Antidepressants No Better Than Placebo for About 85% of People

Researchers can’t predict the 15% who benefit from antidepressants, and the other 85% are unnecessarily exposed to the harms of the drugs.

24
15072

Clinical trials typically find that antidepressants are slightly better than placebo, at least in the short term. However, many researchers argue that this difference—about two points on a 52-point depression scale—is clinically imperceptible.

The question is what this minimal average difference means. There are two possibilities:

  1. Most people experience just a tiny bit more improvement on the drug (a 12-point improvement) than they would on a placebo (a 10-point improvement); or
  2. A small group of people experiences a larger effect from the drug, which is canceled out on average by the larger group of people who experience no effect.

In a new study, researchers have now concluded that it is the latter—in clinical trials, about 15% of people experienced a large effect from the antidepressant drug that they would not have received from the placebo. The authors write:

“The observed advantage of antidepressants over placebo is best understood as affecting a minority of patients as either an increase in the likelihood of a Large response or a decrease in the likelihood of a Minimal response.”

The paper appeared in BMJ. It was led by Marc Stone at the FDA’s Center for Drug Evaluation and Research. Also, it included famed Harvard placebo effect researcher Irving Kirsch, as well as researchers from Johns Hopkins and the Cleveland Clinic.

The study was a participant-level analysis of the double-blind, placebo-controlled trials of antidepressants to treat depression submitted to the FDA. The data included 242 studies that were conducted between 1979 to 2016—a total of 73,388 participants.

The researchers accounted for age, sex, and baseline severity of depression in their analysis.

Consistent with the previous research, they found the usual, minimal, less-than-two-point difference between the drug and the placebo effect, on average, across all 73,388 participants.

“The difference between drug and placebo was 1.75 points,” they write.

(This is the average for adults. For children and adolescents, the average difference between drug and placebo was less than 1 point, at 0.71.)

For both the drug and the placebo group, adults were more likely to get better if they were younger and had worse symptoms at the start of the trial.

However, because this was an individual patient-level analysis, the researchers were also able to break down the statistics further. They found that those who took the drug were a little more likely to experience a large improvement than those in the placebo group.

They write, “About 15% of participants have a substantial antidepressant effect beyond a placebo effect in clinical trials.”

Essentially, the researchers suggest that there is a small group of people for whom the placebo response doesn’t really happen and for whom the antidepressant drugs reduce symptoms.

More Information

The drug and placebo groups both had extremely high rates of symptom improvement: 84.4% of the placebo group found their depression symptoms improved, while 88.5% of the drug group improved. However, in many cases, this “improvement” was small.

More important is the number of people who experienced a large improvement. This improvement is more likely to be clinically relevant. The researchers found that those taking the drug were more likely to experience this level of improvement—24.5% of the antidepressant group experienced large improvement, versus 9.6% of the placebo group.

Based on these numbers, there seems to be a small group—about 15% of people—who experience a large response to the drug who would not otherwise improve to this level.

Unfortunately, the researchers found no way to predict who, exactly, is in this 15%. They write that if everyone with a depression diagnosis is given an antidepressant, about seven people need to be given the drug (and thus be exposed to the harmful effects with no benefit) before one person benefits.

“Further research is needed to identify the subset of patients who are likely to require antidepressants for substantial improvement,” they write.
“The potential for substantial benefit must be weighed against the risks associated with the use of antidepressants, as well as consideration of the risks associated with other treatments that have shown similar benefits.”

Common risks of antidepressants include weight gain, sexual dysfunction, and emotional numbing, and the drugs are challenging to discontinue once started.

Explanations for the Findings

Despite some newer arguments that the placebo effect has been increasing over time—thus making new drugs look worse—the researchers found that the placebo effect has remained stable since the 1980s.

“Depression symptoms” measured on common depression questionnaires include bodily responses like sleep and eating, and the drugs’ sedative and appetite effects could account for some of this improvement.

Another explanation is that some people receive an enhanced placebo effect because they can tell, from the side effects, that they are in the active drug group (breaking the “blind” of the study).

