A Story of Bad Science: How Defenders of Antidepressant Efficacy Make Their Case

As you will see in this article, antidepressants don’t work for depression despite their misleading name.

The story psychiatric leaders have told to the public is the opposite, that these drugs are highly effective. I shall debunk this claim and give advice about the drugs that is in accordance with the science.

Antidepressants don’t work for depression

The first SSRI (Selective Serotonin Reuptake Inhibitor) that came into widespread use was fluoxetine (Prozac), approved by the FDA in December 1987 for depression in adults.

The drug is so poor that the German drug regulator found it “totally unsuitable for the treatment of depression” and noted that, according to the patients’ self-ratings, there was little or no response, in contrast to doctors’ ratings.¹

It is a consistent observation that antidepressants are ineffective when the patients are asked of their opinion, and this has been demonstrated in meta-analyses of the placebo-controlled trials. A convenient way to pool effects that have been measured using different ranking scales is to use the effect size, which is the average effect divided by the standard deviation of the measurements. As it is dimension-less, it allows pooling of results obtained with different scales. If the average effect on the Hamilton Depression Rating Scale, which goes from 0 to 52, is 2 and the standard deviation is 8, the effect size is 0.25. If the scale goes from 1 to 5, and the average effect is 0.16, with a standard deviation of 0.64, the effect size is also 0.25.

In placebo-controlled trials in children and adolescents, the psychiatrists reported an effect size of SSRIs of 0.25 whereas the patients reported an effect size of only 0.05.² A review with more trials and patients found the same, effect sizes of 0.29 and 0.06, respectively.³ Trials in adults of old drugs like amitriptyline had effect sizes of 0.25 and 0.06, respectively.⁴

The main reason why psychiatrists report much bigger effects than their patients is that the trials have not been blinded effectively. Because of the conspicuous adverse effects of antidepressants, the psychiatrist can often guess if a patient is on drug or placebo. This introduces bias. Placebo-controlled trials are virtually always funded by drug companies and psychiatrists are keen to show that the drugs work, as it will ensure that they get additional lucrative contracts with drug companies.

The effects psychiatrists have reported can easily be explained by this unblinding bias.⁵ And if the trials are better blinded, by putting atropine in the placebo, which has similar adverse effects as the drugs, the drug effect is smaller than in other trials.⁶

Recent meta-analyses, with many more patients, have confirmed these results. In adults, the effect sizes, as estimated by the psychiatrists, were 0.26⁷ and 0.30.⁸ These effects correspond to 2 points on the Hamilton Depression Scale, and the least clinically relevant effect using this scale is 5-6.⁹

Thus, even when the effect is measured by biased psychiatrists who are very often on industry payroll,¹⁰ it is so small that it makes no difference for the patients.

It is even worse than this, which I shall describe next. The patients don’t like the pills.

Antidepressants cause more harm than good

When the patients decide whether to continue in a placebo-controlled trial to the end or to drop out, they weigh the benefits against the harms of the drugs. Based on 71 clinical study reports (18,426 patients) we had obtained from drug regulators, my research group found that 12% more patients dropped out while on drug than while on placebo.¹¹

This is a very important result. The psychiatrists believe depression pills are helpful but the patients prefer placebo even though some of them have been harmed by cold turkey withdrawal effects when being randomised from a drug they were already on, to placebo, after a washout period that is too short to avoid such harms. This means that the drugs are even worse than what we documented.

The worst harm of antidepressants is that they double suicides in adults and very likely also in children.^12,13 Moreover, they impair the sex life in half the patients,¹⁴ and the inability to have sex or orgasm can become permanent.¹⁵ When the patients try to come off the drugs, half of them experience withdrawal symptoms, which are severe in half of the cases.¹⁶

So, what does the psychiatric establishment do when facing documentation in the randomised trials that antidepressants don’t work, double suicides, and cause important harms in half the patients, and when the patients prefer a placebo for an active drug?

They don’t draw the only honourable and evidence-based conclusion, which is that these drugs shouldn’t be used by anyone. To protect their harmful specialty, they continue deceiving the public. This is why I have called psychiatry the only medical specialty that survives on lies.¹⁷

Hyping the non-existent benefit

For decades, the non-existent benefit of antidepressants has been hyped in the extreme. The meta-analysis I mentioned above, which reported an effect size of 0.30, was published by Andrea Cipriani and colleagues in The Lancet in 2018.⁸ It got enormous media attention even though the drug effect was the same as in earlier meta-analyses.

