A new federal watchdog report has identified significant weaknesses in the written policies governing how the Drug Enforcement Administration and Food and Drug Administration evaluate and schedule controlled substances, raising questions about procedural consistency at a moment when the federal government is considering one of its most consequential drug-scheduling decisions in decades: whether marijuana broadly should move from Schedule I to Schedule III. The U.S. Government Accountability Office released the 31-page report on September 23, 2026, concluding that DEA lacks comprehensive policies describing who is responsible for key scheduling decisions and exactly how agency personnel should conduct evaluations, while FDA lacks written procedures explaining how staff should perform the scientific analyses used to recommend a drug’s schedule.
The findings require an important qualification. GAO did not conclude that DEA or FDA improperly handled marijuana’s rescheduling review, nor did it find widespread instances in which DEA ignored federal health officials. In fact, GAO found that between 2020 and 2025 DEA considered HHS scientific and medical recommendations in every case where federal law required it to do so, and every finalized decision in that group matched the HHS recommendation. Instead, the watchdog identified a structural problem: much of the scheduling system depends on statutes, regulations, expert judgment and institutional knowledge rather than detailed written procedures capable of ensuring that the process operates consistently when personnel or circumstances change.
GAO Examined 208 Federal Drug-Scheduling Actions
Under the Controlled Substances Act, drugs considered to pose risks of abuse and dependence are divided into five schedules. Schedule I is the most restrictive category and is reserved for substances considered to have high abuse potential, no currently accepted medical use in the United States and no accepted safety for use under medical supervision. Schedule III substances, by comparison, have accepted medical uses and are considered to have lower abuse potential than substances in Schedules I and II. Scheduling can influence federal controls, research requirements and the criminal penalties associated with unauthorized manufacture, distribution or possession.
GAO examined 208 substances that were the subject of DEA scheduling actions from 2020 through 2025. Of those, 95 required DEA to consider a scientific and medical evaluation and scheduling recommendation from the Department of Health and Human Services. DEA did so for all 95. By the end of 2025, final rules had been issued for 84, and DEA’s decision matched HHS’s recommendation in every one. The remaining 11 were under extended temporary scheduling orders at the end of GAO’s review period; GAO notes that all 11 were subsequently placed in Schedule I in 2026, again consistent with HHS recommendations.
That record is significant because the report was requested amid congressional questions about whether DEA gives adequate weight to federal health agencies when making controlled-substance decisions. GAO’s evidence does not show a recent pattern of DEA disregarding HHS. Instead, the report turns attention toward how the agencies reach those decisions and whether enough of that process has been formally documented.
DEA Has No Comprehensive Internal Scheduling Procedures
GAO’s strongest criticism concerns DEA’s lack of detailed written procedures. Outside an interagency memorandum governing information exchange with FDA and the requirements contained in statutes and regulations, DEA told investigators that it does not maintain separate internal guidance explaining how personnel are supposed to carry out major parts of the scheduling process. DEA officials argued that the Controlled Substances Act and associated regulations already provide clear legal requirements and that the scientists involved are experienced subject-matter experts with advanced training.
GAO concluded that the law does not provide the operational detail needed to substitute completely for agency procedures. Among the areas lacking formal guidance are how DEA personnel collect and analyze data before initiating administrative rulemaking, conduct the three-factor analysis used for temporary scheduling, conduct the more extensive eight-factor analysis used for permanent scheduling, evaluate HHS recommendations and assign responsibilities among the Diversion Control Division and other relevant offices. GAO warned that leaving those methods largely undocumented could create consistency problems if experienced employees leave or if institutional knowledge is otherwise disrupted.
That does not mean DEA has been making scheduling decisions without scientific analysis. The agency told GAO that its scientists consider law-enforcement intelligence, drug encounters reported through forensic laboratory databases and other evidence, while formal administrative scheduling requires an eight-factor review. The watchdog’s concern is that the methods experienced staff have developed for carrying out those responsibilities have not been sufficiently converted into durable written policy.
