Mark Horowitz @markhoro.bsky.social

I've written a blog in response to @awaisaftab's misleading blog about Kalfas: "No, the Kalfas Meta-Analysis Doesn’t Prove Antidepressant Withdrawal Is Modest: Nonsense Disguised as Nuance" bit.ly/4ofhd7S

Jack has not permitted comments to be put on his blog so I’ll post this here. Many of the points in his blog are false.

He has latterly corrected the quotation from our press release after I emailed him (although he does not disclose this was the reason for his correction).

As he now acknowledges, we did not say what he had quoted us as saying and he has now corrected the quotation to say “this research calls into question reasonsfor the use of antidepressants.”

Of course our paper calls into question the main rationale given to patients for prescribing these drugs – of course it calls the ‘reasons’ into question. He and others may wish to argue that the drugs are prescribed because they are effective (discussed at length elsewhere) but the rationale presented to many patients was (and is) false.

Naturally, if it were found that Hashimoto’s disease was not caused by low thyroid hormones it would throw into question the rationale for prescribing thyroid hormones. It would throw into relief questions about long term harms and benefits of such a treatment that would not be raised if the treatment was acting in the self-evidently sensible manner of correcting a hormonal deficit.

His central accusation – and the central point of his published article – is that we misleadingly presented our review regarding the aetiology of depression as showing that antidepressants are ineffective. We did not such thing. Jack could find no such quote in our press release so he fudged it. (I believe this is the polite euphemism).

Of course such a finding about aetiology invites reflection on the efficacy of the drugs and so naturally we discussed this in the Conversation piece - our comments in this piece are all supported by evidence.

His statement about the Nordic Cochrane is also false. Nordic Cochrane moved from Righospitalet in Copenhagen to the University of Southern Denmark in Odense where it merged with the Centre for Evidence-Based Medicine and was renamed to Cochrane Denmark. It still functions today and houses the Cochrane Bias Methods Group which advises Cochrane : www2.cochrane.dk
He is performing quite some somersaults, slinging mud from unrelated comments made by Peter Breggin, and Peter Gotzsche, to avoid reading an article published in the BMJ Open. An article I would still encourage him and other to read so that they become aware of the substantive criticisms of Cipriani’s meta-analysis rather than assuming the solidity of claims made about antidepressant efficacy: bmjopen.bmj.com/content/9/6/e0…

I expect that after correcting his recent blog he will also correct his original blog on Awais’ website (where he falsely says: “a press release written by Moncrieff and co-author Mark Horowitz explicitly linked the umbrella review to a lack of efficacy in antidepressants”) and his academic paper were he makes the same false claim (“a press release written by Moncrieff and co-author Mark Horowitz explicitly linked the umbrella review to a lack of efficacy in antidepressants”) as he has now acknowledged that these claims are false.

I would like to be notified when these corrections are made.

Looking forward to speaking at this event.

I've responded briefly to Jack's blog here: mhorowitz.substack.com/p/unexamined-a…

The two major problems with Jack Goulder’s blog (and the paper on which it is based) is, one, that Jack is working from the unquestioning (and largely unexamined) assumption that antidepressants work and anyone who suggests otherwise must have mischievous motivation. And two, he has misrepresented what we have said about our umbrella review.

Jack said to me tellingly in an email:

“I should be clear that my academic background is in history and social science, not epidemiology or pharmacology. I’m not sure I’m qualified to give a detailed analysis of the subtleties of this quite vast literature!”

So without engaging with the debate about efficacy in any meaningful way he has embarked on what sounds like a year-long project on the premise that ‘the drugs work because senior people in my field agree they do’ and that he saw a young woman (with a different condition to the literature he is commenting on) improve with antidepressants ‘after a few months, and with the help of therapy too.’

