|
medpundit |
||
 |
 |
|
Friday, August 02, 2002”After adjustment for known risk factors, there was an increase in the risk of heart failure of 5 percent for men and 7 percent for women for each increment of 1 in body mass index.” This statement is based on a ratio of rates of incidence of heart failure over the years of the study, not on a comparison of the actual incidences. It is one more convolution of the numbers that wasn’t really necessary to get the message across, but which makes the results sound so much more impressive. I’m not even sure how valid it is. What kind of risk goes up? Actual risk? Relative risk? As for the attributable risk, DB thinks that is the meat of the paper: The population attributable risk of heart failure due to overweight was 14.0 percent in women and 8.8 percent in men. The corresponding population attributable risks due to obesity were 13.9 percent in women and 10.9 percent in men. The attributable risk is nothing more than a stastician’s device to make his conjectures seem more dramatic. It is a conjecture with numbers, but numbers don’t lie, so we accept it as truth. But, coming up with that number relies on assumptions about possible contributing factors that the authors had to make. We aren’t privy to those assumptions because they are cloaked behind the calculations. Therefore, I’m not at all sure, as DB is, that 50,000 heart failure cases a year are caused by obesity. My main point about the paper, is not that it didn’t show an increase in heart failure among the obese. It did. My point is that the increase wasn’t as dramatic as we are being led to believe. The data are being maniuplated to give the greatest possible effect, so that the authors can make concluding statements like this: Our findings suggest that obesity is an important risk factor for heart failure in both women and men. Approximately 11 percent of cases of heart failure among men and 14 percent among women in the community are attributable to obesity alone. The contribution of obesity to the risk of heart failure has not been adequately recognized, and our observational data suggest that efforts to promote optimal body weight may reduce the risk of heart failure. Our results are particularly relevant given the alarming trend toward increasing obesity in the United States. The authors have very neatly made it clear that their paper is a significant one on a timely and fashionable subject. It now becomes more likely to be published in a prestigious journal. They are more likely to get another government grant to do yet more obesity research. And, most importantly, they can make the pharmaceutical company that gave them a grant, Roche Laboratories, happy. Roche is the maker of the weight loss drug Xenical. (Once upon a time Servier Amerique, another source of grant money for the study, would have been happy with the results, too. But their drug, Redux, is no more.) Statistics Have Consequences: Statistics are the use of numbers to make a point. They are used not only to validate the money spent by companies and governments on research studies, but to sway public opinion. As such, we should be just as critical of them as we are of ideas. They have consequences. Sometimes big consequences. Take, for example, Newsweek’s story on the whole obesity issue: Now lawyers are filing class-action lawsuits against fast-food makers, charging that deceptive marketing practices encourage obesity. “For years I ate fast food because it was efficient and cheap,” says Caesar Barber, 56, a maintenance worker with heart disease and the lead plaintiff in an anti-fast-food lawsuit filed in New York last week. “I had no idea I could be damaging my health.” This fall, Northeastern University law professor Richard Daynard is holding a closed-door strategy session for nearly 100 lawyers interested in pressing similar claims against Big Fat, or what—in reference to “Big Tobacco”—they’re calling “Big Food.” “Five years ago, when we said we’d take junk-food makers to court,” says Daynard, “people laughed. Forget about personal responsiblity and free choice. Armed with numbers, these lawyers will make sure none of us enjoy a burger again. Don’t think they’ll be satisfied with the hamburger at the fast food restaurant. Hamburger at the butcher counter has to be just as bad, so why not go after the beef industry, too? UPDATE: The Washington Post offers up an example of the consequences of the obesity study statistics. This, from the director of the NHLBI, which provided a grant for the study: Obesity "is almost a primordial risk factor," said Claude Lenfant, director of the National Heart, Lung and Blood Institute. "If you have it . . . you are going to be pulled to all these problems, and eventually . . . to heart failure." Is that right? The study showed a ten year incidence of heart failure in the obese of 6.8%(women) to 10%(men). That means the overwhelming majority of the obese didn't have heart failure in the ten year time frame, not a 100% incidence as the good director implies. UPDATE II: Another consequence of the statistics, this time from the Boston Globe: "Overweight people who gain as little as 4 to 8 pounds can significantly increase their risk of heart failure." They can statistically increase their risk, but not clinically increase it at 4 to 8 pounds. posted by Sydney on 8/02/2002 06:23:00 AM 0 comments 0 Comments: |
|
