2

Foundational Aspects of Human-Subjects Research

INTRODUCTION

During its deliberations, the committee recognized a need to articulate key concepts and definitions over a range of disciplines as part of addressing its statement of task. This chapter is organized in sections that introduce the subject areas that comprise this report, which include air pollution health effects science, air pollution regulation, research involving human subjects, and research oversight. The committee’s approach included the identification of key documents in each of the subject areas that are relevant to the statement of task.

Establishing a clear vocabulary and concepts is intended to set the stage for answering the questions “To what extent have controlled human inhalation exposure (CHIE) studies been valuable to inform and reduce uncertainties in setting pollutant standards?” (Chapter 3), “What are the risks of clinical adverse events posed to CHIE study participants in past studies?” (Chapter 4), “Is it warranted to continue to conduct CHIE studies as part of the U.S. Environmental Protection Agency’s (EPA’s) larger research agenda for air pollutants?” (Chapter 5), and “What methods should be used to characterize and communicate risks associated with participating in CHIE studies?” (Chapters 6 and 7). Below we introduce terms and topics that are presented in greater detail in this chapter and Chapter 3. The purpose here is to orient the reader to our methodologic approach, introduce topics and terms, and describe the organization of the report narrative.

The committee assessed the value of CHIE studies by considering the body of knowledge derived from those studies and the integration of that knowledge with the results of toxicologic and epidemiologic studies in the development of EPA’s Integrated Science Assessments (ISAs)—documents which inform decision making concerning the National Ambient Air Quality Standards (NAAQS), as discussed below. The committee utilized the ISAs to assess how others had utilized CHIE study results and to assess the benefits they provide. Considerations of causality in epidemiology and public health developed by Sir Austin Bradford Hill (Hill, 1965), and as adapted by EPA for its ISAs, were used by the committee as an organizing framework for describing different dimensions of the values provided by CHIE studies (see Chapter 3).

To address whether it is warranted to continue to conduct CHIE studies as part of EPA’s larger research agenda for air pollutants, the committee reviewed recent strategic plans of EPA’s Office of Research and Development and identified some remaining knowledge gaps in air pollution health effects science (see Chapter 5) that influence the issues being discussed herein.

The committee took a more narrow approach in considering risks to test subjects and issues of informed consent. In this instance, the committee utilized the eight CHIE studies provided by EPA to illustrate current practices for articulating study objectives, rationale, and design. Committee members considered in detail the screening processes, inclusion/exclusion criteria, consent procedures, and characterizations of risks to study participants for the Institutional Review Board (IRB). In addition, the committee assessed the adverse events reported to the University of North Carolina IRB for all CHIE studies for the period from January 2009 to February 2015.The committee also considered the ethical analysis framework articulated in the 2004 National Research Council (NRC) report Intentional Human Dosing Studies for EPA Regulatory Purposes: Scientific and Ethical Issues (2004a) in its review of these eight studies.

The committee used requested documentation provided by EPA to develop an understanding of the structure of the CHIE study processes, including its multiple internal oversight steps and opportunities for external input. Through the information provided to the committee, EPA personnel demonstrated a detailed understanding of issues central to the committee’s task. The committee found EPA’s procedures to be consistent with, and indicative of, ethical approaches to human-subjects research, and that finding is consistent with those of EPA’s Office of Inspector General’s report (see Chapter 1).

The committee developed its report in consideration of themes that represent the continuing evolution of scientific knowledge, social and ethical thought, and public policy. The committee also was cognizant of the importance of harmonizing terminology, where possible, to facilitate communication across institutions and disciplines. In general, the committee adopted the language and conventions of human-subjects research and IRBs for describing biologic responses and adverse events. In contrast, the methods for characterizing risk of adverse events to study subjects draws upon language and conventions of epidemiology, toxicology, risk assessment, and communication sciences.

AIR POLLUTION HEALTH EFFECTS SCIENCE FOR MANAGING AIR QUALITY

EPA states that one of its purposes is to ensure that “all parts of society—communities, individuals, businesses, and state, local and tribal governments have access to accurate information sufficient to effectively participate in managing human health and environmental risks” (EPA 2017b). Accurate information about the impact of environmental pollutants on human health is thus a precious commodity, assembled through painstaking, multidisciplinary, iterative efforts. The committee was unanimous in its view of the importance of a robust body of valid scientific information that supports that purpose of EPA.

The committee’s focus is on understanding the health impacts of inhalation exposures to pollutants in the outdoor ambient air in an individual’s immediate surroundings, both when outdoors and when exposed to outdoor pollutants that have penetrated into indoor environments. There is no single approach, no simple oracle that provides this information. Originating in the congressional mandate of the Clean Air Act (CAA), there is a process through which scientists compile and interpret this scientific information on air pollution and health effects, which is then translated into primary (health-related) and secondary (welfare-related) NAAQS (see Table 2-1), which lead in turn to emission control strategies. Because the secondary standards do not focus on human health, they therefore are not within the committee’s scope of work.

