Tuesday, 3 September 2013

When obesity prophecies fail

Remember this policy-based doom-mongering from 2011?

'Tax unhealthy foods or half will be obese by 2030'

Ministers should tax unhealthy foods, experts urge today, as they warn that nearly half of all adults in Britain will be obese by 2030

In the end, the government didn't tax 'unhealthy' foods. Instead they worked with industry on a voluntary agreement that the public health authoritarians hated. Most of those authoritarians have subsequently thrown their toys out of the pram and stormed out of the room.

So what happened next? Did obesity rates continue to spiral? Are we still on course for having a male obesity rate of 48 per cent in 17 years time?

Er, not quite...

Britons have stopped getting fatter

Adults in Britain have stopped getting fatter after years of rising obesity rates, a new study has shown.

The study that has inspired this headline is only catching up with what the rest of us noticed a while ago—that rates of obesity in Britain started levelling off ten years ago. The "half will be obese by 2030" prediction came from a 2011 study in The Lancet which ignored the trend of recent years to produce a policy-based prediction that never looked credible.

The central claim of that study (which I wrote about at the time) was that 48 per cent of men would be obese by 2030. But if one looks at the male obesity rate in England since the turn of the millennium (below), nothing in the current trend suggests that obesity will rise to anything like that level (the female obesity rate shows a similar picture).

The trick The Lancet pulled was to extrapolate from the data of the 1980s and 1990s—when the obesity rate undoubtedly did rise significantly—while ignoring the current trend which is nearly flat since 2000 and completely flat since 2006.

There is, in any case, no reason to assume that obesity rates will rise in a linear way, but even if there was, there's no reason to think that historic data are a better guide than recent data.

Although it's too early to say for sure, The Lancet's prediction looks like being just another in a long line of vastly inaccurate obesity forecasts which are quietly forgotten when prophecy fails. Back in 2006, for example, it was predicted that 33 per cent of men would be obese by 2010. Currently, the male obesity rate is 23.6 per cent—which is exactly what it was in 2006. It's going to have to double in the next 17 years for The Lancet prediction to be correct. Considering that the rate has increased by just 2.6 percentage points since 2000—and by zero since 2006—that seems a tad unlikely.

For more appalling obesity predictions, see this old post.


Fredrik Eich said...

Well, smoking prevalence stalled five years ago! Must have some effect.

Jean Granville said...

Just redefine obesity and the prediction will prove correct.
And yes, I'd be curious to confront smoking rates and obesity rates. Some kind of reversed correlation would't be surprising.

Christopher Snowdon said...

I haven't plotted obesity against smoking prevalence but there's no doubt that nonsmokers weigh more than smokers on average.


Ivan D said...

Klim McPherson lectured Statistics at Oxford so we might expect better accuracy from him.

He does like to bash industry from his ivory tower and I suspect that his calls for punitive measures stem from his personal politics.