Debunking Anti-PsychiatrySkepticism

Why Jerry Coyne is Still Wrong about Antidepressants

A few months ago, Jerry Coyne, Professor in the Department of Ecology and Evolution at the University of Chicago and an staunch supporter of evolution against creationists, made a series of remarkably flawed claims about medical psychiatry in general and antidepressants in particular. He did this after reading a couple of book reviews on a few controversial books on psychiatry and asserted that medical psychiatry was a scam. Needless to say, I confronted his claims in Why Jerry Coyne is Wrong about Medical Psychiatry and shown that Prof. Coyne made several glaring errors: he incorrectly characterized the mainstream view on the causes of depression, he claimed that the effectiveness of a drug was not evidence for the underlying model (thus implicitly agreeing with HIV/AIDS denialists that the effectiveness of antiretroviral treatment is not evidence that HIV causes AIDS), he did not understand the difference between genetic mapping and estimations of heritability, he advocated Big Pharma conspiracy theories, incorrectly claimed (based on Kirsch flawed studies) that antidepressants are no better than placebo and contradicted himself by claiming that mental disorders were not caused by chemical factors in the brain while at the same time claiming that antidepressants cause psychopathology without any evidence.

After this, I stopped regularly visiting his blog, so it is only now that I noticed that he wrote a follow-up article called Peter Kramer defends antidepressants. In it, Prof. Coyne repeats many of the same flawed arguments as before and it reads like an advertisement of Kirsch book on antidepressants. It is now clearer than ever that Prof. Coyne has gone of the deep end with regards to this topic. It is clear that his pseudoskepticism is deepening and that is why I have decided to write another criticism. There will necessarily be some repeats of content that I discussed in previous entries, but will try to keep it to a minimum.

Let’s get started, shall we?

1. Active versus inert placebos

Prof. Coyne’s rejection of double-blind, placebo controlled studies (the gold standard of medical research) that show that antidepressants are more effective than placebo is based on the notion that the placebos that were used were not active placebos (i.e. placebos that does give pharmacological side-effects) but inert placebos (i.e. placebos that do not give any pharmacological side-effects and therefore, supposedly, do not have side-effects at all). This means, according to Prof. Coyne, that people on the antidepressants will know that they are on the active drug because of these supposedly side-effects, thus increasing the placebo effect.

It is interesting to note that Prof. Coyne does not make this same objection to studies that he thinks show that antidepressants are no more effective than placebos (or to studies of other drugs with side-effects), demonstrating that he applies a different and much higher level of skepticism (and demands a much higher burden of evidence) to things that contradicts his position than to things that confirm it, which is a hallmark of pseudoskepticism.

The claim is based on many dubious premises such as (1) the side-effects of antidepressants are severe and (2) inert placebo (like sugar pills), in fact, does not have side effects. Both of these are seriously flawed and will be discussed below. However, there is another glaring problem with his claim that must be investigated first.

Older antidepressants, such as MAO-inhibitors or tricyclic antidepressants (TCAs), are generally considered to be less effective than newer antidepressants such as SSRIs or SNRIs and have more severe side-effects (new drugs are generally developed to increase effectiveness and reduce side-effects). How can this be? If the newer drugs have less side effects, then the additional placebo effects from knowing that you experience side-effects should be less not more, thus giving the appearance of less effectiveness compared with inert placebo, not more.

2. The severity of side-effects of antidepressants

No one is denying that antidepressants have side-effects, but are these side-effects generally severe? I discussed the side-effects of antidepressants in A Critical Examination of Stefan Molyneux’s Claims about Antidepressants, so I will settle with a shorter recap.

According to the Mayoclinic (2011) article on clinical depression, side-effects of SSRIs include decreased sexual desire, digestive problems, jitteriness, restlessness, headache and insomnia. This does not, of course, mean that every person taking SSRIs will experience all, or even most, of the side-effects. It just means that these are experiences a lot of patients have in association with taking SSRIs. Of course, patients are not statisticians and may be subject to correlational and post hoc fallacies, which we have seen in the vaccine/autism debacle. These side-effect must be compared with the natural frequency of digestive problems, restlessness and headaches, which are reasonably quite high.

