Do The Drugs Work? It’s Complicated

By Neuroskeptic | July 11, 2009 9:55 am

Over at Comment is Free a week ago, Ed Halliwell proclaimed that “The Drugs Don’t Work”. The drugs being antidepressants. On this blog I’ve often written about antidepressants and the evidence that they work, or don’t, so I was interested to see what he had to say.

Halliwell begins by noting that antidepressant prescriptions are rising. This, he declares, is a bad thing because antidepressants just don’t work very well – “….A recent review found the SSRIs barely more effective than a placebo pill. Still, the NHS bill for prescribing them runs into hundreds of millions of pounds a year. It’s a crazy situation, and the tide may be turning…”

This invokes the famous Kirsch et al 2008 PLoS review of antidepressants. This analysis concluded that 6 weeks treatment with antidepressants was only slightly better than placebos for depression. But slightly is better than nothing. Kirsch et al is evidence that the drugs do, in fact work.

This despite the fact that the analysis included “suppressed” unpublished drug company data unfavorable to antidepressants. So almost uniquely in medicine, there can be none of the publication bias which plagues all clinical trials. In other words, an exceptionally high standard of evidence shows that the drugs do work. A bit. And in fact they probably work better than that, because Kirsch et al’s paper was biased against antidepressants as a series of classic posts by P J Leonard and Robert Waldmann pointed out.

Halliwell then says that instead of popping pills to ease our troubled minds, we should turn to “…simple, socially based steps everyone can take to improve their wellbeing. These include building good relationships, lifelong learning, being kind to others and exercise – not rocket science, but somehow we seem to have forgotten them.”

Now I don’t know what drugs you would need to take to think that “building good relationships” is “simple” – Ecstasy washed down with alcopops might do it. But once you sobered up and read a novel, or watched a play, or just remembered your last breakup, you would realize that relationships can actually be quite complicated. Not to mention being kind to others and lifelong learning, which are so simple that everyone is has a PhD in Being Nice.

But such nonsense aside, the actual hard evidence that these kinds of things can treat clinical depression as opposed to just “improving wellbeing” is weak. There is some evidence that exercise can treat depression, for example, but it’s often of poor quality, with no placebo, and publication bias could be rampant.

Indeed as luck would have it, today, two Cochrane Reviews were published. One was about antidepressants for the treatment of depression in primary care. The other was about exercise for depression.

Respectively, they concluded that – “Both tricyclic antidepressants and SSRIs are effective for depression treated in primary care” (not massively effective I hasten to add, but it’s something) while the exercise one concluded that “Exercise seems to improve depressive symptoms in people with a diagnosis of depression, but when only methodologically robust trials are included, the effect sizes are only moderate and not statistically significant…”

I’m not endorsing these conclusions. Neuroskeptic readers know that I’ve long been critical of antidepressant trials and sometimes even Cochrane Reviews. I study antidepressants for a living, and I honestly don’t know if they work for the millions of people who take them. (I’m fairly sure that they work in severe depression, but this isn’t a very common disease.) But I do know that it’s a really complicated issue. And I know that simplistic pro- or anti- medication rhetoric helps no-one and insults the intelligence of all.

But there’s more here. Ed Halliwell has a history. He was the lead author of “In The Face of Fear”, a deeply flawed Mental Health Foundation report from a couple of months back. The main message of the report was that anxiety is currently on the rise in Britain. We’re getting more scared and anxious. There is an epidemic of fear – right now.

This claim was supported by two things – a completely unscientific opinion poll (which didn’t really show much of an increase at all), and a reference to government mental health survey data from 1993 and 2007. These indeed show an increase in reported prevalence of anxiety disorders.

Now whether surveys such as these give meaningful data is questionable, but quite apart from that, the MHF report was guilty of a much more glaring error. As I said at the time, it simply failed to mention that we also have data from 2000. And in 2000, rates of anxiety disorders were almost exactly the same as in 2007, higher in some cases.

