarations regarding the ethical treatment of human research subjects include the Nuremberg Code, and the WMA Declaration of Helsinki.
• The description of the Tuskeegee Experiment was a watershed event in the history of US human research ethics, leading to the publication of the Belmont report elaborating the ethical principles in treatment of human subjects research.
• US IRBs have largely functioned as regulatory bodies assuring compliance with federal regulations.
• Determination of risks and assuring that risks are both acceptable and minimized is an important part of IRB review. Minimization of risk includes assurance of appropriate research design.
• Additional new issues facing IRBs include protection of the identities of human subjects, and the minimization of conflicts of interest in research.
1* The World Medical Association Declaration of Helsinki. Ethical Principles for Medical Research Involving Human Subjects. Adopted by the WMA General Assembly, Helsinki, Finland, June 1964. Last amended by the WMA General Assembly, Seoul, October 2008. http://www.wma.net/en/30publications/10policies/b3/index.html.
2* Beecher, H.K. (1966). Ethics and clinical research. N Eng J Med, 274(24), 1354–60.
3 The National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. (1979). The Belmont Report: Ethical Principles and Guidelines for the Protection of Human Subjects of Research. Dept of Health, Education and Welfare.
4 21CFR56.107. Code of Federal Regulations. Title 21, volume 1. Part 56: Insitutional Review Boards. Subpart B: Organization and Personnel. Revised April 1, 2009.
5 Informed consent: 21 CFR 50, Institutional Review Boards: 21 CFR 56.
6* Gawande, A. (2009). The Checklist Manifesto, How to Get Things Right. New York: Henry Holt.
Annas, G.J. and Gordon, M.A., eds (1992). The Nazi Doctors and the Nuremberg Code. Human Rights in Human Experimentation. New York: Oxford University Press.
Code of Federal Regulations. TITLE 45:Public Welfare Department of Health and Human Services, Part 46. Protection of Human Subjects http://www.hhs.gov/ohrp/humansubjects/guidance/45cfr46.htmP.
Jones, J.H. (1983). Bad Blood: The Tuskeegee Syphilis Experiment. New York: The Free Press
Wendler, D., Belsky, L., Thompson, K.M., and Emanuel, E.J. (2005). Quantifying the federal minimal risk standard: implications for pediatric research without a prospect of direct benefit. JAMA, 294, 826–32.