Dr. A is a family practitioner with a special interest in the treatment of chronic obstructive pulmonary disease. He receives a letter from the coordinator of a study to evaluate a promising new treatment for this condition. The letter invites Dr. A to submit the names of potentially eligible patients. He will be paid $100 for each name provided.
Dr. B, a psychiatrist in private practice, is approached by a pharmaceutical company to assist with a clinical trial to test the efficacy of a new drug in the treatment of acute psychosis. The study will enroll acutely psychotic patients with no history of psychosis (or of treatment with antipsychotic drugs) through physicians’ offices and emergency departments. Patients enrolled in the study will be randomly assigned to receive the new medication or a placebo and will remain in hospital for eight weeks. During this time, they will not be permitted to receive antipsychotic medications other than the study drug. Informed consent will be obtained from each participant or a proxy. Patients may be withdrawn from the study if their medical condition worsens substantially.
Research involving human subjects can raise difficult and important ethical and legal questions. The field of research ethics is devoted to the systematic analysis of such questions to ensure that study participants are protected and, ultimately, that clinical research is conducted in a way that serves the needs of such participants and of society as a whole.
Many of the ethical issues that arise in human experimentation – such as those surrounding informed consent, confidentiality, and the physician’s duty of care to the patient – overlap with ethical issues in clinical practice. Nevertheless, important differences exist between research activities and clinical practice. In clinical practice, the physician has a clear obligation to the patient; in research, this obligation remains but may come into conflict with other obligations – and incentives (Hellman and Hellman, 1991). The researcher has an obligation to ensure that the study findings are valid and replicable, and this has implications for the design and execution of the study. For example, the study must be designed in such a way that the research question is answered reliably and efficiently; sufficient numbers of subjects must be enrolled in a reasonable period, and study participants must comply with their allocated treatment. Substantial rewards can accrue to the successful completion of a research project, such as renewed funding, academic promotion, salary increases, respect from colleagues, and, in some cases, fame. Unfortunately, in a number of research studies, including some conducted in the world’s leading research institutions, the welfare of individual subjects has been sacrificed to these competing interests (Elliott and Weijer, 1995; Weijer, 1995). Various ethical principles, legal requirements, and policy statements have been formulated at national, international, and intergovernmental levels in an attempt to ensure that clinical research is conducted in accordance with the highest scientific and ethical standards.
The predominant ethical framework for human experimentation was set out by the US National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research (1979) in the Belmont Report. This report articulated three guiding principles for research: respect for persons, beneficence, and justice. These principles are also found in leading international formulations of research ethics. Respect for persons requires that the choices of autonomous individuals be respected and that people who are incapable of making their own choices be protected. This principle underlies the requirement to obtain informed consent from study participants and to maintain confidentiality on their behalf (Levine, 1988). The principle of beneficence requires that participation in research be associated with a favorable balance of potential benefits and harms (Levine, 1988). In the articulation of the Belmont Report, beneficence included non-maleficence – the principle commonly understood as the injunction to “do no harm.” The principle of justice entails an equitable distribution of the burdens and benefits of research. Researchers must not exploit vulnerable people or exclude without good reason eligible candidates who may benefit from participation in a study (Levine, 1988).
There is an ongoing and important discussion about the sufficiency and continuing relevance of the Belmont Report and its principles (Childress et al., 2005), since the principles it set out do not exhaust the ethical requirements for clinical research (Meslin et al., 1995). Conditions such as the following must also be met.
The researcher’s duty to have informed consent from research subjects is established in almost all of the world’s legal systems. The legal doctrine often described as “informed consent” is better understood as “informed choice,” since a physician’s legal duty is to inform the subject so that he or she may exercise choice – which does not always result in consent. The physician’s duty to disclose information relevant to the choice that the subject is asked to make falls under an aspect of civil (that is, non-criminal) law: the law of negligence. A physician may be found negligent if a subject’s choice (including the choice to forgo medically indicated treatment) is inadequately informed and results in harm (Truman v. Thomas, 1980). Accordingly, subjects who are invited to enter a study must be informed of, among other things, the nature and extent of the known risks of participation, the possibility that participation may present unknown risks, and the intended benefit of the study to participants and others. A subject’s treatment in a trial without consent may be grounds for legal action on the basis of “unauthorized touching,” which is dealt with in two domains in the UK: assault in criminal law and battery in civil law.
