Ethical Considerations in Pediatric Oncology Clinical Trials

Historic Perspective

For hundreds of years, medical research has involved children as subjects. Edward Jenner first used an experimental smallpox vaccine on his own 1-year-old son at the end of the eighteenth century. At the end of the nineteenth century, a 9-year-old child was the first human recipient of Louis Pasteur's rabies vaccine. However it was not until the first half of the twentieth century that regulations governing such experimental therapy began to be established and the importance of consent and parental permission became clearer.

The informed consent of human research subjects was first officially advocated in the Nuremberg Code, a code of ethics that grew out of the trial of German physicians who conducted egregious human experiments in Nazi concentration camps during World War II. However because this important code emphasized the absolute requirement of informed consent, it implicitly excluded children—who cannot give informed consent—from participating in research on human subjects.

The issue was more explicitly addressed in 1964, when the World Medical Association adopted a set of research ethics principles now known as the Declaration of Helsinki. This set of principles established the priority of the human subjects' interests over those of science and society, and it sanctioned the participation of children in research if permission was given by the child's responsible guardian.

Interestingly experiments were performed on children after the publication of the Nuremberg Code, which implicitly excluded children from research, and before the Declaration of Helsinki, which explicitly allowed such research under specific conditions. One of the most important of these experiments was a series of radiation exposure experiments. Among participating institutions was the Fernald School in Framingham, Massachusetts and the Massachusetts Institute of Technology with funding from the Quaker Oats Company, in which children deemed mentally retarded were fed radioactive iron and calcium in their cereal. President Clinton's Advisory Committee on Human Radiation Experiments revealed the ethical consequences of inadequate parental permission in these experiments.

Another important set of experiments on children at that time occurred at the Willowbrook institution (located on New York's Staten Island) and was designed to follow the natural history of hepatitis in children deemed mentally retarded. Newly arrived children were infected with hepatitis virus by housing them with others already known to be infected. Participating children were kept in a special unit with better conditions and nutrition, and children whose parents agreed to participation were admitted more rapidly. The lead investigator did require a thorough consent process that included a 2-week waiting period for full deliberation. However because Willowbrook was so crowded, critics subsequently argued that the expedited admission and special treatment amounted to coercion. These two cases underscored the need for ethical review and oversight of human subjects research involving children, even when parental permission is obtained.

The United States Congress became increasingly concerned about research ethics during the early 1970s, in part because of the syphilis study involving poor black men conducted by the U.S. Public Health Service at the Tuskegee Institute. As a result the National Commission for the Protection of Research Subjects of Biomedical and Behavioral Research was formed, and it published the Belmont Report in 1979. This report embraced three principles that are now familiar and accepted as crucial for research involving human subjects: respect for persons, beneficence, and justice. The principle of beneficence acknowledges the Hippocratic maxim “do no harm” and extends it to include maximizing the possible benefits and minimizing the possible harms to research subjects (minimizing harms is often formulated separately as a fourth principle, nonmaleficence). The principle of beneficence is operative in any discussion weighing the risks and benefits of research on human subjects, especially when the subject is a minor. The principle of justice concerns the right and fair distribution of the benefits of research, as well as the burdens, an issue that has a particular urgency when clinical treatment and research are considered globally. The concept of respect for persons includes two principles: that individuals be treated as autonomous agents and that those with less autonomy are entitled to protection. The latter principle is especially relevant in pediatrics and is somewhat fluid in its definition because of the growing autonomy of pediatric patients as they approach adulthood.

Given the past abuse of research subjects who were harmed in studies that they did not understand and for which they had not given meaningful permission, it is understandable that “informed consent” has become the dominant concept in discussion about ethical research on human subjects. However recent arguments have been put forth that informed consent is not always necessary for ethical clinical research, nor is it sufficient to qualify research as ethical. A recent review of major codes and declarations relating to research with human subjects led to the proposal that there be seven requirements for ethical clinical research. The research must lead to enhancement of health or knowledge. It must be methodologically rigorous. Selection of study sites and of individual subjects should be determined by scientific objectives and the potential for and distribution of risks and benefits. Given standard clinical practice and the research protocol, risks should be minimized, the potential for benefits should be enhanced, and risks must be outweighed by potential benefits to individuals and knowledge for society. Individuals should be informed about the research and provide voluntary consent. Research subjects should have their privacy protected, have the opportunity to withdraw, and have their well-being monitored. These requirements are thorough and persuasive, but they still must be adapted to the conditions under which the research is conducted. Factors of health, economy, culture, religion, and technology will affect how these requirements are translated into concrete, ethical action. For this reason, the specific context in which the research is conducted will affect ethical deliberation.

