Digging Into HHS's Latest FOIA Release

What does it say? What does it mean? What difference does it make?

Yesterday, we posted 250+ pages of documents that HHS released in response to Matt’s FOIA request seeking HHS’s scheduling recommendation and analysis with respect to cannabis. At the time, we hadn’t had a chance to review the documents, so we simply uploaded them so everyone had access to them as soon as possible. Since then, however, I have looked them over carefully and report my initial findings below. Here are the key takeaways:

• Compared to HHS’s past reviews of marijuana’s abuse potential, these documents seem to reveal a heightened interest in marijuana’s abuse potential compared with other substances of abuse. This shift in focus could facilitate a conclusion that marijuana has an abuse potential less than substances in schedules I and II.

• Compared to HHS’s past reviews of marijuana’s scheduling status under the CSA, these documents appear to focus on a broader and more modern data set. This, too, strikes me as a move that would tend to support a decision to transfer marijuana out of schedule I.

• HHS dedicates the bulk of its analysis—well over 150 pages—to the question of whether marijuana has a currently accepted medical use in treatment in the United States. Had HHS applied the usual standards to decide that issue, it could have resolved it in two or three pages (as it did when it addressed the question in 2016, for example). I therefore conclude that it has likely announced and applied a new standard for assessing whether marijuana has a currently accepted medical use.

• Taken together, and for reasons I explain in greater detail below, all this leads me to conclude that HHS has likely recommended that DEA transfer marijuana to schedule III.

In the rest of this post, I’m going to dig into the documents themselves, breaking them down in some detail. In Part I, I provide an overview of what each of the five docs is and how it fits into the bigger picture. In Part II, I analyze each document carefully, highlighting any insights I’m able to glean along the way.

The Documents: What They Are and How They Fit Into the Bigger Picture

We have five docs to discuss. The first, titled “2023 01171 Interim Response Letter,” is a December 7, 2023, letter to Matt from HHS FOIA Appeals and Litigations Director, Alesia Y. Williams, explaining that HHS is responding to his FOIA request with 252 pages of (mostly) heavily redacted materials. The second, titled “Ash To Dea Letter Marijuana 8fa Cder,” is an August 29, 2023, letter from the Assistant Secretary of Health, Rachel L. Levine, M.D., ADM, USPHS, to DEA Administrator Milgram, explaining that (1) FDA conducted the analysis required by section 811(c) of the CSA and made a scheduling recommendation; (2) NIDA reviewed FDA’s analysis and concurred in that recommendation; and (3) the Assistant Secretary of Health reviewed the evidence and FDA’s recommendation and concurs as well. The third is a 20-page document titled “Hhs Foia Enclosure 2” that appears to be a memo from the Office of the Assistant Secretary of Health to the Commissioner of FDA requesting that FDA conduct a scientific and medical evaluation of marijuana to enable HHS to convey recommendations to DEA Administrator Milgram regarding the appropriate scheduling of marijuana. The fourth, titled “Enclosure 1,” is a 77-page document that appears to be FDA’s analysis of the 8-factors HHS is required to consider under section 811(c) of the CSA, FDA’s explanation of the three findings it must make to recommend scheduling a substance in the CSA (abuse potential, medical use, and safety/dependence liability), and FDA’s ultimate scheduling recommendation. And last, but not least, the fifth is a 154-page document titled “Hhs Foia Enclosure 3.” It appears to be an August 28, 2023 “report” prepared by FDA’s Center for Drug Evaluation and Research (in particular, CDER’s Controlled Substances Staff and personnel from its Office of Surveillance and Epidemiology) regarding “Considerations for Whether Marijuana Has a Currently Accepted Medical Use in the United States for Purposes of Section 202(b) of the Controlled Substances Act.”

To understand the significance of these documents, you need to understand the role HHS’s scientific and medical evaluation and scheduling recommendation plays in the scheduling process under the CSA and the internal process HHS uses to develop the evaluation and recommendation. I’ve explained this process in some detail before, so I’ll stick to the basics here.

