Could findings about psychological causes of CFS be overstated or even wrong?
Seven years ago, John Ioannidis detonated a small bomb under the Life Sciences with his paper “Why most published research findings are False”. He made the case that widespread failings such as small sample sizes and flawed data analysis mean that most positive research findings are probably wrong. Judging by commentaries and articles in top journals such as Nature, many scientists broadly agree: science has a serious problem.
Now Ioannidis (pronounced yo-NEE-dees) has taken aim at psychological research. An important new paper has just shown that the claimed relationship between Type D personality and cardiac deaths association is “overstated and most likely false”. Ioannidis argues that the flaws in the original evidence linking Type D personality to cardiac deaths are probably typical of psychological research:
That sounds to me like another bomb going off.
Problems with psychological research
Ioannidis highlights problems he’s identified before in research, such as small samples and flawed data analysis. Too often researchers keep analysing the data in different ways until finally a positive finding emerges - an approach that has been described as ’torturing the data until it tells you what you want to hear’.
However, Ioannidis also thinks another major problem is too much group-think in the psychological sciences. Consequently, any replications of an original study tend to lack the necessary independence for meaningful confirmation of findings.
There is another way that such group-think could lower standards: new papers that fit with the prevailing group-think don’t get properly scrutinised by reviewers or challenged once published. Ioannidis didn’t mention this specifically but it would certainly explain a lot of flaky papers I’ve read .
What's this got to do with CFS?
If Ioannidis is right, it could have profound implications for Chronic Fatigue Syndrome. Here, many studies have reported associations with psychological factors, often arguing that such factors play a causal role in CFS. For example, studies have found psychological ill-health and an 'inappropriate' boom-and-bust response to illness are risk factors for developing CFS. Associations between psychological factors and 'Functional’ illnesses such as Irritable Bowel Disease have been found many times too.
However, one of the most compelling examples of the power of psychological factors was the link shown between Type D personality and subsequent cardiac death. Unlike other outcome measures, cardiac death is not prone to self-report bias or other similar problems of measurement and interpretation. Links between personality and an unambiguous outcome like death provided a dream example connecting psychology to physical health. But it was this association between Type D personality and cardiac death that psychology Professor James Coyne has just meticulously unpicked, sparking Ioannidis’s broadside.
A new era of healthy scepticism?
What I find striking is that these papers by Coyne and Ioannidis were published by the Journal of Psychosomatic Research. This journal has probably published more articles than any other journal claiming that psychological factors are associated with, or cause, CFS. Perhaps this marks the start of a new era of more rigorous—even healthily sceptical—evaluation of claims that psychological factors lead to illnesses such as CFS.
Fresh wind blowing through psychological research
Blog entry posted by Simon, Oct 30, 2012.