Chevron Demise Vis-a-Vis Cannabis Rescheduling

A nothingburger?

MATT JUL 2

READ IN APP

I’m currently working on a longer post on the impact of Chevron’s demise and its impact on drug policy. I’m sure On Drugs alum and admin law pro Shane Pennnington will have a lot to say in the coming weeks as well. In the meantime, I want to publish a short note expanding on my quick take on X:

A Chevron Primer

Depending on who you read, Chevron’s demise is either the first step toward the Republic’s downfall or restoration. Let’s start with the basics. Chevron deference or the Chevron doctrine is a legal framework that courts used to review agency interpretations of ambiguous statutes. Succinctly stated, under the doctrine, courts reviewed an agency’s interpretations of a statute by asking two questions:

1. Is the statute ambiguous; and if so
2. Is the agency’s interpretation reasonable?

These two components of the framework came to be known as “Chevron Step One” and “Chevron Step Two.” If the answer at both steps was “yes,” then a reviewing court would defer to the agency’s interpretation even if that interpretation was not the best interpretation of the statute. Over time, the Supreme Court added layers and nuance to the framework. For example, in United States v. Mead, the Court added a “Step Zero” threshold level question where a reviewing court asks whether the two-step Chevron analysis even applies. Exceptions to Chevron proliferated. For example, under the “major questions” doctrine, the Supreme Court refused to defer to agency interpretations on questions having vast “economic and political significance” where Congress did not clearly empowered the agency with authority over the issue. Whatever that means. The Chevron doctrine rested on a number of fictions, most prominently, “implicit delegation”: the notion that when Congress leaves ambiguities or gaps in a statute an agency administers, Congress implicitly delegated authority to the agency. Pragmatism also supported Chevron—or at least it did as originally conceived. Rather than have generalist federal judges without special subject matter expertise pass on statutory interpretation questions, why not give the experts (i.e., bureaucrats in federal agencies) the benefit of the doubt? Perhaps I’ll dive more in depth in my next post related to Chevron and drug policy more generally. Suffice to say, Chevron gave agencies a tremendous handicap when it came to applying statutes. After all, after enough time, with evolving society and technology, all statutes are bound to give rise to gaps and ambiguities. A statute can only be so precise and prescient Chevron tilted the scales of justice with these necessities and eventualities in favor of the agency. But as the Loper Bright opinion makes clear, decades after Chevron, the framework became unworkable: Exceptions became too many. Courts couldn’t agree when a statute was ambiguous. The nature of the doctrine and a (terrible) case called Brand X allowed different administrations to adopt different meanings of a statute—even when a prior judicial precedent held an ambiguous statute meant otherwise. The Supreme Court reviews many agency interpretations every term, but had not deferred under Chevron since 2016.

Why Overturning Chevron Was a Good Thing for Rescheduling

In the above quick take on X, I’m not stating that overturning Chevron is bad for rescheduling. Hardly. What I’m saying is that overturning Chevron was good for rescheduling. Now, not so clear what impact Chevron has. Rather, before these rescheduling proceedings, overruling Chevron was a clear boon. That is because DEA refused to entertain petitions to reschedule marijuana—for example, in 2016—by relying on a five-part test that most certainly did not constitute the best reading of the statute, specifically the phrase “currently accepted medical use in treatment in the United States.” And, it is precisely that interpretation of the statute—and the erroneous application of Chevron deference—that prevented earlier rescheduling efforts from succeeding. DEA first construed the statutory phrase in a case called Grinspoon v. DEA related to MDMA scheduling. In that case, among other items, Lester Grinspoon (backed by Rick Doblin / MAPS predecessor) challenged DEA’s interpretation of the phrases “currently accepted medical use in treatment in the United States” and “accepted safety for use . . . under medical supervision” in the Controlled Substances Act. DEA contended that MDMA belongs in Schedule I, while Grinspoon et al. argued it belonged in Schedule III. DEA’s position rested on its understanding of the two statutory phrases. DEA contended that both phrases meant that the FDA has evaluated the drug and approved it for interstate marketing, i.e., FDA-approval. Grinspoon et al. argued that DEA misinterpreted the statute and that the two phrases should be interpreted with reference to the medical community. At the outset of the case, the appeals court reviewed DEA’s interpretation under Chevron. But while the appeals court concluded that the statute was ambiguous, it did not adopt DEA’s construction because the court found it to be reasonable. Nonetheless, Grinspoon laid the groundwork for deference to DEA’s interpretation of “currently accepted medical use” in future rescheduling proceedings related to cannabis. And that’s what happened years later, when the D.C. Circuit reviewed DEA’s decision to keep marijuana in Schedule I in the 1991 case Alliance for Cannabis Therapeutics v. DEA. The case essentially begins and ends with Chevron. An administrative law judge (ALJ) in 1988 concluded that the evidence supported rescheduling marijuana from Schedule I because it had a “currently accepted medical use.” DEA superseded that recommendation, however, by interpreting “currently accepted medical use” to require eight factors:

