What Does Science Say You Should Eat?

Most diets aren't realistic or advisable, including the U.S. agriculture department's famous food pyramid. Instead, a Harvard scientist recommends a new way of eating based on the world's largest and longest food study.

By Brad Lemley and Dan Winters
Feb 5, 2004 6:00 AMJun 28, 2023 3:32 PM


Sign up for our email newsletter for the latest science news

America clearly needs dietary guidance.

More than 44 million people are clinically obese compared with 30 million a decade ago, putting them at increased risk for heart disease, stroke, type 2 diabetes, and breast, prostate, and colon cancers. In the meantime, the noun diet seems to attract a different adjective every week, including Atkins, Ornish, Cooper, grapefruit, rice, protein, Scarsdale, South Beach, Beverly Hills, Best Chance, Eat Smart, and Miracle, not to mention Help, I’m Southern and I Can’t Stop Eating. While some of these plans overlap, others seem to specifically contradict each other, notably the meat-intensive regime of the late Robert Atkins versus the near-vegetarian program of Dean Ornish.

No wonder Americans are tempted to follow Mark Twain’s admonition to “eat what you like and let the food fight it out inside.” But still, we wonder: Is there really an optimum way to eat?

Although debate rages, academic nutrition researchers have begun to form a consensus around a plan with an important advantage—it is based on a preponderance of sound science. The regime does not as yet have a name, but it might well be called the Willett diet, after its leading proponent, Walter Willett, chairman of the department of nutrition at the Harvard School of Public Health.

Featuring abundant fruits, vegetables, whole grains, and vegetable oils, as well as optional portions of fish and chicken, Willett’s plan resembles the much-touted Mediterranean diet shown in several studies to reduce the risk of heart disease. Nonetheless, Willett resists the comparison. “The Mediterranean diet is specific to a certain climate and culture,” he says, adding that by focusing on healthy ingredients rather than specific dishes, “anyone can adapt this plan to his own tastes.” The results: stable blood-sugar levels, easier weight control, clearer arteries, and overall better health.

In this case it’s hard science, not just opinion. Willett’s plan is based on the largest long-term dietary survey ever undertaken: the 121,700-participant Nurses’ Health Study, begun in 1976 by Harvard Medical School professor Frank Speizer, with dietary assessments supervised by Willett since 1980. The study isn’t just big: Willett carefully crafted it so that he and others could extract specific recommendations about food intake.

Participants even surrender blood and toenail samples so that Willett can track absorption of trace elements and other nutrients. If a participant reports a major illness, such as heart attack or cancer, “we write for permission to obtain medical records for further details,” says Willett. To ensure that the data include both sexes and two generations, Willett and several colleagues also launched the Health Professionals Follow-Up Study, which includes 52,000 men, and the Nurses’ Health Study II, a survey of 116,000 younger women.

In the past, nutritional scientists have largely relied on studies of animals, small groups of people, and/or petri-dish biochemistry that may not reflect the vagaries of human metabolism, although Willett uses such studies when he deems it appropriate. His access to a unique quarter-million-person pool of humans who carefully track both their diets and their health lends added credibility to his research. “When you put animal, metabolic, and epidemiological studies together and they all point in the same direction, you can be pretty confident about your conclusions,” Willett says.

While soft-spoken and self-effacing in person, Willett isn’t shy about using this formidable database to take on the federal establishment. His Healthy Eating Pyramid differs radically from the Food Guide Pyramid pushed by the U.S. Department of Agriculture. “At best, the USDA pyramid offers wishy-washy, scientifically unfounded advice,” Willett argues in his book, Eat, Drink, and Be Healthy: TheHarvardMedicalSchool Guide to Healthy Eating. At worst, he adds, “the misinformation contributes to overweight, poor health, and unnecessary early deaths.”

The numbers back him up. Men and women in Willett’s studies whose diets most closely paralleled the Healthy Eating Pyramid’s guidelines lowered their risk of major chronic disease by 20 percent and 11 percent respectively, according to an article published in the December 2002 issue of The American Journal of Clinical Nutrition. That compares with reduced risks of 11 percent and 3 percent for those whose diets most closely mirrored the USDA pyramid’s guidelines.