Before initiating the formal rulemaking process to reschedule a substance, the statute requires DEA to “gather[] the necessary data” and “request from the Secretary [of HHS] a scientific and medical evaluation, and his recommendations, as to whether such drug or other substance should be so controlled or removed as a controlled substance.”¹ After the President directed the Attorney General and the HHS Secretary to reconsider cannabis’s scheduling classification under federal law, DEA requested HHS’s evaluation and recommendation. At that point, the Assistant Secretary of Health directed FDA’s Controlled Substances Staff (CSS) to conduct the 8-factor analysis required under section 811(c) of the CSA. This appears to have happened through the third document referenced above.

At that point, the CSS got to work on the 8-factor analysis and develops a draft medical/scientific evaluation and scheduling recommendation, which then eventually goes to the Commissioner of FDA and NIDA for review. Once NIDA has reviewed and the Commissioner has approved the CSS’s draft, it then goes to the Assistant Secretary of Health at HHS for review and approval. Documents four and five appear to be FDA’s analysis. Specifically, the fourth document appears to be FDA’s eight-factor analysis, scientific/medical evaluation, and scheduling recommendation, and the fifth document appears to be an extension of that analysis focused exclusively on one of the scheduling criteria in particular, namely whether marijuana has “a currently accepted medical use in treatment in the United States.”²

Finally, once the Assistant Secretary of Health has reviewed the evidence and approved FDA’s analysis and conclusions, it is forwarded along to DEA. This appears to be the function of the second document.

In short, I believe the documents HHS has released are heavily redacted versions of each of the key pieces of analysis HHS sent over to DEA with respect to cannabis scheduling. These documents are critically important to the scheduling process because (1) they will be part of the administrative record that will likely end up in front of a court one day when any eventual DEA rule on rescheduling is subject to judicial review, and (2) HHS’s views on scientific and medical issues are binding on DEA in the scheduling process.³

Now that you know what each document is and how it fits into the bigger rescheduling picture, let’s dig into what the documents (or at least the portions HHS has permitted us to see) say and what difference (if any) that new information makes.

The Breakdown

Document 1: HHS’s Letter to Matt

I have only one thing to say about the first document—HHS’s letter to Matt attaching the documents in response to his FOIA request: While HHS has disclosed more material than it did initially, its response remains far too heavily redacted. The agency cites FOIA exemption (b)(5) as the basis for the redactions, but suffice to say, Matt and I don’t believe that exemption justifies the sweeping redactions that remain. In the end, however, that’s an issue for another post (and litigation), so I’ll move on for now.

Document 2: The Assistant Secretary of Health’s Letter to Administrator Milgram

The second document, the Assistant Secretary of Health’s letter to Administrator Milgram conveying HHS’s recommendation and evaluation, is more interesting. I’ll highlight two things here. First, the letter explains that the Assistant Secretary is recommending that “marijuana, referring to botanical cannabis (Cannabis sativa l.) that is within the definition ‘marihuana’ or ‘marijuana’ within the CSA be ….” The end of that sentence is redacted (of course), but the fact that the Secretary’s recommendation addresses all botanical cannabis material meeting the CSA’s definition of marijuana is important. For one thing, this detail confirms that HHS’s recommendation does not single out a single cannabis product, compound, or molecule (or class of products, compounds, or molecules) for rescheduling while recommending that DEA leave the “rest” of marijuana in schedule I. Rather, it either recommends that DEA keep all marijuana in schedule I or transfer all of it to another schedule.

The fact that HHS’s recommendation address marijuana as a whole is also significant because, in the past, HHS has insisted that for a number of reasons, it’s almost impossible to analyze marijuana for scheduling purposes. In its 2016 evaluation and recommendation, for example, HHS explained that because marijuana strains and products vary in chemistry and composition, it is not possible to show that “marijuana” as a whole meets the “reproducible chemistry” requirement under DEA’s five-part test for “currently accepted medical use in treatment in the United States”:

Marijuana does not meet any of the elements for having a ‘‘currently accepted medical use.’’ First, FDA broadly evaluated marijuana, and did not focus its evaluation on particular strains of marijuana or components or derivatives of marijuana. Since different strains may have different chemical constituents, marijuana, as identified in this petition, does not have a known and reproducible chemistry, which would be needed to provide standardized doses.⁴

This time, HHS dedicated well over 150 pages to the topic of whether marijuana has a currently accepted medical use. If HHS still believed that variation among strains and products prevented any demonstration of reproducible chemistry and thus of currently accepted medical use, one would have expected its analysis of the issue to be much shorter. While it’s impossible to be sure, this leads me to believe that HHS likely found that marijuana does have a currently accepted medical use and that it likely came to that conclusion based on a different standard than it and DEA have applied in the past. After all, DEA’s five-part test still requires reproducible chemistry, and the variation in chemical composition among various strains and products remains as wide or wider today than it was in 2016.