While it is true that the FDA requires that all antidepressant medications carry black box warnings that notes that it is possible that, in some cases in certain subgroups, patients may have an increase in suicidal thoughts and behaviors when taking antidepressants. However, the increase that media reported was just from 2 to 4% and this may have been due to increase in reports (Hall, 2009). Also, after the prescription rate fell by 18-20%, suicides increased by 18% (Hall, 2009). The Olfson et al. (2006) study that appeared to show an increase in suicidal thoughts and behaviors was problematic, since it made the incorrect assumption that the two groups had the same risk for suicide, whereas it was likely that the group treated with antidepressants had more severely depressed patients and thus a higher risk for suicide. Furthermore, Bridge et. al. (2005) showed that suicidal behavior and thoughts in antidepressant tests were similar to psychotherapy trails and Simon et. al. (2006) showed that suicide rates before starting antidepressant treatment were higher and that this declined progressively after starting medication.

This means that the supposed side-effects of antidepressants in this area are greatly exaggerated and that some levels of suicidal thoughts and behavior may be tolerated while on antidepressants if they are significantly lower on medical treatment than the were before.

3. Do inert placebos have side-effects?

An inert placebo is a placebo that has no pharmacological side-effects. Does this mean that it has no side-effects at all? Not at all. When you take part in a double-blind placebo-controlled study, you are not told whether you are given the drug or the placebo. This means that there will most likely be expectancy effects from taking the placebo. This is entirely uncontroversial. However, just like there will be expectancy effects that improves the condition, so to will patients experience side-effects as if they were on the real drug (another expectancy effect). This is called the nocebo effect. Let us look at a few papers.

Amanzio et. al. (2009) looked at all of the placebo-controlled trials conducted on a specific migraine drug. They revealed that the placebo group who had gotten sugar pills also commonly experienced side-effects of whatever drug they thought they may have eaten e.g. memory difficulty and anorexia from placebo in a trial investigating anticonvulsants. Levine et. al. (2006) showed that if you give patients a placebo sugar pill and tell them it will make them more nauseous after the experimental protocol they not only get more nauseous, but also show a higher prevalence of gastric tachyarrhythmia. Ernst et. al. (2003) are critics of homeopathy and they showed that there was no statistically significant difference in side-effects between the placebo group and the group that were given homeopathy. Both groups experiences side-effects, even though it is clear that neither a sugar pill nor homeopathy has any pharmacological effect at all. Dr. Ben Goldacre has discussed these studies and others in a very interesting article called All bow before the mighty power of the nocebo effect.

Something similar has been shown in many antidepressant trails as well. I have mentioned this study in several other entries, but it is worth repeating. Walkup et. al. 2008 compared the efficacy of sertraline (an SSRI) alone, cognitive behavioral therapy (CBT) alone, placebo, and Sertraline and CBT together and the results were: improvement with CBT alone (59.7%), sertraine alone (54.9%) were both better than placebo (23.7%) and a combination of CBT and sertraine (80.7%) was the best option. Side effects where roughly equal in the group recieving the SSRIs treatment and the group receiving placebo. This study is not were, but it independently converges with other such studies showing that SSRIs are by and large effective compared with placebo. The critical thing to note here is that the side-effects were more or less the same in both the SSRI group and the placebo group.

This means that inert placebos do have side-effects! Not pharmacological side-effects of course, but side-effects due to the standard expectancy effect. More than that, these side-effects are roughly equal in the groups receiving the active drug and those who get placebo. This completely unravels and undermine Prof. Coyne’s insistence on active placebos. In fact, active placebos, may potentially mess with the results as the group on antidepressants will not experience the pharmacological effects induced by the active placebo. So there are both positive evidence and negative arguments against using active placebo. Yet Prof. Coyne rejects the studies that Kramer presents simply because they do not reach Prof. Coyne’s own artificial standards of a good study. A classical case of the sophistication effect and selective rationality.

4. Kramer’s model of depression

Prof. Coyne writes that

Kramer broaches his own theory of depression, which comes from his 2003 book Against Depression:

In 2003, in “Against Depression,” I highlighted research that suggested antidepressants influence mood only indirectly. It may be that the drugs are “permissive,” removing roadblocks to self-healing. That model might predict that in truth the drugs would be more effective in severe disorders. If antidepressants act by usefully perturbing a brain that’s “stuck,” then people who retain some natural resilience would see a lesser benefit.

I haven’t read his book, but his summary above doesn’t make much sense to me.