The absurdity in this can be seen by comparing two articles in The Guardian. Prozac (fluoxetine) didn’t work in 2008 (effect size 0.32),¹⁸ but ten years later, all drugs worked (effect size 0.30).⁸

There was nothing new,¹⁹ but Cipriani claimed in the BBC that there were “big differences in how effective each drug is” and that “at least one million more people in the UK would benefit.”²⁰

Cipriani and the Royal College of Psychiatrists called it the final answer to the long-standing controversy about whether the pills work for depression. When you hear about “the final answer” in a debate about doubtful drugs that has lasted for decades, you can be pretty certain that it is hype propagated by people on industry payroll.

The Cipriani meta-analysis is much cited, uncritically, although it falls apart on closer inspection. Cipriani rewarded the companies that had cheated the most, and some of his findings were contradicted by more reliable research.²¹

Cipriani is the archetype of psychiatric leaders who have disinformed the public in scientific articles, guidelines, psychiatric textbooks and interviews to such an extent²² that if you use Artificial Intelligence, you will get wrong answers. Fluoxetine was the first antidepressant approved for children, and I asked Grok if this was correct, just to test it. I was told that “Fluoxetine stood out as the first with strong evidence from pediatric trials supporting its use for depression.”

Strong evidence supporting its use in children when it doesn’t work and likely doubles suicides? This is statistical alchemy turning sand into gold. I scrutinised the clinical study reports of the two placebo-controlled trials Eli Lilly had submitted to drug regulators to get fluoxetine approved in children with depression and invited psychiatrist David Healy to join me in this research.²³

The two fluoxetine trials were fraudulent because fraud is any activity that relies on deception to achieve a gain.²⁴ We found that essential information was missing; numerical discrepancies were unexplained; new outcomes appeared that were not prespecified in the trial protocol and rating scales and analyses were changed (the Texas sharpshooter fraud; you hit the bull’s eye because you make a new target after you fired the bullet); and the trial protocols were violated in other ways.²³

Even though the efficacy outcomes were biased by differential dropouts and missing data, the effect as assessed by the psychiatrists was only 4% of the baseline score, which is not clinically relevant, and patient ratings did not find fluoxetine effective at all.

The psychiatric leaders have fooled the public in numerous other ways and I shall mention the most common tricks.

Claims that the drugs work for severe depression are false

When psychiatrists admit that the drugs don’t work for mild depression, they always say they are effective for severe depression.

This is false.²⁵ Firstly, the effect measured by psychiatrists in patients with very severe depression is only 2.7 on the Hamilton scale,⁷ considerably lower than the smallest effect that can be perceived, which is 5-6.⁹ Second, the misconception that the drug effect is related to the severity of the depression is due to two mathematical artefacts, one of which is an illegitimate regression analysis.²⁵ The authors of a meta-analysis claiming this relationship made the error of regressing change in symptoms on initial symptom severity.²⁶ Thus, they looked at (x − y) = ax + b, where x is the initial value and y the final value. Since x appears on both sides of the equation, 50% of the variation is already explained. This means that even when two factors are unrelated, the analysis will show a relation, which, however, is spurious.

Claims based on number needed to treat are an illusion

Psychiatrists constantly tell the world how effective their drugs are by referring to the number needed to treat (NNT) to benefit one patient. Technically, NNT is calculated as the inverse of the benefit difference. If, for example, 60% have improved on drug and 50% on placebo, NNT = 1/(0.6-0.5) = 10. But that is just the mathematics. The data such NNTs are derived from are highly flawed.²⁷

When the top among UK psychiatrists in 2014 tried to convince their readers that depression pills are highly effective, they claimed that they have an impressive effect on recurrence, with an NNT of around three to prevent one recurrence.²⁸ I explained in my rebuttal that the so-called maintenance studies, in which patients after successful treatment get randomly assigned to continue with the drug or a placebo, cannot be interpreted as showing that the patients still need the drug because withdrawal symptoms, which can include depression, are inflicted on the placebo group.²⁹

As only two patients are needed to get one with withdrawal symptoms when a drug is stopped,¹⁶ there cannot exist an NNT to prevent recurrence, only a number needed to harm (NNH), which is two.

