There has been something of a tea-cup storm over publication of findings from the TREAD study in the British Medical Journal on June 9th (Chalder et al 2012). TREAD is an inventive acronym for the NIHR funded TREAtment of Depression with physical activity study which was conducted in Bristol and Exeter between August 2007 and October 2010. The BMJ paper was published online on June 6th 2012 and so cyberspace was already buzzing before my own paper copy came through the letter box. The reason why is of course because the findings, as they were presented, were counter-intuitive and challenged practices many GPs and their patients find attractive. The episode also provides insight into ways in which a combination of the press’ and professional’s separate but complementary interests in simplifying the complex can result in misleading impressions.
Martin Robbins provides an account of how it went from a journalist’s perspective. The paper’s abstract reads “The addition of a facilitated physical activity intervention to usual care did not improve depression outcome or reduce use of antidepressants compared with usual care alone.”. A related press release read “New research published today [6 June] in the BMJ, suggests that adding a physical activity intervention to usual care did not reduce symptoms of depression more than usual care alone.”, and included a quote from Prof. John Campbell, General Practice and Primary Care, University of Exeter: “This carefully designed research study has shown that exercise does not appear to be effective in treating depression.”. In the early hours of June 6th, only a short time after the press embargo had been lifted the BBC reported “Combining exercise with conventional treatments for depression does not improve recovery, research suggests.”. Later that morning the Guardian appeared with the headline “Exercise doesn’t help depression, study concludes. Patients advised to get exercise fare no better than those who receive only standard care, researchers argue”. This was followed by a number of online reactions by other journalists writing for the Guardian and the Daily Telegraph, and other commentators. Sharp eyed colleagues spotted all of this and the original BMJ paper was the subject of our clinical work-place journal club on June 11th.
One of the earliest responses posted by the BMJ was from Stephen Pilling and Ian Anderson who led the development of NICE Guidelines for the Treatment and Management of Depression which were published in October 2009 and which recommend the use of exercise as a “treatment” for depression (Anderson et al 2009). On June 8th they posted criticism of the TREAD report drawing specific attention to two perceived shortcomings. The first of these was that the intervention “tested” by TREAD was not in fact exercise itself, but contact with an exercise facilitator. Their NICE recommendations had been based upon understanding of research directly considering structured group exercise and as a result they did not regard the TREAD intervention as comparable. The second was that their recommendation was that exercise should be used as a treatment for “mild” to “moderate” depression, and the mean Beck Depression Score of TREAD subjects on entry into the trial (32.1) suggested that they were better thought of as in the “moderate” to “severe” range. Other postings identify a variety of other technical and anecdotal concerns about the trial.
All of this could be nothing more than a storm in a tea-cup were it not symptomatic of several familiar and consistent strands, and didn’t result in misinforming or confusing vulnerable people who listen to the radio or read newspapers. Three questions are worth considering, and there may be more.
- How did so high profile a medical journal publish findings from so flawed a trial?
- Why did the press pick up upon its conclusions in so misleading and uncritical a way?
- What are the wider implications for understanding how we conduct and disseminate mental health research?
Was the trial flawed? What would Austin Bradford-Hill, credited father of it make of contemporary use of the term “Randomised Controlled Clinical Trial”? Chandler et al acknowledge that “Owing to the nature of the intervention, none of the participants, general practices, clinicians, or researchers performing the outcome assessments could be blinded to treatment allocation.” (page 2). Somehow the strict experimental requirement of double blinding has been allowed to lapse and clinical trials are considered randomised controlled clinical trials and elevated to premier status in the evidence hierarchy even when subjects and those involved in their treatment are aware of their treatment status. The whole purpose of RCT methodology is that it provides a way of controlling for the effects of that myriad of variables, predictable and unknown, which might influence the outcome of a complex phenomenon such as an episode of illness or emotional distress. We know that placebo or expectancy effects are considerable in mental health and yet we have drifted away from strict adherence to research protocols which control for them. Where investigated some 80% of antidepressant trial subjects correctly guessed whether they were taking control or trial medication on the basis of experienced side effects (Rabkin et al 1986). Someone I am clinically involved with who had agreed to take part in another investigator’s trial of a psychosocial intervention informed me that they had been allocated to the control arm. “How did you figure that?” I asked. “Because they told me.” was the reply. I doubt that Austin Bradford-Hill would respect that as an RCT and he would probably have similar difficulty with TREAD, but both are likely to be considered RCTs and their findings respected as such.
