463,396 Teenagers, Tracked for Years, Found the Same Thing: Cannabis Use Roughly Doubles Later Psychosis Risk. The DEA Is Deciding Its Schedule This Month Anyway.

A DEA administrative law judge is, right now, a few miles from adjourning a hearing that will decide whether marijuana moves from Schedule I to Schedule III — a hearing that recessed for the July 4th holiday and resumed this week, with a legal deadline to wrap by July 15. In a courtroom built around that one narrow regulatory question, a study published days ago in JAMA Health Forum is the kind of evidence that rarely gets read into the record, because it doesn’t answer the question anyone in the room is actually asking.

Rescheduling is a question about the drug war. This study is a question about what happens inside a developing brain, and those two questions have almost nothing to do with each other — which is exactly the problem with settling them in the same news cycle.

Researchers from Kaiser Permanente, UCSF, and USC followed 463,396 adolescents, ages 13 to 17, into their mid-twenties, tracking who used cannabis in the prior year and who later received a diagnosis of a psychotic disorder or bipolar disorder. Published as a peer-reviewed cohort study in JAMA Health Forum (DOI 10.1001/jamahealthforum.2025.6839), the finding is about as close to unambiguous as a large observational study gets: teens who used cannabis in the past year carried roughly double the risk of a later psychotic or bipolar diagnosis, with that diagnosis typically arriving 1.7 to 2.3 years after use began. The association held after the researchers statistically accounted for baseline mental health and use of other substances — the two variables most likely to be quietly doing the work behind a correlation like this one.

Here’s the plain-language version, distilled from both the paper and ScienceDaily’s summary of it, because “psychotic disorder” is a clinical label doing a lot of hiding: this is a study about whether smoking or vaping weed as a teenager makes it meaningfully more likely you’ll later experience delusions, hallucinations, or the kind of severe mood swings that get you a bipolar diagnosis — not a moral judgment, a measurable, trackable, roughly-doubled shift in odds, in a sample size large enough that “it was just a few unlucky kids” stops being a credible objection.

It matters that this is a large cohort, not a small clinical sample cherry-picked from people already showing symptoms. Nearly half a million teenagers is the kind of number that survives the standard objections to addiction-science headlines: reverse causation (are kids using cannabis to self-medicate an illness that was already brewing?), confounding by other drug use, and confounding by pre-existing mental illness. The study design directly tested for the first two and found the association persisted. It doesn’t fully settle causation — no single observational study does — but it moves the needle further toward “this is a real signal” than a study this size moving in this direction usually gets credit for in the public conversation about legalization.

That conversation, right now, is almost entirely about something else. The DEA’s rescheduling hearing, which resumed this week after a holiday recess, is adjudicating a narrower legal question: whether cannabis meets the criteria for Schedule III under the Controlled Substances Act, a determination about abuse potential and accepted medical use, not a referendum on adolescent brain development. Marijuana Moment’s rolling coverage of the hearing testimony makes clear the scheduling fight is happening almost entirely on adult-access and criminal-justice terrain — Drug Policy Alliance’s Cat Packer has argued in the same hearing’s public comments that decades of cannabis-related arrests and incarceration deserve as much weight as any pharmacology. That argument has real merit on its own terms. It simply isn’t the same conversation as the one this study is having.

Both conversations can be true at once, and the country keeps acting like they can’t. You can believe — as the evidence on criminalization’s harms increasingly supports — that arresting people for cannabis has done more damage than the drug itself, while also believing that a nearly-half-million-person study showing doubled psychosis risk in teenagers deserves its own separate, serious policy response: youth-specific potency limits, packaging and marketing restrictions aimed at people under 21, and screening protocols in pediatric and school-based health settings that ask about cannabis use the way they already ask about alcohol. Rescheduling doesn’t require ignoring this study. It requires building the adolescent-specific guardrails into whatever regulatory system replaces prohibition, instead of assuming Schedule III alone will do that work.

It requires building the adolescent-specific guardrails into whatever regulatory system replaces prohibition, instead of assuming Schedule III alone will do that work.

If you are a parent, school counselor, or pediatric provider reading this before the hearing concludes: the actionable move this week isn’t waiting for a DEA ruling. It’s asking the same direct, non-judgmental screening question about cannabis that’s already standard for alcohol at every adolescent well visit between now and whatever Schedule the DEA lands on. The rescheduling debate will resolve on its own timeline. A fifteen-year-old’s risk window doesn’t wait for it.

For more on how potency and product changes are reshaping cannabis risk, see our cannabinoids coverage and our earlier reporting on the rescheduling hearing’s opening days.