Ethical decision making in therapeutic and research efforts for children comprises all of the previously discussed considerations, plus the profoundly salient fact that the recipient of the risk or benefit is not the person who makes the decisions. The minor may not be capable of assimilating sufficient information to meaningfully participate in an “informed choice.” One need not enter the arena of international bioethics to find challenges to the role of “autonomy” in ethical decision making: such challenges are the daily fare in pediatric medicine and research, because children's “autonomy” is an issue that changes from day to day.

What is “Consent” in Pediatrics?

The Nuremberg Code provides one starting point for understanding the meaning and importance of informed consent. The statement in this code that “the voluntary consent of human subjects is absolutely essential” has been interpreted as legal capacity, a power of free choice based on knowledge and comprehension. Consent, which was once seen as a single event, has come to be understood to be more of a process. It is not surprising, with the development of guidelines and goals for consent, that the process has become more intentional and more highly scrutinized.

The dominant theoretical framework for morally valid informed consent requires that four criteria be met: disclosure, understanding, voluntariness, and competence. Briefly, the core set of information that must be disclosed includes (1) facts (such as risks, benefits and alternatives) that patients (subjects) and providers believe are relevant to the decision, (2) the recommendation of the professional, and (3) the purpose, nature, and limitations of consent. Understanding goes beyond disclosure because, although the elements disclosed can be objectively stated, true understanding involves many variables and is more difficult to assess. Establishing and documenting understanding remains a great challenge, because information that has been disclosed but not understood contributes little to the ideal paradigm of informed consent. Voluntariness —another complex notion susceptible to interpretation—means at least that a decision is made without constraints of coercion or manipulation. Finally, competence , which is conventionally understood as the ability to perform a task, has also become a complex concept the definitions of which derive from law, psychiatry, and philosophy. At base a person's competence (or perhaps more accurately expressed as cognitive capacity) to make a particular decision relates to the person's ability to understand and think rationally about available choices, to weigh the benefits and burdens and risks within the context of his or her life and values, and then to use that understanding and logic to make a rational decision.

In the context of complex pediatric medical therapy or research, the process of consent is influenced by the fact that parents are making high-stakes decisions on behalf of their own child. The Children's Cancer Group (CCG) conducted studies that showed that a sense of pressure is perceived by parents when making the consent decision for intervention in the context of childhood cancer. However a more recent study indicated that despite significant dissatisfaction with the consent process among CCG clinician-investigators, most parents were satisfied with it. The same study indicated that the parents' satisfaction did not constitute adequate informed consent.

Even among clinicians assessment of the process of informed consent varies. The experience of the clinician is relevant in this assessment. Simon et al. found that clinicians with 10 or fewer years of experience were more likely to say that the most important goal of informed consent is to explain the disease and treatment and were more likely to suggest to parents that other children might benefit from the research. The same study found that in the end, when reports from clinicians and parents are compared, clinicians are dissatisfied with aspects of consent that parents seem far more satisfied with. For example, clinicians expressed concern about information overload and about the fact that the consent discussion often occurs while the parents are still in a state of shock about the diagnosis.

Further insights into this difference have been gained from a number of studies addressing the perspectives of parents of children with cancer. One such study used three focus groups of parents of children with cancer to retrospectively examine their perceptions of the informed consent process. High levels of stress were consistently reported, which were attributed to efforts to cope with multiple demands, including assimilating their child's diagnosis, nurturing and supporting the child, understanding the information given to them about the diagnosis and treatment, getting to know an entirely new group of people involved in the child's care, and participating in the child's treatment. Another important point was that the parents did not consistently distinguish research from their child's medical treatment. This finding underscores the importance of clearly explaining that research is optional during the process of “informing.” If no distinction is made between research and treatment, the goal of informing has not been met. Physician-investigators often experience tension between their role as the patient's physician and their role as a researcher offering participation in a clinical trial designed to contribute to generalizable knowledge. This tension suggests that someone besides the physician-investigator should conduct the consent process. The study also suggested that the nature of clinical research, as well as the difference between research and proven current therapy, might be difficult for parents to grasp.

These conclusions were supported by a second study in which parents of children with newly diagnosed cancer were interviewed. All of the participants recalled the diagnosis, and most (80%) recalled survival statistics. However only slightly more than half knew that the treatment protocol involved research and understood the concept of randomization. This finding is striking when coupled with the fact that three fourths of the parents thought that alternatives to enrollment in a randomized protocol had been insufficiently discussed. Because randomization is such an important tool for answering certain clinical questions, this finding suggests that greater emphasis be placed on explaining unfamiliar concepts during the process of “informing.” Again, in this study, the majority of participants were satisfied with the consent process.

