Sky Digibox
Sky Analogue


Why not buy one of my low-cost "virtual books" that you can download direct to your computer? My books can be read on-screen or printed out on paper. They are full of colour photos and illustrations. Click on "Books", above, for more information.


Why our fears about fat are Misplaced

The war on obesity is based not on sound science but on medical self-interest and cultural hysteria, argues Paul Campos.

New Scientist 11 May 2004

FAT, flab, adipose tissue: call it what you will, it is one of the great obsessions of our age. In the early 1980s, stories about obesity were running in the world's major English-language media at the modest rate of about one per week. BY 2003, reports sociologist Abigail Saguy from the University of California at Los Angeles, the figure had expanded to nearly 20 per day. We are in the throes of an unprecedented "obesity epidemic", doctors, scientists and health organisations repeatedly tell us.

Well, people certainly have been getting fatter in many countries. But here's the conundrum: overall health and life expectancy in these nations continues to improve. Take the US. Between 1990 and 2002 life expectancy here rose from 75.2 to 77.4 years, even though the nation's obesity rate rose, according to the Centers for Disease Control and Prevention, by 61 per cent. Indeed over that same period, the incidence of type 2 diabetes, the supposed bete noire of rising obesity levels, hardly changed, while death rates from heart disease, hypertension and some cancers actually dropped.

Perhaps the timebomb has yet to go off: public health officials often claim we are "about" to see the devastating consequences of obesity. Yet their predecessors were making the same claims 50 years ago - and indeed according to current definitions nearly half the American population was overweight as long ago as 1960.

So what else is going on that might explain the conundrum? In the past few years I have been taking a close look at the claims of those who warn of the supposedly impending global calamity of the obesity epidemic. I have sought out the studies and findings behind the public health pronouncements, and canvassed views from a wide range of experts on what they really reveal. What I have found may prove hard for some to swallow: save for exceptions involving truly extreme cases, the medical literature simply does not support the claim that higher than average weight is a significant independent health risk.

What it actually demonstrates is, first, that the association between increased weight and increased health risk is weak, and disappears altogether when confounding variables are taken into account; and second, that public health programmes which attempt to make "overweight" and "obese" people thinner are, for a variety of reasons, likely to do more harm than good. In short, the current war on fat is an irrational outburst of cultural hysteria, unsupported by sound science.

Sceptical? Here are some statistics from what was, in the 1980s, the world's largest ever epidemiological study. From records of nearly 2 million Norwegians spanning a decade, it found the highest life expectancy among people with a body mass index (BMI) of 26 to 28 - people who were solidly overweight, according to definitions now used by, among others, the World Health Organization and the American public health establishment. Furthermore, the study found people with a BMI of 18 to 20 (almost all of whom these same institutions would classify as "ideally thin") had a lower life expectancy than those with BMIs between 34 and 36: who under current classifications were 60 to 75 pounds (25 to 35 kilograms) overweight, and therefore seriously obese

As my book "The Obesity Myth" describes in detail, these statistics are typical of such studies. Large-scale studies consistently find little or no increase in mortality risk associated with weight, except at statistical extremes. Indeed, many such studies find the lowest mortality rates among supposedly overweight people, and a higher mortality risk among people who are five pounds "underweight" than among those who are supposedly 75 pounds overweight. And even these modest associations disappear when variables other than weight are taken into account. BMI figures correlate with increased mortality risk only among sedentary individuals: people who maintain quite modest activity levels show no such correlation.

But aren't fat people who slim down and stay slim healthier in the long run? Nobody can say, because no long-term study has tested the notion. And why not? For the simple reason that researchers have been unable to produce significant long-term weight loss in statistically meaningful numbers of people. Researchers repeatedly find instead that for the vast majority of people the long-term result of attempts to lose weight is "weight cycling" - repeatedly losing weight and then putting it on again. And that, as numerous studies have found, is no recipe for long-term health. For example, the famous Framingham study in the US found a strong correlation between weight cycling and mortality, and no increased risk among obese people who did not weight cycle.