The press by their own subsequent acknowledgement picked up upon a sensational sound-bite without questioning its source or its provenance. Exercising to improve well being is folk lore. When a prominent medical figure says “This carefully designed research study has shown that exercise does not appear to be effective in treating depression.” news is breaking. “Prominent medical figures” are under pressure to maximise the impact of their research and might be tempted to present findings in a way that attracts press attention. Quoting experts is not the same as conducting one’s own investigative journalism, and it is certainly easier and less risky. Sensational medical stories are popular and so the temptation is understandable. News items reporting medical research findings that promise “a breakthrough” are all too common. However, the cynic doesn’t find it difficult to hear and see the appeal for more research funding or how far away the “breakthrough” actually is if by “breakthrough” what is meant is a radical improvement in everyday experiences of this, that or another condition. Perhaps we should consider the need for a “Medi-Levenson”, to consider the relationship between “prominent medical figures” seeking funding and good impact ratings, and the press seeking newsworthy reports of medical advance.
Two strands of this might be considered particularly relevant to the world of mental health. One is that devaluation and degradation of the strict experimental requirements expected of a definitive randomised controlled clinical trial are a particular problem in this area. The other is that our clientele are vulnerable and desperate by definition, and therefore particularly susceptible to misinformation.
The first of these is a direct consequence of the nature of the field. Expectancy and/or placebo effects do make a powerful contribution to outcomes from all forms of “mental disorder” and so it is particularly important to control for them in the course of evaluating a treatment. Unfortunately that is very difficult to do. All of our drugs have prominent side effects and trialists are obliged to warn participants of their possibility. Psycho-social interventions are even more difficult obscure. The result is that our old friend, the Emperor’s Clothes, becomes a little see-through in this context as well. We should not claim to be offering treatments that are truly tested by exacting RCT methods in the same way much of medicine presents itself as doing, but we do.
The second is self-evident but possibly deserves a little elaboration. A core feature of that most widely accepted psychosocial intervention, cognitive behaviour therapy, is acknowledgment of the reality and relevance of cognitive distortions. The problem is as much how the experience; ruminations, others’ voices, palpitations, dysphoria, anger, etc. is understood as it is the presence of the experience itself. What psychiatrists as authorities in the field tell people about the nature of emotional distress and how it might be mitigated plays an important “public health” role. Possibly one of the most damaging acts of unintended harm things we perpetrate is to claim we have answers, when in fact we don’t. This little story about exercise and depression might be a timely reminder. If you add TREAD to the other research concerned with this question then the answer has to be “this sort of research doesn’t and can’t provide a definitive answer”. If you are feeling low and you think it might help, do exercise.
To quote from Rudyard Kipling:
The cure for this ill is not to sit still,
Or frowst with a book by the fire;
But to take a large hoe and a shovel also,
And dig till you gently perspire;
Will ill-applied RCT methodologies ever truly improve on this?
Hugh Middleton June 2012
Clinical Associate Professor,
University of Nottingham School of Sociology and Social Policy
Honorary Consultant Psychiatrist, Nottinghamshire Healthcare NHS Trust
Melanie Chalder, Nicola J Wiles, John Campbell, Sandra P Hollinghurst, Anne M Haase, et al (2012) Facilitated physical activity as a treatment for depressed adults: randomised controlled trial. British Medical Journal OPEN ACCESS, BMJ 2012;344:e2758 doi: 10.1136/bmj.e2758 (Published 6 June 2012).
Anderson, I. Pilling, S, Barnes, A. et al (2009). Clinical Practice Guideline No.90: Update: Depression in Adults in Primary and Secondary Care (Update). Gaskell/British Psychological Society. London
Rabkin, J. G., Markowitz, J. S. and Stewart J. (1986). How blind is blind? Assessment of patient and doctor medication guesses in a placebo-controlled trial of imipramine and phenelzine. Psychiatry Research, 19, 75-86
Kipling, R. (1902) Just So Stories. How the Camel Got His Hump. Accessed on line June 14th 2012. http://www.boop.org/jan/justso/camel.htm