This post is part of Notice & Comment’s symposium on psychedelics and the law. For other posts in the series, click here.

Dr. Sunil Aggarwal runs an “integrative oncology clinic” in Seattle, Washington, where he treats patients with late-stage cancer. Some of those patients suffer from the anxiety and depression that often accompany a terminal diagnosis. Aggarwal thought that psilocybin, the active ingredient in so-called magic mushrooms, might help.

But psilocybin is a Schedule I controlled substance, meaning that, according to the federal drug laws, it has “no currently accepted medical use.”

So, Aggarwal asked the U.S. Drug Enforcement Administration (DEA) to either (a) confirm that the federal Right to Try Act authorized his use of psilocybin or (b) waive the registration requirements of the Controlled Substances Act (CSA) that apply to research uses of Schedule I drugs.

DEA declined. And, in February 2025, in Advanced Integrative Medical Science Institute v. DEA (AIMS), the U.S. Court of Appeals for the Ninth Circuit upheld DEA’s decision. The court described its review of the decision as “extremely limited” and “highly deferential.”

In other words, the Ninth Circuit held that DEA has a great deal of discretion on the matter of psilocybin access.

Reasonable minds can differ about whether psilocybin and other psychedelics should be widely available for experimental use or whether they should continue to be treated like, say, heroin and methamphetamine (also Schedule I drugs).

This essay doesn’t take sides on what the law of psychedelics access should be. It instead focuses on the structural question of who decides what the law is. In particular, it questions whether the wide discretion the Ninth Circuit granted DEA in the AIMS case was justified as a matter of institutional design.

“Indiscretion about Discretion”

Any legal or scholarly analysis of discretion can, in my view, usefully draw from Judge Henry Friendly’s superlative 1982 lecture, “Indiscretion About Discretion.” Friendly observed in that lecture that the single word, “discretion,” can mean quite different things in different contexts. Nearly five decades later, the framework Friendly sketched can help us think about what went wrong—and what may, eventually, go right—for Aggarwal and his patients.

Friendly’s basic argument was that discretion is not a fixed amount of deference that a reviewing decisionmaker (often an appellate court) owes to an inferior decisionmaker (often a trial court or an administrative agency). Rather, according to Friendly, discretion exists on a spectrum.

At one end sits the kind of decision where first-hand contact with the relevant facts gives the initial decisionmaker a clear informational advantage, and there’s not much “law” to be applied: a trial judge’s determination about a witness’s credibility, say, or a sentencing decision that turns on an assessment of a defendant’s character. At the other end sits the kind of decision where nothing about the initial decisionmaker’s position makes her especially well-equipped to answer the question and where values like consistency and predictability cut in favor of closer review: for instance, a trial judge’s decision to grant a preliminary injunction, which, Friendly observed, is controlled by clear equitable principles (the plaintiff must show irreparable harm, likelihood of success on the merits, and so on) and is often case-dispositive.

In other words, Friendly urged that the scope of permissible discretion on any given issue should be determined by reference to the reason why the decision is committed to the initial decisionmaker’s discretion in the first place.

DEA and Discretion

So, was the Ninth Circuit correct in AIMS to grant DEA wide discretion in denying Aggarwal’s request to use psilocybin? Applying Friendly’s framework, we can start by asking why DEA is the decisionmaker here and by scrutinizing the reasons the agency gave for denying Aggarwal’s petition.

DEA leaned heavily on Congress’s classification of psilocybin as a Schedule I drug and on the statutory criteria for Schedule I status set forth in the CSA: “high potential for abuse,” “no currently accepted medical use,” and “lack of accepted safety for use . . . under medical supervision.” The Ninth Circuit quoted these statutory phrases in its opinion affirming DEA, treating them as a sufficient basis for the agency’s decision.

DEA, in other words, exercised its discretion by parroting the words of a relevant statute. And the Ninth Circuit approved.

The Wrong End of Friendly’s Spectrum

Given the reasonable and fact-driven arguments on both sides of the substantive dispute—whether the benefits of administering psilocybin to Aggarwal’s patients would outweigh the risks of abuse—Friendly would have been the first to say that some deference was appropriate.

But, as the list of psychedelic drugs with therapeutic potential grows, the question of whether a psychedelic should be made available, in limited circumstances, to a specific cohort of seriously ill patients is almost certain to arise again, whether next month, next year, in the Ninth Circuit or in another circuit, involving psilocybin or another drug.

Under Judge Friendly’s framework, this likely recurrence is a reason to tighten judicial review of the agency, not loosen it. Patients, physicians, drug companies, and state regulators need to know what, exactly, the law is. Uniformity and predictability, in other words, are what’s important.

But those aren’t the only considerations relevant to the amount of discretion DEA should (or should not) get in a case like AIMS. DEA’s expertise is in drug enforcement. The question Aggarwal put to the agency, however, was not entirely a drug enforcement question. It was in part about health care: which drugs should be available, under what conditions, to dying patients? On that question, DEA is not the expert. To the contrary, the U.S. Food and Drug Administration (FDA) had previously designated psilocybin as a “breakthrough therapy,” suggesting that the relevant medical authority has a different view.

None of this is to say DEA reached the wrong answer in denying Aggarwal’s petition. Indeed, Aggarwal did himself no favors by offering DEA “no . . . information regarding the efficacy of psilocybin and none regarding its safety.” He also described to DEA neither his treatment plan nor any safety controls he would implement.

Still, and perhaps on a more substantial evidentiary record, one could imagine an agency decision that walks through numerous considerations—the state of the scientific evidence, the risks of abuse, the interaction of the CSA and the Right to Try Act—and reasonably concludes that access should be denied.

But that’s not what happened in AIMS. Instead, DEA gestured at the CSA and the Ninth Circuit affirmed because that gesture was enough to survive its “extremely limited,” “highly deferential” review.

Deference the Other Way

The silver lining for advocates of psychedelic access is that the same deferential standard that made DEA’s refusal of Aggarwal’s petition hard to challenge also makes the agency’s position easier to change. In April 2026, the President signed an executive order directing FDA and DEA to accelerate review of psychedelic drugs, to establish a pathway for eligible patients to access investigational psychedelics under the Right to Try Act, and to facilitate “any necessary Schedule I handling authorizations for treating physicians and researchers.”

If the executive branch’s new posture produces agency actions easing access to psychedelics, and those actions are challenged in court, they should be upheld under the same “highly deferential” standard that shielded DEA in AIMS.

Psychedelics and Social Determinants

If the political capital being spent on psychedelic reform succeeds, the relief will be real for patients who gain access to treatments that work for them. But the population of Americans who suffer from the mental illnesses psychedelics could potentially address—depression, anxiety, post-traumatic stress, addiction, and more—is enormous. And the share of that suffering that traces back to social determinants of health—poverty, housing instability, isolation, inadequate primary care—is huge. Psychedelic therapy, even at its most promising, is administered over months at a cost of thousands of dollars per session. Whatever it turns out to be may not be a solution at scale.

Judge Friendly’s lecture on discretion was not about drug policy. But thinking about the federal law on psychedelic access in light of that lecture raises a harder question than the one the Ninth Circuit answered (or, perhaps more accurately, dodged) in AIMS: whether the political structure built around psychedelic access—agencies wielding discretion shaped by activist coalitions (and commercial interests)—is the structure we would choose if we were thinking carefully about where health resources ought to go.

Paul R. Gugliuzza is the Thomas W. Gregory Professor of Law at University of Texas at Austin School of Law.