A study by Kodish and colleagues confirmed the difficulty of effectively conveying the meaning of randomization during the consent process in pediatric leukemia trials. In this multisite study, informed consent conferences were observed and audiotaped and were then compared with information acquired in interviews with parents shortly after the conference. The investigators found that although randomization was explained in 83% of cases, 50% of the parents did not understand it. Further, parents who did not understand randomization were more likely to consent to the randomized study than did those who did understand it, although this difference was not statistically significant ( P = .07). These findings led to several recommendations for improving understanding, including a clear explanation of the differences between the randomized trial and off-study therapy, assessment of parental understanding of randomization, and further explanation when the idea was not yet grasped.

Other efforts have been made to improve the process of informed consent in pediatrics, especially in the case of complex clinical trials. For example, a study by the Children's Oncology Group (COG) assessed the possibility of staged consent; investigators had the option of obtaining consent over a 28-day period using a staged approach. This option allowed parents and patients more time to discuss and absorb facts about the disease, the purpose of the trial, the design of the study, and the potential risks. Several measures in this study suggested benefit from the staged approach. There was greater understanding of treatment choices and of the distinction between a randomized controlled trial and standard therapy with the staged approach (80% understanding) than in the other studies (62.5%, P = .05).

Contemporary medicine will require the introduction of new methods of informing patients and parents to meet the ethical imperative of informed consent in pediatric trials. Informed consent documents for cancer clinical trials are sometimes long and difficult to understand, with language that serves less to protect human subjects than to decrease institutional liability. Therefore some advances may take the form of technologic and educational methods to explain ever more complex medicine and research to people not formally trained in a medical discipline. One such example is a device known as the Informed Consent Team Link ( Fig. 74-1 ). This device was designed to help families understand phase I trials, which present daunting challenges to fully informed consent. The design was based on information from focus groups of parents, physicians, nonphysician health care workers, and teenaged patients. Among the elements considered important were (1) the need for a “big picture” overview that would create a context for details, (2) delivery of information in small chunks, (3) use of multiple modes of information delivery, and (4) prioritization of information. Such basic observations provided a basis for the design of a device that uses interactive technology, voice, animation, and health education principles to deliver information about phase I trials. Because of children's apparently natural affinity for interactive electronic devices, this approach may be helpful in addressing assent as well.

Concepts about the nature of consent in pediatrics are evolving with changes in contemporary medicine. For example some persons argue that informed consent may be too restrictive a concept and that “valid consent” should be substituted. The three aspects of valid consent are personal competence (Does the patient have the capacity to make the decision?), procedural competence (Is the consent given correctly?), and material competence (Is the procedure consented to appropriate for valid consent?). The concept of valid consent has been explored in the pediatric context by the Society for Industrial and Organizational Psychology Working Committee on Psychosocial Issues in Pediatric Oncology. The value of this notion, they suggest, is that it emphasizes the patients' or parents' understanding of what is being consented to and recognizes that there are both rational and nonrational aspects to the decision-making process that must be understood. The underlying concern is that informed consent has come to mean legally signed documentation rather than real understanding, insofar as that is possible. Because parents come from different backgrounds and because children of different ages differ in their ability to grasp complex concepts, the level of understanding that is attainable may vary from situation to situation. The notion of valid consent acknowledges this fact and counters the idea that a signed document is equivalent to informed consent.

The four elements of informed consent previously mentioned—disclosure, understanding, voluntariness, and competence—are useful for informed consent in adults, but informed consent in pediatrics is complicated by the fact that three parties are involved (parent, child, and clinician/investigator) and that the research subject is the child. In pediatrics, it is not autonomy (the basis of the four components of consent) that takes precedence but rather the best interest of the child. The best-interests notion is clearer in a purely clinical setting than in a research setting, in which interventions are designed to contribute to general knowledge. Despite the difficulty of approximating truly informed consent in the setting of pediatrics, and especially pediatric research, the obligation to advance pediatric medicine lends urgency to continued efforts to offer the closest possible approximation to informed consent.

One important aspect of the difference between informed consent in adult and pediatric medicine and research is the fact that in pediatrics, informed consent is better thought of as a combination of parental permission and (as appropriate) the more complex concept of the assent of the child.