The duty to ensure confidentiality is usually implicit in the physician–patient contract, fiduciary duty, and legislation, and it is now equally applicable to research subjects. Confidentiality is a usually implicit term of the physician–patient contract (that is, the tacit agreement between physician and patient on the rendering of care), and its violation is, therefore, a basis for legal action against the physician. Increasingly, however, as physicians move from fee-for-service payment to salaries or other remuneration systems, confidentiality is addressed under the law of fiduciary duty (Dickens, 1994). Fiduciary duty – the highest standard of duty implied by law – requires that physicians disclose information about a patient only in the patient’s best interests and that they avoid any conflict of interest in the disclosure of patient information (even if that information is contained in records physicians lawfully hold). Unauthorized disclosure is actionable as a breach of fiduciary duty. It may also violate a duty of confidentiality enacted in legislation. For example, in the Canadian province of Quebec, the Civil Code is so protective of patient information that some have claimed that anonymous epidemiological studies may be unlawful without the consent of each person whose medical record is used (Deleury and Croubau, 1994).
A number of international policies guide the conduct of research. Although the Nuremberg Code (1947) and the UN International Covenant on Civil and Political Rights (1958) remain important early statements, the World Medical Association’s Declaration of Helsinki, as amended most recently in October 2000, with clarifications in 2002 and 2004, is probably the most influential document governing research worldwide (World Medical Association, 1964). Many of the requirements set out under “Ethics” in this article reflect the Declaration of Helsinki. The Declaration highlights an important additional requirement: subjects’ participation in research should not put them at a disadvantage with respect to medical care.
Researchers conducting studies funded by the US National Institutes of Health must do so in accordance with the regulations of the US Department of Health and Human Services (1991), and where the studies involve drugs, biological products, or devices that are to be licensed or marketed in the USA, the research must be conducted in accordance with additional regulations from the Food and Drug Administration (US Department of Health and Human Services, 1991). Researchers conducting research in other countries should consult the guidelines of the Council for International Organizations of Medical Sciences (CIOMS) (CIOMS, 2002). Similarly, geneticists, for instance, should consult the guidelines developed by the Human Genome Organization (1996).
Both the Nuremberg Code and the Declaration of Helsinki address how medical research should be conducted, but neither offers a definition or description of what is medical research (Levine, 1988) as opposed to other medical interventions such as therapeutic innovation. The CIOMS 2002 International Ethical Guidelines for Biomedical Research Involving Human Subjects stated:
The term “research” refers to a class of activity designed to develop or contribute to generalizable knowledge. Generalizable knowledge consists of theories, principles or relationships, or the accumulation of information on which they are based, that can be corroborated by accepted scientific methods of observation and inference. In the present context “research” includes both medical and behavioural studies pertaining to human health. Usually “research” is modified by the adjective “biomedical” to indicate its relation to health.
The CIOMS guidelines go on to note that:
Research involving human subjects includes: studies of a physiological, biochemical or pathological process, or of the response to a specific intervention – whether physical, chemical or psychological; in healthy subjects or patients; controlled trials of diagnostic, preventive or therapeutic measures in larger groups of persons, designed to demonstrate a specific generalizable response to these measures against a background of individual biological variation; studies designed to determine the consequences for individuals and communities of specific preventive or therapeutic measures; and studies concerning human health-related behavior in a variety of circumstances and environments.
Research that is not intended to result in generalizable knowledge, such as quality-assurance studies for internal institutional use, and audits of departmental safety and efficacy of care, is generally not considered research. Proposals for research involving human subjects must be submitted to a relevant research ethics review board. Whether a proposal requires review by such a board is a decision for the board itself, which may find a proposal to fall outside its mandate or jurisdiction. Investigators should not decide for themselves not to submit their proposals for review, except in self-evident cases such as department audit not intended for publication.
Empirical studies have much to contribute to our understanding of informed consent and the risks and benefits of participation in research. For example, if the principle of respect for persons is to be upheld, it follows that research subjects must not only be informed of the purpose, nature, risks, benefits, and alternatives associated with their participation but must also understand this information. But how well do research subjects understand information presented to them in the consent process? The answer seems to be “Not well at all” (Reicken and Ravich, 1982). Indeed, because of a phenomenon that Appelbaum and colleagues (1987) referred to as “therapeutic misconception,” patients commonly believe that experimental projects are tailored to optimize their individual care. In its final report, the White House Advisory Committee on Human Radiation Experiments (1995) detailed the results of a survey of 1900 research participants and concluded that serious deficiencies remain in the current system of protecting human subjects of research.
