in Specific psychiatric disorders, Thinking about psychiatry

SSRIs in the doghouse

It’s a bad time to be an SSRI antidepressant.

Strike one:

In January a paper in the NEJM found that the evidence base for the use of these drugs was incomplete, with a large publication bias towards positive results.  When negative results had been published, it was in such a way as to give these results a positive skew.  This publication bias in itself does not mean that a drug is ineffective, but hardly instills confidence.  There have been rumblings about this sort of thing going on for a long time….

Strike two:

A paper published today in the Public Library of Science Medicine Journal entitled Initial Severity and Antidepressant Benefits: A Meta-Analysis of Data Submitted to the Food and Drug Administration collected (via the freedom of information act) what it believes to be all the trial data available before their licence was granted (published and unpublished, positive and negative) on Fluoxetine, Paroxetine, Venlafaxine and Nefazodone, and found via meta-analysis their effects for mild and moderate depression to have no clinical improvement over placebo. 

This raises a number of concerns for psychiatrists and the population at large, several million of whom were taking these medications at the last count.

  • Already not especially convincing, where does this leave the monoamine theory of depression?
  • Have pharmaceutical companies deliberately mislead the doctors and patients as to the effectiveness of their product, thereby costing health providers billions of pounds and subjecting millions to ineffective medication and unnecessary side effects?  If so, shouldn’t we take them to court or something?
  • What should we tell our patients?  There are plenty of people out there on the above medications for mild/moderate symptoms of depression.  See end of piece for a quick summary of how depression is categorized.  Doctors who prescribe antidepressants in these circumstances will be aware that it’s not the best treatment, but wouldn’t wish to give somethings that really doesn’t work at all.  If these antidepressants were working because of a placebo effect then, with all this publicity, they won’t be working now.
  • What are we going to do instead?  NICE already recommends that counselling/therapy is the most appropriate intervention in mild/moderate depression.  However waiting lists are very long and many people crave a quick fix to their problems, which in any case are more ingrained than that which is amenable to a course of cognitive behavioural therapy.

Classification of depression

Depression as an illness is somethings that’s talked about a lot these days.  The first thing to say is that the difference between ‘depression’ as an illness and simply feeling ‘gloomy’ is not qualitiative but quantitative.  There is also no test for depression, but rather a line in the sand that when (methaphroically) overstepped people say that you are depressed.  The diagnosis of depression is made on clinical grounds rather than via a test.  This is not unusual in medicine, epilepsy for instance is diagnosed in the same way.

When we as psychiatrists, (but also as lay people, but in a less formalised way) say that someone is depressed, what me mean is that they are displaying a number of symptoms that suggest to us that their main problem is of low mood.  In order to make diagnoses more consistent and also to aid in professional communication psychiatrists use diagnositic guidelines for their diagnoses.  In America these guidelines are called the Diagnostic and Statistical Manual edition IV and in Europe we mostly use the International Classification of Diseases edition 10. 

(I’m almost there.)

Although mood and hence depression exists on a dimension, i.e. there is a continum from ‘not depressed’ to ‘very depressed indeed’ with no breaks in between, we choose to draw further lines in the sand and make categorical diagnoses – this is where the mild/moderate/severe depression diagnoses come from. 

So, when the above study talks of ‘mild depression’ it’s saying that a person meets the criteria for either the DSM-IV or ICD-10 criteria for depression.  The ICD-10 criteria are:

F32.0 Mild Depressive Episode: Diagnostic Guidelines
Depressed mood, loss of interest and enjoyment, and increased fatiguability are usually regarded as the most typical symptoms of depression, and at least two of these, plus at least two of the other symptoms described above should usually be present for a definite diagnosis. None of the symptoms should be present to an intense degree. Minimum duration of the whole episode is about 2 weeks.
An individual with a mild depressive episode is usually distressed by the symptoms and has some difficulty in continuing with ordinary work and social activities, but will probably not cease to function completely.

If you’re interested in classification of depression then try this site


Bad Science on SSRIs and criticism of media response to PLOS paper

Guardian articles on PloS paper: Ann Robinson – If the drugs don’t work; Allegra Stratton – A bitter pill; Sarah Boseley Prozac, used by 40m people does not work, say scientists; Mark Lawson Something for Nothing



Write a Comment


  1. I find all of this negative press about anti-depressants to be rather irritating. I have taken duloxetine for some time now and I have no doubt at all that it works. Admittedly it has side effects, some of them more worrying than others, but it has taken the edge off of some pretty nasty symptoms just as my CPN said it would. Are these stories not mis-representing the facts… and is it not the case that what is actually being reported is inappropriate prescriptions rather than inefficient medicine?

  2. No one is saying that SSRIs don’t work, just that they don’t work any better than placebo for mild and moderate depression (nb this is better than no effect) If you’ve got a CPN I should imagine that your depression was worse than that. You’re right in thinking that SSRIs have been oversubscribed – and will continue to be so I fear.

  3. I sometimes feel depressed too, but i hate to rely on medications to feel better. It especially scars me when publications come out after these medicines have been prescribed for many years.