Now, the 1993, 2000 and 2007 figures are right next to each other in the government report (page 41 of this publicly accessible pdf): unless Halliwell has some kind of visual defect rendering him unable to see the middle of three columns of numbers, he must have seen this. The best available scientific data is that rates of anxiety have been stable for the last 10 years.

Stable, but very high. And the irony – and the tragedy – is that these massive reported rates of depressive and anxiety disorders in the British population (17.6%, at any one time and this is only for some disorders), are why antidepressants are so widely prescribed. Not these figures alone, but rather the more general belief that mental illness is very common – a belief promoted, inter alia, by organisations such as the MHF.

This belief – the “one in four” myth – is music to the ears of the very drug companies that Halliwell and others lambasts for pushing pills. When tens of millions of people are told that they are ill and need treatment, can you blame them for turning first to pills instead of a wholescale reconstriction of human life and society?

  • CyberLizard

    I just gotta say, they work for me. At least they lighten the 10,000 lb weight on my chest that depression causes enough to actually live life and work on “improving my wellbeing.” This Halliwell person sounds like an asshat that has obviously never been through severe depression. The way he talks it sounds like he's never had to deal with anyone being treated for depression. My wife and I have a joke that we pull out when my D gets bad. One of us will declare, “I've found the cure for depression: just don't be depressed!” That sarcastic joke sounds like what this guy is actually espousing.

    I would suggest that the majority of the people dealing with major D understands that meds are not a magic pill and that they will continue to have to work (hard) on all the aspects of life that Halliwell talks about. The meds just make it possible to actually do that.

  • Pallas

    Neuroskeptic, do you have any thoughts on the theory that antidepressents only appear to work in drug trials because the side effects cause people to suspect that they are not taking a placebo, and that a significant part of people in these studies can actually guess correctly whether they are in the control group?

    David Burns presents this theory in his book When Panic Attacks, and suggests that if a study was done in which placebos were used which cause side effects, we would learn that antidepressents do not work except as a placebo at all.

    I guess to some extent it doesn't matter, because if a antidepressent which is really a glorified placebo works, it works.

  • Neuroskeptic

    Pallas – That's an interesting issue. There actually have been several trials comparing antidepressants to so-called “active placebos” which produce side effects – but they're mostly old, small and of poor quality

    A Cochrane Review of these trials here found that the effect of antidepressants was quite small but it was very dependent upon which trials were included in the analysis…

  • Mike

    This comment has been removed by the author.

  • Mandy

    Hmmm..that all depends on who you talk to or what theorist you follow.

    My experiences with meds have been mostly negative ones but, to add some objectivity in, I don't react too well to drugs for physical problems well either.

    Sometimes, it is a case of different strokes for different folks but I do think that the hype outweighs the majority of expriences (that is of those people I have come across in addition to myself).

    And a medication driven MH service doesn't serve its patients well enough.

  • Neuroskeptic

    Mike – Sure, good idea!

  • Donna B.

    I don't know how much I suffer from depression, but I do know that Lexapro (and Celexa to a lesser extent) reduce arthritic or fibromyalgia pain… whichever school of pain one subscribes too.

    With help from a sympathetic PCP, I did my own experience with Lexapro, Celexa, and nothing over a year's time. The end result is Lexapro works best for me, Celexa next best and I'm in enough pain to restrict movement without either.

    The interesting thing is that the first few weeks off antidepressants, I actually felt more cognitively capable and a little happier. That was not enough to counter the pain that built up gradually. I'd rather be able to actually do something rather than just think about how much I'd enjoy it if it didn't hurt so much.

    What I want to note is that Lexapro and Celexa have a significant effect on pain. Since x-rays (or so I've been told) show a fairly severe amount of arthritis in various joints, I don't think this pain is all psychological.

    I decided after my experiment that the much higher co-pay I have for Lexapro is worth over “settling” for Celexa.