The second feature of the Assistant Secretary’s letter to Administrator Milgram that caught my eye is also evident only when comparing it to the letter the former Assistant Secretary wrote when conveying HHS’s recommendation that DEA keep marijuana in schedule I back in 2016. For one thing, the 2016 letter was much longer and went included a fairly in-depth discussion of the need for further research into cannabis and various obstacles to getting that research done under federal law:

FDA’s Center for Drug Evaluation and Research’s current review of the available evidence and the published clinical studies on marijuana demonstrated that since our 2006 scientific and medical evaluation and scheduling recommendation responding to a previous DEA petition, research with marijuana has progressed. However, the available evidence is not sufficient to determine that marijuana has an accepted medical use. Therefore, more research is needed into marijuana’s effects, including potential medical uses for marijuana and its derivatives. Based on the current review, we identified several methodological challenges in the marijuana studies published in the literature. We recommend they be addressed in future clinical studies with marijuana to ensure that valid scientific data are generated in studies evaluating marijuana’s safety and efficacy for therapeutic use. For example, we recommend that studies need to focus on consistent administration and reproducible dosing of marijuana, potentially through the use of administration methods other than smoking. A summary of our review of the published literature on the clinical uses of marijuana, including recommendations for future studies, is attached to this document.

FDA and the National Institutes of Health’s National Institute on Drug Abuse (NIDA) also believe that work continues to be needed to ensure support by the federal government for the efficient conduct of clinical research using marijuana. Concerns have been raised about whether the existing federal regulatory system is flexible enough to respond to increased interest in research into the potential therapeutic uses of marijuana and marijuana-derived drugs. HHS welcomes an opportunity to continue to explore these concerns with DEA.⁵

The new letter is much shorter, and while there are a few redactions, HHS disclosed enough of it for us to tell that it doesn’t contain a similar vignette regarding a dearth of cannabis research and the related obstacles to research under federal law. While there may be other explanations for this difference, it could be that the Assistant Secretary didn’t include a similar discussion of lacking research in the new letter because HHS found that there was sufficient research to justify rescheduling marijuana this time. And as I’ll explain in my discussion of some of the other documents below, there is good reason to believe that FDA and HHS gathered and analyzed a lot of research and data before drawing their conclusions this time.

Document 3: The Assistant Secretary of Health’s Memo to FDA

This document is so heavily redacted that I can’t tell much of anything about what it says. I find it interesting that what appears to be a simple request from the Assistant Secretary of Health to the Commissioner of FDA for FDA’s scheduling evaluation and recommendation is apparently 20 pages long and contains several tables and figures as attachments. In any case, so far, I haven’t been able to draw any useful insights from this document. If you come up with anything, please let me know in the comments.

Document 4: FDA’s Scheduling Evaluation and Recommendation

Comparing FDA’s just-released medical and scientific evaluation and scheduling recommendation to the one it prepared in connection with DEA’s denial of the 2016 rescheduling petition reveals a few noteworthy points:

1. In footnote 3 of the new recommendation, FDA acknowledges DEA’s authority under section 811(d)(1) of the CSA to override HHS’s scheduling recommendation where doing so is necessary to ensure U.S. compliance with the Single Convention on narcotic Drugs. FDA’s 2016 recommendation was silent regarding the treaty issue.