This is not how science works. I sent him a very clear explanation of the issues with antidepressant efficacy written by a Cochrane group in Denmark – who I’m sure have never read Szasz! – published in the BMJ Open but he seems not to have read this: bmjopen.bmj.com/content/9/6/e0…

To just briefly re-state the issues with the Cipriani review on antidepressants (which provides the basis of Jack and other’s unshakeable belief in the efficacy of these drugs): presenting response rates inflates the small absolute differences between antidepressant and placebo.1 This difference is 2 points on a 52 point depression scale.2 This lies below any analysis of what constitutes a clinically important difference (rather than just a statistically significant difference).3,4 See figure above to illustrate some of these elements. These studies only examine effects at 8 weeks, whilst people take these drugs for years.5 Even the small effects at 8 weeks are likely to wear off due to tolerance – the corollary of any drug that causes withdrawal.6

There are also unblinding effects – people who know they are on the drugs have raised expectations which improve mood and these expectations inflate the apparent effects of antidepressants in placebo-controlled trials.7 For example, in an analysis of the TADS study treatment guess had a substantial effect on outcome, while actual treatment group (antidepressant or placebo) did not.7 The Stone et al meta-analysis did not refute the role of unblinding as Jack suggests, it merely notes that the trials it included did not test the integrity of the double blind design (Stone writes “We cannot fully exclude the possibility that the effects of the drugs are accounted for by functional unblinding”). 8

There is also massive publication bias.9 The studies which find 15% of people benefit at an 8 week time point suffer from all the same limitations as above.8 And if there is a small pharmacological effect in the short term from this class of medication it leaves the question of the utility – or the ethics – of giving people drugs that numb their emotions when they are distressed,10 especially with the risk that for a portion of them this numbing does not wear off when the drugs are stopped.11,12 Jack’s piece ignores almost all of these issues.

A central part of this argument is that Joanna and I put out a press release arguing that our paper had demonstrated antidepressants are ineffective. Notably he does not quote the lines from the press release in his blog or the paper – remarkable when this is a key part of his argument.

Of course the reason he does not quote this is because his assertion is false. We don’t anywhere say in our draft press release that we have demonstrated in our umbrella review of the role of serotonin in depression that we have demonstrated that antidepressants are not effective. The closest we come is when we say: “Therefore, this research calls into question reasons for the use of antidepressants.” The research does indeed undermine a common reason for the prescription of antidepressants.

I think what is most unsettling about this piece is that Jack has inverted the work of Naomi Oreskes. Her work and her book Merchants of Doubt sets out the playbook taken by large industries to undermine critics of their hegemony. When big tobacco was being criticised for causing lung cancer by researchers, it funded researchers who sewed doubt in this hypothesis by pointing to asbestos, stress, etc, anything other than tobacco which might cause lung cancer.

The same playbook was used by fossil fuel companies: as climate scientists started pointing the finger at this industry for driving climate change, scientists associated with industry put out arguments sewing doubt: temperatures natural vary, humans couldn’t be doing it, etc.

This is the identical playbook adopted by the pharmaceutical industry and its supporters when it comes to obscuring the harm of their products. Antidepressant withdrawal isn’t caused by stopping antidepressants, they say, it is caused by patient’s expectations of worsening (nocebo effects), their mental health condition returning, other diseases that co-incidentally occur at the same point (functional neurological disorder) and short-term withdrawal studies show that there is no real issue, etc.13,14

There is an absurdity to painting academics, many of whom are paid by one of the largest industries in the world as representing unimpeachably pure science and Joanna and a bunch of her academic colleagues (including me) as the ‘merchants of doubt’. Of course some of the scientists who defend the efficacy or minimise the harms of antidepressants are not paid by industry, but other interests, such as professional status, are also at play.

I would strongly encourage Jack and anyone reading his essay to go and read criticisms of antidepressant efficacy before jumping to conclusions that anyone putting forward a criticism of the current consensus must be mischieviously motivated.

The current (and poorly examined) consensus on the relative safety and effectiveness of antidepressants in the long-term has led to an avalanche of harm that can be seen spilling out onto social media, newspaper articles, government inquiries, parliamentary groups and, increasingly, concrete NHS projects and directives. Rather than expending further energy trying to shoot (and silence) some of the messengers it would be better for patients if psychiatrists examined the reality of what is occurring to patients and stopped deriving false reassurance from short term studies centrally concerned with commercial objectives.

References available at: mhorowitz.substack.com/p/unexamined-a…

A new study funded by the Wellcome Trust will take people off antidepressants over 4 weeks and will measure serotonin levels (using a brain scan following tryptophan depletion) to predict who will relapse. There are 9 measures of relapse or mood symptoms. There are no measures of withdrawal.

pradam.org