The Dysregulation of Human Subjects Research

The regulatory system for protecting human subjects in research in the United States is increasingly dysfunctional. While many critics assert that research participants are inadequately protected,^{1 - 2} there are increasing concerns that the system is overregulated, with more time and expense devoted to activities of marginal utility in protecting human research participants.^{3 - 6} In some cases these activities actually appear to be reducing protections for participants in research, by diverting energy that could be spent on more fruitful tasks and by creating disincentives to institutional review board (IRB) membership.

Undeniably the early history of research involving human subjects was dominated by scandals, including widespread egregiously unethical behavior by US physician investigators.⁷ In 1973, 7 years after the US government required IRBs for federally funded research, many institutions were found to be performing only perfunctory reviews.⁸

By 1983 a presidential commission⁹ found the IRB system to be working well, with no examples of grossly unethical studies as had been reported earlier.⁷ In the ensuing decades, there were occasional reports of research misconduct.¹⁰ These incidents generally involved behaviors beyond the control of the IRB, although critics' responses sometimes suggested that such deviations could be prevented by imposing more duties on the IRB system.

In this issue of JAMA, Ness¹¹ provides data that expand current awareness of obstacles to the conduct and review of research created by policies designed to protect the rights and welfare of human research participants. In the survey of a large group of US epidemiologists regarding the effect of the HIPAA (Health Insurance Portability and Accountability Act) Privacy Rule on their research, two-thirds thought the rule had made research substantially more difficult, and a majority thought the rule had a negative effect on the protection of human subjects. Others have reported on HIPAA's effect in discouraging legitimate low-risk research. O’Herrin et al¹² reported a 70% reduction in medical records research requiring IRB review in the first year of the HIPAA rule. The reports by Ness¹¹ and Armstrong et al¹³ describe the effect of HIPAA on selection bias (which undermines the validity of research results) in research involving patient registries.

But the problems are not confined to research involving existing medical data, nor is HIPAA the only or even the most important cause of concern. Over the past decade, the oversight of IRBs has been characterized by increasing requirements for meticulous documentation of compliance with narrow interpretations of regulations and policies, often with punitive sanctions, accompanied and perhaps exacerbated by a drumbeat of assertions that the regulatory system is broken.¹⁴ Citations of noncompliance, even when unrelated to harm to research participants, are cited as evidence that the system is broken, leading to demands for more oversight or reform. Critics assert that “rats and mice get greater protection as research subjects . . . than do humans.”^{1 ,15}

Many requirements imposed by federal agencies, and now by the accreditation process,¹⁶ have little relationship to the protection of human research participants. Consider the requirement for a quorum at a convened IRB meeting. According to Robert's Rules of Order,¹⁷ documentation of a quorum at the beginning of a meeting is sufficient, unless a member challenges it. But the Office for Human Research Protection (OHRP), the regulatory branch of the Department of Health and Human Services charged with enforcement of federal research regulations, has insisted that a quorum be documented for every 1 of 100 or more actions items on the agenda of a single meeting of a typical IRB. Institutions have been informed by an accreditation agency that the presence of a “nonscientific member” must be documented for every action item. An audit by another agency threatened a warning letter because the institution was defining a quorum as “more than one-half” rather than “one-half plus one.”

The requirement for documentation of the regulatory justification for approval has been taken to extremes. A project funded by a grant from the National Institutes of Health was delayed by OHRP for nearly a year because the IRB minutes did not correctly identify the section of the regulations that provided the justification for approval, with no claim that any element of the protocol was in any way ethically problematic.

One of the greatest causes of increased IRB workload is the requirement that continuing review of protocols and most of their amendments be conducted in fully convened meetings of the IRB, even if the overwhelming majority of these items raises no significant issues. Systems of review that are arguably more thorough have been rejected without regard to the quality of the review process if they do not conform to OHRP's interpretation and application of the relevant sections of the Common Rule.¹⁸

In addition to increasing requirements for detailed documentation, IRBs are burdened with responsibilities to monitor adverse events, a task for which they are particularly unsuited. Several commentators^{4 - 5 ,19} have argued that this responsibility should be shifted to data monitoring committees that have access to all the needed data (both numerator and denominator) and also have the authority to “unblind” a clinical trial if necessary.

