Exercise, depression and the nature of medical evidence

A quick recap

May 21 Louise Mensch tweeted ‘Self-esteem: exercise is the best anti-depressant and ANYONE can do it. Try this – 3 minute desk-side mini walk’. It caused a mini-storm on Twitter so Mensch followed it up with a blog post. Those who have experience of the illness pointed out that for many people with depression exercise is impossible, for various reasons. @bezukhova and I blogged about it here. Mensch was not particularly impressed with my response and tweeted me about it several times.

Below are some extracts from tweets sent by Mensch on 24 May:

 Apparently Mensch’s “blog gives a link to google scholar with multiple pages of studies that support my contention …”

Whereas my blog “might lead somebody to believe that exercise is not a potent antidepressant, and thereby remove a source of help’”

Further: “the studies indicate it is an antidepressant often as potent as medications. And universally tolerated, no side effects” and: “It’s a lot more than ‘exercise helps’. The studies and evidence go much further than that.”

Thus in sum, Mensch has argued that exercise is a “potent antidepressant” which is “universally tolerated” with “no side effects”. I must, she says, “address the overwhelming weight”; the “mountain of peer-reviewed science that proves the efficacy” the “seriousness of the studies” and the “sheer number of them.” According to her I am being irresponsible in arguing that exercise often helps alleviate depression but may not be possible for everyone.

Given that Mensch has said that I am being irresponsible I do feel that some things are worth clarifying at this point. I am not clinically qualified. I am quite categorically NOT offering clinical advice here. For anyone with mental health issues, treatment needs to be decided by patients and clinical staff together, although I recognise that that relationship can be complex. Nonetheless, if you are experiencing any kind of illness you do need to discuss this with clinicians rather than take evidence from the internet on trust.

I am clinically depressed. Often it is in remission. At times its icy fingers steal around my mind. I have been treated through therapy but not with medication. My academic background is in history of medicine and my PhD was jointly supervised by one person trained in physiology and another who holds a medical doctorate. This is the position from which I write.

Now in order to assess Mensch’s claims I do need to talk about the nature of scientific evidence, since she claims science is in her favour. I’ve kept this as succinct as possible whilst maintaining accuracy. Those versed in science and its history will see that it is a shortened version but that is inevitable in what is a blog post, not an academic article.

Scientific methods and their application to medicine

Given the number of different disciplines under the umbrella term ‘science’, there are variations in scientific methodology. However, science can broadly be characterised as seeking hypotheses that fit observed facts. A hypothesis can then be used to make predictions and experiments can be performed to test if the hypothesis is correct. Science proceeds using observable and measurable evidence (using ‘observable’ broadly rather than restricting it to what one can see with the naked eye). Science tries to find theories which have the best fit with the world around us.

There are some exceptions to this. Theoretical physics for example does not, as far as I am aware, include experimentation, but broadly speaking science attempts to find models which explain and predict the world around us. Models and theories can be and often are superseded by something which is seen to provide a better fit because it has a greater weight of evidence behind it.

The relationship between science and medicine has a long and complex history. There is considerable debate about the extent to which medicine can be held to be a science although increasingly during the twentieth century it became more reliant on scientific evidence. Nonetheless, information does not always transfer readily from the laboratory into the clinic. Science proceeds by experimental testing of theories. Medical experiments are fraught with ethical difficulties. When performing a scientific test it is possible to have a control group and a group that is tested. In medicine it is necessary to ask whether the test is possible and if it is, is it ethical to have a control group who remain untreated?

 In addition to ethical problems, clinicians have another problem when applying scientific methods to medicine. Scientists generally, though not always, have the advantage of controlling their experiments so that they only test one variable, whereas within the human body there lie a multitude of variables which a clinician cannot control. Put very bluntly, if you want to see the effect that an acid has on different metals, you can devise an experiment that tests the acid’s effects and only that. Where medicine is concerned you must consider whether or not an experiment is ethical and then if it is found to be so, your chances of testing for only one variable are limited.

Given the way in which healthcare is delivered, it can appear to be more of an art than a science. Doctors making day to day decisions about patient care will not necessarily do so based on the latest research but may go on their own personal experience and ‘gut feeling’. To counteract this, since the 1970s there has been a movement within medicine to strengthen its evidence base. More information about evidence-based medicine (EBM) can be found here. EBM aims to consolidate data from randomised controlled trials and argues that clinicians “should treat where there is evidence of benefit and not treat where there is evidence of no benefit (or harm)”.

However, the movement towards a firmer evidence base for medicine is troubled. Medicine does not take place in a vacuum. Those offering treatment may not be disinterested. Now, quite categorically I am not impugning the integrity of anyone practising or undertaking research in medicine. There is however a problem within medicine. Some trials that are carried out never see the light of day. They simply are not published. There may be many reasons for this and you can read more about the campaign to have all trial results published here http://www.alltrials.net/about/. Suffice it to say that the information available to clinicians is not comprehensive because not all work is publicly available.