There are two additional problems with using NNT.²⁷ It only takes those patients into account that have improved by a certain amount. If a similar number of patients have deteriorated, there can be no NNT, as there is no benefit. Thus, a totally useless drug, which only makes the condition after treatment more variable, so that more patients improve and more patients deteriorate than in the placebo group, will seem effective based on NNT.

The other problem is that NNT is derived from data on a ranking scale. Statisticians have explained why it is a bad idea to dichotomise such data, which opens the door to additional bias.³⁰ If the chosen cut-off for improvement does not yield the desired result, other cut-offs can be tried till the data confess under torture. Manipulations with the data during data analysis are very common.³¹

Fraud

Lancet Psychiatry reported in 2015 that the ”landmark” STAR*D study had shown that 70% of depressed patients become symptom free.³² However, this study had no control group and if you wait long enough, which the researchers did in this study, years in fact, most patients will become symptom free without treatment. Moreover, the study—the biggest the US National Institute of Mental Health (NIMH) ever funded, at a cost of $35 million—was fraudulent on many counts.^22,33,34

But even after several researchers had repeatedly contacted reporters at The New York Times urging it to set the record straight by writing about the fraud,^35,36 the Times repeated its false claim of a 70% effect in a 2024 article with the preposterous title, “What you really need to know about antidepressants.”³⁷

The STAR*D study is still highly cited in psychiatric textbooks and elsewhere.³⁸ Its fraudulent results are not questioned and have not been retracted despite repeated requests.²² It could very well be the most harmful fraud ever published about antidepressant drugs.

There are none so blind as those who WILL NOT SEE

A widely used tactic to maintain the false narrative about antidepressants is to ignore unwelcome research results.

Nothing illustrates this better than Thomas Insel’s book, Healing: Our Path from Mental Illness to Mental Health.³⁹ From 2002 to 2015, Insel, called “America’s psychiatrist,” was the director of the NIMH.

The NIMH is the most prestigious institution in the world in mental health. Robert Whitaker therefore took a close look at the book in his article, “Thomas Insel Makes a Case for Abolishing Psychiatry.”⁴⁰

The book reflects the thinking of psychiatric leaders everywhere and encapsulates how psychiatry has consistently betrayed public trust and misinformed the public, which will never hear the truth about psychiatric drugs. Being a former NIMH director, Insel had an obvious ethical obligation to tell his readers about the poor long-term outcomes of treatment with psychiatric drugs, including antidepressants, as documented in expensive and prestigious research funded by the NIMH, but he didn’t.

When everything else fails…

As a last resort, if you press leading psychiatrists really hard with hard evidence, they might say that it is reasonable to try antidepressants because some patients respond better than others.

This way of thinking was illustrated by a Cochrane meta-analysis of depression pills in children,⁴¹ which I called a garbage in, garbage out review.⁴² The review demonstrated beyond any doubt how beholden to the drug industry the Cochrane Collaboration has become. This is clear already in the abstract, which is highly misleading, also in relation to the increased risk of suicide the pills cause, which is downplayed.

The authors conclude in their abstract that “most newer antidepressants may reduce depression symptoms in a small and unimportant way compared with placebo” but nonetheless argue that the drugs might be recommended “for some individuals in some circumstances.”

Such wishful thinking can be used about all ineffective treatments. The Cochrane authors apparently don’t know what statistical variation is. We use average effects to draw conclusions about whether we should use a drug or not.

Some years ago, at a meeting at my hospital, a clinical pharmacologist acknowledged that the effect of antidementia drugs is so small that it is irrelevant.⁴³ But he added that the drugs could be tried because some patients respond better than others. I told him a little about statistical variation. If a trial is repeated in the same patients, other patients will falsely seem to respond than those who had the biggest effect the first time. This is called random variation, a key concept in statistics.

Imagine your old car is causing trouble and your mechanic tells you that his fix doesn’t work, but he hopes it will work for your car. Why is it so difficult for doctors to realise that the way they use drugs, they would not want their mechanic to copy?