1. Scientifically determined and accepted knowledge of its chemistry;
2. The toxicology and pharmacology of the substance in animals;
3. Establishment of its effectiveness in humans through scientifically designed clinical trials;
4. General availability of the substance and information regarding the substance and its use;
5. Recognition of its clinical use in generally accepted pharmacopeia, medical references, journals or textbooks;
6. Specific indications for the treatment of recognized disorders;
7. Recognition of the use of the substance by organizations or associations of physicians; and
8. Recognition and use of the substance by a substantial segment of the medical practitioners in the United States.

Although, as the ALJ found, evidence showed the usefulness of cannabis with patients undergoing chemotherapy, applying the eight factors, DEA concluded that the evidence did not establish a “currently accepted medical use.” On appeal, the D.C. Circuit explained as follows:

The petitioners renew their argument that the Administrator unreasonably rejected the evidence they presented (largely anecdotal) that a number of physicians believe marijuana is medically useful and, instead, improperly predicated his determination on the absence of demonstrated scientific evidence that the drug is medically useful and safe. The difficulty we find in petitioners’ argument is that neither the statute nor its legislative history precisely defines the term “currently accepted medical use”; therefore, we are obliged to defer to the Administrator’s interpretation of that phrase if reasonable. See NLRB v. United Food Commercial Workers Union, 484, U.S. 112, 123, 108 S.Ct. 413, 420, 98 L.Ed.2d 429 (1987); Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 843-45, 104 S.Ct. 2778, 2781-83, 81 L.Ed.2d 694 (1984) … The determination as to how much weight to place on scientific uncertainties as opposed to anecdotal evidence in applying the statutory phrase “currently accepted medical uses,” then, is very much a policy judgment which we have no authority to challenge. See generally Chevron.

To summarize, after an evidentiary hearing an ALJ concluded cannabis had an accepted medical use based on evidence from the medical community. DEA then changed its interpretation of the law. Under a different standard, that same evidence did not establish medical use. Only FDA approval would meet the mark. The reviewing appeals court then deferred to DEA’s interpretation of the statute under Chevron and was essentially obliged to accept the agency’s findings about the evidence under the APA. The 1991 ACT court did go on reject three of the test’s factors because those conditions were impossible to meet and sent the case back down to the agency for an explanation. On remand, however, the agency simply changed the test again resulting in the “five-part test.” This time, as the recent Office of Legal Counsel memo explains the reviewing court rejected the claim that DEA’s order “rested on an ‘unreasonable interpretation of the statute’”: In other words, the DEA’s order rejecting rescheduling rested on an interpretation of “currently accepted medical use” that passed muster under Chevron, even if that interpretation was not the best interpretation of the statutory phrase. As Shane and I argued here and elsewhere, extensively, the decision to afford DEA Chevron deference in this regard was error for a variety of reasons. One of those reasons is an important one: the CSA delegates responsibility over medical/scientific matters to HHS, as the Supreme Court held in Gonzales v. Oregon. It is impossible to square Chevron deference to DEA to interpret “currently accepted medical use” in light of that delegation to HHS and the holding in Gonzales. We made that argument in the highlighted citation above from the case we litigated, Sisley v. DEA. In view of those petitioners’ arguments, Judge Watford’s concurrence states:

