Just in time for your New Year’s resolutions, the obesity scaremongers at the Harvard School of Public Health cooked up another statistical stew that’s just as hard to swallow as their previous broadsides. In a widely publicized study published last week by The New England Journal of Medicine, they purport to prove that physical activity does not negate the supposedly adverse health effects of obesity.
Of this study’s many flaws, one stands out: its measurement of physical exercise accounted for only the following activities: running, jogging, walking (outside, but not on a treadmill), biking, swimming laps (but not swimming in a lake or the ocean), calisthenics (rowing counts, but not yoga), tennis, and squash and racquetball. So let’s get this straight. A team of Ivy League PhDs used a handful of sports including rowing, racquetball, and squash to measure the total physical activity of the American public.
It appears the authors have never heard of basketball or other team sports, not to mention yard work, work-related exercise, or any of the other physical activities that are part of our daily lives. And considering that the study’s data was collected from more than 116,000 nurses, you might think the authors would have included the often strenuous activities associated with the nursing profession.
Many of this laughable study’s co-authors have a history of allying themselves with food cops and receiving financial support from the manufactures of diet pills. Their collective achievements include:
Writing editorials for The New England Journal of Medicine defending the deadly diet drug Redux without disclosing their financial support from the drug maker.
Serving on the scientific advisory board of the Center for Science in the Public Interest (CSPI), despite being listed on CSPI’s website as a financially conflicted researcher.
Participating in CSPI press conferences.
Authoring the infamous and seriously flawed study that concluded obesity costs the U.S. economy $117 billion each year.
The new Harvard study explicitly set out to undermine a canon of far more powerful and persuasive studies that show physical fitness is a more important indicator of health than fatness. Study after study after study has reported that fitness essentially nullifies the mortality risk of being overweight . Although the Harvard team references a small fraction of these studies, they make no effort to account for the seemingly contradictory results. And for good reason. These studies use actual fitness indicators (like treadmill tests). They don’t rely on self-reported time spent on a handful of arbitrary activities. Click here to see the survey that the Harvard authors used.
To calculate each person’s physical activity, the authors simply added up the total number of self-reported hours spent on all specified activities combined. They made no distinction, for example, between intense running and walking at a moderate pace. An accompanying editorial published in The New England Journal of Medicine makes a similar point:
Physical activity of light-to-moderate intensity, which constituted the bulk of the energy expenditure in most people’s lives, was not included in the analysis by [the Harvard researchers]. This omission might also result in the understatement of the true influence of physical activity on the risk of death. [emphasis added]
For more information on this dubious study, consider these five additional reasons to reject the Harvard team’s apparent claim that obesity is more important than physical activity:
Self-reporting leads to underestimation of the importance of physical activity.
The Harvard authors admit: “Measurement errors in self-reported levels of physical activity might have biased the association between physical activity and mortality.” The accompanying editorial goes further:
[T]he measurement of physical activity may have caused some misinterpretation of the data. A group of fit women with a high body-mass index would be harder to identify with the use of a self-reported physical activity questioner than with a treadmill test … A more difficult issue is the possibility that overweight subjects understated their weights and overstated their levels of physical activity. These biases would tend to cause an underestimation of the influence of physical activity on the reduction of both fatness and the risk of death … one would not expect the level of physical activity, particularly that which is subjectively recalled with at most a moderate level of validity in the best hands, to explain more than a moderate amount of the effect of body-mass index on the risk of death. [emphasis added]
A sedentary lifestyle is worse than being overweight.
According to the Harvard study itself, sedentary “normal” weight people are at greater risk than those who are “overweight” but engage in more than 1 hour per week of the listed activities. Moreover, the study reports no statistically significant difference between overweight and “normal” weight people who spend less than 3.5 hours per week engaging in the specified activities. By lumping together both obese and overweight individuals, the authors were able to overlook these important caveats.
Not all excess deaths among the obese are caused by their obesity.
The authors’ bias becomes apparent when they try to explain away the inconvenient fact that the skinniest people seem to be at a high risk. It’s not being slender that’s the problem, they insist, but some underlying factor like a disease. The authors never acknowledge, however, that the same logic applies to obese people. As we’ve explained before, any number of factors can increase the risk of death among the obese aside from their obesity. In last week’s Washington Post, University of Virginia professor Glenn Gaesser points to two of these factors: “Much of the health problems of the overweight could very well be due to weight fluctuation or the use of weight-loss medication,” he noted.
In fact, the widespread misuse of diet pills throws a monkey wrench in any epidemiological attempt to link obesity with mortality. Addressing an obesity conference in July, acting Food and Drug Administration Commissioner Lester Crawford noted that amphetamines were heavily used in the 1960s and 1970s to control weight. “They were widely prescribed, and there were even manufacturing plants devoted only to them by major pharmaceutical companies,” he pointed out. In her book, Losing It, Laura Fraser includes an astounding statistic: In 1970, eight percent of all pills prescribed in America were amphetamines. As Crawford concluded, “This turned out to be a mini-disaster.” But the Harvard authors make no attempt to account for the “mini-disaster” of amphetamines.
The authors relied on the flawed Body Mass Index (BMI) standard — which (mis)classifies people like Arnold Schwarzenegger and Tom Cruise as obese — to calculate the relationship between weight and mortality.
The editorial accompanying the study says it best:
We need to heed the concept put forth by [Ancel] Keys that body-mass index is a gross measure that does not tell the whole story of fatness. Indeed, we believe that researchers in the field are only now scratching the surface in regard to the physiology of fat and muscle as it relates to human health and disease. It is clear that, if the body-mass index is held constant, the amount of muscle, fat, and fat distribution around the body will vary considerably among individual persons.
The study was limited to a disproportionate number of white women.
The women examined by the study were largely Caucasian. As a result, the authors admit that their “results may not apply to other racial or ethnic groups.” According to several peer-reviewed studies (including at least one published in The New England Journal of Medicine), the connection between obesity and mortality is dramatically smaller for African Americans. A more representative sample would have lowered the overall risk from obesity.