Assent in Childhood Cancer Trials

The ethics of assent is one of the most difficult issues faced in pediatric clinical trials. Many adolescents, and even some younger children, possess the elements of competence or health care decision-making capacity (setting aside for the moment the legal definitions, which hinge on age rather than capacity). This is especially true, perhaps, of children who are exposed to long-term clinical trials and to the environment of a children's hospital for long periods. Certainly younger children are not developmentally capable of comprehending complex protocols, but they do have some level of understanding that increases with age and experience. How should assent be understood and when should it be required? But first, what do we actually mean by “assent” and how can it be distinguished from consent?

Consent refers to an active affirmative agreement to something by an autonomous agent (i.e., a person). It is capable of being given by anyone who meets the minimal criteria of being in the state of mind and body to be an autonomous agent; generally, most adults would qualify as long as they were not placed in a situation in which their decisions might be suspected to be the result of some form of coercion. The ability to give consent is also context and subject dependent in the sense that most people are generally not competent in all spheres of their lives: for instance, many persons who are perfectly capable to make health care decisions or even agree to participate in research may not be competent to make (informed) decisions about automobile repair. For the purposes of this discussion, the relevant area in which cognitive capacity is important is in the domain of health care (and associated human subjects research). In the United States the legal age of consent is 18 years for most realms (nearly all jurisdictions, however, permit older adolescents who have not yet reached the age of 18 years to make some types of health care decisions [i.e., give consent], usually in regard to reproductive medical issues). Assent is a more complex concept; in general it describes a less active form of agreement to a proposed procedure or action, rather than the more active and involved consent. It is used in an attempt to engage children and adolescents in the process of consent that must be given by parents or other authorized surrogate decision makers. However what may qualify as acceptable assent (a grudging or sullen nod of the head or a more energetic and participatory form of positive interaction) remains an area of reasonable concern.

Far less is known about assent than about consent. Much remains to be learned about the practices of institutional review boards (IRBs), the perceptions of parents, clinicians, and children about consent, and the methods that might be most effective in the process of assent. Several studies have looked into these issues and documented the need for further work to address variability in IRB practices and the implementation of assent. One study found that half of IRBs have a method they require investigators to follow when determining which children are capable of assent, whereas half have no such method, and the majority of IRBs rely on investigators' judgment. IRBs need guidance on the implementation of requirements regarding assent. Authors of a second study compared standards for assent, as well as consent forms approved by 55 local IRBs, when reviewing three standardized multicenter research protocols. Standards varied widely; 35 had separate forms and simplified language for assent, and 31 specified lower age ranges for obtaining assent in three studies. For a hypertension study, the age at which assent was required ranged from 6 to 15 years;, for a pain study, the age range was 6 to 12 years; and for a respiratory failure study, the range was 7 to 12 years. It is not clear why some IRBs consider a child of 6 years to be capable of assent while others do not require assent until a child is 15 years old.

Do the regulations help us understand assent? Most research on children is governed by subpart D of 45 Code of Federal Regulations (CFR) 46, which provides additional protections for children (as vulnerable subjects) beyond those specified in subpart A of 45 CFR 46. Subpart D, added in 1983, outlines three categories of research that a local IRB can approve: research not involving greater than minimal risk to participants (45 CFR 46.404), research involving greater than minimal risk but with the prospect of direct benefit to individual participants (45 CFR 46.405), and research involving no greater than a minor increase over minimal risk (with no prospect of direct benefit) but likely to provide generalizable knowledge of the subject's condition or disorder that is vital to understanding or ameliorating it (45 CFR 46.406; Table 74-2 ). Local IRBs cannot approve research that does not fall into one of these categories, and performance of the research is possible only if two conditions are met: (1) the local IRB finds that the research provides a reasonable opportunity to further the understanding, prevention, or alleviation of a serious problem affecting the health and welfare of children, and (2) the protocol is approved by the Secretary of Health and Human Services, after soliciting the opinions of an expert panel and providing for a period of public comment (45 CFR 46.407). According to these regulations governing research on children, assent means “a child's voluntary affirmative agreement to participate in research. Mere failure to object should not, absent affirmative agreement, be construed as assent” (45 CFR 46.402). If the IRB determines that the participants in certain categories of age and maturity are capable of providing assent, investigators must obtain it to proceed. When assent is required, a child's refusal is generally binding.

Can an IRB waive the requirement for assent? Yes, but only if (1) the intervention or procedure offers the possibility of direct benefit that is important to the health or well-being of the child and this intervention or procedure is only available in the research context (45 CFR 46.408[a]) or (2) the IRB determines that children below a certain age in a certain situation or with a certain condition have such a limited capacity to participate in the decision that they cannot reasonably be consulted.