Doctors start patients on ineffective drugs and see how it goes. In contrast to cars, many patients become better with time, and they then draw the false conclusion that the drug worked. Drug authorities are equally unreasonable. Numerous package inserts recommend trying a drug and seeing how it goes. But drug regulators know perfectly well that clinical experience is grossly unreliable, which is why they demand randomised trials as a prerequisite for approving new drugs.

Conclusions

There are plenty of good reasons why I called my most recent, freely available psychiatry book, Is Psychiatry a Crime Against Humanity?²² In my view, some of the psychiatric leaders should be in jail because they knowingly spread falsehoods about psychiatric drugs that kill so many people that psychiatric drugs are the third leading cause of death.⁴⁴

Antidepressants are the major killer. Not just because they can cause suicide^1,12,13 and homicide,⁴⁵ but because elderly people may lose balance, fall, and break their hip, which kills one-fifth of them within the next year.⁴⁴

I cannot see any role for antidepressants in treating depression. These drugs should be taken off the market as they are harmful.

References

1 Gøtzsche PC. Deadly psychiatry and organised denial. Copenhagen: People’s Press; 2015.

2 Spielmans GI, Gerwig K. The efficacy of antidepressants on overall well-being and self-reported depression symptom severity in youth: a meta-analysis. Psychother Psychosom 2014;83:158–64.

3 Hetrick SE, McKenzie JE, Cox GR, et al. Newer generation antidepressants for depressive disorders in children and adolescents. Cochrane Database Syst Rev 2012;11:CD004851. I calculated the effect size based on standard deviations from other reviews.

4 Greenberg RP, Bornstein RF, Greenberg MD, et al. A meta-analysis of antidepressant outcome under “blinder” conditions. J Consult Clin Psychol 1992;60:664-9.

5 Gøtzsche PC. Deadly psychiatry and organised denial. Copenhagen: People’s Press; 2015; Hróbjartsson A, Thomsen AS, Emanuelsson F, et al. Observer bias in randomised clinical trials with binary outcomes: systematic review of trials with both blinded and non-blinded outcome assessors. BMJ 2012;344:e1119; Hróbjartsson A, Thomsen ASS, Emanuelsson F, et al. Observer bias in randomized clinical trials with measurement scale outcomes: a systematic review of trials with both blinded and nonblinded assessors. CMAJ 2013;185:E201-11.

6 Moncrieff J, Wessely S, Hardy R. Active placebos versus antidepressants for depression. Cochrane Database Syst Rev 2004;1:CD003012.

7 Jakobsen JC, Katakam KK, Schou A, et al. Selective serotonin reuptake inhibitors versus placebo in patients with major depressive disorder. A systematic review with meta-analysis and Trial Sequential Analysis. BMC Psychiatry 2017;17:58.

8 Cipriani A, Furukawa TA, Salanti G, et al. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: a systematic review and network meta-analysis. Lancet 2018;391:1357-66.

9 Leucht S, Fennema H, Engel R, et al. What does the HAMD mean? J Affect Disord 2013;148:243-8.

10 Whitaker R, Gøtzsche PC. The pervasive financial and scientific corruption of psychiatric drug trials. Institute for Scientific Freedom 2022;March 23.

11 Sharma T, Guski LS, Freund N, et al. Drop-out rates in placebo-controlled trials of antidepressant drugs: A systematic review and meta-analysis based on clinical study reports. Int J Risk Saf Med 2019;30:217-32.

12 Gøtzsche PC. Observational studies confirm trial results that antidepressants double suicides. Mad in America 2025;Feb 8.

13 Gøtzsche PC. Antidepressants are major drivers of suicides. But authors of suicide prevention articles ignore the elephant in the room. Mad in America 2026;Aug (in press).

14 Montejo A, Llorca G, Izquierdo J, et al. Incidence of sexual dysfunction associated with antidepressant agents: a prospective multicenter study of 1022 outpatients. Spanish Working Group for the study of psychotropic-related sexual dysfunction. J Clin Psychiatry 2001;62 (suppl 3):10–21.

15 What is PSSD? The PSSD Institute (undated; there are references to the most relevant literature).

16 Davies J, Read J. A systematic review into the incidence, severity and duration of antidepressant withdrawal effects: Are guidelines evidence-based? Addict Behav 2019;97:111-21.