Clinical trials also usually hand-pick their participants, searching for those with no other conditions and who are not suicidal. This makes them very different from the individuals most often treated with the drugs in real life.

Indeed, in a study this year, other researchers found that response to treatment is much lower in real life. For example, in a study where over a thousand people with depression were treated with antidepressant drugs—more than half on multiple drugs—as well as therapy and hospitalization, less than a quarter responded to treatment.

In another paper, those same researchers also found that those with more severe depression, those with comorbid anxiety, and those who were suicidal were least likely to benefit from the drugs.

 

 

****

Stone, M. B., Yaseen, Z. S., Miller, B. J., Richardville, K., Kalaria, S. N., & Kirsch, I. (2022). Response to acute monotherapy for major depressive disorder in randomized, placebo-controlled trials submitted to the US Food and Drug Administration: individual participant data analysis. BMJ, 378(e067606). http://dx.doi.org/10.1136/ bmj-2021-067606 (link)

24 COMMENTS

  1. Actually a 15% response is not that bad when compared to the benefits of some other drugs. For example, statins, which are widely prescribed, will not extent lifespan or prevent cardiovascular disease for 98% of those without preexisting heart disease. Insulin given to those with type 2 diabetes will reduce blood sugar levels, but a major study showed marginal non statistically significant reductions in heart attacks, strokes, and maybe microvascular disease among those given insulin after 10 years of treatment.

    The point is Americans take numerous drugs whose benefits are minimal or nonexistent. Why this is the case is a good question.

  2. Even these sorts of reviews are misguided and superficial.

    No matter what “clinical scale” is used – a down-stream perception, not anything objective – there is little meaningful insight that can be gleaned from this sort of evaluation. And of course, we have worked backwards an in circles, asserting etiologies for a vague disease entity based on (flawed) evaluations and then asserting their validity and necessity from the presumed etiology. (Oh right, it’s not “chemical imbalances” anymore, but other chemical imbalance stand-ins like “it’s all inflammation” or “it’s more than one chemical!”.) At the end of the day, the thing being targeted is mental state and perception and any attempt to deny that is delusion and misdirection. It doesn’t matter if there is or is not some day a discovery of a biological factor; there are always biological factors, depending on how silly we want to be, like having enough oxygen to be alive. If someone does not want chemical interference with their mind, their consciousness, then that has to be respected. And if they are fed misinformation or coerced into compliance with a treatment or narrative, then there is certainly no consent. It’s just violence. Period. All this business about clinical scales or “chemicals” and the rest of it is distracting from the crux of the issue.

    It’s invalid to say “here we (think) we observe a behavior or feeling, and so we say it is a disease defined as such” while also saying “if the thing of concern is explained by other means or is otherwise gone, the disease must remain (and be of the form we have asserted) because it is necessary to justify what we do”.

    There are many reasons a given clinical scale or individual or clinician might report one way or another for these reviews. Some of them can be controlled for and others cannot and still others will remain unknown. It’s especially difficult with this sort of thing because a person who is expecting a medical fix may see any psychological effect – which they would not see with a sugar pill or no pill – as “working”. It’s more difficult to have a simple placebo experiment. Broad meta reviews are worse still, since they will examine data that includes people “smiling and nodding” in order to be seen as compliant (probably more so for other diagnoses more likely to involve involuntary commitment) or clinicians “just saying what they’re doing is good and works” – I’ve seen that many times. I’ve heard psychiatrists blatantly say they’ll “just say things” in records or in court to get a desired order or “cover my rear”. And then we have iatrogenic effects, either from the drugs or from the confusion, confinement, loss of rights, and other general treatment and circumstances. Objectivity from an observer’s perspective is extremely difficult and just about always biased toward presumed or convenient or desired narratives. I’ve never seen any study or review pay more than lip service – if any consideration at all – to these problems.