This is because, as I discussed in the previous critical entry, Prof. Coyne has an incorrect view on the the mainstream model for the causes and treatment of depression. Dr. Steve Novella explains: “Despite the commercials – treating depression pharmacologically is not treating something as simple as a low level of a single neurotransmitter. Measuring serotonin levels, even if practical, would likely be of no clinical value. Depression is a result of poor emotional regulation among various brain regions. Drugs are a blunt tool by which we can nudge brain function in a direction which, for some people, can change this regulation and reduce depression. It’s not really about the levels.” (Novella, 2011). This change might also help to explain why antidepressants together with psychotherapy is better than either of them alone. Antidepressants nudge brain function in a beneficial direction, which is further aided by psychotherapy.

5. Side-effect of psychotherapy?

Towards the end, Prof. Coyne makes the following very peculiar assertion:

In the end, Kramer does take a more judicious stance. He initially treats depressive patients with psychotherapy and says “I am to use drugs sparingly. They have side effects, some of them serious.” Indeed, for the side effects of psychotherapy are nil.

Are the side-effects of psychotherapy really “nil” (meaning non-existent)? Not really. Barlow (2010) discusses a few, such as intense anxiety during exposure trials, Bergin’s deterioration effect, negative effects from critical incident stress debriefing and deaths resulting from rebirthing techniques in oppositional children. To that list we can add false memory syndrome (e. g. believing that one has been a victim of satanic ritual abuse) and sexual abuse of patients by therapists. Even if some of these are rare, they certainly do not qualify as “nil”.

6. Kirsch and effect size

Prof. Coyne continue to reference Kirsch and I have discussed dissections of his studies elsewhere on this blog, especially in Harriet Hall on Kirsch and Efficacy of Antidepressants (discussing Harriet Halls criticism of Kirsch on the Science-Based Medicine blog), so I will just summarize the arguments.

  • The effect size of all drugs tested were, compared with placebo, positive (replicating findings of earlier studes).
  • None of the calculated confidence intervals overlapped zero, meaning that it is very unlikely that antidepressants tested and placebo are no different in efficacy.
  • Kirsch made an unfortunate interpretation of clinical significance, concluding that because the effect size was under the arbitrarily selected cut-off value by NICE (National Institutes for Clinical Excellence) of 0.5, Kirsch concluded that there was little or no clinical significance. While it is true that a glass that is 1/3 full is not 1/2 full, a 1/3 glass is not empty. NICE no longer uses this 0.5 effect size cut-off.
  • If Kirsch’s interpretation was reasonable, we would have to reject psychotherapy as a treatment as well antidepressants, because psychotherapy has an even lower effect size.

7. Summary and Conclusions

Prof. Coyne is delving deeper into the anti-psychiatry movement becoming increasingly more pseudoskeptical. He claims that the difference between the group getting antidepressants and placebo can be attributed to the severe side-effects of antidepressants breaking the blind. However, studies show that the placebo group experience common side-effects that the group receiving active drugs have and this is true of antidepressant trials as well. Also, using an active placebo may mess with the results as the group receiving the active placebo gets additional pharmacological effects that the group receiving the drug that is being tested do not. Moreover, the difference in placebo and antidepressants should decrease over time (making antidepressants appear less and less effective) as the newer drugs have less side-effects than the older ones, but this is not the case. Newer antidepressants are usually more effective than older, or at any rate not less effective. The severity of the side-effects of antidepressants, such as suicidal thoughts and behaviors are greatly exaggerated and not supported by the evidence. Depression is a result of poor emotional regulation among many regions of the brain, making drugs are a blunt tool to nudge brain function in a beneficial direction. The Kirsch study did find that antidepressants were statistically more effective than placebo (which replicated earlier findings). Kirsch assumed an arbitrary cut-off standard for clinical significance that is no longer used, then argued that because the glass was 1/3 full, this was less than half-full, so therefore, it must be empty. Finally, there are well-documented side-effects of psychotherapy.

8. References and Further Reading

Fournier JC, DeRubeis RJ, Hollon SD, Dimidjian S, Amsterdam JD, Shelton RC, Fawcett J. Antidepressant drug effects and depression severity: a patient-level meta-analysis. JAMA. 2010 Jan 6;303(1):47-53.