2. When considering factors associated with marijuana’s abuse potential and associated public health risks, the new evaluation and recommendation seems to place greater emphasis on how marijuana compares to other substances that are abused.⁶

3. Emphasizes focus on new data and research, meaning that which has become available since FDA completed its previous evaluation and recommendation in connection with DEA’s denial of the 2016 rescheduling petition.⁷

4. Lists various epidemiological databases it reviewed and consulted, including NSDUH, BRFSS, RADARS, NMURx, MTF, YRBSS, and ICPS.⁸ Compared to its 2016 evaluation and recommendation, the new list drops two databases (TEDS and DAWN) and adds five others (BRFSS, RADARS, NMURx, YRBSS, and ICPS). This strikes me as a particularly important shift, but I’m not currently familiar enough with the various databases in play to comment intelligently on what it means. If you’re familiar with these databases and what this change might mean, please let me know in the comments.

5. Drops the following sentence that appeared in its 2016 evaluation and recommendation when introducing its approach to evaluating marijuana’s “actual or relative potential for abuse”: “The data include a discussion of the prevalence and frequency of use, the amount of the substance available for illicit use, the ease of obtaining or manufacturing the substance, the reputation or status of the substance ‘on the street,’ and evidence relevant to at-risk populations.”⁹ Instead, the new analysis focuses on things like “chemistry, receptor binding, behavioral effects …, pharmacokinetics [etc.] ….”

6. Flags Congress’s amendment of the CSA’s definition of marijuana to exclude hemp as an important development impacting its analysis since its previous evaluation and recommendation in connection with DEA’s denial of the 2016 rescheduling petition. Notes that hemp includes some intoxicating products and derivatives and emphasizes that it did not include CBD products in its analysis.¹⁰

In sum, FDA’s analysis appears to include a broader and more current data set. It also appears to have shifted to focus less on the sheer amount of illegal marijuana use and more on marijuana’s abuse potential and safety profile compared to other substances of abuse.

Document 5: Considerations for Whether Marijuana Has a Currently Accepted Medical Use in the United States

Document 5 is heavily redacted. We can see that its purpose is to “analyze and present relevant data and make a determination as to whether marijuana is considered to have a CAMU in the United States under the CSA.”¹¹ There is also a stray footnote that evaded redaction on page 52. It says, “Two of the eight observational studies were included in other indications (i.e. anorexia and anxiety) as pain was a secondary outcome.” Why is that solitary footnote the only sentence that escaped redaction in the last 133 pages of the document? Beats me. If you have ideas, again, please share them in the comments.

While HHS kept the vast majority of the substance of this document away from us for now, we can nevertheless tell that it is significant just from (1) its existence, (2) its topic, and (3) its length. Put simply, the mere fact that FDA produced a 154-page standalone exegesis on whether cannabis has a currently accepted medical use is obviously important. Why? Because it shows that the agency’s extraordinary focus on a scheduling factor that, under the standards that DEA and FDA have applied for years when addressing this question, would have been simple.

Since about 1987, DEA and FDA have insisted that a substance has a currently accepted medical use in treatment in the United States only if (1) FDA has approved it for interstate marketing as a drug under the Federal Food, Drug, and Cosmetic Act (FFDCA) or (2) it satisfies a five-part test that DEA made to mimic the standard for FDA approval under the FFDCA. FDA obviously hasn’t approved marijuana for interstate marketing as a drug under the FFDCA, and both DEA and FDA have repeatedly explained that marijuana does not and will never satisfy DEA’s five-part test because, among other reasons, it’s a plant and therefore cannot satisfy the “known and repeatable chemistry” requirement. That’s why FDA’s recent assessments of the question have been straightforward and brief. In its evaluation and recommendation in connection with DEA’s 2016 denial, for example, FDA spent less than 2 pages on the issue.¹²

Given that FDA has spilled over 150 pages on the issue—double the length of FDA’s entire scheduling recommendation and evaluation—this time around, I think we can safely assume something has changed. But what, exactly? Well, as I’ve already explained, we know that marijuana still can’t pass DEA’s five-part test, and we also know that it still hasn’t obtained FDA approval for interstate marketing under the FFDCA. Logic says that if marijuana hasn’t changed, then the standard FDA is applying to the question must have. And there is some evidence to support that explanation. We know, for example, that FDA requested and received state-level data on the medical use of cannabis as it developed its evaluation and recommendation. In the past, FDA and DEA have insisted that that sort of data was irrelevant to the currently accepted medical use inquiry. The fact that FDA has apparently changed its mind on that issue could be a clue that the agency has in fact recognized a new path for demonstrating currently accepted medical use.

