What's Wrong With Fat? (24 page)

In fact, news reports on the Fat-OK study are
more
likely, than news reports on the Eating-to-Death study, to discuss the health risks of “extreme
”
or “morbid
”
obesity, typically arguing that, while the health risks associated with overweight and moderate obesity may have been overblown, there remain serious health risks associated with higher levels of obesity: “The report from Centers for Disease Control and Prevention scientists didn’t challenge the long-held view that being at least 50 pounds overweight is risky. But it did suggest that being even as much as 40 pounds overweight seems to protect people more from an untimely death than being at a normal weight.” 75 In other words, the news media tend to maintain that weighing “too much” is unhealthy but that what constitutes “too much” may be higher than previously thought.

Importantly, news reports on the Fat-OK study frequently air scientific debate over the health risks associated with overweight/obesity and over the appropriate cutoff for these categories. In contrast,
none
of the news reports on the Eating-to-Death study mention any controversy over BMI categories, despite the fact that there was already considerable debate about this among researchers. 76 Specifically, 49 percent of news reports on the Fat-OK study explicitly reject the estimate that overweight and obesity are associated with 400,000 annual excess deaths, 36 percent discuss scientific debate over the health risks associated with overweight/obesity, and 39 percent discuss scientific debate over where the cutoff for overweight or obesity should be drawn. Moreover, 33 percent of news reports on the Fat-OK study discuss that “being very thin increased the risk of death, even if the thinness was longstanding and not due to illness,” 77 and 13 percent suggest that “it is fitness not fatness that matters.” 78 In contrast,
none
of the news reports on the Eating-to-Death study discuss either of these topics. A typical news report on the Fat-OK study states: “Published in the Journal of the American Medical Association last week the study reports that excess weight accounts for a mere 25,814 deaths in the United States each year. In 2004, the CDC placed that estimate at 400,000 deaths.” 79
Emphasizing scientific controversy, as in the following quote, is common in the new reports on the Fat-OK study: “Health experts increasingly are faulting a recent study by scientists at the federal Centers for Disease Control and Prevention that concluded obesity is not nearly as dangerous as was thought and that being a little plump might actually lower the risk of death.” 80

In other words, the publication of the Fat-OK study provided an opportunity for a different kind of news media reporting—one that emphasized scientific uncertainty and debate over the health risks of higher body mass—to emerge. These analyses demonstrate that when a study challenging the public health crisis frame was published by high-status scientists in a leading medical journal and endorsed by a public authority like the CDC, this, in turn, led to a shift away from the typical news media framing of obesity as a public health crisis. That is, the news media are responsive to scientific research. However, the ultimate impact of specific scientific studies is also likely to be shaped by whether the news media present them as credible or suspect, definitive or tentative. The news media have the potential either to shore up or to undermine scientists’ credibility before a lay public. In reporting on this controversy, how did the news media weigh the evidence and evaluate credibility?

WEIGHING CONTROVERSY, ASSESSING CREDIBILITY

News media reporting is more likely to question the credibility of the Fat-OK study, despite the fact that the CDC ultimately deemed it to be the better study. As is shown in figure 4.4, the news media are significantly more likely to present the Eating-to-Death study as providing additional evidence of what we all know: that obesity and overweight are deadly, while being significantly more likely to present the findings from the Fat-OK study as surprising. This matters because it serves to introduce doubts about the validity of the Fat-OK study that are likely to affect its ultimate impact. Specifically, more than 75 percent of news reports on the Eating-to-Death study suggest that the findings confirmed what was already known, compared to less than 10 percent of the news reports on the Fat-OK study. In contrast, more than one-third of the news reports on the findings from the Fat-OK study, but less than 3 percent of the Eating-to-Death study, frame its findings as surprising.

Part of this is surely a product of sequencing. Whereas the first study tended to be evaluated on its own terms, the second was read in relation to the first. That is, whereas the first study attracted attention because it strengthened the powerful narrative of obesity as a public health crisis, the second was newsworthy because it created a scientific controversy. In other words, the findings of the Fat-OK study were more likely to be seen as surprising because they ran counter to the Eating-to-Death study. This is important, as the representation of the Fat-OK study as running counter to previous expectations invites skepticism, which, in turn, could temper its social impact.

In fact, the Fat-OK study was consistent with a huge body of epidemiological literature showing that risk of mortality follows a U-shape curve, with lowest risk in the middle of the distribution (which, in the United States, falls in the overweight range) and highest risk at the high and
low
extremes of BMI. For instance, a 1998 article based on a 20-year prospective study of a nationally representative sample of U.S. adults aged 55 to 74 years suggests that the lowest mortality occurs in the BMI range of 25 to 30. After adjusting for smoking status and preexisting illness, lowest mortality occurred at a BMI of 24.5 in white men, 26.5 in white women, 27.0 in black men, and 29.8 in black women. 81 Scores of other studies of U.S. and non-U.S. populations, published before and since the Fat-OK study, have similarly concluded that there is either reduced or no excess mortality risk among those with a BMI greater than 25 and less than 30 (the overweight category) and sometimes in the obese category as well. 82

Figure 4.4:
Percentage of News Samples Quoting Skeptical or Supportive Scientists and Presenting
Study as Confirming What Is Known or as Surprising.

