Why the New York Times Is Wrong About Meat
By Adam Campbell
Men’s Health Features Editor
Apparently, the New York Times hates hamburgers and steaks. I’m not sure how else to explain Jane Brody's recent article, “Paying a Price for Loving Red Meat," which is the number one viewed story on the newspaper's Web site.
My beef with the piece begins with this passage: “. . . a new study of more than 500,000 Americans has provided the best evidence yet that our affinity for red meat has exacted a hefty price on our health and limited our longevity."
That’s a strong statement. And while the story that follows seems to support it, the actual data from the study isn’t nearly as clear cut. But don’t just take it from me: "I think most scientists would consider this evidence extremely weak if not completely inconsequential," says Richard Feinman, Ph.D., professor of biochemistry at SUNY Downstate Medical Center.
Here’s an overview of the study: In 1995, NIH scientists mailed 3.5 million people aged 50 to 71 a nutrition questionnaire that listed 124 foods. The subjects were asked to mark their typical consumption of each item, and return the survey. More than 500,000 people followed through. Then, for 10 years, the researchers tracked who died, and recorded their cause of death—for example, heart disease, cancer, or accident. In all, 71,252 people died.
Now let’s examine the findings. From Ms. Brody’s article: “Extrapolated to all Americans in the age group studied, the new findings suggest that over the course of a decade, the deaths of one million men and perhaps half a million women could be prevented just by eating less red and processed meats, according to estimates prepared by Dr. Barry Popkin, who wrote an editorial accompanying the report.”
Sounds bad, right? Well, maybe it all depends on your perspective—and how powerful you believe this study data to be.
First, you need to remember that we’re talking about, at most, a difference of 3 to 4 ounces of red meat a day. More importantly, you must accept that the scientists have been able to separate the effect of red meat from all other factors. That’s because the people who ate the most red meat also consumed the most total calories, ate the fewest fruits and vegetables, had the highest body mass index, were less physically active, and were more likely to be smokers than those who ate the least red meat. All of which, of course, are associated with an increased risk for heart disease and cancer.
Ms. Brody points out the problem of these confounding variables. However, she’s also quick to discount them: “. . . in analyzing mortality data in relation to meat consumption, the cancer institute researchers carefully controlled for all these and many other factors that could influence death rates.”
In other words, the researchers realized that all of the aforementioned factors affected the data. So they ran the numbers in a such a way to exclude all variables except red meat. Which makes one wonder, Is this even possible? Consider how interconnected higher calorie intake, obesity, and sedentary behavior are with each other and with disease rates, and it simply seems unlikely that they could be extracted in any meaningful, real world way—especially when you add in smoking.
You also have to wonder how accurate the survey data was in the first place. It’s well-documented that people under-report their food intake on nutrition questionnaires. You might not want to admit to drinking three sodas a day, so you write down two instead. Or you feel like you eat healthy most of the time, so you respond with the way you eat on an ideal day. Or you simply aren’t very good at estimating your portion sizes. Why would you be? You’re not a nutritionist.
The scientists know all of this, and try to adjust accordingly. But there’s no way to know how precise the food intake records are. It’s simply a limitation of this type of study. However, it’s not the only problem with the nutrition data.
Ask yourself this: Has your diet changed in the past 10 years? Chances are that it has, and this is one major study limitation that the study authors don’t discuss. The assumption is that all the participants ate the same between 1995 and their time of death, which might have been a decade later. But what are the odds that this is the case? In a European study of more than 27,000 people, subjects reported that they changed their diets within the first two years of the trial. The study authors say that this was most often due to public health recommendations, and the most frequent changes were that people ate more fruits and vegetable and less fat. That’s the very message that was promoted by most health officials in the U.S. from 1995 to 2005.
This isn’t to suggest that changing your diet overnight would alter your mortality risk, but to demonstrate that there are many not-so-obvious factors that make this a hard study to glean any real take away from.
One other detail, straight from the study: “There was an increased risk associated with death from injuries and sudden death with higher consumption of red meat in men . . .” You read right: Eat more red meat, and apparently, you’re at greater risk of being murdered or dying in a car accident. Perhaps too many people are eating quarter-pounders while driving.
I could go through the other studies Ms. Brody mentioned in a similar fashion (including the ones citing long-established scientific facts, many of which are addressed in this story about saturated fat), but it simply becomes an exercise in redundancy. What it comes down to is this: The study here only provides correlations, which means we can’t draw any conclusions from it. Not one. A classic example that I learned in statistics class is the link between ice cream consumption and crime rates. The more ice cream people eat, the more crimes that are committed. Is the ice cream making people crazy? Of course not. Turns out, people eat more ice cream and commit more crimes when the weather is warm. (In case you’re wondering, crime presumably rises because more people are out and about, increasing the interaction of people and boosting the potential for crime-related incidents.)
So while ice cream consumption and crime have nothing to do with each other in reality, they can be statistically related thanks to a third variable. And even though you may be able to adjust for that third variable, you’re still only observing a statistical association, and may be missing some key factor that hasn’t been considered.
It’s also worth noting, as Michael Eades, M.D., author of Protein Power points out, that the same week this paper was published, so was a new report from a long-running British study of more than 64,000 people. It showed no difference in mortality between vegetarians and meat eaters. This contradiction in findings matches the first sentence of the study that Brody reported on. It reads: “Meat intake varies substantially around the world, but the impact of consuming higher levels of meat in relation to chronic disease mortality is ambiguous.”
What does all this mean? Well, Dr. Feinman may have put it best: The study in question is simply “inconsequential." Trouble is, the publicity from the New York Times’ story has made it consequential after all.
May 01, 2009
mhtoday.menshealth.com |
