Telling the antidepressant story
What's new about the Kirsch et al study?1 It's already established that the medical literature represents a selective and biased subset of study outcomes.2 NICE has also already agreed that the size of the difference favouring SSRIs over placebo as measured by the Hamilton Depression Rating Scale is unlikely to be clinically significant.3 And yet the media have widely reported that the new study has found that antidepressants do not work.
Doctors do not generally tell patients about the small effect size and substantial non-response rate of antidepressants for fear of undermining the effectiveness of medication. Discontinuing antidepressants should be expected to cause a nocebo effect because the placebo response to treatment is so large.
This study raises again the question of whether the small difference between antidepressant and placebo in clinical trials could be totally explained by placebo amplification due to methodological bias in clinical trials, such as unblinding.4 If this were the case, antidepressants are pure placebo. Ignoring this issue reinforces the belief that antidepressants work, to the extent that it has been regarded as malpractice not to prescribe antidepressants in major depression. The debate about whether antidepressants work needs to be legitimised.
- 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;5(2):e45 doi: 10.1371/journal.pmed.0050045.[CrossRef][Medline]
- Chan A-W and Altman DG. Identifying outcome reporting bias in randomised trials on PubMed: Review of publications and survey of authors. BMJ, doi:10.1136/bmj.38356.424606.8F (published 28 January 2005)
- Double DB Lack of clinical significance of antidepressants. bmj.bmjjournals.com/cgi/eletters/330/7486/267#95940, 8 February 2005 [Full text]
- Moncrieff J & Double DB. Double blind random bluff.Mental Health Today 2003; Nov: 24-26
Competing interests: None declared
Competing interests: No competing interests