Ignorance and Forecasting in Obesity

Tim Hartford of the BBC hosts a show called “More or Less” that fact-checks ostensibly outlandish claims making headlines with “a sock full of statistics,” more or less. The perp (or victim) this week was a report released by the UK-based National Obesity Forum that contradicts other reports to find that the problem of obesity is worsening. Titled “The State of the Nation’s Waistline,” it warns:

The Foresight Report (2007) concluded that half the UK population could be obese by 2050 at a cost of £50 billion per year. However, upward trends in obesity levels suggest these conclusions could be optimistic and could be exceeded by 2050.

Hartford took to task this statement in an episode that interviewed both an author of the Foresight Report and one of the National Obesity Forum. The 2050 forecast was generated using data taken between 1993 and 2004, when obesity and overweight was without a doubt increasing. Since then, though, the 2012 Health Survey of England found obesity rates to be increasing at a slower rate. Not that this is terribly less alarming:

Between 1993 and 2012, there has been a marked increase in the proportion of adults that were obese. This increased from 13 per cent of men in 1993 to 24 per cent in 2012 and from 16 per cent of women in 1993 to 25 per cent in 2012. The rate of increase in obesity prevalence has been slower in the second half of the period than the first half, and there are indications that the trend may be flattening out, at least temporarily. However, obesity in both men and women peaked in 2010 and was at its highest level since 1993.

This finding leads one of the Foresight authors to the reflection that the 2050 forecast was most likely an overestimate. That the author of the report itself makes this note is what motivated Hartford to mug the National Obesity Forum (NOF) with his sock full of statistics. The polite confrontation with the NOF representative yields almost immediately that the basis for sensationalizing the 2050 prediction as an underestimate is predicated on observation and anecdotal evidence, on the order of increasing claims for obesity-related comorbidities like heart disease and hypertension (I would guess these are lagging indicators for the “peak” obesity mentioned by the survey) and the jab that health statistics for obesity project BMI, which he claims underreport obesity.

The issue Hartford takes with the NOF is not so much that they’re wrong, but that they misrepresent the data by not making clear that their projections are based on anecdotal observations. Sidenote: Seldom have I heard stats-savvy folk speak in terms of right or wrong. In a methodology that aims to say something about a population using a sample of data, the cardinal sin is not being incorrect, it is misrepresenting the data, to say anyways what cannot be said about a population based on the sampled data.

Pummeled with socks of stats, the NOF guest is…unperturbed. Maybe stats weigh less than rocks. He thinks the problem of obesity is a major one, a claim hardly in dispute by Tim, the Foresight Report, or the Health Survey. To him, despite its characterization, the NOF report is not a vehicle for reporting data. It is a tool for advocacy. He accepts the criticism, saying “a little exaggeration forces the message home, and that’s what we wanted to do.”

I’m reading anthropologist Adriana Petryna’s ethnography of the Chernobyl meltdown Life Exposed. Tucked into one of her arguments about the way in which the Ukrainian state was founded on claims of biological citizenship is an insight into the instrumentalization of certainties and uncertainties of radiation science.

Ignorance does not represent a negative state of knowledge. Nor does it imply a simplistic lack of access to or unwillingness to recognize the truth. It refers to “a praxis, a method, a path to a certain sort of attitude” … the modern idea of scientific progress will be the sum total of something like ignorance, knowledge and imprecision as an important “intervening phase between simpler truths and more complex ones.”