“Nutrition used to be like religion. Everyone said, ‘I have the truth, everyone else is wrong,’ and there wasn’t much data to refute that,” says Willett. “Now we’re starting to have a real scientific basis for understanding what you should eat.”

Just inside the door of Willett’s office at the Harvard School of Public Health in Boston sits his bicycle, mud-spattered from his daily commute over the Charles River from his home in Cambridge. Past that, on top of a pile of medical journals, perches a plastic bag full of plump, homegrown cherry tomatoes, a late-season-harvest gift from his administrative assistant. Willett knows good tomatoes. As a member of a fifth-generation Michigan farming family, he paid his undergraduate tuition at Michigan State by raising vegetables, and today he grows “as much as possible” in his tiny urban backyard.

Behind the cluttered desk sits Willett himself, trim, toned, and turned out in a sharp gray suit. “All you have to do is take a look at Walter to see the value of his research. The proof is in the pudding,” says David Jenkins, a nutrition researcher at the University of Toronto. Willett vigorously follows his own plan and at age 58 reports that his weight, cholesterol, and triglycerides are all where they should be. He is, in short, the picture of where applied nutritional science might deliver us all, if we had the proper information.

That’s the problem. In recent years, Willett says, the American public has been victimized by dodgy advice. Not only has obesity skyrocketed but “the incidence of heart disease is also not going down anymore. It has really stalled.”

What happened? In Willett’s view, things began to go awry in the mid-1980s, when a National Institutes of Health conference decreed that to prevent heart disease, all Americans except children under 2 years old should reduce their fat intake from 40 percent to 30 percent of their total daily calories. The press touted the recommendation as revealed truth, and the USDA’s Food Guide Pyramid, released in 1992, reflects this view, calling for 6 to 11 servings of bread, cereal, rice, and pasta daily, while fats and oils are to be used “sparingly.”

Too bad, says Willett. “The low-fat mantra has contributed to obesity. The nutrition community told people they had to worry only about counting fat grams. That encouraged the creation of thousands of low-fat products. I call it ‘the SnackWell revolution.’” Blithely consuming low-fat foods full of carbohydrates is a prescription for portliness, says Willett, adding that any farmer knows this. “If you pen up an animal and feed it grain, it will get fat. People are no different.”

The problem with overeating refined carbohydrates such as white flour and sucrose (table sugar) is that amylase, an enzyme, quickly converts them into the simple sugar called glucose. That goads the pancreas to overproduce insulin, the substance that conducts glucose into the cells. But excessive sugar is toxic to cells, so after years of glucose and insulin overload, the cells can become insulin resistant and may no longer allow insulin to easily push glucose inside them. That keeps blood glucose levels high, forcing the pancreas to make even more insulin in a desperate attempt to jam the stuff through cell membranes. Willett likens the effect to an overworked, undermaintained pump that eventually wears out. Type 2 diabetes can be one result, but Willett contends that insulin-resistant people who don’t develop full-blown diabetes still face significant health risks.

Other researchers agree. Stanford endocrinologist Gerald Reaven coined the term Syndrome X to describe the constellation of health problems that spring from insulin resistance. Until the late 1980s, Reaven says, “the common scientific view was that insulin resistance only mattered if it led all the way to type 2 diabetes. Looking at the data, it’s clear that most people who are insulin resistant don’t get diabetes but are greatly at risk for coronary heart disease, hypertension, non-alcoholic-type liver disease, polycystic ovary syndrome, and several kinds of cancer.”

In the case of heart disease, Reaven says that high blood concentrations of insulin and glucose can damage the endothelium that lines coronary arteries and set the stage for the formation of plaques. “A big problem is the lack of drugs to treat this problem,” he adds. “A lot of doctors’ education comes from drug companies. They know about cholesterol because everyone is pushing their statin. They know about hypertension because there are multiple hypertensive drugs. But they know a lot less about insulin resistance and its consequences, and that’s unfortunate.”