Several events over the past decade have contributed to the dramatic increase in federally inspired concern with minutia and punitive sanctions. The Office of Protection from Research Risks was created in 1974 as an educational office. Under the leadership of a new director in the 1990s, it took on an increasingly aggressive regulatory role.²⁰ The reason for the change is unclear, possibly abetted by a “zero tolerance” policy promulgated by President Clinton in 1994.²¹

In 1995 a committee appointed by President Clinton reported on widespread studies involving administration of or exposure to radiation.²² In some studies the doses of radiation exposure were toxic, and among the subjects were children, pregnant women, and patients with lethal diseases; some of these studies would have been considered egregiously unethical by contemporary standards. All of these studies occurred before 1974, the year in which the Department of Health, Education and Welfare first promulgated regulations for the protection of human research subjects. Nonetheless, President Clinton linked the report with the creation of a wider inquiry into the adequacy of protection of human research participants.^{23 - 24}

In 1999, the Office of Protection from Research Risks (later replaced by OHRP) began a series of widely publicized shutdowns of clinical research at leading academic medical centers, including Duke and Johns Hopkins universities, due to deficiencies in IRB review of research.²⁵ The deficiencies primarily involved failures of documentation of compliance,²⁶ or failure to follow required procedures. In the Duke case, which was typical, there were no claims of unanticipated harms to human research participants, or even approval of unethical research protocols.³

Similarly, a review of US Food and Drug Administration (FDA) warning letters to IRBs concentrated on procedural problems and failure of documentation, without regard to the relationship of these problems to episodes of harm to participants, or claims that research protocols were substantively unethical.²⁷ Consent forms were commonly found to be deficient, without regard to whether the prescribed changes would have affected the individuals' willingness to participate in the study.

In 1998, a report on the national IRB system by the Office of the Inspector General (OIG) of the US Department of Health and Human Services was subtitled “A System in Jeopardy” even though the OIG's investigations had turned up no recent evidence of exploitation of or harm to subjects that could be attributed to IRBs. After intense criticism that the evidence did not support the subtitle, it was changed to “A Time For Reform,”¹⁴ but many commentators continued to refer to the “system in jeopardy.”²⁸ OHRP Director Ellis stated that “when you set aside the language of danger and menace,” the OIG report offers no evidence that patients have been harmed or are at risk. Noting that every clinical trial goes through many layers of ethical review, Ellis said he considered the likelihood of a “catastrophic failure” to be “slight.”²⁸

The increase in the IRBs' burden is not entirely the responsibility of federal oversight agencies. Part of the problem is self-inflicted, as academic medical centers shifted responsibility for IRB structure and function to senior institutional officials, often with little IRB experience, who made a political judgment that, in order to avoid sanctions, the prudent course was to impose requirements on the system that are even more stringent than those of the regulatory agencies.

In addition, a small number of unanticipated deaths of research subjects at prestigious medical centers, including the University of Rochester, Johns Hopkins University, University of California-Los Angeles, and the University of Pennsylvania, became causes cèlébres. Even though the relationship of these unfortunate events to IRB responsibilities is uncertain at most, their reporting reinforced cries that the entire system was broken.²⁹ Clearly, the recent demands for increased bureaucratic procedures and their documentation would not have prevented any of these episodes. To the contrary, the increasing focus on minutiae has been distracting IRBs from more important substantive issues. The implementation of the regulations and—even more importantly—the federal and institutional interpretations of the regulations' requirements has gone too far.

Inflexible requirements for adherence to narrow interpretations of every word in regulations and other policies have led to a system that is more concerned with protection of the institution than protection of human research participants. The cost of the system is increasing without evidence of a return on investment with regard to protection of patients and other research participants.^{30 - 31} The burdens of IRB review discourage young investigators. Senior faculty commonly avoid serving on IRBs because the work is perceived as primarily bureaucratic and focused on unimportant minutiae; IRB members feel constrained in their efforts to exercise independent judgment.⁴ As one former director of OHRP concluded, “The end result is an undesirable form of ‘reactive hyperprotectionism’ that can have a stifling effect on research productivity without meaningfully enhancing the safety and well-being of research participants.”³²

The national system for the protection of human research participants is indeed a system in jeopardy. The major source of the threat to its proper functioning is the increasing pressure to perform tasks that either do not require doing, could be done better by others, or could be done more efficiently using expedited review procedures.

The sources of these problems include OHRP and the FDA because they appear to threaten institutions with draconian penalties for minor infractions; institutional (university and other) administrators acting out of fear that their institution could be the next to have its entire research operation suspended by “getting caught” in one of these minor infractions; and credentialing and certifying agencies for supporting these excesses by including them in their criteria for accreditation. A satisfactory resolution for this situation is urgently required. Resolution will necessarily require participation by federal oversight agencies, institutional officials, representatives of credentialing and certifying agencies, researchers, and research participants.