As mandated by the CAA, NAAQS have been established for each of the six current criteria pollutants and the NAAQS for each pollutant must be reviewed periodically using a well-defined process. Included in this process is the development by EPA’s Office of Research and Development (ORD) of an ISA. The ISA is an extensive review and summary of peer-reviewed literature relevant to the NAAQS (see Chapter 3).

The composition of ambient air and the concentration of its specific constituents are determined by multiple natural and anthropogenic sources and factors, including emission sources related to mobile sources (such as vehicular traffic), area sources¹ (such as dry cleaning facilities), and larger stationary sources (such as electricity generation facilities). Each human individual has a personal exposure profile, determined in part by his or her regular activities. Decades of research, as summarized in the ISAs and other reports, show that to a remarkable degree, exposures to pollutants in the local outdoor air (or ambient air) influence a broad range of health outcomes (EPA, 2009, 2013; NRC, 2004b).

The scientific studies contributing to the knowledge of air pollution–related health effects, and hence establishing the basis for regulation, include toxicologic studies (using whole animals or cellular cultures), observational studies of health-related responses to exposures while humans are engaged in their regular activities, and CHIE short-term inhalation exposure studies of volunteer subjects to individu-

__________________

al criteria pollutants or mixtures. Each study type has strengths and limitations, but together they can provide a basis for informing EPA decision making (Brown et al., 2007; EPA, 2009, 2013).

Observational studies, such as epidemiologic studies and panel studies, are used to identify associations between ambient pollution concentrations and adverse health effects by examining changes in various health end points, within an entire population or particular subgroups. These studies examine effects associated with short- or long-term exposures to pollutant mixtures, which are generally not controlled by the researcher. However, they are unlikely to elucidate biologic mechanisms or establish causal relationships.

TABLE 2-1 National Ambient Air Quality Standards for Six Criteria Pollutants (as of December 2016)

^aThe Clean Air Act requires EPA to set two types of NAAQS: primary NAAQS to protect public health, and secondary NAAQS to protect the public welfare from known and anticipated adverse effects (such as crop damage from pollutant exposure).

^bThe form defines the air quality statistic that is to be compared to the standard in determining whether an area attains the NAAQS.

^cIn areas designated nonattainment for the Pb standards prior to the promulgation of the current (2008) standards, and for which implementation plans to attain or maintain the current (2008) standards have not been submitted and approved, the previous standards (1.5 µg/m³ as a calendar quarter average) also remain in effect.

^dThe level of the annual NO₂ standard is 0.053 ppm. It is shown here in terms of ppb for the purposes of clearer comparison to the 1-hour standard level.

^eFinal rule signed October 1, 2015, and effective December 28, 2015. The previous (2008) O₃ standards additionally remain in effect in some areas. Revocation of the previous (2008) O₃ standards and transitioning to the current (2015) standards will be addressed in the implementation rule for the current standards.

^fThe previous SO₂ standards (0.14 ppm 24-hour and 0.03 ppm annual) will additionally remain in effect in certain areas: (a) any area for which it is not yet 1 year since the effective date of designation under the current (2010) standards, and (b) any area for which implementation plans providing for attainment of the current (2010) standard have not been submitted and approved and which is designated nonattainment under the previous SO₂ standards or is not meeting the requirements of a SIP call under the previous SO₂ standards (40 CFR 50.4(3)). A SIP call is an EPA action requiring a state to resubmit all or part of its State Implementation Plan to demonstrate attainment of the required NAAQS.

SOURCE: Adapted from EPA (2016c).

Toxicologic studies of experimental exposures to whole animals provide the capability of assessing possible biologic mechanisms for a broad range of health end points in response to specific pollutants or ambient mixtures. In addition, toxicologic studies can provide relevant information on causality when the biologic responses observed are similar to responses expected in humans under ambient exposure conditions. However, uncertainty is introduced when results are extrapolated to humans. Results from in vitro studies using cell cultures that probe interactions of chemicals with cellular components could provide mechanistic information or support for results from whole-animal studies.

CHIE studies of human volunteers are used to help understand relationships between short-term exposure to a known concentration of a defined pollutant or pollutant mixture and an adverse health effect. They also can be used to evaluate the relevance of biologic mechanisms observed in animals, and they can provide information on the biologic plausibility of associations observed in epidemiologic studies. Controlled studies provide the potential for demonstrating causality for the conditions of the particular exposure protocol (for example, specific exposures and the health status of the study subjects). Although some CHIE studies have included health-compromised study subjects, such as those with certain respiratory or cardiovascular diseases, they are not likely to represent the most sensitive individuals in the population (as discussed in subsequent chapters).

