The origins of pediatric research ethics

The unique vulnerability of children as research subjects came to light when Henry Beecher published a landmark article in the New England Journal of Medicine entitled, “Ethics and Clinical Research.” Beecher’s article cataloged several research studies that he argued were ethically unacceptable. One of them was a study by Saul Krugman and colleagues, conducted at the Willowbrook State School, a residential facility for children with neurocognitive problems. In Krugman’s studies, some children were deliberately infected with hepatitis to study the natural history of hepatitis and better characterize different types of hepatitis, with the ultimate goal of developing a vaccine against hepatitis. Although these studies were done with parental permission, Beecher argued that they were unacceptable because the risks to the children were too high and the informed consent process lacking: “Artificial induction of hepatitis was performed in an institution for mentally defective children in which a mild form of hepatitis was endemic. The parents gave consent for the intramuscular injection or oral administration of the virus, but nothing is said regarding what was told them concerning the appreciable hazards involved” (p. 1359).

Other critiques of Krugman’s studies soon followed. In 1970, theologian Paul Ramsey wrote of the Willowbrook studies: “Such use of captive populations of children for purely experimental purposes ought to be made legally impossible … stopped by legal acknowledgement of the moral invalidity of parental or legal proxy consent for the child to procedures having no relation to a child’s own diagnosis or treatment.” Five years later, the Lancet published an exchange of letters about the Willowbrook studies. Steven Goldby wrote that “it was indefensible to give potentially dangerous infected material to children, particularly those who were mentally retarded, with or without parental consent, when no benefit to the child could conceivably result.” Krugman (1971) himself wrote back, defending the studies on the grounds that the children involved did benefit, because:

(1) they were bound to be exposed to the same strains under the natural conditions existing in the institution; and (2) they would be admitted to a special, well-equipped, and well-staffed unit where they would be isolated from exposure to other infectious diseases which were prevalent in the institution—namely shigellosis, parasitic infections, and respiratory infections—thus, their exposure in the hepatitis unit would be associated with less risk than the type of institutional exposure where multiple infections could occur.

This debate polarized the research community. The editors of the Lancet criticized the Willowbrook studies and suggested that they would no longer publish Krugman’s papers. In contrast, the editors of the Journal of the American Medical Association published Krugman’s follow-up studies, along with a laudatory editorial criticizing what they called the pious tone of the Lancet editorial and suggesting that Krugman’s studies were ethically justifiable. A decade later, medical historian David Rothman (1982) wrote of Willowbrook that the parental consent was meaningless because “[t]he consent form that parents signed to allow their children to be infected with the virus read as though their children were to receive a vaccine against the virus.” Furthermore, he argued, parents often consented to the studies to get their children out of the overcrowded wards at Willowbrook and into the superior accommodations of the research wing. Thus, he argued, the supposed benefit of better care in the research wing of Willowbrook was a coercive inducement to participate in the studies. Such a justification for research, he suggested, would specifically put poor children at risk of participating in the most risky research.

The debate about Krugman’s Willowbrook studies exemplifies a central feature of the ongoing debates about pediatric research, specifically which studies are justifiable, for which populations of patients, and with what safeguards and oversights. Many children cannot participate in decisions about their participation in research. Even those who can participate in such decisions may not be able to fully understand the risks and benefits. Parents, physicians, and scientists all have an obligation to protect children from the harms of research, but they may also hope to discover new treatments or cures for childhood illnesses. The tension between the 2 goals of protection and progress is inevitable.

Reasonable people can disagree about the proper balance in any particular intervention. Those disagreements occasionally attract public scrutiny. Specific controversial studies become paradigm cases that serve as the basic building blocks of the unique body of moral philosophy that is the foundation of pediatric research ethics. This article reviews the current federal guidelines for research in children as well as some of the controversies that test the application of those guidelines.

Ethics and regulatory oversight of pediatric research

Research that involves children has always been ethically problematic. The fundamental reason is straightforward. Research, by its nature, uses subjects as a means to the end of creating knowledge that can be generalized. In adults, the solution to this fundamental ethical problem is to get the voluntary, informed consent of the research subject. Children cannot consent on their own behalf. Instead, researchers, parents, and regulators must determine whether the risk/benefit ratio is acceptable to permit the research to go forward.

The regulations governing research conduct in the United States have always included special requirements and considerations for pediatric research subjects and other vulnerable populations. These special requirements call for a higher level of scrutiny and more stringent thresholds of protection than for less vulnerable populations.

Although the reasons for extra scrutiny of pediatric research are straightforward, the arguments for the necessity of doing research in children are also compelling. Children, and particularly infants, respond differently to drugs and other medical treatments than do adults. There are many stories of drugs that, although safe in adults, have serious and even fatal side effects in children. Studies in children often require longer follow-up than studies in adults to determine whether innovative treatments have any long-term developmental effects.