Syndrome X, also known as metabolic syndrome or insulin-resistance syndrome, is largely unknown to the public as well. While many people avoid cholesterol and fat-laden foods, few understand the threat posed by carbohydrate excess. That needs to change, says Willett. “Cholesterol is relevant, but the danger is overblown,” he says. “Syndrome X is the global public-health problem of the 21st century. Almost certainly the vast majority of Americans have a higher degree of insulin resistance than is optimal.”

The Willett plan aims to even out the glucose roller coaster through an emphasis on foods with low glycemic loads—foods that convert to glucose slowly—like whole grains, plant oils, and vegetables. This keeps blood glucose levels relatively constant, sparing the pancreas overwork. Steady blood glucose also helps keep the appetite in check, which makes maintaining a healthy weight easier, says Willett. So instead of high carb, low fat, one might summarize the Willett plan’s directive as good carb, good fat.

“People are being told to reduce fat and eat more carbohydrates. For many people, particularly overweight people with a high degree of insulin resistance, that produces exactly the opposite of what they need,” says Willett. Randomized trials, he says, show that people on low-fat diets generally lose two to four pounds after several weeks but then gain back the weight even while continuing the diet. “Most of them would be better off reducing carbs, switching to better carbs, and increasing their intake of healthy fats.”

Willett, like virtually every other nutrition researcher, advises eating vegetables in abundance, consuming alcohol in moderation, and taking a daily multivitamin to cover nutritional gaps. He also touts fish as a source of protein and heart-protective n-3 fatty acids, which are also known as omega-3 acids. (Those who worry about mercury contamination in fish got some good news recently: In one study conducted in the Seychelles, a group of islands in the Indian Ocean, scientists from the University of Rochester Medical Center tracked pregnant women who ate an average of 12 fish meals a week, about 10 times the quantity of fish eaten by the average American. “We’ve found no evidence that the low levels of mercury in seafood are harmful,” said lead author Gary Myers. Moreover, various tests indicated that the women’s children suffered no adverse cognitive, behavioral, or neurological effects.)

High on the list of food ingredients Willett counsels avoiding are hydrogenated fats, often referred to as trans fats, which are found in shortening, margarine, deep-fried foods, and packaged baked goods. That advice was controversial when Willett published a groundbreaking paper on the subject in 1991, but it has since become close to dogma. “Both controlled-feeding studies that have examined the effects of trans fat on blood cholesterol and epidemiological studies of trans-fat intake in relation to the risk of heart disease and diabetes indicate they are considerably worse than saturated fats,” he says.

Daily exercise is essential, Willett adds, and he confirms the often-cited advice that walking is the best choice for many people. The Nurses’ Health Study revealed a “very strong link” between walking and protection against heart disease: Women who walked an average of three hours a week were 35 percent less likely to have a heart attack over an eight-year period than those who walked less. It may seem odd that Willett includes exercise in his Healthy Eating Pyramid, but he is adamant that exercise and diet cannot be teased apart. “It doesn’t have to be extreme. I run along the Charles for 25 minutes most mornings.” A half hour daily of moderate activity offers “impressive health benefits,” he says, but there is “added benefit for greater intensity for longer times.”

Willett’s more iconoclastic conclusions include the heretical notion that soy—touted as a miracle food that fights cancer, obesity, and virtually every other human ill—may have “a dark side.” He points to a British study in which 48 women with suspicious breast lumps were randomly assigned to receive either no supplement or one containing soy isoflavones (a compound in soybeans molecularly similar to estrogen) for 14 days. Those taking the supplement showed substantially more cell growth in the tissue removed than the women who were not taking the soy. Another troubling study showed memory loss and other cognitive declines in elderly Japanese men in Hawaii who stuck to their traditional soy-based diet, as opposed to those who switched to a more of a Western diet. “In moderation, soy is fine,” says Willett. “Stuffed into everything, you could get into trouble.” And soy isoflavone supplements, he counsels, should be regarded as “totally untested new drugs.”

Willett also counsels that dairy products—which supply concentrated calories and saturated fat—are not the best way to get calcium and that the recommended daily intake of 1,200 milligrams daily for adults over 50 appears to be more than what’s needed. His advice: Eat calcium-bearing vegetables, including leafy greens, take calcium supplements if you’re a woman, and exercise. “The evidence for physical activity being protective against fractures is huge,” he says.

