article is headlined to ask why we shame people for taking antidepressants. I don't think this is what those who are critical of the evidence for the effectiveness of antidepressants are doing. All that is being asked is why there is such a small difference in clinical trials between active and placebo treatment and whether this could be an artefact (eg. see previous post).
The paper that led to the Guardian article (Hieronymus et al published online 25 July 2017) argues against the theory that antidepressants outperform placebo solely or largely because of their side effects. I'm not sure if this is precisely the reason why the argument is made that antidepressants may be merely amplified placebos (see another previous post). There may be other ways in which patients can become unblinded in clinical trials. True, the evidence I summarised from Irving Kirsch's book in that post focuses on side-effects. And, it is also true that there is an expectation that side-effects may be the most common way of unblinding in clinical trials and therefore that the more side-effects that patients experience on the active drug the more they improve in clinical trials. As far as I know, the paper, Barbui et al, referred to in my previous post was never published and I'm not sure why.
Furthermore before Hieronymus et al (2017), there was a BJPsych paper (Barth et al 2016) that also did not find an association between adverse events and efficacy. These authors concluded, "Our results do not support, but also do not unequivocally disprove, the hypothesis that adverse events lead to an overestimation of the effect of SSRIs over placebo". I would agree with this conclusion. And, Hieronymus et al (2017) also admit that their "study does not allow any firm conclusions".
In fact, for citalopram, they found that the rating on the Hamilton Depression Rating Scale (HDRS) for those that reported adverse events was not statistically significant from placebo. This may be why they focus in the paper on the depression rating within HDRS, which did find a significant difference. For the paroxetine trials, the association with adverse events goes in the other direction, although, for some reason which I don't understand, in the paper the authors suggest this observation for paroxetine (do they mean citalopram?) "could be interpreted as support for side effects exerting some impact on the response to the drug through unblinding".
Anyway, I agree papers of this sort do not seem to completely corroborate the amplified placebo hypothesis. But, as Barth et al (2016) point out, reporting of side-effects in clinical trials is extremely variable. Reporting of side-effects may, therefore, not be the best proxy measure of unblinding. We do need studies that systematically check for unblinding (Even et al 2000). There is also the issue about whether depressed people may do better over the long-term without medication (see previous post). I am afraid this debate is likely to keep running and can't be closed down by the Guardian article.
Monday, September 04, 2017
Subscribe to: Posts (Atom)