Outdoor air quality is a societal resource, protected by the CAA and regulations imposed by EPA. The CAA requires EPA to periodically review the NAAQS for each of its criteria pollutants (specified at 5-year intervals, but often needing more time for completion of its review) and either retaining them or revising them in the light of new information. EPA is currently in the process of reviewing the 2009 NAAQS for particulate matter (PM) (EPA, 2009) in order to take into account the hundreds of peer-reviewed research and review papers that have been published since completion of the prior review. As noted in the chapters of this report that follow, many of these more recent publications confirm and extend the conclusion of the 2009 ISA that ambient air PM has been shown to influence a broad range of pulmonary and cardiovascular health outcomes.

The CAA also mandates that EPA establish and enforce National Emission Standards for Hazardous Air Pollutants (NESHAPs) to protect public health from exposures to known toxicants emitted from definable point sources. Each NESHAP establishment or revision is reviewed by the EPA Administrator following the finalization of the supporting scientific documents by the Environmental Health Committee of EPA’s Science Advisory Board that met in public for its deliberations. The regulatory process for NESHAPs focuses on emissions, rather than ambient air concentrations. However, because the intent of NESHAPs is to protect public health, there is a need for a scientific base of knowledge about the human toxicity of these pollutants at ambient concentrations. Although CHIE studies are most often used to inform NAAQS decision making, they potentially could be used to inform NESHAPs decision making. CHIE studies involving NESHAP pollutants would be subject to the same considerations and reviews as those articulated for NAAQS pollutants.

One goal of our committee’s report is to inform a national conversation about the role of CHIE studies in developing accurate air pollution health-effects information. This conversation, in turn, is part of an evolving and lively discussion about what constitutes ethical research, whether it involves humans or animals. The conversation is informed by social norms and sensibilities, by acknowledgment of past mistakes and transgressions, by divergent views about the risks and benefits of this type of research, and by discussions about what constitutes informed consent of human study subjects. In this chapter, we provide a framework for our deliberations, and reference documents that are benchmarks in this continuing conversation.

HUMAN-SUBJECTS RESEARCH

CHIE studies have been conducted for many decades within and outside of EPA to investigate relationships between air pollution exposure and human health. The 2008 ISA for sulfur oxides (EPA 2008) cites CHIE studies that were undertaken in the 1950s and 1960s to investigate effects of sulfur dioxide (SO₂) exposure on short-term lung function. For example, Amdur et al. (1953) and Frank et al. (1962)

were cited, among other studies, to have observed human respiratory effects at SO₂ exposure concentrations greater than 1 ppm, including increased respiration rates, decrements in peak flow, bronchoconstriction, and increased airway resistance. EPA has been conducting human experimental exposure studies since 1973. The agency has used results of CHIE studies of O₃ exposure and respiratory symptoms, conducted within and outside of EPA, to inform the O₃ NAAQS decisions in 1979, 1997, 2008, and 2015. In the 2000s, there was greater use of PM CHIE studies to investigate possible mechanisms of health effects observed in epidemiologic studies.²

As stated in Chapter 1, EPA asked the National Academies of Sciences, Engineering, and Medicine to assess whether it is warranted to continue to conduct CHIE studies as part of EPA’s larger research agenda for community air pollutants and, if so, to recommend guidance to improve methods of characterizing risks to study subjects and improving informed consent. Underlying that task is the question of whether the CHIE studies meet the requirements for ethical human-subjects research and whether the data and scientific benefits derived from those studies can be ethically balanced against the risks posed to the human subjects who participate in the CHIE studies. This risk–benefit balance is imposed by the standards that govern human research in the United States: the Federal Policy for the Protection of Human Subjects, better known as the “Common Rule,” which governs human research conducted by EPA as well as most other federal agencies. Therefore, the committee’s task has not been to reinvent the ethical structure of the inquiry but rather to consider how that ethical structure should be interpreted in the context of CHIE studies, where the potential risks are borne by the subjects involved in the studies. Volunteers for CHIE studies routinely undergo physical exams to determine their eligibility to participate, and in the process might discover health information of value to them. Such information should not be considered a benefit of participation, because such health information is not the aim of CHIE studies and is fortuitous. Instead, the benefits are enjoyed by society and not by the individuals participating, except within their lives as members of society.

During its information-gathering session on August 24, 2016 (see Appendix B), the committee heard concerns that CHIE studies involving any exposures of study subjects to PM concentrations greater than the NAAQS concentration limit are unsafe and unethical. However, reliance only on a concentration for evaluations in this context is not consistent with the concept of the NAAQS. Each of the NAAQS is stated in terms of an averaging time and a statistical form as well as a concentration (level) (see Table 2-1). A NAAQS concentration for PM_2.5 averaged over 24 hours is permissible on 2% of days averaged over 3 years. The underlying rationale is based on the relationship between cumulative dosages delivered from short-term exposures (24 hours for PM) and acute responses, allowance for extreme meteorological conditions favoring a temporary buildup of pollution, and the margin-of-safety factor built into the setting of a NAAQS. Therefore, evaluation of CHIE studies with respect to the applicable NAAQS needs to consider the exposure concentration and duration. It is also important to consider the adequacy of the process used to screen the volunteer subjects for entry into the study and the medical monitoring of the subjects during the course of the inhalation exposures.