    Also, an anecdote about fibromyalgia. I did not know until my mid-30s that everyone did not suffer pain when their shoulders and knees were squeezed. I just thought a lot of people were cruel.

  • OutEast

    If we accept that the bulk of the impact of antidepressants is meaning effect – something the evidence does seem to support – then doesn't that make taking antidepressants a 'simple, socially based step' in itself? Making the decision to seek treatment, taking medical advice, etc etc are all contributory factors in achieving meaning effect, and these factors are certainly 'socially based'.

    Regarding Pallas's comment, I don't recall offhand any study 'in which placebos were used which cause side effects', but Moerman (in The Meaning Effect) covers a study where placebos were given a bitter taste to make them seem identifiable as 'real medicine'. I don't have the book to hand so can't cite, but the impact was definitily noticeable. And of course there have been multiple placebo surgery studies, which point in the same direction – albeit for surgery and not for treatment of depression.

  • Neuroskeptic

    OutEast – Heh, good point. And as others have noted, if antidepressants really are working, but only as placebos, then articles decrying antidepressants could be seen as a public health hazard.

  • Ed Halliwell

    Dear Neuroskeptic,
    Thank you for your comments about In The Face of Fear and The Drugs Don't Work. You make some good points.

    My intention with In The Face of Fear was to write a report that would promote confidence, suggesting ways we could relate with fear and anxiety more effectively. I think the report contained some useful messages – however, I accept that by drawing attention to increased anxiety disorder prevalences and perceptions of increased fear, there was a risk of fuelling rather than dispelling them.

    You call into question the fact that I didn't mention the 2000 psychiatric morbidity survey data. I tried to take a long view on some of the trends that might be affecting fear and the perception of fear, touching on the effects of changes over at least the last century (eg urbanisation). As far as possible, I tried to illustrate with data that also spanned the longer term. The psychiatric morbidity surveys went back to 1993, and I compared that data with the most recent figures, from 2007. Perhaps I should have also referred to the 2000 figures, clarifying that an increase in anxiety seems to have occurred between 1993 and 2000, and that the prevalence then remained at that higher level in 2007. I might also have been clearer in my language, emphasising that the government data indicates anxiety disorder prevalence HAS risen, rather than allowing the impression that this data shows they ARE rising now. However, including the 2000 data would not have substantially changed my analysis – that there appears to be more anxiety now than 15 years ago, and that we could take greater steps to alleviate the suffering it creates.

    As for the opinion poll we conducted, this was just that – a poll, nothing more, nothing less. I analysed the results quite extensively and, I think quite clearly, pointed out that people's answers were likely to be influenced by their own perceptions and biases. Given that perception is so important in the experience of fear, I felt those results were rather interesting – especially people's strong perception that others had become more fearful over the last decade (77% agreed with this, twice as many as felt they themselves had become more afraid). I suggested there might be a feedback loop going on – we receive powerful fear-inducing messages, and while most of us do not feel more afraid ourselves, we seem to take on board the idea that people are generally more fearful. But as our emotions are inevitably woven into our perception of the world around us, that sense of greater general fearfulness must inevitably affect us to some extent – this is perhaps borne out by the (albeit much smaller) increase in self-perceived fear that our survey found.
    I was disappointed that so little of the media coverage picked up on these more complex points, or on the positive messages about confidence, and instead focused primarily on the story about anxiety levels having increased since the early 1990s. In hindsight, I could have made more effort to ensure the messages about perception bias and confidence took centre stage. I aim to be more mindful of these kinds of issues in the future, and to present them in ways that will resonate more effectively. I appreciate you bringing them to my attention.


