# Lesson 1: Psychedelics and Psychedelic-Assisted Therapy: Definitions, Effects and How to Weigh Claims

*Companion-podcast transcript • Sarah & Kiffer*  
*Office Hours episode to listen to after working through the lesson*

---

**Sarah:** Here's a sentence from the website of a Canadian psilocybin store, one of the statements that researchers reproduced in a 2025 study: "Psilocybin is a potent tool to cure severe cases of depression and PTSD." So which drug did the phase 3 trials for PTSD that the FDA reviewed in 2024 actually use?

**Kiffer:** MDMA. The application that an advisory committee of the United States Food and Drug Administration reviewed in June 2024 rested on phase 3 trials of MDMA-assisted therapy, with three eight-hour sessions set inside weeks of preparation and integration. MDMA acts through the serotonin transporter, and in Canada it sits in Schedule I, while psilocybin sits in Schedule III.

**Sarah:** So for PTSD, the sentence seems to lean on evidence about a different drug, given in a very different way.

**Kiffer:** And the word psychedelic is what makes the borrowing easy. Much of the work of weighing a claim in this field is finding out which molecule, which setting and which study the claim actually rests on.

**Sarah:** Welcome to Office Hours. I'm Sarah.

**Kiffer:** And I'm Kiffer. That problem of working out what a claim is really about runs through Lesson one, Psychedelics and Psychedelic-Assisted Therapy: Definitions, Effects and How to Weigh Claims, so let's start there.

**Sarah:** Question one. When a headline says psychedelics, how do you tell which drugs it means, and why does it matter?

**Kiffer:** Start with the narrow meaning. The classic psychedelics, psilocybin, LSD, DMT and mescaline, form one class because they activate the serotonin 2A receptor, even though they come from three chemical families. MDMA is the case that confuses people. It's a phenethylamine, like mescaline, so on paper it looks like a relative. Yet it binds weakly or not at all to serotonin receptors, and it works by being carried into neurons by the serotonin transporter, which then releases stored serotonin. Its structure suggests one family, and its mechanism puts it in another.

**Sarah:** And mechanism is what matters to a clinician, because it shapes the effects, the risks and the drug interactions, which Lesson three takes up. The other complication is that the law and the national surveys each sort the same drugs in their own way.

**Kiffer:** Right. So here's one for everyone. Ketamine has been in clinical use as an anesthetic since 1970, and it sits in Schedule I of the Controlled Drugs and Substances Act, the schedule that also holds heroin and fentanyl. Psilocybin has no approved product in Canada, and it sits in Schedule III. Does a drug's schedule tell you how dangerous it is, or how useful it is as a medicine? Take a few seconds, and pause the audio if you'd like more time.

*(Pause)*

**Sarah:** It doesn't directly tell you either one. Schedules are legal categories, and this pair shows that they line up with neither danger nor medical use. A drug with more than fifty years of clinical use sits in Schedule I, and a drug with no approved product sits in Schedule III. What a schedule does affect is the penalties and the conditions a researcher has to meet before running a trial.

**Kiffer:** And the surveys add a third system. The 2019 Canadian Alcohol and Drugs Survey counted PCP, a dissociative, among hallucinogens, while the 2025 Canadian Substance Use Survey counts salvia, whose active compound works at kappa-opioid receptors, among psychedelics. So a survey's figure for psychedelic use can include a drug that pharmacology places well outside the classic psychedelics. When I read a claim, I ask three things. Which molecule? Which system of categories? And in which setting was the drug used?

**Sarah:** That third question is where the storefront sentence falls apart, which brings us to question two. When a trial of psychedelic-assisted therapy reports a benefit, what exactly has been shown to work?

**Kiffer:** That depends on what each arm received. In the MDMA trial reported by Mitchell and colleagues in 2021, people randomized to placebo got the same preparation, the same long sessions and the same integration, with an inactive pill. So the trial compared MDMA plus therapy with therapy plus placebo. What it estimates is the added effect of the drug inside that particular therapy.

