Omega-3 fatty acids, found naturally in oily fish, frequently taken as supplements, and lauded for their anti-inflammatory properties, were found to increase the risk of high-grade disease by 71%. 'Taking omega-3 is associated with a 44% greater chance of developing low-grade prostate cancer. Overall, the fatty acids raise the risk of all prostate cancers by 43%.’
That’s how it was reported on Friday on Sky news, and on the BBC. It would seem bad news, for two reasons: first the high percentage for the reported risk; and second, that omega-3 fatty acids have in recent years been touted as the wonder-food that cures all ills – it can allegedly boost our brain power, keep our hearts healthy, strengthen our bones, protect against schizophrenia and for all I know cure male-pattern baldness as well.
Both reasons are statistically doubtful. Take the percentages first: note that the 71% greater risk of getting a high-grade prostate cancer, and the 44% greater risk of getting a low-grade one. Both have that magic word ‘greater’ in them. These are not the absolute risks: they are the relative increases in risk; and reporting relative rather than absolute risks is a well-known and horribly common journalistic bad practice. It makes things look much more frightening than they actually are. What it emphatically does NOT mean is that you stand a 71% or a 44% risk of getting prostate cancer if you take omega-3. In fact, if you are not told what the absolute risk is, being told how much greater it has grown is wonderfully (and absolutely) useless to you.
If you do not take omega-3 (and if you are a male), your chances of getting prostate cancer are about one in 750. (Neither news source quoted that. I have just had to work it out for myself, from the web: the Cancer Research UK website gives the crude incidence rate as 133.7 new prostate cancer cases for every 100,000 males in the UK, which works out at one chance in 750 – or a probability of 0.00133.) It is a pretty rough figure. But the point is that it is also a pretty low figure. It means you are unlikely to get the disease. Any horse coming in at odds of 750 to one against would be a rank outsider, and would make headline news.
So suppose you do take omega-3 supplements. The report says your chances of getting high-grade prostate cancer go up by 71%. Which means that it goes up from 0.00133 to 0.00228 – which last figure is roughly one in 440. Which is still pretty low. And that is the figure that matters. Again you wouldn’t bet on a horse at those odds and expect to go home rich. Take a very low figure and double it and you still get something pretty low.
The health message from that is: even if you are taking the supplements – don’t panic yet. It may well be wise to stop taking the supplements – but you are still very unlikely to get the disease.
The statistical message is that relative risks in news stories are meaningless, unless the actual risks are reported as well. The same statistical fallacy surfaced a few years ago in the ‘bacon sandwich affair’: a UN health organisation seemed to say that you should avoid eating processed meats – bacon in other words – because there was “convincing evidence” that they increase the chances of bowel cancer. The result was mockery in the press – and the mockery was deserved. You can eat a bacon butty a day and still be very unlikely to get bowel cancer. See Significance for more.
According to the evidence we have so far, omega-3 does not seem to help for cancer prevention or treatment; with children's learning or behaviour; with cognitive function; or in preventing cognitive decline with age or mental health problems, including bipolar disease, schizophrenia.
– Dr Lee Hooper. British Medical Journal.
So much for the supposed big risk factor of omega-3. What of its positive health benefits? More than 12,000 scientific studies on the benefits of omega-3 have been published in recent years. Dr Lee Hooper carried out a systematic review of them. Her work appeared in the British Medical Journal in 2006, and was reported – to a rather higher standard this time – by the BBC. In the words of the BBC, the claims for omega-3 are often as fishy as the omega-3-rich foods themselves.
As Dr Hooper put it, "According to the evidence we have so far, omega-3 does not seem to help for cancer prevention or treatment; with children's learning or behaviour; with cognitive function; or in preventing cognitive decline with age or mental health problems, including bipolar disease, schizophrenia."
Similarly, there's "no evidence that the fatty acids assist with cystic fibrosis, allergies, asthma, ulcerative colitis, Crohn's disease, or kidney disease". And Dr Hooper's study found evidence omega-3 improves children's learning abilities and behaviour to be "very poor".
On the plus side, omega-3 probably does help with arthritis, pain and stiffness. And it definitely seems useful for people recovering from a heart attack. "I would very much want anyone I know who has had a heart attack recently to be consuming omega-3" she said.
But that does not mean that those who have never had a heart attack should take omega-3 in order to reduce their chance of having one. "There's no evidence that omega-3 reduces the risk of death or heart attack or stroke or anything like that in those of us who have not recently had a heart attack," says Dr Hooper.
The lesson here is that omega-3 is probably good for a number of things. And probably harmless or useless for a number of things, and probably harmful for a few things. Just don’t get carried away by the hype – the negative hype or the positive ones.