CDC Study Overestimated Annual Obesity-Related Deaths by ‘Tens of Thousands,’ Wall Street Journal Reports
An internal CDC investigation has determined that a "widely quoted" agency study saying that obesity is "gaining fast on tobacco as the leading cause of preventable death and could soon become No. 1," inflated the annual obesity-related death toll by "tens of thousands" because of "statistical errors," the Wall Street Journal reports (McKay, Wall Street Journal, 11/23).
The study, published in the Journal of the American Medical Association in March, is based on U.S. mortality data for 2000. Ali Mokdad, chief of the behavioral surveillance branch at CDC, and colleagues reviewed studies on the role that lifestyle factors have on the development of such conditions as diabetes or stroke to estimate how many of the deaths could be attributed to lifestyle. They also compared their data with a similar study conducted using 1990 U.S. mortality data.
Researchers found that an estimated 400,000 people died from causes related to poor eating habits and sedentary lifestyles -- 33% more than in 1990, when obesity was a factor in 300,000 deaths. In comparison, 435,000 people died from smoking or exposure to tobacco in 2000, up from 400,000 in 1990. Tobacco's share as the cause of total U.S. deaths between 1990 and 2000 dropped from 19% to 18.1%. Researchers found that 16.6% of 2000's preventable deaths could be attributed to poor diet and physical inactivity, up from 14% in 1990.
According to the study, obesity would become the leading cause of death by next year, surpassing 500,000 deaths annually to rival the number of annual deaths from cancer. The study led to an HHS ad campaign on obesity and an increased focus on obesity in research at NIH, which increased its obesity research funding from $378.6 million in 2003 to $400.1 million in 2004 (California Healthline, 3/10).
The study's calculations in determining the number of deaths attributable to obesity have "been the subject of debate and concern" among researchers at CDC and outside experts for more than a year, the Journal reports.
According to the Journal, two CDC researchers completed papers prior to the report's release stating that the traditional method for calculating obesity-related deaths -- which the study's authors used -- is flawed because it does not properly take into account such factors as age and smoking. Even before the papers were published in July and September, "several scientists" at CDC "expressed misgivings" about the study because, among other potential inconsistencies, it used different approaches to calculate deaths related to smoking and obesity, the Journal reports.
According to the Journal, researchers have said CDC's conclusions were "particularly alarming" because the study found only a "modest increase" in tobacco-related deaths. An internal review of the study by CDC found that certain mathematical mistakes, such as using a total mortality number from the wrong year, might have incorrectly added 80,000 to the total estimated 400,000 obesity-related deaths. Such a mistake could have increased the growth rate of obesity mortality by 23 percentage points.
Terry Pechacek, associate director for science in CDC's Office on Smoking and Health, wrote in an April e-mail: "I am worried that the scientific credibility of CDC likely could be damaged by the manner in which this paper and valid, credible and repeated scientific questions about its methodology have been handled. I would never clear this paper if I had been given the opportunity to provide a formal review."
The Institute of Medicine -- at the request of CDC -- in December will hold a two-day workshop on how to factor in risky behaviors in determining mortality causes.
Dixie Snider, CDC's chief of science and the official leading the internal inquiry into the study, said the agency plans to reduce its estimated death total by a yet-to-be-determined number. Snider added that the new results would still show an increase in obesity-related deaths over the 300,000 estimated in 1990. The correction would not change obesity's ranking as the second-leading cause of preventable death in the nation.
CDC Director Julie Gerberding said that warnings from dissenting CDC scientists had not been properly heeded and that the internal review panel has recommended that the agency modify its review process to better address concerns. Snider said that CDC will submit the erratum or correction to JAMA.
Rep. Henry Waxman (D-Calif.) also has asked the Government Accountability Office to conduct its own investigation, which is pending until after CDC releases the results of its internal review.
According to the Journal, the revision to the study, which has been cited "repeatedly" by lawmakers, health officials and drug firms as evidence of an obesity "epidemic," will "further confuse an already roiling national debate" and could "undercut" obesity-related litigation and activists seeking Medicare coverage for obesity-related surgeries.
Gerberding, who co-authored the study but did not conduct the statistical work, said the human errors in the report do not diminish the threat of obesity to public health. "The bottom line is that obesity is a leading cause of death. This paper in and of itself is a very minor contribution to our knowledge of obesity," Gerberding said. She added, "I regret that the internal scientific concerns didn't come to light before the paper was published. We're going to fix" the process.
Snider, noting that CDC is developing an electronic system to make the review process more transparent, said, "It's clear that communication was suboptimal. We don't want to ignore legitimate concerns of scientists." Snider said CDC does not believe the errors were a result of "scientific misconduct," adding, "There's no allegation that anybody had any intent to falsify data. It's an argument about what are the best methods and were the calculations done correctly or not" (Wall Street Journal, 11/23).