Three Challenges for Risk-Based (Research) Regulation: Heterogeneity among Regulated Activities, Regulator Bias, and Stakeholder Heterogeneity

Abstract

On July 26, 2011, federal regulators issued an advance notice of proposed rulemaking (ANPRM) outlining proposed changes to the regulations that govern human subjects research, which have been adopted by 18 federal agencies and departments (and are better known as the Common Rule). As suggested by its subtitle — Enhancing Protections for Research Subjects and Reducing Burden, Delay, and Ambiguity for Investigators — the ANPRM attempts to appease both those who claim that human subjects research is currently overregulated and those who claim that it is underregulated. It proposes to do so by shifting scarce regulatory resources from “low risk” studies, where they unnecessarily burden research, to studies that “pose risks of serious physical or psychological harm,” which currently suffer from insufficiently rigorous review and thereby endanger participant welfare. That is, like regulators in many other sectors responding to charges of over- or under-regulation, the ANPRM’s architects seek to render research regulation “risk-based,” by making the kind and extent of IRB review and agency oversight proportionate to the riskiness of the research. In this respect, the ANPRM is just the United States’s contribution to a global trend toward risk-proportionate regulation of a wide range of activities, including food safety, medicine, work safety, environmental protection, and financial regulation. In pressing for increased risk-based research regulation, the ANPRM cleverly exploits what is perhaps two camps of critics’ only common ground: a shared faith in the possibility of regulators making objectively “correct” risk-benefit assessments, and their too-frequent failure to do the same, as demonstrated in part by widespread variation in IRB decisions regarding similar and even identical protocols. (Here the two camps part ways, with one camp emphasizing Type I errors, in which “unreasonably risky” research is allowed to proceed, and the other emphasizing Type II errors, in which important but “low-risk” research is rejected, altered or delayed.) Like the opposing camps of critics it seeks to appease, risk-proportionate research regulation itself assumes a meaningful way for regulators to distinguish “low-” from “high-risk” research. That is, it requires a basis on which some social planner can, in advance and with respect to all prospective participants in a particular study or category of research, deem some research-related harms insufficiently probable or significant to warrant the full panoply of protections afforded participants in other studies.

This chapter calls that assumption into question. Because prospective research participants are heterogeneous in their preferences and other circumstances, the same protocol will offer a different risk-benefit profile for different participants. (IRB variation, then, should not surprise us in the least; much of it is likely simply a reflection of the fact that all individuals, including those who serve on IRBs, vary in their preferences regarding research risks, benefits, and trade-offs between the two.) Before discussing this challenge from participant heterogeneity, I provide an overview of risk-based regulation and discusses two other notable challenges in applying such a regulatory framework to human subjects research: heterogeneity among regulated research activities and regulator biases. I conclude by suggesting an alternative way to redistribute scarce regulatory resources that embraces, rather than ignores, all three challenges. Although I focus on U.S. governance of human subjects research, the analysis is more broadly applicable in light of the fact that the U.S. has essentially exported this system to many other nations.

Keywords: regulation, heterogeneity, human subjects research, cognitive bias, preferences, research ethics, IRB, OHRP, FDA