17 Gøtzsche PC. The only medical specialty that survives on lies. Brownstone Journal 2025;Sept 8.

18 Kirsch I, Deacon BJ, Huedo-Medina TB, et al. Initial severity and antidepressant benefits: A meta-analysis of data submitted to the Food and Drug Administration. PLoS Med 2008;5:e45.

19 Timimi S, Moncrieff J, Gøtzsche P, et al. Network meta-analysis of antidepressants. Lancet 2018;392:1011-2.

20 Therrien A. Anti-depressants: Major study finds they work. BBC News 2018;Feb 22.

21 Gøtzsche PC. Rewarding the companies that cheated the most in antidepressant trials. Mad in America 2018;March 7.

22 Gøtzsche PC. Is psychiatry a crime against humanity? Copenhagen: Institute for Scientific Freedom 2024 (freely available).

23 Gøtzsche PC, Healy D. Restoring the two pivotal fluoxetine trials in children and adolescents with depression. Int J Risk Saf Med 2022;33;385-408.

24 Association of Certified Fraud Examiners. Fraud 101: What Is Fraud? (accessed 3 Jan 2024).

25 Gøtzsche PC. Do antidepressants work against severe depression? No, it is an illusion that comes from two mathematical artefacts. Gøtzsche’s Perspective 2026;Jan 24.

26 Fournier JC, DeRubeis RJ, Hollon SD, et al. Antidepressant drug effects and depression severity: a patient-level meta-analysis. JAMA 2010;303:47-53.

27 Gøtzsche PC. Number needed to treat with a psychiatric drug to benefit one patient is an illusion. Mad in America 2022;Dec 13.

28 Nutt DJ, Goodwin GM, Bhugra D, et al. Attacks on antidepressants: signs of deep-seated stigma? Lancet Psychiatry 2014;1:103-4.

29 Gøtzsche PC. Why I think antidepressants cause more harm than good. Lancet Psychiatry 2014;1:104-6.

30 Royston P, Altman DG, Sauerbrei W. Dichotomizing continuous predictors in multiple regression: a bad idea. Stat Med 2006;25:127-41.

31 Chan AW, Hróbjartsson A, Haahr MT, Gøtzsche PC, Altman DG. Empirical evidence for selective reporting of outcomes in randomized trials: comparison of protocols to published articles. JAMA 2004;291:2457-65.

32 Sharpe K. The silence of Prozac. Lancet Psychiatry 2015;2:871-3.

33 Pigott HE, Kim T, Xu C, et al. What are the treatment remission, response and extent of improvement rates after up to four trials of antidepressant therapies in real-world depressed patients? A reanalysis of the STAR*D study’s patient-level data with fidelity to the original research protocol. BMJ Open 2023;13:e063095.

34 Whitaker R. The STAR*D scandal: scientific misconduct on a grand scale. The American Journal of Psychiatry needs to retract study that reported fraudulent results. Mad in America 2023;Sept 9.

35 Whitaker R. The New York Times is now engulfed in the STAR*D scandal. Mad in America 2024;April 27.

36 Leventhal A. STAR*D: the harms of orchestrated psychiatric fraud. Mad in America 2024;Sept 18.

37 Caron C. What you really need to know about antidepressants. New York Times 2024;April 25.

38 Badre N, Compton J. STAR*D: It’s time to atone and retract. Psychiatric Times 2024;May 25.

39 Insel T. Healing: our path from mental illness to mental health. New York: Penguin Press; 2022.

40 Whitaker R. Thomas Insel makes a case for abolishing psychiatry. Mad in America 2022;Apr 30.

41 Hetrick SE, McKenzie JE, Bailey AP, et al. New generation antidepressants for depression in children and adolescents: a network meta-analysis. Cochrane Database Syst Rev 2021;5:CD013674.

42 Gøtzsche PC. Garbage in, garbage out: the newest Cochrane meta-analysis of depression pills in children. Mad in America 2021;Aug 19.

43 Gøtzsche PC. Whistleblower in healthcare (autobiography). Copenhagen: Institute for Scientific Freedom; 2025 (freely available).

44 Gøtzsche PC. Prescription drugs are the leading cause of death. And psychiatric drugs are the third leading cause of death. Mad in America 2024;April 16.

45 Gøtzsche PC. Antidepressants can cause homicide. Gøtzsche’s Perspective 2026;June 3.