A Brief History of Progress and Ethical Lapses in Human Subjects Research

The history of progress in defining ethical research is summarized in Table 2-2. An accounting of unethical experiments of the past provides a sobering reminder of the paramount importance of careful attention to protocols and procedures, including consent by participants. The committee’s detailed review of eight CHIE studies, including protocols, procedures, and consent forms, reflects a recognition of the importance of this attention.

Behind each of the milestones listed in Table 2-2 lies an event or a pattern of events that demonstrates the need for the current ethical oversight. Researchers forget these precipitating events at their peril. Both in terms of the safety and well-being of research participants, and also in terms of the social un-

__________________

derstanding and support for scientific research, it is important to remember the lapses that led to the current requirements for ethical research practices. For example, the Nuremberg Code was written in response to the atrocities perpetrated by the Nazi physicians during World War II (Nuremberg Military Tribunals, 1949). The code’s emphasis on informed consent, avoiding harms to subjects, and the freedom of subjects to withdraw at any time speak to the absence of all three of these features in the Nazi experiments and provides a basis for preventing unethical experiments in the future. The Declaration of Geneva, dating from 1949, recognizes that researchers are often physicians; it marks the start of a much deeper understanding of research ethics as having different and more stringent ethical requirements than those reflected in ordinary medical ethics codes. Likewise, development of the Belmont Report and U.S. Department of Health and Human Services regulations during the 1970s were a direct result of the revelations of deceit, abuse, and racism in the Tuskegee Syphilis Studies conducted over a 40-year period by the U.S. Public Health Service. After many decades of learning reactively from missteps in research, it is important to be proactive, anticipating problems and providing ethical safeguards against them. In preparing this report, the committee sought to build upon this history of progress by providing more refined and particular ethical guidance for the kind of research done by the EPA, and others, in CHIE studies and other research that enrolls human volunteers. In so doing, the committee hopes to advance the societal collective approach to human-subjects research.

As presented in Table 2-2, guidance from the Food and Drug Administration (FDA) indicates that clinical trials should be registered on the U.S. National Institutes of Health website: www.ClinicalTrials.gov. Editorial policies for scientific journals generally require that this be done in order for study results to be considered for publication. The committee supports the practice of registering CHIE studies on this website. Although many clinical trials have a therapeutic end point as their primary goal, the committee considers the practice of registering CHIE studies on the website as a way to encourage the harmonization of human-subjects protection practices. In addition, registering CHIE studies will help make other investigators aware of this proposed research, potentially preventing duplication and providing opportunities for studies to build upon one other.

Requirements for Ethical Research

The committee reviewed the basis for prior deliberations which concluded that CHIE studies were examples of ethical research. Table 2-3 provides seven requirements for determining whether a human research trial is ethical and is particularly valuable because it presents a coherent principle-based framework for evaluating the ethics of clinical research projects. This framework distills the requirements and insights of a large range of both historical and contemporary ethical statements from U.S. and international sources. The seven requirements are social or scientific value, scientific validity, fair subject selection, favorable risk–benefit ratio, independent review, informed consent, and respect for potential and enrolled subjects. None of the requirements can stand alone as a sufficient justification for a human-subjects research project. While not all are necessary to all forms of research, for example, some emergency research has been exempted from an informed consent requirement, all seven are ethically necessary for EPA’s CHIE studies. Table 2-3 provides additional details, including the ethical values that justify each of the requirements. For example, the social or scientific value of a project is required by considerations of distributive justice (not wasting the scarce resources available for research) and by nonexploitation (not placing human subjects at risk for research that has no scientific or social value). Table 2-3 also indicates the kind of expertise needed to make an evaluation that each of the seven requirements is satisfied. For example, for the informed-consent requirement, expertise in scientific knowledge about the purpose of the trial and the risks and benefits to which participants will be exposed is essential. Equally important, legal and ethical expertise is needed to ensure best practices in the consent process, described in Chapter 7 as disclosing information, encouraging deliberation, and enabling authentic decision making. Table 2-3 provides the ethical foundation on which the deliberations and recommendations of this report are built. Further discussion of the ethical meanings and uses of the concepts of informed consent, risk, and benefit are provided in Chapter 7, along with recommendations about communication of these concepts to research participants.

TABLE 2-2 Selected Laws, Ethical Codes, and Other Milestones Related to the Protection of Human Subjects

Source: Adapted from Sparks 2002.

TABLE 2-3 Seven Requirements for Determining Whether a Research Trial Is Ethical^a

^aEthical requirements are listed in chronological order from conception of research to its formulation and implementation. Source: Emanuel et al. (2000). Reprinted with permission; copyright 2000, Journal of the American Medical Association.