    I’m not even saying no one should ever use these substances. But no one should ever have them imposed for any reason, especially not disingenuous excuses based on bogus science. But the problem for psychiatrists is two-fold: First, that means they have to admit things are more complicated and they don’t have the expertise to navigate their jobs without relying on the above; Second, they may be relegated to a seldom-employed therapeutic role, rather than the center of all things, harming job prospects. Overhanging everything is the prospect of damage to reputations or even liability for past actions. Most people don’t want to attack the foundations and institutions they are a part of and their own legitimacy. It’s Ironic because that defensiveness is what makes them illegitimate. There is probably a role for real psychiatrists and proper psychiatry, people like Joan Moncrieff, but instead those genuine figures are demonized.

    • Seriously. As much as I generally want to say “you do you” and respect “personal truths” or the possibility that someone is ‘objectively’ deriving benefit… What percentage of people say booze or pot has are an objective biomedical treatment in some form? There may be some ways to contrive an argument, moreso for pot, but then there always is. Same goes for LSD or psybicilin, which have gotten a lot of attention lately.

      And still, more power to you. But in terms of a biomedical narrative, if all these things have a place or are interchangeable, the biomedical justification falls apart. It’s difficult to say, “you have a biomedical need for this” based on an assumed etiology in turn rooted in a narrative on efficacy (which is scientifically bogus, but…), but be unable to juxtapose that narrative with analogous but incompatible ones for totally unrelated psychoactive substances as well as psychotherapies, other modalities, or indeed nothing.

      • The idea that ingesting substances to alter consciousness is neither new nor particularly scientific. Pretending that a response to a drug means anything about cause or constitutes “treatment” is the central sophistry used by psychiatry to “explain” its outrageous corrupt actions. I have NO trouble with people ingesting substances to alter their consciousness, and I’m sure in some cases something good can come out of it. I have BIG trouble with calling such ingestions “medical treatment” and charging people and governments ridiculous amounts of money to run an uncontrolled experiment on our brains.

      • Thanks for introducing me to the new idiom “you do you” which seems to want to hang out with “it is what it is”. As a quick aside you dont do you as much as probably hoped. So much is mimicry. mirroring, copying, conforming, following trends, fitting in… others do you is more honest much of the time.

        That’s one reason that ssri drug use is so popular. Others do you. So pop the pills and fall in line with the social norn.

        You refuse the drugs and accompanying narratives? Then we’ll properly get to work on doing you. We’ll other you as a conspiracy nut, antipsychiatry, antiscience. Anti-us!

        It’s true that any person doing drugs that alter consciousness can claim the ritual as personally beneficial. However, it gets said that the loner drug user is on a path to destruction. That it is healthier to drug with others. In bars, private members clubs, and crack dens.

        GPs tried to resist doing the dirty work of psychiatry but after the closure of the asylums and the movement of the mad back to society, there was first a demand from society that the mad were drugged, for health and safety and jason vorhees reasons, and then later bizarely from pressure groups comprised of the mad themselves. And so gradually over many decades GPs became relucantly and then feverishly frontline drug pushers.

        Ask any so-called healer how much ego boost they get from exploiting the placebo effect and most will lie and claim none.

        When reallythe placebo effect is the name of the game and the ego boost is the addicting sauce and it is most intense when others are doing you and you want it so bad you’re willing to see any face looking back at you in the mirror so long as it isn’t frowning.

        When I first resisted SSRIs the GPs were still cynical, openly, about psychiatry as a legitimate medicine and would at times steer lost souls away. So back then it was easier because smart and respectable and kind and honest GPs would back you up, even if only privately. Fast forward to now and everyone is frightened of career ending honesty, or more often, has fallen under the social spell of bullshit.

        Never forget that patient pressure played a big role in sewing up mouths against speaking openly and honestly about SSRIs.

        It is a peculiar reality that others make for us.

  3. My doctor teased me unmercifully when I raved about my incredible improvement less than 2 weeks after starting Prozac, decades ago. “You’re just experiencing a placebo effect. It can’t work so soon”. Yeah. Freaking Hypomania. A couple weeks later, suddenly can’t stomach the dose. Eventually, after over 2 months of exhausting bliss. Crash like a long cocaine binge. I’m still triggered to score some. Like an addict. I know it’ll kill me. Getting by on mood stabilizers and Buspirone. Bipolar2 and induced mania is a hard way to go.