The requirements are difficult to apply, especially in the context of clinical trials for childhood cancer. Most of the children are enrolled in research studies in which the therapy is prolonged and carried out in the context of profound physical, spiritual, and psychological stress. The Bioethics Committee of the COG convened an international multidisciplinary task force to address assent in 2003. This group identified a number of problems with the regulatory framework. In brief the regulations do not take adequate account of the dynamic moral and cognitive development of children and thus allow a child either to have no formal role in decision making or to have the power of veto. Nor do the guidelines make clear what constitutes meaningful assent, and so they leave IRBs and investigators with uncertainty about when they have or have not hit the mark (although other groups have provided guidance in this area). The regulations do not take into account the fact that some clinical research is more complex than other research, and thus some research may be more accessible to the understanding of children than other research. The regulations say that permission from the guardians and assent from the child are distinct decisions, and they do not take into account the manner in which parents and children make decisions together. Finally, cultures in which autonomy is less emphasized are not accommodated by a regulation that potentially sets parental prerogatives against a child's veto power.

Taking into account these factors, the COG Task Force offered three principles governing children's participation in research decisions ( Box 74-1 ). A number of specific recommendations followed from these principles. The Task Force acknowledged that in some settings the principles offered will sometimes conflict. Therefore the recommendations emphasize the importance of establishing procedures for the resolution of conflict rather than attempting to define universal rules. In that way the absence of universal rules governing decisions in families with, perhaps, different cultural backgrounds and assumptions does not end conversation and negotiation. This approach is especially valuable in an increasingly pluralistic society and a global culture of information sharing. Much work remains to be done to understand the processes of communication and learning and the interactions that comprise the permission-assent process.

The flexibility of the approach using conflict resolution acknowledges that the obligation to obtain assent may change, depending on the time and situation, even for the same child. For example assent considerations may differ between a 9-year-old child with standard-risk leukemia who may obtain considerable benefit from enrollment in a phase III leukemia trial and a 9-year-old child who has leukemia that has relapsed multiple times and for whom a phase I trial is being considered. Similarly, a child at the beginning of a 2 and one half-year course of therapy for leukemia will be different in age, maturity, vocabulary, experience, and capacity to understand by the end of that course of therapy ( Fig. 74-2 ). Developmentally appropriate approaches to help the child understand the experience do not have the restrictions and limitations that accompany the binding aspects of assent.

For older adolescents assent should be approached in the same way as adult informed consent, even though parental permission may be required. Some adolescents may have the capacity for consent but may choose to have their parents make decisions about their participation in research. Several studies have recently assessed assent in adolescents. Our understanding of the differences between parents' and adolescents' perceptions of research is growing as we learn more about the assessment of risk (i.e., concern for physical safety) versus aversion (psychological discomfort). More than 3000 adolescents in the United States alone die of chronic illness and cancer each year. In such situations when adolescents meet the criteria for functional competence, the broad ethical consensus is that decisional authority should be granted to them regardless of their legal decisional status. However this is an area in which a wide divergence exists between what ethics may demand but the law allows. Most states and jurisdictions do not permit minors to be sole participants in medical decision making, including consenting to participate in research, unless they have been emancipated according to the laws of the state in which they reside. This restriction does not imply that adolescents should not fully participate in the decision-making process, and there certainly is no legal proscription against their doing so; it simply underlines the difference between good ethical clinical (and research) practice and the legal framework. It might perhaps be better stated that in these situations, if the adolescent is going to be a research subject, assent is necessary but not sufficient.

Pediatric Clinical Trials

The majority of children with cancer are enrolled in clinical trials. This systematic evaluation of interventions is largely responsible for the dramatic improvement in survival during the past 40 years. A number of important ethical concepts influence the design and execution of these trials. Permission and assent have already been discussed. Other important issues include the distinction between therapeutic and nontherapeutic research, the various concepts of minimal risk and clinical equipoise, and the considerations that distinguish phase I and II trials from phase III trials.

As previously described and in Table 74-2 , reports published by the National Commission for Subjects of Biomedical and Behavioral Research formed the basis of current regulations governing research with human subjects, including children. One of the most important distinctions made was between research that does, versus does not, offer the prospect of direct benefit to the child. Some elements of a clinical trial may offer the prospect of direct benefit (for example, chemotherapy for acute lymphoblastic leukemia), whereas other elements may not (e.g., biologic and pharmacokinetic studies designed to improve future trials). What are the general criteria used to determine whether the latter sorts of research are permissible?