Antidepressants and the placebo effect

When I was in middle school, I went to Lake Tahoe with my best friend on a ski trip. I had only been skiing a few times before, and even though my friend was only twelve, she was already a talented and experienced skier.

She was patient with me, but I became increasingly frustrated as I had trouble going down even the most basic runs without falling. At one point, after falling several times in a row, I gave up, and plopped down onto the snow, completely discouraged.

My friend maneuvered over to me and sat down. She explained that she knew a technique that would help my skis work better. She reassured me that once she fixed my skis, I wouldn’t have any more problems.

Then, she took a handful of snow and carefully rubbed it on the bottom of my skis.  ”This will help you get better traction,” she explained. She worked meticulously, as if she were engaged in a very important task. As I sat there, I began to feel relieved. She finished, and I got up, encouraged. We both made our way down the mountain, with minimal falling on my part.

The powerful placebo effect

Of course, it makes absolutely no sense that rubbing snow on the bottom of skis would help them work better (I must have been kind of slow in my middle school days), but what matters is that I believed it would. What matters is that I trusted my friend and felt reassured by her attention. What matters is that her actions helped transform my discouragement into motivation to get down the mountain.

Antidepressants and the placebo effect

For years, psychiatrists prescribed antidepressants only for the most severely depressed patients, as the early antidepressants (such as tricyclic antidepressants, or TCAs, and monoamine oxidase inhibitors, or MAOIs) had many side effects. Then Prozac, the first selective seratonin reuptake inhibitor (SSRI), came out in the late 1980s, and was much safer and better tolerated than the older-generation antidepressants. Prescriptions for antidepressant medications skyrocketed.

For a while, the efficacy of antidepressants was not questioned. Doctors would see their patients get better, with few side effects, and keep writing scripts. But over time, as more studies were done, it became evident that antidepressants were not much more effective than the sugar pills they were compared with in clinical trials.

In 2009, the psychologist and researcher Irving Kirsch published a book titled The Emperor’s New Drugs: Exploding the Antidepressant Myth. He cited clinical evidence suggesting that the most widely-prescribed drugs in psychiatry were not as effective as previously believed. In January 2010, the cover of Newsweek read, “Do Antidepressants Work?” The feature article had the sensationalist title, “The Depressing News About Antidepressants: Studies Suggest That the Popular Drugs Are No More Effective Than a Placebo. In Fact, They May Be Worse.”

Summaries of the evidence demonstrate that antidepressants are effective for depression about 31-70% of the time, while placebos are effective 12-50% of the time, for an average antidepressant-placebo difference of 20%. So, while antidepressants appear more effective than placebo, only a fraction of the benefit we see in patients is likely from a direct neurobiological impact of the medication itself.

One meta-analysis (a statistical summary of the evidence) argues that 25% of the benefit we see from antidepressants is due to a direct impact of the medication, while 25% is due to spontaneous remission (people who would have gotten better anyway), and 50% is due to the “expectation of benefit,”—also known as the placebo effect.

Interestingly, the antidepressant-placebo difference seems to be decreasing over time—not because antidepressants are becoming less effective, but because placebos are becoming more effective, at a rate of 7% a decade. How on earth could placebos be becoming more effective, when, by definition, they do not have any therapeutic value?

Researchers have suggested many possible explanations for this phenomenon, including that more people with mild-to-moderate depression are being including in clinical trials than in the past. The evidence suggests that antidepressants are quite effective for people with severe depression, but less effective, or not effective at all (when compared with placebo), for people with mild-to-moderate depression.

Why is ‘placebo’ a dirty word?

Okay, so the evidence shows that antidepressants are not as effective as we’d like to think. As psychiatrists, we need to be honest that more of what we do is “the placebo effect” than we’d like to believe.

But—dare I say it—so what?

I prescribe antidepressants because I see people get better with them. Yes—researchers need to pay attention to the placebo effect in randomized controlled trials, but do clinicians? Who cares what is making patients feel better if they’re feeling better? Who cares what is alleviating their suffering, if they are getting relief? Why is placebo such a dirty word?

Did you know that some soldiers in World War II who were given saline injections instead of morphine(because of depleted morphine stocks) experienced relief from their pain? Did you know that a placebo can lead to airway dilation in asthma, when the person is told they received a bronchodilator, like albuterol? Did you know that in one study, 50% of people with osteoarthritis reported decreased knee pain with a high-tech surgery, while 50% experienced relief with… sham surgery?

Did you know that there are case reports of women given ipecac, known to induce vomiting, who reported a relief in their nausea when they were told the ipecac was an anti-nausea medication? Did you know that physical symptoms of hypoglycemia (sweating, increased heart rate, tremor) have been induced by placebowhen a patient is told they are getting insulin? Did you know a flavored drink can suppress a person’s immune system (as measured by biochemical tests), when in the past that flavored drink was paired with the immunosuppressive drug cyclosporine?