And he defends eggs. Although cholesterol fears have caused American per capita egg consumption to drop from 400 to 250 per year, “no research has ever shown that people who eat more eggs have more heart attacks than people who eat fewer eggs,” Willett says. A 2001 Kansas State University study identified a type of lecithin called phosphatidylcholine in eggs that interferes with cholesterol absorption, which may explain why many studies have found no association between egg intake and blood cholesterol level. If the breakfast menu option is a white-flour bagel or an egg fried in vegetable oil, says Willett, “the egg is the better choice.”

Perhaps the most comprehensive studies Willett has assembled compare the health consequences of eating saturated versus unsaturated fat. The term saturated means that every available site along each fat molecule’s carbon chain is filled with a hydrogen atom; such fats—including butter and animal fat—are solids at room temperature. There are two types of unsaturated fats: monounsaturated fats such as olive oil, which are missing one pair of hydrogen atoms, and polyunsaturated fats such as soy, corn, and canola oils, which lack more than one pair. Both sorts are liquid at room temperature.

Some researchers have questioned whether saturated fat is dangerous. In his book, The Cholesterol Myths: Exposing the Fallacy That Saturated Fat and Cholesterol Cause Heart Disease, Swedish physician Uffe Ravnskov asserts that as of 1998, 27 studies on diet and heart disease had been published regarding 34 groups of patients; in 30 of those groups investigators found no difference in animal fat consumption between those who had heart disease and those who did not. “Anyone who reads the literature in this field with an open mind soon discovers that the emperor has no clothes,” Ravnskov writes.

Willett turns to his Nurses’ Health mega-study for the definitive word. “The amounts of specific fats did make a difference,” he says. “Women who ate more unsaturated fat instead of saturated fat had fewer heart problems.” Willett calculated that replacing 5 percent of saturated fat calories with unsaturated would cut the risk of heart attack or death from heart disease by 40 percent. Other studies—notably the French Lyon Diet Heart study, begun in 1988—show a similar correlation.

A healthy diet plan is worthless if people won’t stick to it, and Susan Roberts, director of the energy metabolism laboratory at Tufts University, contends that Willett’s regimen is too severe. “Most people would say his recommendations are healthy but that other, less difficult diets are healthy too,” she says.

Difficult is in the palate of the eater. The last half of Willett’s book aims to dispel any taint of Calvinism with recipes that verge on the sybaritic, including pork tenderloin with pistachio-gremolata crust, chicken enchilada casserole, and grilled salmon steaks with papaya-mint salsa. On the other hand, some resolve might be required to soldier through a few of the other dishes listed there, including hearty oat–wheat berry bread or the onion-crusted tofu-steak sandwich. But most people, Willett believes, can summon the willpower to substitute whole-wheat flour for white and plant oils for shortening or lard, and eat less sugar overall. “I think what I suggest is not severely restrictive, because it can be achieved mainly by substitution,” rather than slavishly following recipes, Willett says. In any case, “it does not mean you cannot eat any of those foods but rather that they should be de-emphasized.”

So take heart. Even Willett has a little chocolate now and then.

Willett vs. Ornish vs. Atkins

Walter Willett’s dietary recommendations are similar in many ways to those advanced by another doctor-nutritionist, Dean Ornish, who pioneered an ultralow-fat, near-vegetarian regime that has been shown to halt or reduce coronary blockage in most heart patients. Both Willett and Ornish emphasize whole grains, fruits, and vegetables, and both minimize animal proteins. But they part ways on fats: Willett recommends replacing saturated fats in the American diet with unsaturated ones, while Ornish suggests sharply cutting fat intake altogether, especially for those at risk for heart disease. “No one has shown that the kind of diet that Walter Willett recommends can reverse heart disease,” says Ornish.

For his part, Willett insists that “replacing saturated fats with unsaturated fats is a safe, proven, and delicious way to cut the rates of heart disease.” He says the Lyon Diet Heart study, a French trial that tracked heart-attack survivors on an oil-rich Mediterranean diet versus those on the low-fat American Heart Association diet, showed a significant drop in second attacks for the Lyon group. Ornish responds that the drop in deaths in that study was most likely due to increasing heart-healthy omega-3 fats and decreasing intake of omega-6 fats, saturated fats, animal protein, and cholesterol, not to high overall consumption of fat. Ornish recommends that everyone consume three grams of omega-3 fats daily, either through eating fish or taking supplements.

