But it’s the “data-free” dietary advice that Willett says gets him most upset. During the low-fat craze of the 1980s and ’90s, he railed against experts who didn’t distinguish between the good fat in things like nuts and the bad in red meat and some margarines. From what he saw, “nut consumption went down about 50 percent in the 1990s based on the data-free advice that nuts weren’t good for you,” he says. “People will die because of that bad advice.”
If Willett looks like a Congregationalist minister, he often sounds the moralistic themes of the abolitionist preachers. Like them, he believes the consequences of inaction are unforgivably high. Americans are getting fatter dangerously fast. An obesity rate that was about 13 percent a half-century ago is now 36 percent and climbing, a surge that is bringing runaway rates of diabetes and other maladies.
When Willett sees the CDC study on obesity, his mind flashes to the picture of what he is sure this confusion will lead to: patients packing on more weight while their doctors become less willing to broach the subject. People will get fatter, die sooner, and all the medical bills could cripple the health care system. Willett feels he has no choice but to go on the attack.
“It’s important to push back strongly against the promotion of ideas and analyses that are based on faulty data,” he says.
In May, the respected scientific journal Nature called Willett out, saying he was pushing back too hard and giving him a rare slap on the wrist for his neck-wringing of Flegal. It editorialized that scientists must avoid the temptation to oversimplify, offering black-and-white pronouncements like his “rubbish” comment, based on the fear that more nuanced messages might confuse people into inaction. Gray, the journal said, is often the color of science.
Yet Willett, who presides over decades worth of data on more than a quarter of a million people, sees patterns that make him feel he can’t wait around for things to get sorted out. In that way, he sometimes calls to mind the scientist character from one of those disaster movies. The guy who looks up from his data to offer urgent warnings of doom, only to see them ignored until just before the asteroid hits.
PERCHED AT THE HEAD of a U-shaped table on the ninth floor of the Harvard School of Public Heath, Willett traces his finger along a printout while the dozen people sitting around him remain silent. Then he removes his wire-rimmed reading glasses, folds them up to turn them into a prop for gesturing with his hand, and asks, “OK, is there a prostate cancer representative here?”
This meeting is an update for the Health Professionals Follow-Up Study, one of the three so-called prospective cohort studies that Willett and his team have been running for decades. The data, collected from nearly 300,000 people, are the foundation on which just about all of Willett’s pronouncements about diet have been built.
This study started following more than 50,000 male dentists, veterinarians, pharmacists, and the like in 1986. (About 19,000 of them have died by now, prompting Willett’s colleague Eric Rimm to crack, “It’s a morbid business.”) But it’s not even the biggest study. That honor goes to the Nurses’ Health Study, which began following more than 121,000 female nurses in 1976 and added 116,000 in 1989.
Every two years, the participants in each of Willett’s studies are sent questionnaires asking them about their health, drinking, and smoking habits — with detailed dietary questions coming every four years. This rich corpus of data gives Willett uncommon insight.
The idea behind cohort studies is that all the participants have something in common, which allows researchers to test hypotheses about how they might be different, such as whether the nurses who eat high-fiber diets have lower rates of colon cancer than nurses who don’t. During today’s meeting, one researcher reports that the data show people who drank more coffee lowered, quite markedly, their risk of developing Type 2 diabetes.
After the meeting breaks up, Willett and Rimm tell me they’ve found java to be pretty much uniformly associated with reduced risk of lots of bad stuff. The most dramatic, Willett says, was the finding that women who drank two to three cups of coffee a day had a 60 percent lower rate of suicide than those who drank none. When I ask Willett why he’s not out there recommending that everyone drink several cups a day, he says it’s only because some people will become jittery or suffer insomnia. Continued...