If FDA has decided that state-level use of medical marijuana is relevant to the currently accepted medical use question, it wouldn’t be the first time. As I have pointed out elsewhere, in 1982 HHS recognized that drugs can “obtain[] ‘accepted medical use’” for purposes of § 812(b)(1)(B) “by virtue of totally intrastate production and use.”¹³ Driving FDA’s deference to the practice of medicine at the state level is a core appreciation for the states’ traditional authority in that area. Indeed, the CSA “presume[s] and relies upon a functioning medical profession regulated under the States’ police powers,” prohibits the federal government from making “anterior judgment[s]” about what constitutes accepted medicine or medical treatment, and “manifests no intent to regulate the practice of medicine generally.”¹⁴ Even DEA has acknowledged that it lacks authority over traditional state regulation of the medical profession: “DEA does not act as the Federal equivalent of a State medical board overseeing the general practice of medicine. State laws and State licensing bodies … collectively regulate the practice of medicine. In contrast, the scope of the CSA (and role of DEA) is much narrower.”¹⁵

In short, while I cannot know for sure without seeing the entirety of FDA’s analysis, I strongly suspect that the reason FDA has dedicated so much attention to the currently accepted medical use question is because it needed to explain and justify concluding that marijuana has a currently accepted medical use despite the fact that (1) FDA has not approved it for interstate marketing and (2) it cannot satisfy DEA’s longstanding five-part test. Finally, I view FDA’s past recognition of the possibility that a substance could demonstrate accepted medical use “by virtue of totally intrastate production and use” and its more recent interest in state-level data regarding widespread medical use of marijuana under state law as fairly powerful clues about what FDA’s newly-minted standard might look like. Of course, all this is at best an educated guess. Only time will tell how accurate my prediction is.

Conclusion

Having reviewed HHS’s most recent disclosure carefully, I’m more convinced than ever that HHS has in fact recommended that DEA transfer marijuana to schedule III. The agency’s shift in focus to marijuana’s abuse potential relative to other substances of abuse and its inclusion of a broader and more current data set strike me as moves that could easily support a conclusion that marijuana in fact has an abuse potential lower than substances in schedules I and II. And FDA’s recent interest in state data combined with the extraordinary amount of attention it has given to addressing whether marijuana has a currently accepted medical use convinces me that it has likely concluded that the answer to that question is yes.

1. 21 USC 811(b).

2. 21 USC 812(b)(1)(B).

3. 21 USC 811(b). There is an exception to this rule, however. Under section 811(d) of the CSA, DEA need not defer to HHS’s findings on scientific and medical issues if DEA deems it appropriate to place cannabis in a schedule other than the one HHS recommends to ensure compliance with U.S. treaty obligations. Matt and I have explained why this provision of the CSA is unconstitutional here.

4. 81 Fed. Reg. 53688, 53707 (Aug. 12, 2016).

5. 81 Fed. Reg. 53688, 53689 (Aug. 12, 2016).

6. See, e.g., FDA Evaluation and Recommendation at 1-2, 5-6, 28, and 37.

7. See, e.g., FDA Evaluation and Recommendation at 5 and 32.

8. See FDA Evaluation and Recommendation at 32.

9. Compare FDA Evaluation and Recommendation at 6 and 81 Fed. Reg. 53688, 53690 (Aug. 12, 2016).

10. See FDA Evaluation and recommendation at 3.

11. See Considerations for Whether Marijuana Has a Currently Accepted Medical Use at 8.

12. See 81 Fed. Reg. 53688, 53706-07 (Aug. 12, 2016).

13. 47 Fed. Reg. 28141, 28150-51 (June 29, 1982).

14. Gonzales v. Oregon, 546 U.S. 243, 270, 272 (2006).

15. 71 Fed. Reg. 52716, 52717 (Sept. 6, 2006). See also 57 Fed. Reg. at 10505 (DEA acknowledging that the CSA “does not authorize … the DEA Administrator[] to make the ultimate medical and policy decision as to whether a drug should be used as medicine. Instead, he is limited to determining whether others accept a drug for medical use. Any other construction would have the effect of reading the word “accepted” out of the statutory standard”).