Oh, and here’s my favorite: Did you know that despite the evidence that acupuncture is no more effective than sham acupuncture, in China—where people believe strongly in the power of acupuncture—it has been used in lieu of traditional anesthesia during open-heart surgery? I didn’t believe it until I read the study myself.

In psychiatry, we know that placebos not only work clinically, but lead to similar functional brain changes as antidepressants.

We also know that the placebo effect has an equal opposite: the nocebo effect. Just as the suggestion of positive benefit can help, the suggestion of negative outcome can harm. In antidepressant clinical trials, for example, about 25% of people report side effects from the placebo—side effects that match the ones they were told could happen with the active medication during the informed consent process.

How to harness the placebo effect

Yes, there are ethical considerations when it comes to placebos. Clinicians are obligated to give informed consent, and can not ethically lie to patients about treatments they are getting. No, we should not offer treatments that have no scientific benefit when they might cause harm. I’m not saying we should go around giving people ipecac for nausea or performing fake surgery for knee pain.

Clearly, though, the placebo effect is much more than “in our heads.” Hope and expectation can cause biological and functional changes in our brains and bodies. Shouldn’t physicians be interested in this phenomenon? Shouldn’t we pay attention to it? If it has the potential to heal, shouldn’t we harness it?

And more importantly, why do we argue ad nauseam about evidence-based medicine (should I give my patient an SSRI or an SNRI? Cognitive behavioral therapy or psychodynamic psychotherapy?) when the evidence suggests that the specific intervention we choose is a fraction as important as the way we deliver it?

What is lumped together as “the placebo effect” is probably a collection of multiple factors. It is the therapeutic alliance a person has with their healthcare provider, it is their belief in the power of the treatment, it is our society’s cultural expectations about sickness and health. It is the personality and style of the physician and their ability to demonstrate compassion and instill confidence. It is reassurance, it is ritual between doctor and patient, it is ceremony.

The evidence about placebos tells us that we clinicians need to do a lot more than just write scripts to serve our patients and our communities. We need to:

  • Respect individual beliefs and treatment preferences.
  • Study the placebo effect with the same scientific rigor with which we study the medications we use.
  • Listen empathetically and communicate compassionately.
  • Share our hope and positive expectations with patients.
  • Not attribute all of a patient’s improvement to medication, and instead reinforce the patient’s self-healing mechanisms.
  • Be honest about the possible side effects of medications, but emphasize the potential benefits.
  • Consider carefully the cultural message that pharmaceutical companies are spreading in their pervasive consumer advertising (“Take our brand-name medication and you’ll be running through a field of daisies in no time!”).
  • Consider that unreputable sources on internet, unfortunately, might be creating a nocebo effect.
  • Strive to be encouraging and supportive of our patients. We clinicians should be placebos for our patients, not nocebos.

Yes, I use drugs, but this is only one tool in my arsenal. I use words, I use hope, and I use heart. I care about science, but I also care about our humanity. Isn’t that what medicine is all about, anyway? When my patients tell me that antidepressants help, who am I to tell them they’re wrong?

Elana Miller is a psychiatrist who blogs at Zen Psychiatry.

Comments are moderated before they are published. Please read the comment policy.

  • ninguem

    “… China—where people believe strongly in the power of acupuncture—it has been used in lieu of traditional anesthesia during open-heart surgery? I didn’t believe it until I read the study myself….”

    Now wait a minute. It is very misleading to say that acupuncture is used “in lieu of general anesthesia”, fentanyl and midazolam were still used, and the “very low dose” alluded to in the abstract was not specified.

    All we have is the abstract. Unknown from that, how much drug was actually given, whether local anesthetic was used as well. The language of the abstract refers to it as “combined acupuncture-medicine anesthesia”

    Even the authors themselves are not claiming that acupuncture “replaces” general anesthesia. It is part of an overall anesthetic plan that still includes conventional anesthetic agents. That trope of doing surgery with acupuncture alone, goes back to the Nixon in China days. When you looked more closely, it was clear they were using acupuncture AND conventional anestheic agents, not acupuncture INSTEAD OF conventional anesthetic agents.

    • Elana Miller

      Hi Ninguem, thanks for your comment. The original article, not the abstract, details the information you’re asking about. If you would like a copy, email me through my blog and I’ll send it to you, or you can purchase it through the journal’s website. The acupuncture group received 13% of the dose of fentanyl as the control group, and were not intubated, while the control group was. Equal doses of morphine were used preoperatively. They did use lidocaine in the acupuncture group prior to median sternotomy. Yes, acupuncture and anesthetic agents were combined – I did not mean to imply that no anesthesia was used, only that this was an alternative to a traditional anesthesia procedure done typically done during open heart surgery. I’m no anesthesiologist, but I was impressed that they were able to control pain with a fraction of the dose of narcotic medication and have better outcomes.

  • LeoHolmMD

    There is a strong anti-placebo effect coming from pharmaceutical companies and many psychiatric professionals trying to convince the planet that they are depressed and need treatment. It makes sense that a placebo would work to reduce this effect. In the meantime, suicide rates are higher, and trolling for depression seems to be on the rise. I think you have a reasonable approach to the whole issue.