RISK–BENEFIT FRAMEWORK

As noted earlier in this chapter, ethical clinical research must be bounded by a favorable risk–benefit ratio. Indeed, not only must the ratio be favorable, but IRBs—and researchers—are required to ensure that risks are minimized to the extent possible even within a favorable risk–benefit ratio. The 2004 NRC re-

port on intentional human dosing, introduced in Chapter 1, presented a framework for comparing risks and societal benefits for EPA studies that involve deliberate exposures of human subjects to specific chemicals and/or other potentially toxic materials, such as pollutant mixtures. In that 2004 report, the authoring committee determined that it is ethically acceptable to involve volunteers in studies that expose them to somewhat higher risks than studies that pose no identifiable risks or for which there is a reasonable certainty of no harm, if the studies have the potential of providing a clear health or environmental benefit to society and the studies are expected to cause no lasting harms to study participants (NRC, 2004a, p. 105). It provided examples of two extremes, one involving a type of study that was almost certainly acceptable and the other involving a study that was almost certainly unacceptable (NRC, 2004a, p. 107). The former involved a study where the exposure was designed to investigate pharmacokinetic information, including absorption of the chemical and the subsequent metabolism of the chemical, where the exposure has no known biologic effect on the participant. The latter involved a study in a medically vulnerable population with the potential for lasting adverse effects. While these examples present clear boundaries for what would be ethical and unethical research, the 2004 report notes that cases between these two extremes are more difficult to evaluate. Such cases might require sophisticated risk assessments and understanding about the short- and long-term effects of the chemicals involved; this might be beyond the capability of a typical IRB.

Since the publication of the 2004 report, there has been considerable scholarly attention to the risk–benefit ratio as part of the broader ethical framework for biomedical research. In 2011, Annette Rid and David Wendler offered a comprehensive step-by-step method for making risk–benefit evaluations (Rid and Wendler, 2011). This system is not a substitute for the judgment that is required by such an evaluation, but it ensures that appropriate attention is given to all facets of the evaluation.

• Step 1 requires that the proposed study achieve a minimum level of social value. Rather than doing that analysis after determining risks, Rid and Wendler believe that analyzing the expected value is an essential first step that focuses the research more appropriately. In the context of the CHIE studies, this means that the studies are expected to provide information for regulatory decision making that cannot be obtained by other means.
• Step 2 is to identify the research interventions and to ensure the safety of those interventions. Such interventions are expected to be focused on the question posed by the study and aligned with the social values identified in Step 1. In addition, they are expected to yield information likely to be important and nonduplicative. It may require alternatives to be considered to enhance safety.
• Step 3 is to evaluate and reduce or minimize any risks posed by the study. The first element of this step of course is to identify any potential risks, and we discuss this process in the next section on foreseeable risks. Risks are characterized by the probability, magnitude, timing of the onset, and duration of the potential harm. This step also requires another look at alternative procedures that might mitigate risks. Changes to inclusion and exclusion criteria might also mitigate risks. Here Rid and Wendler introduce an important distinction: the purpose of such a risk–benefit evaluation is “not to protect participants from risks … but to reduce the extent to which participants experience harm from participating in a research study” (p. 149).
• Step 4 is to evaluate and enhance the potential societal benefits for participants. In the CHIE studies, there are no anticipated medical benefits for participants. (See Chapter 7 for a discussion of other potential benefits.)
• Step 5 is to evaluate the extent to which potential clinical benefits might offset the risks of undergoing the intervention. (However, this step is not applicable to studies that do not offer clinical benefit and therefore it is not applicable to CHIE studies.)
• Step 6 is to evaluate the extent to which the net risks of some of these interventions might be justified by the potential clinical benefits of other interventions included in the same study. (Like Step 5, this step is not applicable to CHIE studies.)

• Step 7, the final step, determines whether the net risks are justified by the study’s social value. Because of the iterative process of the evaluation, this step might provide a far more thorough analysis of the question than would occur if done less systematically.

One controversial issue that has not been resolved by bioethicists is whether there is an upper limit on how much risk can be tolerated in a study that may have very high social value but provide no direct benefits to participants. The Common Rule does not place any limits on such risks so long as the risk–benefit ratio is favorable. The Nuremberg Code tentatively endorses research that might result in death or disability, but only in those rare cases in which physician-researchers are also the research subjects. In its other provisions, the Nuremberg Code is lucid in its expressed opposition to research that incurs such serious risks. Some people argue that if informed consent is appropriately done, there should be no upper limit in a study involving competent adults. Others have less faith in the informed-consent process and prefer a more paternalistic approach. Resnick (2012) suggests that studies that pose more than a 1% risk of serious harm (defined as death, permanent disability, or severe illness or injury) should not ordinarily be allowed, absent a compelling public health interest and, even then, the acceptable risk should not exceed slightly more than 1%. Although the committee believes that Resnick’s framework could have value by providing a level of acceptable risk, the framework presupposes that risks to CHIE study subjects can be quantified in a reliable manner. As discussed in Chapter 6, the committee recommends an alternative approach to characterizing risk.