Medical evidence and mental health

There are further problems when considering how we find out about the best way to treat mental health problems. Malaria is caused by a parasite and can be diagnosed by a blood test. Influenza is a viral infection, usually diagnosed by symptoms but it can be diagnosed by blood tests. There are a myriad of illnesses which can now be diagnosed by using laboratory tests to find the cause, meaning treatments can be targeted. Mental health problems are much more difficult to diagnose.

Diagnosis of depression is generally arrived at after a clinician asks a patient various questions about how they are feeling. There are standardised tests which clinicians use to give a score to depression. Nonetheless, the decision about the extent of someone’s depression is based on something relatively subjective i.e. the patient’s knowledge of how they feel, via a process that is also complex and reliant on the clinician’s interpretation. Many health care professionals are amazing and very understanding. Some patients are less fortunate in the clinicians that treat them. One of the problems with depression is this subjective diagnosis. Sufferers may feel fraudulent and the attitude of  others who do not really understand the seriousness of the illness may exacerbate this sense of fraudulence.

There are, for those who are interested, several debates about the diagnosis of mental illness. One can start with Foucault’s Madness and Civilization. But there are extensive debates over for example the diagnosis of schizophrenia, not least because diagnosis may be racially biased (this is frequently discussed, one example can be seen here). One can also turn to Elaine Showalter’s The Female Malady for an in-depth historical analysis of gender bias in diagnosis.

So, the relationship between science and medicine is problematic. Even as medicine draws more on scientific methodology, there are doubts about the extent of evidence available. Evidence is even more difficult to handle where mental health is concerned. When taking part in a discussion about treatments and mental illness one ought, before making sweeping claims, to be aware of all these problems.

Are Mensch’s claims supported by the evidence she cites?

According to a link on Mensch’s blog she googled on “exercise effective antidepressant”. As any good undergraduate who has read Karl Popper will tell you, this is not the right way to do science. Remember philosophy 101: just because you have seen 10,000 black ravens, doesn’t mean your lecturer won’t walk into the room with an albino corvid. As history of science shows us, you can have a mountain of evidence in your favour and still be superseded by another theory which more accurately explains observed phenomena. Even theories which have been verified by substantial evidence can be falsified by other evidence.

We could just stop there. But let’s just test this further. Do the studies Mensch cites back up her claims? No. No, they really don’t.

This is one of the studies Mensch draws on: http://www.psychosomaticmedicine.org/content/69/7/587 The study examined 202 adults with major depressive disorder over a period of 16 weeks. The participants were divided into a control, a group on sertraline, a group with supervised exercise and a placebo group. After 4 months of treatment, 41% of patients had achieved remission. Medication had the highest remission rates, the placebo the lowest. So in this case exercise was not quite as potent as medication and the placebo actually scored reasonably well. At no stage do the authors of this paper make any claims as grand as Mensch’s. They mention no side effects, they are positive about exercise but they also say that the relative success of the placebo group suggests that ‘a considerable portion of the therapeutic response is determined by patient expectations, ongoing symptom monitoring, attention, and other nonspecific factors’. Also bear in mind that this was 202 patients over 16 weeks. I am in no way knocking the study, but the authors themselves do not make the sweeping claims that Mensch does.

Moving onto another article Mensch cites, from 1999: http://archinte.jamanetwork.com/article.aspx?articleid=485159 This study examined 156 patients with major depression, again over 16 weeks. The authors’ conclusions are careful and worth reading in full. Amongst other things they state:

An exercise training program may be considered an alternative to antidepressants for treatment of depression in older persons. Although antidepressants may facilitate a more rapid initial therapeutic response than exercise, after 16 weeks of treatment exercise was equally effective in reducing depression among patients with MDD.

In discussing studies that show the effectiveness of exercise in treating depression, the authors of this study state:

These studies have been plagued by methodological problems, however, including limited sample sizes, lack of randomized designs, uncontrolled concurrent therapies, failure to document exercise training effects, and imprecise diagnosis of depression


Although these studies are suggestive, to our knowledge the therapeutic effects of exercise on clinical depression have not been evaluated systematically.

This bears repeating. One of the studies Mensch cites in her favour states that as far as the authors are aware, the effects of exercise on clinical depression have not been evaluated systematically. Further on in their conclusions, the authors state that follow ups will be needed to assess the impact of recurrence in the long term. They also state that:

Patients also were sufficiently motivated to volunteer for a study of exercise training, tended to be highly educated, and were healthy enough to participate in an exercise program. The extent to which these findings may be generalizable to other older depressed individuals will need to be studied. [emphasis added]

I.e. the authors realise that the patients under study were able and willing to exercise whereas not all patients with depression might be. Further, they point out that their studies were undertaken with patients in group settings and that the nature of the setting may be part of the effectiveness. I.e. jogging around your desk at home on your own might not be quite so much fun.