**Sarah:** Let's get a sense of how much therapy that is. In that protocol there were three eight-hour experimental sessions and nine integration sessions of ninety minutes each. Counting just those, how many hours of scheduled sessions did each participant have? And compare that with the escitalopram arm of the psilocybin trial by Carhart-Harris and colleagues, who took a pill on about forty-two days. Take a few seconds.

*(Pause)*

**Kiffer:** Three sessions of eight hours is twenty-four hours. Nine sessions of ninety minutes is thirteen and a half hours. Together that's thirty-seven and a half hours, before you add the three preparatory sessions. The escitalopram group took a daily pill on about forty-two days, and the psilocybin arm of that trial took its active drug on only two days. That means much of what psychedelic-assisted therapy delivers is measured in hours of supervised contact with trained staff.

**Sarah:** Which raises an awkward question. If both arms got all of that therapy, the trial can't tell us how much the therapy itself contributed.

**Kiffer:** That's right, and it was one of the concerns that members of the FDA advisory committee raised in June 2024: there was little evidence about the contribution of the psychological component. The design also can't tell us what the drug does with no therapy at all, which is the situation of a person buying capsules in a store. Esketamine is the mirror image. The FDA approved the nasal spray in 2019 for use under supervision in a certified office or clinic, and psychotherapy is not part of the approved product.

**Sarah:** The setting has a legal side too, which matters for anyone heading into counselling or nursing. As of October 2026, the lawful routes in Canada for therapy with psilocybin or MDMA are clinical trials, the Special Access Program, whose route for these drugs has been governed since the first of October 2026 by the Controlled Substances Regulations, and exemptions under subsection fifty-six one of the Act.

**Kiffer:** And here's a point that matters for future clinicians. Because these drugs are scheduled, section four of the Act makes possessing them an offence except as authorized, and section twenty-one of the Criminal Code makes anyone who aids or abets an offence a party to it. So a regulated professional who gives preparation or integration support for unauthorized use may be at legal risk. That's a general description of the law, offered as information only, and Lessons seven and eleven work through it in more detail.

**Sarah:** Question three. What can an account of a psychedelic experience establish, whether it comes from a research questionnaire or from a website?

**Kiffer:** Take the questionnaire first. In the validation data for the MEQ30, ninety-four of one hundred eighty-four psilocybin sessions, about half, counted as a complete mystical experience. That figure comes from a rule: at least sixty percent of the maximum score on every one of the four factors. The items follow W. T. Stace's 1960 account of mysticism, which comes from one philosophical and religious tradition, and participants rated them about seven hours after taking the capsule.

**Sarah:** So a person whose experience doesn't fit that tradition's categories could score below the line and still find the session deeply meaningful. And someone at fifty-nine percent on one factor is classified with someone at ten percent.

**Kiffer:** Both true. The strength of a scale is that everyone answers the same items, so you can compare doses, drugs and people. Its limit is that researchers chose those items in advance. User-reported sources have the opposite profile.

**Sarah:** Here's where I'd push. Erowid's Experience Vaults hold more than forty-five thousand reports. Surely a collection that size tells you something about how common an experience is?

**Kiffer:** Size doesn't fix the problem of who writes. Let's make up a community to show it, and every number here is hypothetical. Imagine ten thousand psychedelic sessions, and suppose one in ten involves a challenging experience. Now suppose that people who had a challenging experience are five times as likely to write about it, so five percent of them write a report, compared with one percent of everyone else. What share of the written reports describe a challenging experience? Take a few seconds.

*(Pause)*

**Sarah:** One in ten of ten thousand is one thousand challenging sessions, and five percent of those is fifty reports. The other nine thousand sessions produce reports at one percent, which is ninety. That's fifty out of one hundred forty, or about thirty-six percent. So the reports would make challenging experiences look more than three times as common as they really are.

**Kiffer:** Now the harder part. Suppose instead that challenging experiences are only twice as likely to be written up. Keep the same ten thousand sessions.

**Sarah:** Then it's two percent of one thousand, which is twenty reports, and still ninety from everyone else. Twenty out of one hundred ten is about eighteen percent.

**Kiffer:** So the same true frequency of ten percent shows up as eighteen percent or as thirty-six percent of the reports, depending on a writing rate that nobody observes. And notice what didn't help. A bigger collection multiplies both counts by the same amount, so forty-five thousand reports carry the same distorted share as one hundred forty. Treating a large archive as if it were a representative sample is the trap.

**Sarah:** And you still couldn't check what anyone took. The microdosing study by Szigeti and colleagues listed the lack of verification of substance and dose as a key limitation, and Health Canada warns that illegally produced drugs aren't tested.

**Kiffer:** Which is why the course uses these sources for what they do well. They record the words people use, like tracers or thought loops, and the range of what people describe, including rare experiences a questionnaire would never ask about. Researchers who analysed Erowid reports found preliminary signs that drugs with similar receptor profiles are described in similar language, which gives laboratories a hypothesis to test. Many clients arrive speaking this vocabulary, so I think practitioners should know it. They should also keep in mind that it describes experiences and can't show that an effect is common, safe or helpful.

**Sarah:** Question four. Is psychedelic use in Canada going up?

**Kiffer:** The honest answer is that the national survey can't tell us yet. Past-year use went from 6.0 percent in 2023 to 7.4 percent in 2025, but the questionnaire moved from one grouped question to specific ones, and lifetime use jumped by more than ten points in two years, which is hard to believe as a change in behaviour. The survey is on firmer ground within a single year, because everyone in that year answered the same questions.

**Sarah:** Let me try that with the mental health figures, because the 2023 numbers add something. In 2025, past-year use was 13.1 percent among people who rated their mental health as fair or poor, and 6.0 percent among people who rated it good or better. That's about 2.2 times as high. In 2023 the figures were 12.5 and 4.5 percent, which is about 2.8 times. So the gap narrowed.

**Kiffer:** Check where the change came from, though. The fair or poor group barely moved, from 12.5 to 13.1, and its two confidence intervals overlap. Most of the change was in the larger group, from 4.5 to 6.0, where the intervals don't overlap. A change in wording could prompt people in some groups to report use more readily than people in others, so we can't say whether the gap really narrowed. The gap within 2025 is the more solid finding, since both groups answered the same questions that year.

**Sarah:** And even that gap can't tell us the direction. Distress may lead to use, use may affect mental health, or stress, trauma and discrimination may raise both.

**Kiffer:** Agreed. The practical point for clinicians is simpler. People who rated their mental health as fair or poor were more than twice as likely to report past-year use, so I'd expect practitioners in mental health settings to meet clients who arrive with their own experience of these drugs and their own reading about them.

**Sarah:** Question five is the checklist, and I'd like to try it on a newer claim. In March 2026, JAMA Network Open published a pilot randomized trial in which eighty-two adults who smoked received either a single dose of psilocybin or a nicotine patch, each with thirteen weeks of cognitive behavioural therapy. At six months, forty-one percent of the psilocybin group and ten percent of the patch group had biochemically verified prolonged abstinence.

**Kiffer:** Before Sarah rates it, try the six items yourself: design, comparator, blinding, sample, follow-up and funding, and say cannot tell wherever the description leaves you guessing. Take a few seconds.

*(Pause)*

**Sarah:** Okay. Design meets the standard, because it's randomized, though it's a pilot. The comparator meets it too, because the nicotine patch is an established treatment and both groups got the same therapy. And blinding also meets the standard, since people were randomized, so neither group chose its own treatment.

**Kiffer:** Hold on to that last one. What problem does randomization solve, and what problem does blinding solve?

**Sarah:** Right. Randomization decides who goes into each group, so the groups start out alike. Blinding is about what people know after they're assigned. This trial was unblinded, so everyone knew whether they'd had psilocybin or a patch. Blinding falls short.

**Kiffer:** And why does that matter here, given that abstinence was checked biochemically?

**Sarah:** The biochemical check stops people from misreporting whether they smoked. It can't stop knowledge of the treatment from changing how hard people actually try. Someone who received a novel treatment might feel hopeful and work harder at quitting, and someone handed a patch they may already have tried might be disappointed and give up sooner. Both would widen the gap between the groups.

**Kiffer:** Good. That's the difference between protecting the measurement and protecting the behaviour that gets measured. Finish the list.

**Sarah:** Sample falls short, because eighty-two people make a pilot and the estimates will be imprecise. Follow-up at six months is reasonable for a pilot of this kind. And I can't tell anything about funding from what I've read, so I'd rate that one as cannot tell from the evidence given. The difference is thirty-one percentage points, which looks large, and it rests on a small, unblinded study.

**Kiffer:** That's a fair summary. Compare it with the open-label pilot from 2014, where twelve of fifteen people, eighty percent, were abstinent at six months. It's tempting to say that randomizing cut the success rate in half. Check the outcomes first, though. The eighty percent meant no smoking in the seven days before the six-month check, a less strict measure, and on that measure the newer trial's psilocybin group was at fifty-two percent. The samples and designs differed too, and the eighty percent came from a single group with nothing to compare it with. What the newer design adds is a comparison group that also received thirteen weeks of therapy, and ten percent of that group achieved prolonged abstinence.

**Sarah:** Okay. Now the part where we disagree, and it's about a result from Section four. In June 2025, Compass Pathways reported to investors that a single 25 milligram dose of its psilocybin product beat placebo by 3.6 points on a sixty-point depression scale at six weeks. I think the field is still inside the hype bubble that Yaden, Potash and Griffiths described, and a result like that draws far more excitement than it has earned.

**Kiffer:** I'd put more weight on the other end of the hype cycle, the trough, where modest but real benefits get thrown out with the exaggerated ones. These were adults with treatment-resistant depression, after one dose, in a double-blind trial with a placebo. Then in February 2026 the company reported a second phase 3 trial, in which two 25 milligram doses three weeks apart beat two 1 milligram doses by 3.8 points at six weeks. Two trials that point in the same direction give more confidence than one.

**Sarah:** Both company-reported, both from the same sponsor, and both open to the problem of functional unblinding from the lesson. A strong psilocybin session is hard to mistake for a placebo. And look at the confidence interval for the first trial, from 1.5 to 5.7 points. The data are compatible with anything from 1.5 points to a benefit nearly four times that size, and at the small end, 1.5 points on a sixty-point scale is very little.

**Kiffer:** That's fair, and the MDMA story supports your caution. A large difference from placebo still failed to persuade the advisory committee in 2024, partly because of blinding. Where I land is in the middle. I'd describe the psilocybin evidence for treatment-resistant depression as promising and modest, and I'd wait for the full peer-reviewed reports and the FDA's review before calling it established. In July 2026 the company reported that, in the second trial, the 25 milligram group stayed separated from the 1 milligram group through twenty-six weeks, which speaks to the follow-up item, but that is still the sponsor's own summary.

**Sarah:** I'm more skeptical than Kiffer, but we agree on the practical rule. When you summarize a result like this, give the size of the difference with its interval, name the comparator and the length of follow-up, and say who reported it. A summary with those four details avoids both common errors, calling the result a cure and calling it a failure.

**Kiffer:** Let's pull it together with three things to take away.

**Sarah:** First, before weighing any claim about psychedelics, find the molecule, the category system and the setting. Mechanism groups these drugs one way, the schedules another and the surveys a third, and evidence about one drug can't be carried over to another.

**Kiffer:** Second, a trial of psychedelic-assisted therapy tests a package. When both arms receive the same hours of therapy, the result is the drug's added effect inside that therapy, and it says little about the drug on its own or the therapy on its own.

**Sarah:** And third, check what protects each claim. Randomization protects the starting groups, blinding keeps knowledge of the group from shaping what happens next, and a denominator protects any claim about how common something is, which user-reported collections lack however large they grow.

**Kiffer:** If you'd like more practice, rework the made-up reporting example from this episode with challenging experiences three times as likely to be written up, and check how far the share of reports moves from the true ten percent.

**Sarah:** Next time, it's Lesson two, Histories: Ceremony, Canadian Psychiatry, Prohibition and Revival, where we go back to ceremonial traditions and to the psychiatric research of the 1950s in Saskatchewan and British Columbia.

**Kiffer:** Take care, everyone.

**Sarah:** See you in Lesson two.