Why, if the results of the Fat-OK study were consistent with numerous other studies, are they treated as surprising? Why, given this large body of literature, did the Eating-to-Death study not elicit more skepticism? One reason is that the non-association between overweight and mortality is rarely highlighted in scientific studies or brought to the attention of journalists. Rather these findings tend to be buried in the article. For instance, table 2 in a 2006 study published in the
American Journal of Epidemiology
shows that the relative risk of all-cause mortality for people in the overweight category (BMI greater than 25 but less than 30) was not significantly different than those in the “normal weight” category (BMI greater than 18.5 but less than 25). 83 However, neither the abstract nor press releases, on which journalists heavily rely, mention this information. 84 This is consistent with a more general pattern, in which null findings are less likely to be mentioned in the abstract, when they are published at all. 85 This allows the Eating-to-Death study findings to seem unsurprising and the Fat-OK findings to seem surprising, despite the fact that the latter were, in fact, consistent with a huge body of literature. Indeed, the Fat-OK study was novel, not because it demonstrated mortality rates are lower among those in the overweight range, but because it reported these results completely evenhandedly. Moreover, this study was innovative in that it translated the lower relative mortality risk associated with being in the overweight category into a positive finding: an estimate of almost 90,000 lives
saved
in the year 2000 due to overweight.

This finding, however, was subject to greater scrutiny than the findings from the Eating-to-Death study. Specifically, as is shown in figure 4.4, more than 30 percent of reports on the Fat-OK study, but
none
of the news reports on the Eating-to-Death study, quoted outside researchers who contested the study’s validity. Journalists may have been more likely to seek out researchers who were skeptical of the 2005 study to develop a compelling narrative of scientific controversy. However, several researchers also made a concerted effort to discredit the Fat-OK study in the news media.

Most notably, Harvard’s School of Public Health, home of Walter Willett and JoAnn Manson, issued a press release calling the Fat-OK study “flawed.”
Drawing on the press release, a news report quoted Manson dismissing the research findings and reaffirming the assertion that obesity is indeed a deadly epidemic: “‘We really can’t afford to become complacent about this epidemic of obesity and certainly not based on findings from an analysis that is flawed,’ said Dr. JoAnn Manson, chief of preventive medicine at Brigham and Women’s Hospital in Boston.” 86 Manson and Willett also held a one-day conference on May 26, 2005, dedicated to criticizing the Fat-OK study and conducted a poll showing that three-fourths of Americans “rate obesity as an ‘extremely’ (34 percent) or ‘very’ (41 percent) serious public health problem in the United States.” The poll showed that “the majority of Americans believe that scientific experts have been portraying accurately (58 percent) or even underestimating (22 percent) the health risk of being obese” and that “very few Americans reported believing that the health risks were being overestimated by scientific experts (15 percent).” 87 These Harvard initiatives generated considerable news media coverage. In other words, part of the framing of the Fat-OK study as controversial and possibly flawed reflected the ability of competing researchers to influence journalistic reporting.

The fact that the Fat-OK study was subject to greater skepticism than the Eating-to-Death study may also be due to what cognitive psychologists call confirmation bias. 88 Confirmation bias leads people to be less critical of evidence that supports their preconceptions and more critical of evidence that challenges them. To the extent that both scientists and journalists are steeped in an obesity-crisis narrative, they are likely to uncritically accept evidence that is consistent with this narrative and to be skeptical of findings that run counter to it. In that the simplicity of the obesity-crisis narrative (fat equals bad) is part of its appeal, the more nuanced argument presented in the Fat-OK study may also account for some of the skepticism it met among reporters.

Research on confirmation bias also helps make sense of how many claims about a looming obesity public health disaster have been uncritically accepted and disseminated, even among scientists themselves, despite their shaky empirical basis. For instance, one of the most widely circulated claims, that the steady rise in life expectancy during the past two centuries will soon come to an end because of current trends in obesity, is based on what one of the studies’ authors has called “back-of-the-envelope, plausible scenarios” that were never meant to be portrayed as precise. 89 Despite the fact that this report offers no statistical analyses to back up its estimates and relies on a handful of convenient, but false, presuppositions about the BMI of Americans and mortality risk, it has been cited more than 1,000 times in the academic literature, based on Google scholar. 90 This prediction, that “unless effective population-level interventions to reduce obesity are developed, the steady rise in life expectancy observed in the modern era may soon come to an end and the youth of today may, on average, live less healthy and possibly even shorter lives than their parents,” continues to be treated as fact. 91 The “fact” that this generation of children will be the first to live shorter lives than their parents is cited in countless news media accounts, by political figures including Michelle Obama, and in television shows including Jamie Oliver’s
Food Revolution
.

THE LINK BETWEEN CRISIS AND BLAME

The corollary of Rubin’s claim that the search for a cause reveals a desire to eliminate the underlying phenomenon is that there is less discussion of possible causes when there is more acceptance of a phenomenon. 92 Indeed, as is shown in figure 4.5, news reports on the Fat-OK study are considerably less likely to discuss a variety of different possible causes of obesity, including individual-level factors in general (13 versus 89 percent) and, more specifically, bad food choices (7 versus 89 percent) and sedentary lifestyle (7 versus 86 percent), but also sociocultural factors in general (8 versus 71 percent) and the food industry specifically (3 versus 54 percent). The only exception, which is not shown in figure 4.5, is genetic contributors, which are discussed in 5 percent of news reports on the Fat-OK study and 3 percent of news reports on the Eating-to-Death study, a difference that is not statistically significant, meaning it could have been the product of chance.