In the twilight of the Trump Administration, the Environmental Protection Agency issued an internal procedural rule entitled “Strengthening Transparency in Pivotal Science Underlying Significant Regulatory Actions and Influential Scientific Information.” The rule which is focused on EPA’s consideration of dose-response data, is widely predicted to have a limited shelf life. Congressional veto under the Congressional Review Act; EPA reconsideration, revocation and amendment or successful court challenges in already pending litigation can be anticipated. Nevertheless, the goals enunciated in the rule are worth pursuing, as they point the way to better EPA decision-making through greater science transparency. As the references cited and listed in the preamble reflect, the call for greater science transparency pre-dates the Trump Administration and extends well beyond EPA and the federal government.
The rule establishes how EPA will consider dose-response data that are crucial to decisions in issuing significant regulatory actions and influential scientific information, such as Integrated Risk Information System (IRIS) profiles, which characterize hazards from chemical exposures. The most vehement and widely-cited criticism of the rule is that it will prevent EPA from considering important dose-response studies because they cannot be made publicly available or submitted for independent validation due to privacy considerations preventing human health data from being shared. In the rulemaking, EPA responds to this concern initially by stating that all relevant scientific studies are in the scope of its review.
The rule identifies nine non-exclusive factors EPA will take into account when evaluating the consideration to afford studies with underlying dose-response data unavailable for independent validation. EPA’s rule calls for giving “greater consideration to pivotal science where the underlying dose-response data are publicly available in a manner sufficient for independent validation.” The preamble notes that if based on these factors two studies are “relatively equal,” both may be considered; but “other things being equal” greater consideration should be given to the one whose underlying data and models are publicly available and can be independently validated. Even this requirement is flexible in that the rule authorizes the EPA Administrator to grant on a case-by-case basis an exemption from any part of the rule if any of five listed circumstances exists. Furthermore, the rule requires EPA, whenever it does make available to the public dose-response data, to do so in accordance with legal protections on privacy, confidentiality and confidential business information and with a sensitivity to national security. Lastly, human health privacy concerns are irrelevant to a vast quantity of non-human dose-response data, such as animal feeding or aquatic toxicity studies, on which EPA bases IRIS toxicological profiles, water quality criteria and standards, drinking water standards and other decisions.
While the rule may encourage the public sharing of data by researchers, the rule does not necessarily result in disclosure of dose-response data to the public. The rule requires EPA to ‘identify the science that serves as the basis for informing a significant regulatory action and to make it “publicly available to the extent permitted by law.” However, as EPA notes in the preamble, “the rule does not obligate it to “collect, store or publicly disseminate dose-response data underlying pivotal science.”
Making scientific data publicly available lies at the core of the science transparency movement, and it is unfortunate the rule does not do more to make data publicly available. To a large degree, EPA has not analyzed data when issuing rules and influential scientific information. Rather, it relies on studies by researchers who summarize their methodology, data and analysis and who present their conclusions. Published, peer-reviewed studies are regarded as the gold standard, but a peer-reviewed study is not necessarily pure gold. Researchers rarely submit the underlying data with their manuscripts and these data are not seen by peer reviewers.
As the preamble notes, “Peer review does not typically include reanalysis of the underlying data . . . and thus peer review is not considered a replacement for the data availability requirements of this rule.” Moreover, as good as peer review may be, something more can be gained when any interested person can review data, replicate the researcher’s analysis or conduct additional analysis. This process may lead to greater confidence in the researcher’s own conclusions, produce additional findings or present a new hypothesis and additional research. Especially where data are generated with EPA’s funds, EPA should make data publicly available. Other federal and state research-funding agencies should do the same in the ordinary course. “Data,” according to Lawrence Summers, former Secretary of the Treasury and president of Harvard University, “is the ultimate public good.” If the public paid for the data, they should be able to see it when a public agency uses it.