Determining Reasonably Foreseeable Risk

The Common Rule requires that studies involving human-subjects research minimize risks to the extent possible, pose risks that are reasonable in relation to the benefits presumed (in the case of EPA CHIE studies, the knowledge expected to be gained that could inform EPA exposure standards), and that risks are appropriately communicated to the subjects. The statement of task for this committee specifically asks it to provide “a template to characterize reasonably foreseeable risks, in terms of the nature, frequency, and magnitude of possible risks, which could be used in obtaining informed consent from potential study participants.” Of course, before the committee can consider how such risks should be communicated, it is important to consider how the potential clinical adverse effects of concern can be identified and characterized.

The term “reasonably foreseeable risks” appears in the Common Rule only as a requirement for informed consent, but it is implicit in the other requirements in the Common Rule for risk determination and assessment. The term is not defined in the rule, nor have any of the agencies adhering to the Common Rule attempted to define the term despite the fact that the issue has garnered considerable scholarly attention in recent years. Moreover, HHS issued an update to the Common Rule on January 19, 2017 (82 Fed. Reg. 7149 [2017]) but it does not deal directly with the issue. In fact, the OIG recommendation to EPA in the context of CHIE studies is but one of several recent calls for better guidance on this issue. For example, the HHS Secretary’s Advisory Committee on Human Research Protections called on FDA to provide a better definition for studies involving FDA-regulated therapies. HHS has issued a draft guidance document designed to help researchers in studies evaluating standard of care treatments to present the reasonably foreseeable risks in the informed consent (OHRP, 2014). While that guidance is not directly applicable to CHIE studies, its logic can be extrapolated to them. It is beyond the scope of the committee’s report to provide a complete framework for all studies covered by the Common Rule, but we attempt to provide such a framework for EPA’s CHIE studies.

The term “reasonably foreseeable risks” as used in the regulations probably has its origins in legal tort law. As EPA notes, a general legal definition of the term is “a danger which a reasonable person should anticipate as the result from his/her actions” (Hill and Hill, 2014, p. 10). In negligence law, if one knows, or should know, of such a danger, one has a duty to avoid or mitigate that danger. Unfortunately, legal scholars have had as much trouble with the term as biomedical researchers, finding it vague and in-

determinate. Two prominent legal scholars note that “in one sense, everything is foreseeable, in another nothing” (Hart and Honore, 1985).

Nonetheless, some aspects of the term can be gleaned from its legal origins. First, it is not a strict liability standard. It does not require researchers to anticipate any hypothetical risk, no matter how remote, that might occur. On the other hand, as an objective standard, it requires researchers to undertake reasonable inquiry to determine what risks may exist. When Ellen Roche, a healthy volunteer, died during a study that sought to understand the pathophysiology of asthma using inhaled hexamethonium at Johns Hopkins University, the federal Office of Human Research Protection (OHRP) suggested that the researchers had failed to find earlier results of research in which subjects had experienced adverse events that exceeded those expected (Kennedy 2001). Although hexamethonium had not been used clinically for some decades, there was literature from the 1950s and 1960s that indicated adverse pulmonary reactions to the drug. Thus, determining “reasonably foreseeable risks” requires a thorough examination of existing literature and all available data to determine any potential adverse consequences.

In 1996, Hoiyan (Nicole) Wan died from a lethal dose of lidocaine that she received while participating in a M.I.T. sponsored medical research project at the University of Rochester Medical Center to investigate if the mutations in the bronchial cells of nonsmokers’ lungs are principally caused by exposure to airborne pollutants. The lung cells were obtained using bronchoscopy, a procedure that involves the insertion of a flexible tube to gather lung cells for analysis. Lidocaine, an anesthetic, was given to make the study subject more comfortable during the procedure. The New York State Department of Health found that the University of Rochester had violated its own guidelines by increasing the dose of the anesthetic (Rosenthal 1996). This case is an example of how procedures and medications used in the procedures pose risks, and therefore need meticulous control and checks.

In its draft guidance, Disclosing Reasonably Foreseeable Risks in Research Evaluating Standards of Care, the Office for Human Research Protections (OHRP) states the following:

In the context of CHIE studies, the sense of that guidance can be extrapolated to mean risks posed to a participant might be different than risks posed during that person’s ordinary daily life. We agree with the approach taken by EPA in adopting the formulation posed by Resnik (2013): “A risk is reasonably foreseeable if we have some credible evidence to expect that it [a potential harm] may occur” (italics supplied). It is important to note that the formulation means there is an “expectation” that the harm may occur. That presents a more stringent criterion than for a hypothetical risk.³ Because CHIE studies involve volunteers who will derive no therapeutic benefit from participation in the research, it is tempting to compare this criterion to “first in human studies” such as Phase 1 drug studies or even studies such as the hexamethonium study described above. But that is generally not an apt comparison because, unlike Phase 1 drug studies, or a study using a relatively novel agent as the challenge, EPA has a great deal of data about the various pollutants’ biologic effects gleaned from epidemiologic studies, animal studies, and previous CHIE studies. The first step in determining “reasonably foreseeable risks” in the controlled-exposure studies is therefore to thoroughly study those data.