The current regulatory guidelines for pediatric research define 4 levels of risk in research studies, each of which is subject to a different level of regulation and oversight ( Table 1 ). The lowest level of risk is minimal risk. Minimal risk is defined as “the probability and magnitude of physical or psychological harm that is normally encountered in the daily lives, or in the routine medical or psychological examination, or health children” (National Commission, Research Involving Children ). Studies that involve only minimal risk can be performed, even if they do not offer any prospect of direct benefit to the research subjects. For such studies, researchers only need the permission of 1 parent and the assent of the child, if the child is old enough and cognitively capable of giving assent.

However, this categorization is problematic because it is vague. First, the definition does not specify whether these risks should be interpreted relative to normal, healthy children, or relative to the sick children who are like those to be enrolled in the study. The normal daily lives of sick children might include invasive procedures, treatments, and discomforts. Second, the daily life of a child can be risky. Children are at risk of injury when they ride a bike, play competitive sports, take ballet lessons, or climb trees, but these risks are different from those to which a child is exposed in a research study.

This vagueness creates variable interpretations by investigators and institutional review boards (IRBs). In a survey study of IRB chairmen, Shah and colleagues identified marked variation in assessment of the level of risk associated with different procedures; for example, allergy skin testing was found by 23% of those surveyed to be minimal risk, by 43% to convey a minor increase compared with minimal risk, and by 27% to impose more than a minor increase compared with minimal risk; similar variation was observed in assessment of potential for direct benefit.

The second level of risk is a minor increase compared with minimal risk. Research with this level of risk, and with no prospect of direct benefit to the research subjects, may still be approved by an IRB, but only if the research is likely to yield knowledge that is of vital importance to understanding or ameliorating the child’s disorder or condition. The research risks are acceptable if they are commensurate with those in the child’s actual or expected medical, dental, psychological, social, or educational situations. For research in this category, the permission of both parents is required, as is the child’s assent (again, if the child is developmentally capable of providing it).

The third risk category is for studies that involve risks that are greater than a minimal or even minor increase compared with minimal risk but that also include the prospect of direct benefit to the child. These studies were previously referred to as therapeutic research, but that term has gone out of favor. In such studies, the task of the IRB is to conduct a risk-benefit assessment and to determine whether the potential anticipated benefit justifies the risk. Although tangible, quantifiable information about the likelihood of these risks and benefits may be available, this assessment may be subjective. Conduct of these studies requires the consent of 1 parent and the assent of the child.

Implicit in the ethical conduct of studies that are in this third risk category is the assumption of clinical equipoise. First described by Benjamin Freedman, clinical equipoise is defined as “a state of genuine uncertainty on the part of the clinical investigator regarding the comparative therapeutic merits of each arm in a trial.” However, as Freedman also pointed out, these studies cannot be absent of merit: “Equipoise is an important concept in the conduct of ethically sound pediatric research, which must be conducted in pursuit of findings that are of scientific merit.” According to Freedman, clinical equipoise exists when there is “an honest, professional disagreement among expert clinicians about the preferred treatment.” An ethically sound clinical trial should offer reasonable hope of resolving this disagreement. In contrast, if existing evidence clearly favors one treatment or another, or if the proposed trial is unlikely to disturb the state of equipoise, the trial should not proceed as designed. Research must be done in authentic pursuit of answers to valid clinical questions, and patients should not be subjected to research risks in the absence of genuine belief that the answer to the research question is unknown, which also applies to research that involves children. In the initial interpretation of the National Commission’s recommendations for research involving children, Albert Jonsen described the need for this research to be “valuable and necessary for the health and wellbeing of children.”

The fourth risk category is the most complex and the most unusual. It is for studies that involve more than a minor increase above minimal risk, no prospect of direct benefit for the child, but that are judged to be so important in terms of the knowledge that they might yield, that they ought, perhaps, to be conducted anyway. Studies that meet these criteria must be likely to yield knowledge that will prevent or alleviate a serious health condition in childhood. IRBs cannot approve these studies. If an IRB determines that a proposed study meets these criteria, it must refer the study to the Federal Government, which will convene an expert panel to review the proposed study and to decide whether it may go forward. Such studies, if approved, require the permission of both parents and the assent of the child.

Ethics and regulatory oversight of pediatric research

Research that involves children has always been ethically problematic. The fundamental reason is straightforward. Research, by its nature, uses subjects as a means to the end of creating knowledge that can be generalized. In adults, the solution to this fundamental ethical problem is to get the voluntary, informed consent of the research subject. Children cannot consent on their own behalf. Instead, researchers, parents, and regulators must determine whether the risk/benefit ratio is acceptable to permit the research to go forward.