Given all this, how do the public health officials who wage the increasingly intense war on fat support their arguments? The answer should disturb anyone who believes science is immune to the effects of economic and ideological bias. For on close inspection, the current panic over obesity is based on a severe distortion of scientific data.

The distortion tends to take one of three common forms. The first goes to the heart of the nature of epidemiology, which is, in the words of Harvard researcher Charles Hermekens, "a crude and inexact science" . Epidemiologists, he says, "tend to overstate findings, either because we want attention or more grant money". (Ironically it is Hermekens whose work has been exploited by some of the worst fomenters of fat panic.)

Large-scale observational studies can never control for more than a few of the many factors that might explain the associations they observe between risk factors and disease, so epidemiologists who are careful not to overstate their findings will point to the dangers of attributing causal significance to small risk associations. A common rule of thumb is to view with suspicion any factor that fails to at least double relative risk, especially when the baseline risk is small.

Yet in the case of obesity, such caution is routinely abandoned. Health officials and researchers all too often treat small risks as presumptively causal. One recent study, for example, found a 13 per cent increased risk of post-menopausal breast cancer associated with being overweight. It added up to just one extra death per 10,000 "overweight" women per year. The authors still treated the finding as strong evidence of a causal relationship between weight and cancer.

A second way in which health officials - and some researchers - routinely distort the facts on obesity is through ignoring confounding variables. There is overwhelming evidence that many causes of ill-health disproportionately affect the heavier than average person - sedentary lifestyle, nutrition, weight cycling, poverty, access to and discrimination in healthcare, social discrimination generally. Yet many prominent obesity researchers prefer instead to pin the blame exclusively on body fat.

A particularly egregious example of this is a famous study that appeared in the Journal of the American Medical Association YAMA) in 1999, which concluded that excess weight was killing 300,000 Americans per year. This 'fact" has been cited more than 1700 times in the major English language media over the last two years alone. In their statement of methods, the authors noted: "Our calculations assume that all (controlling for age, sex and smoking) excess mortality in obese people is due to their adiposity"! While assuming the validity of one's conclusion certainly simplifies the process of scientific investigation, research ought to be about more than confirming hypotheses through circular reasoning.

The final distortion is to adopt flexible standards of proof. When those prosecuting the war on fat encounter findings that appear to confirm their views, they often dismiss any attempt to question them as ignorant, irrational or biased. Yet with findings that contradict their views, this confident positivism vanishes. The contradictory findings are explained away by an almost endless assortment of methodological caveats.

This leaping back and forth between uncritical faith and profound scepticism is particularly striking when researchers perform interpretive acrobatics with their own findings. For example, the authors of a 2003 JAMA study concluded that it provided compelling evidence of the deadly effects of higher than average weight. The study actually found negligible mortality increases among whites with BMIs up to the mid 30s, and no evidence of elevated mortality rates among black Americans across the "overweight" range (BMI 25 to 30). Indeed, among black women, no extra mortality risk was observed until a BMI Of 37. But an accompanying editorial, by JoAnn Manson of the Harvard Medical School, commented that "It would be a great disservice to blacks if these results were used to promulgate the concept that excess weight is not harmful to them"

Ultimately, the current panic over increasing body mass has little to do with science, and everything to do with cultural and political factors that distort scientific enquiry. Among those factors are greed (consensus panels put together by organisations such as the WHO that have declared obesity a major health crisis are often made up wholly of doctors who run diet clinics), and cultural anxieties about social overconsumption in general.

Consider this. While the average American is about eight pounds heavier than in 1990, the average American car now weighs several hundred pounds more than it did in that year. Which statistic has more relevance to the world's long-term health?

Paul Campos is a professor of law at the University of Colorado. The Obesity Myth is published by Gotham Books (

1 May 2004 - NewScientist 121