Observation of the Common Rule in the Development of EPA CHIE Studies

In this section the committee describes EPA’s development and conduct of CHIE studies with respect to the Common Rule. The purpose of this section is to orient the reader to the terminology and us-

__________________

ages of the review and consent process. The presentation uses an adapted version of the Rid and Wendler risk-benefit framework described above. As noted previously, two of the steps in that framework are inapplicable to studies that do not offer a medical benefit and are therefore inapplicable to the CHIE studies.

Step 1: Determine If the Proposed Study Is Designed to Achieve a Minimum Level of Social Value

The study designs are developed by a qualified Principal Investigator (PI) and undergo in-house review by other EPA staff to determine the extent to which the expected results of the study would support the review of NAAQS or other air quality–related decisions. The proposed study is reviewed by the branch chief who has a detailed knowledge of EPA ORD’s strategic plans and can assess the study’s expected value to society. Additional reviews by agency statisticians and controlled-exposure researchers, and outside experts also focus on the value of the study.

Step 2: Identify the Research Interventions and Ensure the Safety of Those Interventions

This step is completed through the submission of an application to the IRB. The University of North Carolina (UNC) Institutional Review Board is the governing IRB for CHIE studies which are conducted by EPA at its own Human Studies Facility on the campus of the University of North Carolina at Chapel Hill. IRBs at UNC are overseen by the director of the university’s Office of Human Research Ethics. That person reports to the university’s Vice Chancellor for Research, who is the authorized institutional official for UNC.

The IRB process has evolved over the 50 years since it was first created, with definitions that guide the review and reporting process. For the past 50 years, submission of an application for IRB review has been required when an activity includes both research and human subjects. Thus all CHIE studies are required to undergo IRB review and to receive approval prior to conducting the studies.

IRBs use the terms “minimal risk” and “more than minimal risk” within a triage mechanism to organize their review process, but use of those terms does not conduct quantitative risk assessments. IRBs assign submitted protocols for an expedited review if they are considered to pose minimal risk to study subjects. They assign a submitted protocol for full review if it is considered to pose “more than minimal risk.” According to the Code of Federal Regulations (CFR) 46.102(i), minimal risk means that the probability and magnitude of harm or discomfort anticipated in the research are not greater in and of themselves than those ordinarily encountered in daily life or during the performance of routine physical or psychological examinations or tests.

The study protocols provided to the IRB for review will specify study-related risks posed by the controlled exposure (“exposure-related risks”) and those posed by the procedures to be used to measure the selected health-related end points being measured (“assessment-related risks” or “procedure-related risks”). The protocol provides inclusion and exclusion criteria for participant recruitment, and specifies how these criteria will be assessed. The committee focused on factors that could alter participant susceptibility to study-related risks, such as preexisting medical conditions.

The IRB process provides an opportunity to demonstrate that the research methods focus on the question posed by the study, align with the social values identified in Step 1, and are nonduplicative. The IRB application provides the opportunity to discuss alternatives to be considered to enhance safety. A detailed study protocol that is submitted in addition to the IRB application enhances the transparency of this step of the process.

Step 3: Evaluate and Minimize Study-Related Risks to Participants

This step of the Rid and Wendler framework specifies the need to evaluate and minimize study-related risks. The committee recognizes that there are many dimensions to this step, as discussed below.

The ultimate ethical responsibility for all phases of the study lies with the PI. EPA has responsibility for oversight, and the IRB of record has the additional responsibility for monitoring the progress of the study and withdrawing its approval, if indicated. EPA has the responsibility for reporting serious adverse events, or any serious protocol noncompliance, to the reviewing IRB and to federal officials, should such occur. Additionally, prior to proposing the study, and while the study is ongoing, the PI is required to consider prior human, animal, and other biologic effects data that are available, to ensure that study-related risks to the study subjects are outweighed by the utility of the study results for informing air-quality management decisions. As mentioned previously, there is no medical benefit to the study subjects. With the foregoing in mind, Chapter 3 considers the value of the CHIE studies for informing EPA decision making.

Each proposed CHIE study is reviewed by two external experts to evaluate aspects of safety of study subjects, scientific rigor, and adherence to ethical principles (Personal communication, T. Schonfeld, EPA, July 2, 2015). CHIE studies involve placing study subjects in specially designed exposure rooms, where controlled concentrations of air pollutants are introduced and monitored. The subjects are enrolled after it is determined, through EPA’s preexposure health evaluations, that there is no reason to believe that their participation in the study might lead to a clinically adverse effect.

The committee strongly supports the CHIE study practice of using a medical assessment to ascertain health status. The assessment protocol that is developed for a particular CHIE study is approved by the relevant IRB. Elements of the assessment include questions posed to the subject about the presence of medical conditions previously diagnosed by a physician, a physical examination, and selective testing. The medical assessments of study subjects might discover health information of value to them.