FDA Also Lacks Written Standards for Crucial Scientific Decisions
FDA has somewhat more internal guidance, but GAO found important gaps there as well. FDA’s Center for Drug Evaluation and Research has policies governing consultation with its Controlled Substance Staff, procedures for evaluating the abuse potential of new drug applications and standard operating procedures covering several scheduling-related activities. Yet the agency does not have written policies detailing how staff should perform the complete eight-factor analysis or develop a final scheduling recommendation.
One particularly important gap involves the meaning of “potential for abuse.” That concept is central to deciding which schedule is appropriate, but the Controlled Substances Act does not define the phrase precisely. FDA officials told GAO that the absence of a clear written definition for “potential for abuse” and “high potential for abuse” can make comparisons between substances difficult. Staff currently rely partly on professional judgment when applying the statutory factors. GAO recommended that FDA document criteria and processes for assessing abuse potential, including how one drug’s potential should be compared with another’s.
FDA defended some of its current flexibility by explaining that every drug presents different scientific questions and that rigid procedures could interfere with case-by-case evaluation. GAO agreed that one inflexible formula is not necessarily required. Instead, it recommended written standards broad enough to preserve scientific judgment while giving staff a clearer and more reproducible framework for conducting analyses and developing recommendations. FDA told the watchdog that such documentation could be useful and agreed to develop new policy.
A Key FDA-NIDA Agreement Is More Than 40 Years Old
GAO uncovered another unusual weakness in the federal scheduling system: the memorandum describing how FDA works with the National Institute on Drug Abuse when developing scheduling recommendations is more than four decades old. The document references FDA offices that no longer exist and contains procedures that no longer accurately reflect how the agencies operate. It also lacks a provision requiring periodic review.
NIDA contributes scientific expertise concerning drug abuse when FDA develops scheduling recommendations. According to GAO, the agencies now use a more streamlined process than the memorandum describes, and both FDA and NIDA agreed that the document should be updated. HHS told GAO that FDA and NIH have already begun revising the agreement.
The outdated agreement does not prove that current recommendations are scientifically unsound. DEA and FDA officials told investigators that the agencies communicate frequently, including weekly calls among senior officials and informal discussions to resolve scientific questions. Disagreements over scheduling do occur, but officials described them as rare and said they generally exchange information and attempt to resolve them before a proposed rule reaches the public-comment stage.
The Findings Arrive in the Middle of Marijuana’s Historic Scheduling Review
Although GAO’s audit covered the scheduling system broadly, marijuana occupies a prominent place in the report because federal officials have recently changed the way they determine whether a substance has a “currently accepted medical use.” Marijuana had remained in Schedule I since enactment of the Controlled Substances Act in 1970. Following a federal review begun in 2022, HHS recommended in August 2023 that marijuana be transferred to Schedule III.
The recommendation relied in part on a newer two-part framework for determining accepted medical use. Under that approach, HHS considers whether licensed health professionals widely use a substance within state-authorized medical programs and whether credible scientific evidence supports at least one of those medical uses. FDA officials told GAO that this framework does not replace every previous method for determining accepted medical use; FDA drug approval and DEA’s older five-part test remain additional pathways.
The Justice Department’s Office of Legal Counsel addressed the issue in 2024 and concluded that DEA’s prior approach to accepted medical use was too narrow. According to GAO’s summary of that opinion, satisfaction of HHS’s two-part inquiry can establish accepted medical use even when the substance itself has not received FDA approval and does not satisfy DEA’s older five-part standard. OLC also concluded that marijuana could be placed in Schedule III while additional federal controls are imposed to satisfy international treaty obligations.