    • Elana Miller

      Hi Leo, thanks for your comment and for sharing your thoughts. I think it’s definitely worth considering how our cultural view of depression as always being an illness outside of our control could be worsening outcomes. Depression can be very biologically based, but sometimes people experiencing normal life stress are made to feel they have a psychiatric illness when they don’t.

  • candy clouston

    As someone whose upbeat psychiatrists relied on placebo effects and minimized the potential harm of various drugs, I now enjoy hemorrhoids (from constipation related to Anafranil), a hiatal hernia (from straining) that led to reflux, weight gain, IBS that responds to med changes, a history of seratonin syndrome, with a past escalation to a mixed bipolar episode concurrent with a change in medication. If I knew then what I know now, I would not have been so patient in tolerating a long list of side effects with little improvement in symptoms. I have genuinely enjoyed stable moods for the last nine months with no medication but weekly supportive psychotherapy (which I have had for over 20 years) and lifestyle changes that are difficult to effect from a depressed state, but not extreme (e.g., more social support, less processed food, regular walking), and that support general good health. Truly informed consent may be at odds with the placebo effect, but I think it is a moral imperative.

    • MabelMabel

      Good for you for getting psychotherapy! I am glad you have found it helpful.
      When I was in graduate school for psychology, a very wise supervisor told me to wait for 3 sessions to see if my new client, who was severely depressed, improved without the need for medications. To my surprise, and relief, the client rallied quickly through our therapeutic relationship and when I reflected to her that her feelings regarding her circumstances were normal. We continued therapy for about a year, during which time she was able to psychologically process a history of terrible childhood trauma, which had been causing lifelong episodes of depression. Since that time in grad school, hundreds of clients later, that word “normal” is still a great healer when clients think they are “crazy” for reacting normally to depressing situations by feeling depressed. (I would, however, never dissuade someone from getting medications if they wanted to, because the placebo effect is powerful.)

    • Elana Miller

      Hi Candy, thanks for your comment, I agree that informed consent is crucial, and it’s not helpful to give medications that have no benefit and can cause harm. There is good evidence that antidepressants are helpful for people who are more severely depressed, so it becomes an issue of weighing potential benefit with potential side effects. I do think it is a problem that medications are too quickly prescribed for people with mild depression. How to weigh informed consent with being encouraging and optimistic is a tough issue.

  • CortneyM

    Fun story – I was two weeks overdue and my OB/GYN ordered an induction. I was dilated 2 cm and hadn’t had any contractions whatsoever at 42 weeks. I get to the hospital, the nurse puts the IV in and hooks me up to some monitors, and my husband and I wander around the halls and watch TV and twiddle our thumbs. After two hours, I looked at the screen and saw that the sensors had been picking up some small contractions. I hadn’t noticed anything, but I wasn’t going to complain. The nurse came in and said, “Oh, I forgot to hook up the Pitocin. Sorry about that.” Then the contractions stopped and didn’t start again for another five hours. Not sure if it was the magic dose of Pitocin, actually seeing the number 30 on the Pitocin pump, a mental daze from the Stadol, or hearing that the doctor had already performed four c-sections that day (Hospital does a lot of high-risk pregnancies), but my lovely daughter was born six hours later in perfect health and without scapel involvement.

    Since that experience, I haven’t been entirely convinced or unconvinced of any medical intervention being effective or ineffective. Motrin sure isn’t as as effective as it used to be, let me tell you…

  • Alan Byron

    After graduation in 1964 I started work as a psychiatrist in a large hospital. There I used Unilateral ECT for severely depressed patients with a real risk of death from suicide. The results plus supportive outpatients were highly satisfactory. Since then as a GP I saw antidepressants – at first the Tricyclics and the serotonin types – used for a gradual downscale of severity until I saw them used for social problems, family difficulties, boyfriend problems, financial difficulties and as a panacea for all social problems that should have been addressed by personal assertive action. The result was a plethora of side effects and the only benefit I could see was that my patients saw me regularly. Eventually I found that a very small dose of amitryptilline at night as a ‘sleeping tablet’ plus encouragement was enough to see people through temporary difficulties. I once read in the Psychiatric Journal that the ‘Prokaletic Challenge’ to be the most useful for people with emotion/depressive problems of a social nature. That is to simply say, sorrowfully, after each intervention failed to benefit, ‘I am so sorry, I cannot find anything else to help you; is now up to you to find a way.’

  • rbthe4th2

    Thank you for the post. The bulleted list was quite enlightening. As someone who has had docs who simply passed off things as mental vs. doing a real H&P, hands on exam, and blowing off worsening symptoms, medical research that they blew off, the above list do a lot towards reminding docs that not everything is mental.
    It actually helps when a doctor investigates systematically symptoms based on common sense and EBM, rather than just giving the impression they don’t want to deal with solving a problem and tossing ‘psych’ pills when fixing the physical problem is what is needed.

Most Popular