The regulations governing research conduct in the United States have always included special requirements and considerations for pediatric research subjects and other vulnerable populations. These special requirements call for a higher level of scrutiny and more stringent thresholds of protection than for less vulnerable populations.

Although the reasons for extra scrutiny of pediatric research are straightforward, the arguments for the necessity of doing research in children are also compelling. Children, and particularly infants, respond differently to drugs and other medical treatments than do adults. There are many stories of drugs that, although safe in adults, have serious and even fatal side effects in children. Studies in children often require longer follow-up than studies in adults to determine whether innovative treatments have any long-term developmental effects.

The current regulatory guidelines for pediatric research define 4 levels of risk in research studies, each of which is subject to a different level of regulation and oversight ( Table 1 ). The lowest level of risk is minimal risk. Minimal risk is defined as “the probability and magnitude of physical or psychological harm that is normally encountered in the daily lives, or in the routine medical or psychological examination, or health children” (National Commission, Research Involving Children ). Studies that involve only minimal risk can be performed, even if they do not offer any prospect of direct benefit to the research subjects. For such studies, researchers only need the permission of 1 parent and the assent of the child, if the child is old enough and cognitively capable of giving assent.

However, this categorization is problematic because it is vague. First, the definition does not specify whether these risks should be interpreted relative to normal, healthy children, or relative to the sick children who are like those to be enrolled in the study. The normal daily lives of sick children might include invasive procedures, treatments, and discomforts. Second, the daily life of a child can be risky. Children are at risk of injury when they ride a bike, play competitive sports, take ballet lessons, or climb trees, but these risks are different from those to which a child is exposed in a research study.

This vagueness creates variable interpretations by investigators and institutional review boards (IRBs). In a survey study of IRB chairmen, Shah and colleagues identified marked variation in assessment of the level of risk associated with different procedures; for example, allergy skin testing was found by 23% of those surveyed to be minimal risk, by 43% to convey a minor increase compared with minimal risk, and by 27% to impose more than a minor increase compared with minimal risk; similar variation was observed in assessment of potential for direct benefit.

The second level of risk is a minor increase compared with minimal risk. Research with this level of risk, and with no prospect of direct benefit to the research subjects, may still be approved by an IRB, but only if the research is likely to yield knowledge that is of vital importance to understanding or ameliorating the child’s disorder or condition. The research risks are acceptable if they are commensurate with those in the child’s actual or expected medical, dental, psychological, social, or educational situations. For research in this category, the permission of both parents is required, as is the child’s assent (again, if the child is developmentally capable of providing it).

The third risk category is for studies that involve risks that are greater than a minimal or even minor increase compared with minimal risk but that also include the prospect of direct benefit to the child. These studies were previously referred to as therapeutic research, but that term has gone out of favor. In such studies, the task of the IRB is to conduct a risk-benefit assessment and to determine whether the potential anticipated benefit justifies the risk. Although tangible, quantifiable information about the likelihood of these risks and benefits may be available, this assessment may be subjective. Conduct of these studies requires the consent of 1 parent and the assent of the child.

Implicit in the ethical conduct of studies that are in this third risk category is the assumption of clinical equipoise. First described by Benjamin Freedman, clinical equipoise is defined as “a state of genuine uncertainty on the part of the clinical investigator regarding the comparative therapeutic merits of each arm in a trial.” However, as Freedman also pointed out, these studies cannot be absent of merit: “Equipoise is an important concept in the conduct of ethically sound pediatric research, which must be conducted in pursuit of findings that are of scientific merit.” According to Freedman, clinical equipoise exists when there is “an honest, professional disagreement among expert clinicians about the preferred treatment.” An ethically sound clinical trial should offer reasonable hope of resolving this disagreement. In contrast, if existing evidence clearly favors one treatment or another, or if the proposed trial is unlikely to disturb the state of equipoise, the trial should not proceed as designed. Research must be done in authentic pursuit of answers to valid clinical questions, and patients should not be subjected to research risks in the absence of genuine belief that the answer to the research question is unknown, which also applies to research that involves children. In the initial interpretation of the National Commission’s recommendations for research involving children, Albert Jonsen described the need for this research to be “valuable and necessary for the health and wellbeing of children.”

The fourth risk category is the most complex and the most unusual. It is for studies that involve more than a minor increase above minimal risk, no prospect of direct benefit for the child, but that are judged to be so important in terms of the knowledge that they might yield, that they ought, perhaps, to be conducted anyway. Studies that meet these criteria must be likely to yield knowledge that will prevent or alleviate a serious health condition in childhood. IRBs cannot approve these studies. If an IRB determines that a proposed study meets these criteria, it must refer the study to the Federal Government, which will convene an expert panel to review the proposed study and to decide whether it may go forward. Such studies, if approved, require the permission of both parents and the assent of the child.