I agree that the petitioners in this case failed to exhaust their administrative remedies and therefore join the court’s opinion dismissing their petition for review. I write separately to note that, in an appropriate case, the Drug Enforcement Administration may well be obliged to initiate a reclassification proceeding for marijuana, given the strength of petitioners’ arguments that the agency has misinterpreted the controlling statute by concluding that marijuana “has no currently accepted medical use in treatment in the United States.” 21 U.S.C. § 812(b)(1)(B).

So Why Overturning Chevron Is A Nothingburger For Rescheduling Going Forward?

Put simply, Chevron stitched together DEA’s five-part test that resulted in the denial of all rescheduling petitions. But it doesn’t stitch together the current HHS test endorsed by OLC. The starting point for understanding why Chevron’s demise isn’t some watershed event for rescheduling starts with the decision overturning Chevron itself—Loper Bright—and some fundamental points about Chevron. First, what Chevron never did and its overruling does not do: Chevron never affected the standard of review courts applied to review agency determinations and findings of fact. Chevron required courts to defer to agency interpretations of ambiguous statutes. But under the APA, courts have always had to effectively defer to factual findings of agencies under the highly deferential “arbitrary and capricious” standard: Second, consider the Loper Bright opinion that overruled Chevron. Loper Bright says that § 706 shown above, in which the “reviewing court shall decide all relevant questions of law,” incorporates the “traditional understanding of the judicial function,” i.e., that courts say what the law is, not executive agencies. But a statute may expressly give an agency discretion as to a particular matter or empower an agency to fill the details of a statutory scheme. Per the Supreme Court, “[w]hen the best reading of a statute is that it delegates discretionary authority to an agency, the role of the reviewing court under the APA is, as always, to independently interpret the statute and effectuate the will of Congress subject to constitutional limits.”¹ As it happens, the CSA isn’t “silent” on delegating and doesn’t rely on a Chevron fiction of implied delegation. Rather, as the Supreme Court stated in Gonzales v. Oregon, where the Supreme Court held that the Attorney General, via a rule interpreting the Controlled Substances Act, could not bar doctors from prescribing a drug to be used in aid-in-dying:

• “The CSA gives the Attorney General limited powers, to be exercised in specific ways…. As is evident from these sections, Congress did not delegate to the Attorney General authority to carry out or effect all provisions of the CSA. Rather, he can promulgate rules relating only to ‘registration’ and ‘control,’ and ‘for the efficient execution of his functions’ under the statute.”
• “The Attorney General does not have the sole delegated authority under the CSA. He must instead share it with, and in some respects defer to, the [HHS] Secretary, whose functions are likewise delineated and confined by the statute. The CSA allocates decisionmaking powers among statutory actors so that medical judgments, if they are to be decided at the federal level and for the limited objects of the statute, are placed in the hands of the [HHS] Secretary. In the scheduling context, for example, the Secretary’s recommendations on scientific and medical matters bind the Attorney General.”

The recent April 2024 memo from the Office of Legal Counsel (OLC)—the “secretive” office within the Department of Justice that “can sway the whole government”—offers a similar take and elaborates. The executive branch interprets these OLC memos to be binding interpretations of law. And, just as we argued in Sisley, the OLC memo concurs in concluding that DEA’s five-part test construction of “currently accepted medical use in treatment in the United States” is an “impermissibly narrow” construction of the statutory language based on the plain meaning of the text:

• “[N]either FDA approval nor DEA’s five-part test examines whether health care practitioners are actually using a drug to treat a condition or whether the entities regulating those practitioners allow the drug to be so used. Instead, FDA approval and DEA’s five-part test rely exclusively on certain scientific evidence and the views of some experts and FDA. Simply put, ignoring widespread clinical experience with a drug that is sanctioned by state medical licensing regulators when evaluating whether a drug has a CAMU is at odds with the plain meaning of section 812(b)”
• “Limiting the [currently accepted medical use] analysis to whether a drug has been approved by FDA or meets DEA’s five-part test also conflicts with the text of section 812(b) by erroneously equating identification of an “accepted” medical use under the CSA with the “approval,” or potential approvability, of the drug under the FDCA.

