Sunday, 19 May 2013

Angelina Jolie and BBC spin

Michelle Roberts. Not a clue.

Michelle Roberts, the BBC's worst health journalist, has taken to the airwaves to talk about Angelina Jolie's double mastectomy. Jolie decided to undergo the operation because she has inherited genes that make her very prone to breast cancer. Her mother died of the disease at the age of 56 and doctors say her DNA means that she has an 87 per cent chance of developing the disease.

I have written before about the BBC's determination to portray smoking, drinking and obesity as the leading causes of breast cancer, despite these being, at most, quite modest risks for the disease. Michelle Roberts was personally responsible for a particularly awful piece of journalism in 2011 which claimed, quite wrongly, that obesity was the "leading driver" of breast cancer.

Roberts could not allow the Jolie story to pass without bringing "lifestyle factors" into it and so appeared on camera—carefully flagged up as "a qualified doctor"—to talk her nonsense.

"In [Jolie's] case, DNA she inherited from her mum meant she was at 87 per cent increased risk of developing the disease in her lifetime and that's why she decided to have a double mastectomy."

No, no, no. She did not have an increased risk of 87 per cent. She had an absolute risk of 87 per cent, as the BBC themselves reported:

She said her doctors estimated she had an 87% risk of breast cancer and a 50% risk of ovarian cancer... Her chances of developing breast cancer have now dropped from 87% to under 5%, she said.

It beggars belief that the BBC's online health editor does not understand the difference between relative risk and absolute risk. Having made this pitiful error, she then moves on to her pet belief that "the vast majority of breast cancers" are caused by "behaviours such as smoking and conditions like obesity".

"Most cases of breast cancer are not caused by faulty genes such as these. Many more cases of breast cancer are caused by things like smoking and obesity—things that we all have the power to prevent."

There are lots of risk factors for breast cancer, including delayed childbirth and not breastfeeding (as this more considered BBC article points out). So why specifically mention smoking and obesity? It is questionable whether smoking increases the risk of breast cancer at all. The NHS doesn't include it in its list of risk factors and a comprehensive meta-analysis of 53 studies published in the British Journal of Cancer concluded that "smoking has little or no independent effect on the risk of developing breast cancer" (results shown below).

A relationship between obesity and breast cancer has been found more regularly, but the risk seems to be relatively small. This meta-analysis found that obesity was associated with a 25 per cent increase in postmenopausal breast cancer and this meta-analysis found a 15 per cent increase in postmenopausal breast cancer and no association with premenopausal breast cancer.

Even if the 87 per cent risk Roberts mentions was relative rather than absolute it would still be much larger than the 15-25 per cent increased risk associated with obesity and obviously much larger than the "little or no" risk associated with smoking. But since the 87 per cent risk is absolute, Jolie's breast cancer risk was off the scale compared to smokers and fatties.

The guiding principle of the Beeb's health reporting is that people shouldn't engage in bad habits so it's okay to misrepresent the facts in an effort to make us change our ways. If that means downplaying major risk factors and exaggerating minor risk factors, then so be it. It is the triumph of morality over medicine. It is not journalism. It is propaganda.


Lysistrata said...

From Rosemary Williams.

Quite. This reporting was woefullly scandalous, grossly insulting to Angelina Jolie's and her partner's decision, and does absolutely nothing to advise women on their risks of breast cancer. In fact, it gives wrong advice.

Carl V Phillips said...

Nice analysis, Chris. The author seems completely oblivious to the fact that the RR from something like BRCA is in the order of 100, compared to the 1.25 you cite for obesity. I suspect that if someone tried to explain that to people like her, their overly coifed but entirely innumerate talking heads would explode.

Many people want health science to be a morality play, but the reality just does not cooperate with them. (Of course, same is true about macroeconomics ;-P.)

Junican said...

Don't these people justify their exaggeration by multiplying the very low risk by the total number of females in the world?
That is very clever, but it does not explain why it is that out of, say, 100 young women who all have similar lifestyles, only one of them actually gets breast cancer. That is the real problem and not lifestyles.

Mark Wadsworth said...


I can't help thinking that the 87% and 50% figures equate neatly to 7/8 and 1/2, which is sort of basic genetic probability predictions (which usually go in 1/4, 1/2 or 1/8 steps), so the 87% ought to be expressed as 87.5%.

I'm guessing that the "5% risk" is either 1/16 or 1/32 but I'm not clever enough to work that out.

Kin_Free said...

According to this multivariate analysis, smoking has a protective effect on breast cancer in those with hereditary BRCA1 and BRCA2 gene mutations. (Brunet et al 1998)

"The magnitude of the risk reduction was large, and the possibility that this finding was due to chance is remote. The risk reduction was greater for women with more than 4 pack-years of smoking than for women with less than or equal to 4 pack-years of smoking. “

Ivan D said...

Roberts places great faith in CRUK spin-doctors so their less than objective output may have influenced her already blinkered view.

Amazingly, CRUK classifies smoking as a "definite" risk for breast cancer on the basis that:

"There is limited evidence that smoking may be linked to breast cancer in women who started smoking before the age of 20 or before the birth of their first child."

It seems that even the feeblest of evidence constitutes "definite" risk in the eyes of CRUK. Either that or their press office have redefined definite. In my experience, "limited evidence" is public health speak used when there is no significant evidence at all but the author desperately wishes that there was.

Having read their whole section on breast cancer I find it worrying that CRUK appears to have no true understanding of risk or the limitations of observational epidemiology.

I agree 100% with your comments on the BBC. The article is appalling and sadly consistent with what we have come to expect from Roberts.

Jonathan Bagley said...

The problem is she relies on Wikepedia, which even lists passive smoking as a cause of breast cancer. The page is controlled by the anti tobacco industry and it's impossible to edit it. There is a new breast cancer research organisation based in Manchester,which has a very good section on risk and does not list smoking as a risk factor.

I think one of those involved must have been the person interviewed on R5L a couple of years ago. The presenter was shocked and disappointed when the professor insisted smoking wis not linked to breast cancer.

Anyone who wants to contribute to a breast cancer charity and doesn't wish to give money to CRUK could consider Breast Cancer Care which, last time I looked, made no mention of smoking on its website.

Seán O'Nilbud said...

After reading yet another below par ghastly piece of quackery on the BBC website I went to the trouble of searching out the "journalist"'s name. Which brought me here. It's terrible what has happened to the BEEB, I loved it so.