Two lessons follow from the empirical studies on informed consent to participation in research. Firstly, researchers need to establish and maintain effective strategies to ensure that research subjects comprehend the information they are given during the consent process. In an elegant review of this topic, Silva and Sorrell (1988) listed a wide range of methods available to improve participants’ understanding. Secondly, although such additional measures are important, the empirical data highlight the inadequacy of consent alone to protect study participants. Consent is a necessary component of this protection, but not sufficient. A research study also must present an acceptable balance of risks and benefits (US Department of Health and Human Services, 1991).
Empirical studies on the risks and benefits of research participation have also made an important contribution to research ethics. For many years, participation in research was viewed as a risky endeavor, one from which people ought to be protected (Levine, 1994). However, a number of studies in the late 1970s and early 1980s showed that the risks associated with study participation were, in reality, relatively small (Weijer, 1996). Indeed, recent empirical work in oncology suggests that patients with cancer who participated in clinical trials received – apart from the specific study treatment – a net benefit, namely, improved survival (Hjorth et al., 1992; Freedman and Shapiro, 1994; Weijer et al., 1996). If further study establishes conclusively that trial participation in itself is associated with a higher probability of benefit, it may be that prospective study participants should be informed of this fact.
Ethical issues in research must not be addressed by researchers as an afterthought. Ethical issues permeate research and must guide research design. What should be used as a control treatment? Who should be included or excluded from a study? How large should the sample be? All of these questions have an ethical component (Freedman and Shapiro, 1994). Researchers ought, therefore, to consider ethical issues from the first stages of planning.
A critical component in assuring the protection of human subjects in research is the prior review and approval of any study by an ethics review committee. Although these institutional committees are referred to by different names–in the USA they are institutional review boards, in Canada they are research ethics boards, and in other countries they are ethics review committees – they have a common commitment to ensuring the protection of the rights and welfare of subjects. These committees are increasingly being consulted prior to the design and conduct of research and have expertise in research ethics.
What resources are available to researchers to guide them in ethical matters? Clearly, all physicians involved in research ought to be familiar with the key documents outlined above, particularly the Declaration of Helsinki, CIOMS guidelines, and comparable guidelines. Though directed primarily toward an American audience, a number of excellent reference texts are available (Levine, 1988; US Office for Protection from Research Risk, 1993). Many peer-reviewed journals now discuss ethical issues in research, and two in particular focus exclusively to research ethics: IRB: Ethics and Human Research, which has been available many years (formerly as IRB: A Review of Human Subjects Research), and a new journal, the Journal of Empirical Research on Human Research Ethics. Both are excellent sources for the researcher in an ethical quandary. Finally, and perhaps most important, clinicians should routinely consult with colleagues who have expertise in the ethics of research, particularly where relevant in developing countries (Benatar, 2002), including members of research ethics boards.
Dr. A is offered a financial reward if he will provide the names of patients to a third party who is coordinating a research study. Such “finders’ fees” are ethically and legally objectionable (Lind, 1990). Physicians act in breach of fiduciary duty and in conflict of interest if they use their professional knowledge of a patient’s medical or other circumstances for their personal benefit. Firstly, names may not be given to third parties without patient consent. A physician who believes that entry in a study may benefit an eligible patient should inform that patient and let the patient decide whether his or her name may be given to the investigator. Secondly, physicians must not accept a fee based on the number of names provided. If a physician is asked to consult patients’ records or to do other searches, he or she may be remunerated for the time required to perform that service, whether or not any patients are identified and consent to participate.
Dr. B is invited to enroll his patients in a placebo-controlled study of a new antipsychotic drug. Is it ethical for him to recommend the study to his patients? No. As we have discussed, consent alone is an insufficient ethical basis for enrolling patients in a study: the study must present a favorable balance of benefits and harms. A physician may recommend participation in a study only if the treatments being studied are in a state of “clinical equipoise,” that is, if there is “genuine uncertainty” within “the expert clinical community about the comparative merits of the alternatives to be tested” (Freedman, 1987b). In other words, genuine uncertainty must exist in the community of expert practitioners as to the preferred treatment (Freedman, 1987b). When effective standard treatment exists for a disease, as it does for schizophrenia (Kane, 1996), it is unethical (since placebo is an inferior “treatment”) to expose patients to the risk of “treatment” with placebo alone. Practicing physicians may be told that placebo controls are necessary in clinical research for scientific, ethical, or regulatory reasons. Freedman and colleagues have reviewed these claims comprehensively and concluded that practitioners should regard them with skepticism (Freedman et al., 1996a, b).
An earlier version of this chapter has appeared: Weijer, C., Dickens, B., and Meslin, E. M. (1997). Research ethics. CAMJ 156: 1153–57.