Rather than use DEA’s five-part test to conclude that marijuana had a “currently accepted medical use,” HHS used a two-part inquiry:

1. Whether licensed health care providers have “widespread current experience with medical use” of the drug “in accordance with implemented state-authorized programs, where the medical use is recognized by entities that regulate the practice of medicine.”
2. If so, whether there is “some credible scientific support for at least one of the medical uses.”

Next, OLC concludes that this new HHS’s two-part inquiry was “sufficient” to determine whether a drug has a CAMU. Notably, the HHS currently accepted medical use finding is a determination based on scientific and medical evidence:

• “[T]o determine if a drug has a [currently accepted medical use], section 812(b) requires an analysis of whether, at the present time, the medical community widely understands that a drug has a ‘use in treatment in the United States.’ Although there is no single right answer as to how specifically DEA should make this determination, the text of the CSA establishes certain basic parameters to guide the inquiry.”
• “[N]othing in the text of the CSA suggests that establishing that a drug has a [currently accepted medical use] requires the medical community to believe that the drug is the best way to treat a condition. So long as there is widespread understanding in the medical community that a drug is a permissible and reasonable way to treat a condition, it has a CAMU. That reflects a basic reality about the medical profession: that ‘in medicine there is often a range of reasonable treatments[.]’”

Finally, OLC explained in the memo that under the express delegations in the CSA, DEA had to give significant deference to HHS’s scientific and medical determinations—which should include whether a drug or substance has a “currently accepted medical use”:

• “[W]e conclude that HHS’s overall CAMU recommendation is not binding on DEA. We also conclude that the scientific and medical determinations that underlie HHS’s CAMU recommendation are binding, but only until the initiation of formal rulemaking proceedings. Once DEA initiates formal rulemaking, HHS’s determinations no longer bind DEA, but DEA must continue to accord HHS’s scientific and medical determinations significant deference, and the CSA does not allow DEA to undertake a de novo assessment of HHS’s findings at any point in the process.”
• “DEA may not simply cast aside HHS’s scientific and medical recommendations once it initiates formal rulemaking proceedings by issuing an NPRM. The categorical use of the word ‘binding’ in section 811(b) suggests that Congress intended HHS’s scientific and medical views to at least be a very significant input in the scheduling process. And there would seem to be little reason to make the HHS’s views binding at any stage in the process if DEA eventually could discard HHS’s determinations and review scientific and medical matters de novo.”

Conclusion

So where does that leave us? All in all, Chevron’s demise was an unquestionably positive sign for rescheduling back when DEA used its five-part test to reject rescheduling petitions. That is because DEA used Chevron to stich the five-part test together. Notably, HHS used that test as well as recently as in the July 2016 petition denial process. Now, however, HHS used a new test. That test follows from the plain meaning of the statutory text. HHS applied that test according to s delegation in the CSA that has it making findings/determinations on medical and scientific matters. That express delegation likely still gets some dap even after Loper Bright as well as the OLC memo that binds DEA going forward. And, of course, deference to agency factual findings under the APA remains undisturbed. This is not to say that Chevron’s demise won’t impact rescheduling at all or in other spheres DEA or FDA regulate. Quite the contrary. More on that in the next essay. I’m sure Shane—an admin law/cannabis law pro—will have a lot more to say about the above as well, wherever it is he writes these days. But suffice to say, as far as I can tell, this take that Schedule will be “canceled” … …isn’t a fully baked and ought not to be give much deference.