Kirsch I, Deacon BJ, Huedo-Medina TB, Scoboria A, Moore TJ, Johnson BT. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. PLoS Med. 2008 Feb;5(2):.

Erick H. Turner, M.D., Annette M. Matthews, M.D., Eftihia Linardatos, B.S., Robert A. Tell, L.C.S.W., and Robert Rosenthal, Ph.D. Selective Publication of Antidepressant Trials and Its Influence on Apparent Efficacy. N Engl J Med 2008; 358:252-260 January 17, 2008

John T. Walkup, M.D., Anne Marie Albano, Ph.D., John Piacentini, Ph.D., Boris Birmaher, M.D., Scott N. Compton, Ph.D., Joel T. Sherrill, Ph.D., Golda S. Ginsburg, Ph.D., Moira A. Rynn, M.D., James McCracken, M.D., Bruce Waslick, M.D., Satish Iyengar, Ph.D., John S. March, M.D., M.P.H., and Philip C. Kendall, Ph.D. Cognitive Behavioral Therapy, Sertraline, or a Combination in Childhood Anxiety. N Engl J Med 2008; 359:2753-2766 December 25, 2008

Hall, Harriet. (2011a). Antidepressants and Effect Size. Science-Based Medicine. Accessed 2011-07-19.

Hall, Harriet. (2011b). Angell’s Review of Psychiatry. Science-Based Medicine. Accessed 2011-08-10

Hall, Harriet. (2009). Psychiatry-Bashing. Science-Based Medicine. Accessed 2010-06-26.

Tuteur, Amy. (2010). Study shows antidepressants useless for mild to moderate depression? Not exactly. Science-Based Medicine. Accessed 2010-06-26.

Passer, M., Smith, R., Holt, N., Bremner, A., Sutherland, E., & Vliek, M. (2009). Psychology: The Science of Mind and Behavior. New York: McGraw-Hill Education. p. 865.

Seth G. Disner, Christopher G. Beevers, Emily A. P. Haigh & Aaron T. Beck. Neural mechanisms of the cognitive model of depression. Nature Reviews Neuroscience 12, 467-477 (August 2011)

Furmark T, Tillfors M, Marteinsdottir I, Fischer H, Pissiota A, Långström B, Fredrikson M. Common changes in cerebral blood flow in patients with social phobia treated with citalopram or cognitive-behavioral therapy. Arch Gen Psychiatry. 2002 May;59(5):425-33.

Mayo Clinic. (2011). Depression (major depression): Treatments and Drugs. Mayo Clinic. Accessed 2011-08-10.

Olfson M, Marcus SC, Shaffer D (2006). Antidepressant drug therapy and suicide in severely depressed children and adults. Arch Gen Psychiatry 63:865-872

Bridge JA, Barbe RP, Birmaher B, Kolko DJ, Brent DA (2005). Emergent suicidality in a clinical psychotherapy trial for adolescent depression. Am J Psychiatry 162: 2173-2175

Simon GE, Savarino J, Operskalski B, Wang PS (2006). Suicide risk during anti-depressant treatment. Am J Psychiatry 163:41-47.

Cuffe, Steven P. (2007). Suicide and SSRI Medications in Children and Adolescents: An Update. DevelopMentor. Acessed 2011-08-10.

Amanzio, M., Corazzini, L. L., Vase, L, Benedetti, F. (2009). A systematic review of adverse events in placebo groups of anti-migraine clinical trials. Pain. 146 (3) 261-269.

Levine, M. E., Stern, R. M. and Koch, K. L, (2006). The Effects of Manipulating Expectations Through Placebo and Nocebo Administration on Gastric Tachyarrhythmia and Motion-Induced Nausea. Psychosom Med May 1, 2006 68:478-486;

Goldacre, B. (2009). All bow before the mighty power of the nocebo effect. Bad Science. Accessed: 2011-08-21.

Novella, S. (2011). Comment on Antidepressants and Effect Size. Science-Based Medicine. Accessed: 2011-08-21.

Barlow, D.H. (2010). Negative effects from psychological treatments: a perspective. American Psychologist, 65, 13-19.


Debunker of pseudoscience.

7 thoughts on “Why Jerry Coyne is Still Wrong about Antidepressants

Comments are closed.

%d bloggers like this:

Hate email lists? Follow on Facebook and Twitter instead.