Cannabis Never Treated Your Anxiety — It Just Made Tonight Quieter

Say it’s 11 p.m. and the apartment is quiet in the way that makes your chest tight — no distractions left to hide behind, just you and whatever your brain wants to replay. You pack a bowl. Twenty minutes later your shoulders have dropped an inch, the loop in your head has gone slack, and you can finally fall asleep instead of lying there rigid. You’ve done this most nights for two years. As far as you’re concerned, cannabis is the thing that treats your anxiety. It’s the only thing that’s ever reliably worked.

Feeling better tonight and getting better over time are not the same thing, and the entire medical-cannabis-for-mental-health economy is built on people not noticing the difference.

Here’s what a team led by Jack Wilson at the University of Sydney’s Matilda Centre just found after combing through every randomized controlled trial on cannabis and mental health ever published — 54 trials, 2,477 people, 45 years of data, released in The Lancet Psychiatry this March: there is no trial evidence that cannabis treats depression. None. There’s no significant effect for anxiety or PTSD either, the two conditions right behind depression as reasons people get a medical cannabis card in the first place, according to coverage of the review. Roughly 27% of adults in the U.S. and Canada report using cannabis medicinally, and about half of them name a mental health condition as the reason, per the University of Sydney’s summary of the findings. That’s tens of millions of people using a substance for a job the best available evidence says it doesn’t do.

This isn’t a piece about whether you should feel bad for what got you through. It’s about what’s actually happening in your head when relief this real produces a result this consistently absent from the data — and what to do with that gap if you’re the person mid-loop tonight.

Your brain is a very good storyteller, and it’s not lying to you on purpose

Start with the plainest version of what researchers call an “expectancy effect”: if you believe a thing will calm you down, some of the calming happens because you believed it, not because of what’s chemically inside the thing. This isn’t willpower or placebo-as-weakness — expectancy effects are measurable, physiological, and they stack on top of whatever a drug actually does pharmacologically. You’ve built two years of nightly evidence that cannabis works. Every single night it has “worked” reinforces the belief, which primes the next dose to work a little better, regardless of what THC and CBD are doing to your amygdala.

Layer onto that what Carrie Cuttler’s lab at Washington State University has documented and what researchers call the “stress-misattribution” problem: a lot of what people label as their anxiety being treated is actually generalized daily stress being temporarily dulled, which gets filed under “anxiety relief” because that’s the story you already have on hand. Cuttler’s research on cannabis and negative affect found short-term reductions in self-reported anxiety and depression that reversed into worse symptoms over extended use — the thing that feels like your treatment is, longitudinally, correlated with the condition getting harder to live with, not easier.

Then there’s the mechanism researchers call negative reinforcement, which is simpler than the term sounds: any behavior that makes a bad feeling go away gets stamped into your brain as “do this again,” whether or not it addresses why the bad feeling showed up. Scratching an itch is negative reinforcement. So is checking your phone when you’re bored. Smoking when your chest goes tight is the same loop, and a 2016 study on anxiety sensitivity and coping motives found that people who are least able to tolerate distress in the first place are the ones most likely to lean on cannabis specifically because it removes the feeling fastest — which also means it’s teaching your nervous system to avoid the sensation of anxiety instead of learning it can be survived. Avoidance coping doesn’t shrink the underlying fear. It just gets very efficient at helping you not look at it, which is a different, much smaller job than treatment.

It just gets very efficient at helping you not look at it, which is a different, much smaller job than treatment.

The relief has a bill, and it comes due on a delay

If cannabis only ever helped, this wouldn’t be a story. What Wilson’s team flagged, and what researchers have been circling for years, is that chronic use tends to blunt the same short-term relief you started with — tolerance builds, so the dose that used to quiet your chest at 11 p.m. stops being enough, and some people escalate use to chase a baseline that keeps receding. Wilson told reporters covering the review that medicinal cannabis “could be doing more harm than good,” citing worsening mental health outcomes and increased risk of cannabis use disorder, on top of the more immediate cost: every night you treat the symptom is a night you’re not pursuing something with an actual evidence base, according to ScienceDaily’s writeup of his comments.

That’s the rebound problem in one sentence: the thing you’re using to manage anxiety about your anxiety is, for a meaningful number of people, quietly generating more of it between doses — and PTSD in particular carries dose-dependent risk, where higher-THC use is more likely to trigger the hypervigilance and dissociation it’s supposedly treating.

The self-medication story didn’t start with you

This isn’t a new pattern. The template for “I use cannabis to treat X because nothing else worked” was built in the late 1980s and early 1990s, when AIDS patients and advocates — many in San Francisco, facing a government actively hostile to them — found cannabis eased the nausea and wasting that era’s antiretrovirals couldn’t touch, and built a movement around that lived experience. It worked: California’s Prop 215 passed in 1996, the first state medical-cannabis law, on patients testifying they knew their bodies better than the DEA did. The underlying pharmacology for nausea held up. But the narrative — I felt better, therefore this treats my condition — got exported wholesale to conditions where the pharmacology doesn’t hold the same way. The feeling of relief became its own evidence, three decades before anyone ran 54 trials to check.

What’s happening in Washington while you read this

That history is colliding with live policy this week. The DEA’s hearing on whether to move marijuana from Schedule I to Schedule III wraps up July 15 — tomorrow — after opening June 29. DEA announced its seven hearing witnesses on June 18, and by NORML’s count, every one of them opposes rescheduling; the agency rejected participation requests from the Drug Policy Alliance, NORML, the Marijuana Policy Project, and half a dozen other groups that wanted to argue the other side. Drug Policy Alliance’s Cat Packer, shut out of the hearing room, wrote afterward that there was no discussion of the lasting harms of marijuana criminalization inside it, according to Marijuana Moment’s coverage. Meanwhile two state-licensed operators, MedPharm Iowa (doing business as Bud & Mary’s) and Tri-Mountain Pure, filed to intervene in the D.C. Circuit litigation trying to undo the DOJ’s April rescheduling order, arguing they’d suffer direct economic and regulatory harm if it’s reversed — a reminder, per Marijuana Moment, that there’s now a multibillion-dollar industry with a direct stake in cannabis staying framed as medicine, mental health claims included, regardless of what the trial data says. None of that changes what the Lancet review found. It does explain why you’ll keep hearing “cannabis treats anxiety” stated as settled fact in the same news cycle where the closest thing to a rigorous answer says otherwise.

If tonight is one of the nights

You are not imagining the relief. If a joint takes the edge off tonight, that relief is real, your nervous system did something measurable, and nobody researching this is telling you you’re weak or making it up. What the data says is that relief and treatment are not the same thing, and confusing them can cost you — in tolerance, in rebound anxiety, in the months or years you spend managing a symptom instead of addressing what’s underneath it. The conditions cannabis isn’t shown to treat have treatments that are shown to work: trauma-focused cognitive behavioral therapy and EMDR both have a real evidence base for PTSD, and SSRIs remain the best-studied pharmacological option for anxiety and depression, imperfect and slow as they can be. None of that requires you to quit anything tonight. It just means the next time cannabis “works,” it’s worth asking what it actually did — bought you eight hours of quiet, or moved you an inch closer to not needing the quiet bought at all.

The DEA hearing closes tomorrow with a verdict about a drug schedule. It won’t touch the question that actually matters at 11 p.m., which was never legal and always psychological: not whether cannabis is medicine, but why relief this convincing can be evidence of so little. For more on the policy fight shaping access either way, see our earlier coverage of the rescheduling fight, and browse more science and medicine reporting and our full cannabis archive.

The DEA hearing closes tomorrow with a verdict about a drug schedule.