It’s not looking great so far. The studies Mensch cites are very careful in their claims about exercise and depression and make their own limitations clear. Let’s randomly go for a third: http://www.sciencedirect.com/science/article/pii/S0091743504000842

You probably won’t be able to get access to that unless you can log in via an academic or public library. In this third study, 73 participants were recruited from ‘a heterogeneous sedentary employee population’. The incidence of depression within this group was slightly higher than average within the general population from which it was drawn. Around 40% dropped out, citing lack of time as their major problem. The authors do not draw strong conclusions about the effect of exercise on depression and they clearly state that longer-term follow ups are needed.

From these three studies, two highlighted by Mensch and one chosen at random from her list, it is clear that they do not make the strong claims that Mensch reads into them. They are careful and circumspect. They are open about their sample sizes, about drop-out rates and about their possible limitations. Those limitations include the fact that volunteers in such trials maybe more motivated and more able to exercise. Those with depression, who have problems exercising, are unlikely to volunteer for a study which could involve exercise. (I’ve been invited to take part in trials before. These invitations are very clear about what I might be asked to do, ethically this is required of them).

In Conclusion

I could go on. And on. I could search for information about exercise, depression and side effects. In particular, it is worth noting that for anyone with bipolar, or those who have depression combined with eating disorders or ME, exercise can be problematic. But in essence, the science does not back up Mensch’s claims, and neither do all those people with depression who have told her she’s wrong.

Now one could at this stage say that those with depression are producing anecdote whilst clinical researchers are producing evidence. But if we do that, we take away the patient’s voice, and what good is that to someone who already has very little sense of self worth? Instead, consider this. Even scientific theories with a mountain of evidence in their favour can be superseded (if you are interested, read T. S. Kuhn’s Structure of Scientific Revolutions). Medicine can and should draw on scientific methods where it can, but humans are not chemicals to be experimented upon. Medical evidence is by its very nature partial and particularly so where mental health is concerned. The studies that Mensch cites make it clear that trials are limited, that longer term follow ups are needed, that volunteers for trials are those who are willing and able to exercise, that patients drop out from trials and that exercise is easier when it is structured. And this is crucial to realise – depressed patients who find exercise difficult are unlikely to be included in trials. That is the nature of medical evidence. So I stand by what I said. Exercise has many, many health benefits but for people with depression, being told to jog around for 3 minutes is just insulting.

5 thoughts on “Exercise, depression and the nature of medical evidence

  1. Hear, hear!
    Not that I suffer from depression (other than manageable reactive depression when external circumstances induce it), but I know and have known lots of people who do or did.

  2. A brave and thoughtful post. For me, exercise – principally cycling – is my way of keeping depression at bay. But there are times when I feel so low that I have to force myself to get on the bike. In a good year, I’m out every weekend, doing Audax (long distance) events and generally feeling great (even when riding through torrential rain at 3 in the morning with 100 miles to go!). In a bad year – like this one – I use the bike to commute but I just can’t work up the enthusiasm to do any more. There’s no logic to it – actually, there’s no logic to depression – but I can understand how researchers may have found it difficult to run robust experiments.

    1. Thanks Rob. It’s an interesting question about how robust the studies are. They’re not robust in the sense that a study in experimental physics would be. However, I noticed that the good studies make their limitations clear and I think that does give them a kind of robustness. If they pretended to prove more than they do there would be a problem with them. But if they clearly state drop out rates and discuss the selection of patients then I think that makes them sound studies.
      That is, they need to be judged on their own terms. They can’t be robust in the way that many other scientific studies can be but the researchers can do everything possible to be aware of limitations and correct for them where possible.
      My GP repeatedly told me to keep cycling and keep seeing my horse. Maybe as much as anything those things kept me grounded.

  3. Well said, I’m glad not everyone takes these studies at face value.

    The way I explain this to people is to say that if it was a drug, it would not be approved for treatment. Why? Because it is not a blinded trial. Non blinded trials are subject to many different biases that are difficult to control for. One is social desirability bias – many people answer questionnaires or interviewers with answers they think they want to hear (not deliberately mind you). It gets worse with exercise studies which do not objectively measure activity levels. People may be reporting that they are exercising and feeling better, even if they are not actually increasing their activity levels.

    I agree with the point that the association between depression and a low level of exercise may simply be a side effect of not having the health/energy or time to exercise, rather than a consequence of exercising itself. Perhaps simply being too busy to exercise may be associated with depression, but the aforementioned studies were not designed to uncover deeper explanations.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s