In contrast with both Willett and Ornish, the late Robert Atkins recommended a meat-intensive, protein-rich regime. “Studies at Duke University, the University of Cincinnati, and the University of Pennsylvania all show that people can lose significant weight, lower their triglycerides, and improve their HDL [high-density lipoprotein] cholesterol levels by consuming protein and limiting carbohydrates,” says Stuart Trager, an orthopedic surgeon who assumed the spokesman’s mantle for the diet after Atkins’s death in April 2003. Trager believes the real strength of the Atkins diet is that “it is something people are willing and able to do.”

Willett concedes that Atkins “was really onto something. He believed, correctly, that most people can better control their weight by reducing the glycemic load of the diet than by other means. But there is evidence that the traditional Atkins diet, which is high in animal fat, is not optimal. There are benefits to having cereal in one’s diet. There is relief from constipation, and we do see [in the Nurses’ Health Study] some benefit for heart disease and diabetes. This is probably partially from the fiber in whole grains, and also partly from the other minerals and vitamins that come along with whole grains that are in short supply in many people’s diets.”

While at first blush the three approaches seem sharply divergent, Trager sounds a conciliatory note. “No one has ever bothered to point out that we are compatriots on many points,” he says. All three nutritionists share an emphasis on reducing blood-sugar spikes by reducing the glycemic load. Moreover, all three condemn trans fats, white flour, and sugar. “There really is universal agreement that you should cut those things out of your diet,” Trager says. —Brad Lemley

Good Carbs/Bad Carbs

The glycemic index (GI) is a way of measuring how quickly the carbohydrate in a given food raises the level of blood sugar. So eating a low-GI food causes a slow, mild rise, while the same quantity of carbohydrate in a high-GI food will trigger a faster, bigger rise. A GI of 55 or less is considered low, 56 to 69 is medium, and 70 or more is high.

But the GI is of limited use in the real world of pears, pork, and pudding because it ignores how much of that food a person eats. A few years ago, Walter Willett pioneered the concept of the glycemic load (GL), a measurement that factors in the quantity of carbohydrates eaten in a single serving of a particular food. The carbohydrates in parsnips, for example, are quickly converted to glucose, so parsnips have a rather high index of 97, plus or minus 19 (the numbers are sometimes imprecise because they are based on feeding foods to test subjects and monitoring their blood-sugar response, which can vary for many reasons). But parsnips have a GL of just 12, because a single 80-gram serving contains a relatively small amount of carbohydrate. A GL of 10 or less is considered low, 11 to 19 is medium, and 20 or more is high. Consistently eating low-GL foods evens out blood-sugar peaks and valleys, which Willett says helps keep appetite and weight under control. Eating low-GL foods also reduces the risk of developing type 2 diabetes. When Willett says “good carbs,” he is essentially referring to fiber-rich, low-GL foods.

Generally, whole grains have lower glycemic loads than refined grains. For example, a 150-gram serving of brown rice has a GL of 18, while the same serving of quick-cooking white rice has a GL of 29. Although the photographs in this story tally the “sugar equivalence” of the carbohydrates in various American foods, the glycemic index and glycemic load of each of these foods needs to be considered as well. The glycemic numbers accompanying the photographs in this article are from Janette Brand-Miller of the University of Sydney, based on a table published in the July 2002 issue of The American Journal of Clinical Nutrition.

1 free article left
Want More? Get unlimited access for as low as $1.99/month

Already a subscriber?

Register or Log In

1 free articleSubscribe
Discover Magazine Logo
Want more?

Keep reading for as low as $1.99!


Already a subscriber?

Register or Log In

More From Discover
Recommendations From Our Store
Shop Now
Stay Curious
Our List

Sign up for our weekly science updates.

To The Magazine

Save up to 40% off the cover price when you subscribe to Discover magazine.

Copyright © 2024 Kalmbach Media Co.