Food for thought: will that hot dog kill you?

Wednesday 14 March, 2012
Fantastic news yesterday: swap your breakfast bacon for a kipper and you might live forever.
You think I'm kidding? Well it's in the papers and it comes out of a Harvard study, so it surely must be true.
Replacing red meat with poultry, fish or vegetables cut the risk of dying by up to a fifth
writes the Telegraph's medical editor.
Replacing one serving of red meat with an equivalent serving of fish reduced mortality risk by 7 per cent.
reported the Daily Mail.
So, too, in various forms did the Guardian, the Independent, the Chicago Sun-Times, the Sydney Morning Herald and the American news networks.
But you'd be forgiven for missing this earth-shattering gem as it generally appeared halfway through a story about how eating roast beef will kill you - and how eating sausages will kill you faster. A story that the Telegraph splashed under the heading
Red meat is blamed for one in ten early deaths

Now I always thought that death was 100 per cent certain. So the news that a bit of fish
might cut that risk to 93 or 80 per cent (depending on which paper you're reading) is pretty amazing.
On the other hand, it seems that if you have bacon for breakfast or a hotdog for lunch every day you'll be dead in a fortnight:
For each serving of processed meat...the risk of dying from heart disease rose by 21 per cent.

The risk increases by 21% with every serving? Surely not?
Of course I'm being facetious. But these linguistic infelicities demonstrate the dangers in translating scientific research into a news story - and there were many pitfalls in this story.

It's understandable that the findings of Frank Hu's team at the Harvard School of Medicine should be afforded international coverage. A huge study - covering 120,000 people over more than twenty years - by the world's foremost university demonstrates that a fundamental element of the Western diet could be lethal. Who could question the results?
But how many reporters actually read the paper produced by Mr Hu and published in the Archives of Internal Medicine along with 30-odd references and footnotes? And how many simply rewrote agency copy or a press release? The chances are that at least some were relying on a precis of a precis of a precis.
Did they speak to Mr Hu? Did they raise any questions? Well, the Telegraph, the Guardian, the Mail and the American and Australian news media all have the self-same quote from Mr Hu, so either he's extraordinarily consistent in his use of language or it came from a handout.

The conclusions reported across the media were that eating 3oz of red meat every day meant you were 13 per cent more likely to die "early" than if you didn't; that if you ate processed meat such as sausages or bacon, that risk was even greater. If you ate fish, poultry or vegetables instead, you would live longer. And if one in ten men had halved their meat intake they wouldn't have died.
The reader was not told who took part in this study, what age or race they were, what their general health was, how the research was conducted, what might constitute an "early" death.

And so to the research.
The Harvard team observed 37,698 male health professionals from 1986 (when they were aged between 40 and 75) until 2008. It also studied 86,644 female nurses from 1980 (then aged between 30 and 55) until 2008. None showed signs of cancer or heart problems at the start of the investigation.
Over the period of the study 8,926 of the men and 15,000 of the women died. In each case about a third of the deaths were from heart disease or cancer.
So what we know for sure is that coronary disease or cancer killed nearly 6,000 men before they reached the age of 97 and about 9,500 women before the age of 83.
We also know from the published paper, though not from the press, that the participants were overwhelmingly white.
The sample is therefore 120,000 white Americans who all work in the same industry - one that is known for stress, long hours and disturbed sleep. They may therefore start off with a raised (or, indeed, reduced) risk of heart disease or cancer. And while having a huge sample from the same background may be useful as a constant, how can we assume that the results will apply to white Americans in all walks of life, to black bankers, Latino factory workers or Asian businessmen, let alone the entire Western population?

The methodology was to ask the participants to fill in a questionnaire every four years, the key questions being how often they consumed various foods of a "standard portion size", with possible answers ranging from "never" to "more than six times per day".
For red meat, such as roast beef, the standard portion size was set at 3oz - which happens to chime roughly with health officials' view of the recommended maximum daily intake.
I'm not sure how I would have answered that. I don't eat meat every day, but when I do I'd look on 3oz as a pretty dismal portion. So say you have 6oz of roast lamb on Sunday, 3oz of cold meat in sandwich on Monday and an 8oz steak on Wednesday, how do you report that?
The answers were analysed with allowances made for "variables" such as age, weight, whether the subject smoked or exercised or took vitamin pills or aspirin. This led the researchers to the assertion: "Men and women with higher intake of red meat were less likely to be physically active and were more likely to be current smokers, drink alcohol and to have higher body mass intake."

From the analysis, the researchers concluded:
We estimated that substitutions of 1 serving per day of other foods (including fish, poultry, nuts, legumes, low-fat dairy and whole grains) for 1 serving per day of red meat were associated with a 7% to 19% lower mortality risk. We also estimated that 9.3% of deaths in men and 7.6% in women in these cohorts could be prevented at the end of follow-up if all the individuals consumed fewer than 0.5 servings per day of red meat.
So they are not saying that red meat is to blame for one in ten deaths. They are saying that if all 37,000 men had cut their meat consumption to half the recommended maximum level, 538 of them - some of whom could have been in their nineties - might not have died over the course of 22 years.

And what of that generalisation about smoking, drinking and general couch potatoery? Can you just discount all those factors as "variables" that have been taken into account without explaining how?

I once worked for a national newspaper editor who wearied of the endless Westminster soap opera being seen as the default splash. "Haven't we got a good health scare?" he would ask.
He knew what sold papers. The thousands of comments appended to the various online reports of this study bear that out, so you can't blame the Telegraph for latching on to it; the only surprise is that the Mail and Express went with the drought instead. It's just that even the most authoritative research may not be all that it seems.

Mr Hu, however, seems to know when he's on a winner. He simultaneously published another study - in another journal - that found that one can of fizzy pop a day increased the risk of heart disease in men by 20%. The subjects of his study? 38,698 health workers who were observed between 1986 and 2008. Fancy that!
Now the thing is, were these cases of heart disease really the result of fizzy drinks - or the consequence of eating too many hotdogs?
Or maybe that sweeping generalisation in the meat study was the true nugget of common sense: if you're a lardarse boozer who eats too much junk food, drinks too much Coke and doesn't move about enough, you may not make it to 97.