The health status of subjects is monitored throughout, shortly before and immediately after the controlled exposures, and again about 24 hours later. If, during the study, there is any evidence that an individual is being or has been harmed by the study, the person is referred for medical observation (or treatment). An individual can participate as a subject in up to six studies per year, provided that it is determined such enrollment will not be harmful to that individual.

Step 4: Evaluate and Enhance the Potential Benefits for Participants

As previously stated, there are no medical benefits for participating in CHIE studies. As discussed in Chapter 7 there are societal benefits and several types of personal benefits.

Step 5: Compare Risks and Societal Benefits

The final step of the Rid–Wendler framework asks whether the net risks are justified by the study’s social value. IRB approval means that the net risks have been determined to be justified by the study’s social value.

Terminology and Usages

Dimensions of Risk

We use the term “exposure-related risks” to refer to risks related to inhalation of the pollutants, and “procedure-related risks” to refer to risks related to measurement of study end points. In general, our primary focus is on exposure-related risks. Potential adverse outcomes associated with experimental procedures used during a CHIE study (for example, bronchoscopy) typically are well characterized through extensive experience in many kinds of clinical studies, and this information could be directly communicated to the IRB and the participants as part of communicating the risks associated with the conduct of the CHIE study.

Temporal Dimensions of Exposure-Related Risk

Focusing on exposure-related risks, the committee distinguished between risks of clinically adverse effects that might occur over the short term (for example, within 1 or 2 days) and risks of chronic conditions (such as cancer) that might develop over the long term. There is no bright line that separates short-term from long-term risks. Clinical practitioners stratify common health conditions based on symptom duration. For example, the National Institutes of Health (NIH) considers a cough to be acute if it lasts less than 3 weeks, subacute if it lasts 3-8 weeks, and chronic if it last longer than 8 weeks (NIH, 2010). Low back pain is considered acute if it lasts less than 6 weeks, subchronic if it lasts 6-12 weeks, and chronic if it lasts over 12 weeks (Chou, 2014). (See Chapter 4 for a discussion of temporal dimensions of CHIE-study exposure-related risk.)

Adverse and Serious Adverse Events

There are many acceptable definitions of adverse events. The committee has adopted the definition provided by HHS guidance (OHRP 2007) designed for adverse event reporting. The definition also is used by the IRBs at UNC Chapel Hill (UNC, 2014). An adverse event is

The committee also adopts the definition of “serious adverse event” that has been provided by the same HHS guidance:

1. results in death;
2. is life threatening (places the subject at immediate risk of death from the event as it occurred);
3. requires inpatient hospitalization or prolongation of existing hospitalization;
4. results in a persistent or significant disability/incapacity;
5. results in a congenital anomaly/birth defect; or
6. any other adverse event that, based upon appropriate medical judgment, may jeopardize the subject’s health and may require medical or surgical intervention to prevent one of the other outcomes listed in this definition (examples of such events include allergic bronchospasm requiring intensive treatment in the emergency room or at home, blood dyscrasias or convulsions that do not result in inpatient hospitalization, or the development of drug dependency or drug abuse).

Note that the duration or persistence of a biologic response is an important consideration in determining whether an adverse event is serious. The definition of a serious adverse event calls for a judgment as to whether an effect results in “a persistent or significant incapacity or substantial disruption of the ability to conduct normal life functions.” For example, a cough lasting only 1 or 2 days after participation in a CHIE study would likely be considered an adverse event, while a cough that began soon after participation in a CHIE study and lasted many months would likely be a serious adverse event.

Informed Consent

For an individual considering participation in a CHIE, the study is presented during an informed-consent process in which risk communication and risk perception play important roles. Informed consent is a process with three sequential elements, described in detail in Chapter 7. Information about a research study is presented to a potential participant in a disclosure process, the potential participant considers the information in a deliberative process, and the participant finally makes a decision to participate or not to participate.

Risk Perception

Risk perception is a subjective assessment resulting from a person’s beliefs regarding the probability of a potential hazardous event or activity and how it will affect him or her. Individual philosophies, principles, and past experiences can shape one’s beliefs about perceived risk. The severity of the risk and the overall public opinion of the risk can also affect individual risk perceptions (Beecher et al., 2005; Slovic, 1987).

Risk Communication

Risk communication is “any purposeful exchange of information about health or environmental risks between interested parties” (Covello et al., 1987). This information incorporates understanding, ideas, and actions as they relate to risks (Anderson and Iltis, 2008).

Exposure Comparators

The use of exposure comparators involves comparing experimental exposure concentrations and durations with ambient concentrations of similar magnitude and duration experienced by a population in everyday life at a certain location. That information is provided to individuals or IRBs for the purpose of enhancing their deliberation about the risks to participants involved in a CHIE study (see Chapter 6).

As IRBs are “consumers” of risk information, their deliberations are likely to be influenced by the individual risk perceptions of its members and of the board. The use of standard terminology for reporting adverse events to the IRB is intended to provide a common language to facilitate the IRB’s work in balancing risks and benefits.