September 10, 2013

Tell us what you really think

From the Herald (and originally the Daily Mail)

People who bottle up their feelings are at least a third more likely to die young than those who regularly express what they’re thinking.

Not what the study says — there is a hazard ratio of 1.35 but even the researchers only claim this is a 3 year reduction in life expectancy and they don’t say anything about how likely people are to die young.

But when the researchers looked at specific causes of death they discovered that the risks increased by 47 per cent for heart disease and 70 per cent for cancer.

This was a fairly small study, with only 34 heart disease deaths and 37 cancer deaths, and the authors make it pretty clear the evidence for heart disease was too weak even for them to be comfortable — they say

Findings for CVD death were in a direction and of a magnitude consistent with prior work though they did not reach statistical significance.

How do I know all this? Well the story doesn’t give the names of any of the researchers, let alone link, but it does give the journal name and it turns out the paper appeared only a month ago, so it didn’t take too much work.

Assuming the correlation is real, it could still easily be due to other factors.  The researchers deliberately didn’t attempt any control for health risks, in case these were due to emotion suppression. For example, they didn’t distinguish smokers and non-smokers, because of the concern that smoking might be due to emotion suppression.  They didn’t even ask if people already had a cancer diagnosis at the start of the study.

The story does quote an independent expert. Well, sort of.

London-based business psychologist Voula Grand said: “It has long been thought that cancers are partly the result of suppressed emotions.

Judging from their web sites about cancer prevention, neither the American Cancer Society nor Cancer Research UK seems to know about this. They might be more reliable than the Daily Mail.

 

avatar

Thomas Lumley (@tslumley) is Professor of Biostatistics at the University of Auckland. His research interests include semiparametric models, survey sampling, statistical computing, foundations of statistics, and whatever methodological problems his medical collaborators come up with. He also blogs at Biased and Inefficient See all posts by Thomas Lumley »