Marijuana Has Already Been Partially Moved to Schedule III
The federal marijuana picture became more complicated in April 2026. Acting Attorney General Todd Blanche issued an order moving FDA-approved marijuana products and marijuana covered by qualifying state medical marijuana licenses into Schedule III, using federal authority connected to international treaty obligations. At the same time, DOJ restarted an expedited administrative proceeding to consider moving the remainder of marijuana from Schedule I to Schedule III.
DEA held hearings on that broader proposal from June 29 through July 15. The proceeding specifically concerns marijuana that was not already covered by the April medical-marijuana Schedule III order. DEA’s public rescheduling page now contains the 11 days of hearing transcripts and orders governing transcript corrections and post-hearing briefs. As of September 24, the agency’s public docket does not show a final decision completing the broader rescheduling process.
That distinction is critical when considering the new GAO report. The watchdog’s review of agency performance covered scheduling actions through the end of 2025, so it did not audit the completed 2026 marijuana hearing or judge the merits of the pending broad Schedule III proposal. Marijuana appears primarily as an example of how the federal scheduling framework has evolved and as a major contemporary reason why questions about the process carry unusually high stakes.
Drug Scheduling Can Take Years
GAO also documented just how lengthy the federal process can become. Among 75 substances for which DEA requested an HHS evaluation and later published a final scheduling decision during the period studied, an average of about three years elapsed from DEA’s request to publication of the final decision. The median was 2.3 years, while individual cases ranged from roughly 1.6 years to more than 14 years.
Federal law requires HHS to provide requested scientific and medical evaluations within a “reasonable time,” but neither DEA nor FDA has a specific written definition of what that means. FDA officials told GAO that the appropriate timeframe depends heavily on the type of substance and the available evidence. A temporarily scheduled emerging drug might need an evaluation within roughly 12 to 14 months, while a public petition lacking comprehensive scientific information could require years of additional work.
The marijuana review itself illustrates how legal, scientific and procedural questions can extend the process. HHS issued its recommendation in August 2023, DOJ proposed broader Schedule III placement in May 2024, earlier administrative proceedings encountered delays, and a new hearing was held in summer 2026.
GAO Issued Three Recommendations—and the Agencies Agreed
GAO ultimately made three recommendations. It called on DEA to establish written policies covering its roles, responsibilities and procedures across administrative, temporary, new-drug and treaty-based scheduling. It directed FDA to develop procedures for eight-factor evaluations and scheduling recommendations, including clearer standards for judging abuse potential. Finally, it recommended that FDA and NIH rewrite their decades-old agreement governing coordination with NIDA and require the document to be periodically reviewed.
The Department of Justice and Department of Health and Human Services agreed with the recommendations. HHS told GAO that FDA intends to develop new eight-factor-analysis policy and that FDA and NIH have already started updating the NIDA agreement. GAO currently lists all three recommendations as open, meaning the watchdog has not yet verified their full implementation.
What the Report Means for Marijuana Rescheduling
The clearest takeaway is not that GAO discovered evidence that marijuana’s Schedule III recommendation was manipulated or scientifically defective. It did not. The watchdog actually found a strong recent record of DEA considering HHS recommendations when legally required and reaching the same final scheduling outcome in every completed case it examined.
Instead, GAO identified a governance problem with potentially important consequences for future controversial drugs: the federal scheduling system depends heavily on experienced specialists applying professional judgment through procedures that are not always formally documented. Marijuana makes those gaps especially visible because its rescheduling has required federal agencies to address difficult questions about medical use, relative abuse potential, state medical programs and the relationship between HHS’s scientific findings and DEA’s law-enforcement responsibilities.
Whatever final decision DEA reaches on broader marijuana rescheduling will still have to rest on the Controlled Substances Act, HHS’s scientific and medical findings, the evidentiary record developed during rulemaking and the administrative hearing. GAO’s report does not dictate that outcome. It does, however, call for the agencies responsible for future scheduling decisions to put far more of their methodology into writing—so that decisions involving marijuana, emerging synthetic drugs and future controlled substances can be made through a process that remains consistent even when the people carrying it out change.