  4. I would guess this landmark study will not be welcome with the mainstream psychiatric establishment in the U.S. I would also suspect that there will be opposition papers published to discredit it in the near future. This study in unlikely, unfortunately, to change the praticice habits of providers anytime soon. Psychiatrists are not trained to find the 15% who are most likely to benefit by using the most important diagnostic validators, course (i.e. onset of symptoms and peridocity over time) and detailed family history including blood relatives who may not have been treated). These validators are not used in the DSM but are in other specialities. The absence of blood tests in the field is a long and sad story. There is no other medical speciality that has not come up with at least one new blood test in the past 40-50 years and often newer ones replace older ones.
    https://medicalmodelredux.com/

  5. Excellent article….I believe now is the time to flood media with facts (again) regarding the specious messaging, self-defined expertise & presumed legitimacy of the dangerous Pharma/Psych industry.

    This is the first, thrilling time I have felt seismic movement in the fight…there’s a palpable ripple in the force field of conventional marketing, social, & the confluence of post-covid Gen Z curiosity towards options to addressing mental distress.

    Michael Jeffrey Jordan said “You either win or you learn”.
    Mr. Whitaker and so many other fighters are on the edge of winning.
    The industry’s startling “We never said ‘chemical-imbalance’ pushback is a pure gift.

    I didn’t think I’d live to see this. I wish Paula Caplan was here for it.
    Carpe omnia.

  6. It is possible that the 15% of responders may have recurrent severe depressive episodes consistent with the classic disease entity of Manic-Depressive Illness of which Bipolar is a smaller subset. MDI was always defined as recurrent depressive episodes with or without mania. MDI was confirmed in over 40 years of family studies but was not acknowledged by DSM-5. The 85% of non-responders may have chronic low grade depression, previously called “neurotic depresion” or mixed depressive symptoms with subsyndromal mania (i.e. mild mania) or mood temperaments of hyperthymia or cyclothymia for which antidepressants do not help. There is a historical opposition by the makers of DSM against research for biological markers.

  7. Feargus, Psychiatry must not exist.

    Psychology and Psychoanalysis are things we should not allow our government to license. No law against these, but not licensed by the government. And then there needs to be court supervision when these on done on minors. Otherwise the therapist / analyst becomes an accomplice of the parents.

    Joshua

  8. Peter, correct me if I am wrong here. I take it 15% of the drug treated had a greater reduction of depressive symptoms than those who were on placebo. However, did the investigators compare the top 15% of the antidepressant group (who experienced greater reduction of symptoms than the bottom 85% of the same group) with the top 15% of the placebo group who had greater symptom reduction than their placebo peers? It would also be good to know the distribution of symptom reduction in the antidepressant and placebo groups.

  9. To be a placebo effect needs a person to almost hypnotically have faith that what they are drinking or nibbling or glugging or quaffing or sipping IS miraculously going to be effective very quickly. Most placebo “cures” require that fudging of the ordinary truth. The ordinary truth seldom inspires a desperate person with the high hopes needed to keep existing if suicidal, a high hope needed by the placebo healing to work. Perhaps this innate understanding of how placebo upliftment is borne aloft on wings of almost shamanistic fantasy is why for so many years ordinary doctors sold antidepressants as more effective than they later turned out to be.

    A new breed of pill is coming that is more in alighnment with psychaedelic medicine. Will it also be pushed with placebo faith fervour?

    Inspiring hope in the hopeless person is a form of nectar in itself that healing types of people can become addicted to. Not necessarily doctors fall into this performance of a magician trick with a pill, but all of us do! It is innately human to want to clump together like chimps and caress the frightened chimp into feeling much better again.

    I blame nobody for having the human trait of yearning to heal the distressed with a goovy bright idea. It is instinctual. The problem arises when the addictive nectar filled idea of healing becomes so important to a healer that there grows a silencing of the wounded of that idea.

LEAVE A REPLY