July 9, 2018
The Honorable Andrew Wheeler Acting Administrator
U.S. Environmental Protection Agency
1200 Pennsylvania Avenue, N.W.
Washington, DC 20460
Subject: Strengthening Transparency in Regulatory Science (Docket ID No. EPA-HQ-OA-2018-0259)
Dear Administrator Wheeler:
I am pleased to respond to the proposed rule, “Strengthening Transparency in Regulatory Science on behalf of the American College of Occupational and Environmental Medicine (ACOEM). ACOEM is the premier national medical association representing more than 4,000 occupational and environmental medicine physicians and other related health professionals who champion the health of workers, the safety of workplaces, and the quality of the environment.
ACOEM has significant and multiple concerns about the proposed Rule. Although we concur that environmental regulatory policy should be based on sound science which has been subjected to careful scrutiny, we do not agree with the Agency’s specific proposals regarding the concepts of “transparency” or “reproducibility.” As presented, we believe this proposal is inconsistent with the proper conduct of scientific inquiry. Furthermore, we do not believe that additional, burdensome, “peer review” steps are likely to strengthen confidence in the evidence base for new environmental regulations.
The proposed rule makes a fundamentally invalid assumption that the creation of public policy typically has been and should be based on specific studies targeted to policy application. This is an unwarranted assumption because it suggests that individual studies must be policy-relevant and therefore subjected to a standard of evaluation laid down by a risk manager that is different from the study’s authors or other scientists. The proposed rule appears to ignore the practical problems of conducting human epidemiologic studies, which are often of value only in retrospect, especially toxicological studies dealing with exposure-response relationships and low-dose extrapolation.
The proposed rule would allow the Agency, acting as a risk manager, too much latitude in deciding in advance which studies to examine and which to ignore. It creates in effect a list of potential study disqualifiers in advance of rigorous review and contextualization of the results. The strictures created by the proposed rule will not be seen as a set of guidelines (such as the Daubert Criteria. Rather, these strictures could be used to disqualify reasonable studies on superficial and technical grounds when in reality the value of a study may often be to provide context. There are hazards inherent in confining one’s attention to a selected body of work while ignoring other studies that may show anomalies or unexpected findings.
Furthermore, the proposed rule, by putting unnecessary emphasis on scientific replication, reframes any risk management decisions as a situation in which prevention of harm must be based on scientific certainty. such an approach would be new and troubling. Requiring scientific certainty as a precondition for decision-making in practice will mean that the difficulty of demonstrating the risk of harm will always outweigh the demonstrable benefit of prevention measures. This locks in a one-sided bias toward doing nothing even when the weight of evidence points toward significant harm from the status quo.
Finally, similar proposals have been made by certain stakeholders in the past with the intention of manipulating EPA’s regulatory process by limiting the flow of information to the agency. A 1996 memo from an attorney to RJR Reynolds Tobacco Corporation proposes to “construct explicit procedural hurdles the Agency (EPA) must follow in issuing scientific reports.” The strategy was written to pre-empt EPA’s contemporary efforts to analyze the health effects of environmental tobacco smoke (ETS). It explicitly lays out the talking points of “sound science,” “transparency” and “reproducible science.” These concepts are also foundational to the proposed Rule, which is concerning and reminiscent of past attempts by outside influences to obscure the truth.
Our experts in environmental medicine have the following specific concerns:
1. Data “transparency” and protected information: The proposed Rule places undue emphasis on the need for “transparency” in the assessment of published studies. On a positive note, the Rule identifies certain circumstances, such as the protection of confidential patient data or the protection of trade secrets, in which “transparency” should yield to privacy concerns. However, the proposed Rule does not establish clear guidelines about how such balancing would be achieved and instead permits the Agency to disregard studies that may have important relevance for risk assessment. The proposed Rule should state explicitly that data “transparency” must necessarily be waived regarding the medical status of individuals who may have suffered adverse health problems as a result of occupational and environmental exposures. That is, the propose Rule should clarify how medical confidentiality will be protected under considerations of expanded disclosure.
2. The concept of “best available science” is illusory: In practice it is very difficult to define when available scientific studies can be considered “best.” Evaluation of any individual study, as noted above, depends strongly on context and judgment, and the quality of the current scientific database is nearly always in flux. The term “best available” should not be applied to a body of scientific works, but instead it is more reasonable to set an expectation among scientists that a body of work has been found to be valid and persuasive. The Agency has a duty “to rationally execute and adequately explain agency actions.” However, such a duty does not require that all bodies of work under consideration be proved to be “best.”
3. The “Replication Crisis”: The so-called “replication crisis” (affecting publication of clinical trials in the medical literature) has only tangential relevance to the issue of environmental science publications because of the great differences related to publication bias, the differing motivations in these two disparate fields, as well as the difficulty in managing confounders in clinical trials or case series.
4. Independent peer review: The proposed Rule’s requirement for “independent peer review” of all studies representing “pivotal regulatory science” is concerning in that it appears to mean any scientific results or findings that might “drive” regulations with a cost of $100 million or more, as based on Office of Management and Budget (OMB) Final Information Quality Bulletin for Peer Review (70 FR 2664).
As proposed, the Rule would require “independent peer review” of any such studies. In practical terms, we wonder how the Agency could exclude independent peer review of any study without first requiring a cost-benefit analysis which had been certified by the OMB. That is, before considering any study as part of an environmental risk assessment, the agency would first have to know what kind of regulatory approach it might later take, then analyze the cost of that approach to determine that such a regulation would not cost more than $100 million, and finally analyze the study. This is a circular approach. Furthermore, conducting independent peer review on the hundreds, and sometimes thousands, of studies that underpin many key environmental regulations would be burdensome. For studies that had already undergone rigorous peer review it would be unnecessary.
In conclusion, the main provisions of this proposed rule are either unnecessary, duplicative, burdensome, and likely to be confusing in their application. The proposed rule will not deliver significant improvements in EPA’s current regulatory processes.
While my comments have been almost exclusively negative, please be aware that ACOEM has many members with vast experience in environmental health, who stand ready to work constructively with EPA in crafting future rules which might assure a high level of public trust and confidence in the evidence basis for critically important environmental regulatory policy. Please contact me if you would like our assistance.
William G. Buchta, MD, MPH, FACOEM
1. Di Q, et al. Air pollution and mortality in the Medicare population. NEJM. 2017;376:2513.
2. Berger RE, et al. Air pollution still kills. NEJM. 2017;376:2591.
3. Kravchenko J. et al. Long-term dynamics of death rates of emphysema, asthma and pneumonia and improving air quality. Inter J Chron Obstruct Pulmonary Dis. 2014;9:613.