  • Ed Halliwell


    Re: The Drugs Don't Work – I don't write the headlines, so I didn't actually say they don't work. Clearly antidepressants DO work for some people – the question is how much of that effect is placebo. I have the same figures as everyone else to go on, and you will know how difficult it is to tease out the various biases that may be influencing them. Of course, if the effects are mostly placebo, that raises interesting ethical questions – as OutEast points out, is it ethical to prescribe them on the basis that they do help people feel better (perhaps a means for delivering the real medicine – hope, support, confidence and so on) even though it might involve a form of deceit to do so? Would we rather feel better, or know the truth? I know what my choice would be, but having been through some pretty unpleasant periods of depression and anxiety, I understand why some people might feel differently. I am not anti-medication in principle – if I became depressed again and there was an antidepressant that worked for me, and whose side effects were worth putting up with, I would probably take it.

    I enjoy your blog, Neuroskeptic, and look forward to you feeling able to write under your real name at some stage in the future.

    Best Wishes


  • Paul

    Further to the comments regarding active placebos, given the fact that participants, due to informed consent, are told of the potential side-effects of a particular medication, the chance of individuals breaking-blind may be very high. Take zopiclone for example. Participants will be told that they may experience a bitter taste in their mouth. If they get this bitter taste, then expectations are raised, the placebo effect is enhanced, and results are biased in favour of the active pill.

    This article by Joanna Moncreif is of interest:

  • Generic Viagra

    Drugs really works not only as a medication but also as therapy, because I've heard some people smoke marijuana in order to get calm, for that reason I think drugs are very necessary.

  • Anonymous

    I'm inclined to agree with with Richard Ashcroft on drugs. The worst antidepressant I've taken was Citalopram. It made me very suicidal. I know it was the drug because I didn't feel this way before being on it. So I changed to Paxil and was very depressed on that. Changed again to Prozac, went cold turkey and experienced severe mania. Also developed diabetes as a result of Risperidone use(antipsychotic). A scientist might say, well oh but, we need double blind studies to prove this, and we do, but when self reporting of thousands of people as to the damage these drugs are causing takes place, and causing serious illnesses like Diabetes, surely someone should sit up and take notice. It's pretty depressing that antidepressants cause major sexual dysfunction as well, hardly antidepressing, really. I am not denying that psychiatric drugs may be good for people in acute situations, they are highly effective as sedatives and behavioral control, but they do not correct chemical imbalances or clear up irrational thinking, they are all, basically, crude tranquilizers that dull and numb your senses, and damage your brain. I believe that the DSM manuals have a major blind spot believing in the chemical imbalance theory-it does not take the context of the person's suffering, life circumstances or sociological factors into account at all, because it believes things like depression and psychosis are basically genetic, which simply has never been conclusively proven. It's funny that a number of studies on public opinion show that the public believes in psychosocial causes of human suffering more than the profession that causes the labelling and stigma in the first place. I think fish oils, in fact, can be better antidepressants than SSRIs, and the high quality ones are very safe, provided they're reflux free. Studies from fish eating cultures such as Eskimos and Greenlanders also show very low rates of depression. Yet they live in cold conditions in which it is dark for about half the year, you'd think they would get very depressed due to a lack of natural light. People get depressed for a number of reasons, sometimes I think sadness and grief get diagnosed as depression as well. I'm sure antidepressants are a godsend to some people, like CyberLizard, but to say they work for everyone is not true, they frequently “poop” out, I've heard that from many depressed people. And because I've been on 3 different ones, all of which were ineffective, I'd have to say I doubt that any SSRIs would work for me. I feel less depressed off them, I have sad days and times of pain, but it's just part of life, and giving myself permission to feel that way is more therapeutic for me than dulling and numbing my senses.

  • viagra

    Deppends of the person, the mind state and the organism I guess, I have taste many drugs, cocaine, heroine, marihuana, but I'm addicted to pain killers… but I don't buy them because I know it's not good, you can see that I command my body, my body doesn't command me, I made exercise and eat healthy, maybe for some weird reason I can choose about what to do in this situation.

  • Jasmine Waring


    I know this is an old post, but it's interesting to see how things actually haven't moved on as much as what you would hope in the past couple of years. Anti depressants are still very much being dished out like water from GP's without having first looked at if just a change of lifestyle would work just as well.



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar