evidence·start here
the evidence library
most mental-health content asks you to trust it. this asks you to check it. every claim on this site traces to a source with a date on it, every source carries a score with a stated method, and the two things we counted ourselves ship as downloadable data so you can recount them and tell us we were wrong.
if you are a reporter, a researcher, or just someone who got handed a prescription and wants to know what it rests on, this is the map. it goes from the raw records upward.
two things we counted ourselves
these are the only pages here that are not a summary of somebody else's work. both are counts over the same fixed set of study records, both publish every row, and both download as CSV under CC BY 4.0 — reuse the numbers anywhere, including commercially, as long as you say where they came from.
dataset · csvhow long were these drugs actually tested?57 of 72 randomised trials in the library state how long they ran. 47 of those measured their headline result at three months or sooner. people take these medications for years.dataset · csvwho paid for the research behind your medication?we could read the funding statement for 148 of 235 records. 17 were funded by a company with a product in the study. the other 87 sit behind a paywall — which is its own finding.
what each approval was actually based on
37 medications have an approval-history page: the indication the FDA actually approved, the year, the trials the approval rested on, how long those ran, and the label text quoted rather than paraphrased. this is public information that is technically available and practically unreadable. these pages are the readable version.
37 medicationsevery approval journeythe full list, one page each. what the FDA approved it for, in what year, on which trials, and what the label says where the record states it.examplesertraline (Zoloft)one of the most prescribed antidepressants in the world. approved 1991 — see what for, and on what.304 recordsthe scored libraryevery study we have read, with its trust tier, its funding, its design and its stated limitations. the two audits above are counts over this.
six questions, answered study by study
each of these takes one question people actually type at 2am and walks the evidence for it — including the trials whose conclusions are weaker than their reputations, and the places two good studies flatly contradict each other.
Sertraline (Zoloft): what the evidence actually showsSertraline is one of the most prescribed antidepressants in the world. Here is what the trials found about how well it works, how long it takes, what it does to sexual function, and what happens when you stop — each claim linked to the study behind it.Trazodone for sleep: what the label actually saysTrazodone is one of the most prescribed sleep medications in America. It is not approved for sleep. Here is what the trials actually found, how long they ran, and the questions worth taking to your prescriber.Benzodiazepines and long-term use: what the evidence says about dementia, memory, and coming offThe trials that got benzodiazepines approved ran eight weeks. Many people take them for years. Here is what the long-term evidence actually shows — including two good studies that flatly contradict each other.Stopping antidepressants: what the evidence says about withdrawal and relapseTwo different things happen when people come off an antidepressant, and they get confused constantly: withdrawal symptoms and relapse. Here is what the trials actually measured, and what a regulator made manufacturers put on the label.Ketamine vs esketamine (Spravato): what the trials actually comparedThey are not the same drug, they were not tested the same way, and the trial that made esketamine an approved medicine is not the trial most people think it is. Here is the evidence, study by study.Does psilocybin work for depression? What the trials actually foundPsilocybin trials produce dramatic headlines and genuinely interesting results. They also share a specific set of problems — broken blinding, sponsor-run studies, and effects that fade. Here is each major trial, what it found, and what is wrong with it.
why the room is arranged this way
the pages above are about whether a treatment works. these are about the machinery around it — how clinicians get paid, what an FDA approval does and does not certify, where a diagnosis comes from, and who funds the organisations that speak for patients. aggressive about the subject, conservative about every claim: each figure traces to a government record, a tax filing, or a peer-reviewed count, and the counter-evidence is on the page next to it.
charts · annotatedhow childhood ADHD got countedthe definition changed in 2013 and the longest clean federal series barely moved. every survey redesign moved it. four segments, every break annotated with the agency's own words.charts · annotatedhow autism got countedthe “1 in 31” figure is a records-based count from 16 surveillance sites, not a national rate — the same year spans 9.7 per 1,000 in laredo to 53.1 in california. every method break annotated with CDC's own words, and the decomposition literature with its confidence intervals intact.primary sourcesthe DSMthe manual that decides who has what: every edition and what changed in it, the committee that writes it, why no diagnosis in it has a confirming lab test, and the money — anchored on insiders, with the american psychiatric association's own rebuttal printed beside every charge.primary sourceswhat the DSM is for227 different symptom combinations all qualify as major depressive disorder, and two people can carry it without sharing one. the companion page: what a criteria set actually is, what the manual says about its own limits, why 1980 was the break, how a category becomes a billing code, whether the label even stays put (0.90 for schizophrenia, 0.29 for schizophreniform — so both slogans are wrong) — and the honest verdict on access (yes) and outcomes (unproven), with the strongest defence printed at the same size.public recordswho funds the advocates?pharma money to advocacy organisations and professional associations, from disclosures, a senate committee record and tax filings — with the organisations' own responses beside it.charts · construct-labeledLGBTQ+ mental health, in the datathe published federal series by orientation and identity, with the measurement named on every number — plus the crosses no agency publishes, and the SOGI data collections that were switched off, documented from agency artifacts.claims gradedthe mental-health lexiconnineteen terms the internet uses fluently — trauma, dopamine detox, adrenal fatigue, the TikTok tics, “on the spectrum” — each graded against its actual paper trail, with contested science cited from both sides and the entries where the science beats the meme.spend · annotatedpharma and the news$6.0B a year in drug ads, 76% of it to television — and the number everyone asks for, pharma's share of news revenue, published nowhere. the money, the measured coverage record, and the honest gap between them.the accreditor’s own bookswho funds doctors’ education$841.9M of drug and device money in accredited continuing medical education in 2025 — a fifth of its income, down from about half twenty years ago, and 80% of it going to for-profit education companies rather than medical schools. plus the regulation that keeps almost all of it out of the federal transparency database.thirteen annual reportswho funds the nonprofits?we went looking for ten mental-health nonprofits with named drug-company funders and found five. every sponsor one of them names is a manufacturer — twelve of twelve — while another publishes constantly about antipsychotics and refuses the money entirely. sorted by disclosure quality, because the three biggest are the three we can say least about.Why your doctor gets paid for procedures, not outcomesMost American healthcare is paid for by the item — a visit, a test, a procedure. Not by whether you got better. Here is what the research shows about how that changes what happens in the room, what it does not show, and what to ask.What FDA approval actually means — and what it doesn'tFDA approval is a real bar and a narrow one. It says a drug beat a comparator for one indication, in a defined population, over a defined period. Here is what that covers, what it leaves open, and what the FDA itself says about the limits — in its own words.The serotonin story: what the evidence said, and what patients were toldA generation of patients was told depression is caused by a chemical imbalance in serotonin. A 2022 umbrella review found no consistent evidence for that. Here is what the research actually shows, what the rebuttals argue, and why none of it means your antidepressant doesn't work.Where diagnoses come fromThere is no blood test for depression, ADHD or anxiety. Psychiatric diagnoses are agreements about which symptoms cluster together — written down, revised by committee, and published in a book. Here is how that book gets made, in the words of the people who ran it.
how any of this is decided
scoringhow we score evidenceevery resource carries a trust badge. this is exactly how that number is calculated, what it can tell you, and what it cannot.the productwhat actually happens when you postthe community numbers, including the flattering figure we published first, audited a week later, and replaced with the smaller true one.