RSS feed of this blog

Sift through the evidence on diets

Editor’s note: This column by Jennifer Motl was published in the Healthy Living section on March 4, 2012.


WHEN YOU READ so many conflicting scientific reports about diets—Paleo, Mediterranean, low-cholesterol, South Beach, carbohydrate-counting, Atkins, to name a few—how do you decide which is best?

In this column, I’ll discuss how to understand the studies, how some advertisers lie with statistics, and how to make up your own mind.

After a recent column about the Paleo Diet, several readers wrote impassioned letters questioning my analysis. When analyzing eating plans, I look for changes that most people can sustain for life. I realize that a small group of folks is passionate about the Paleo lifestyle. I respect their commitment, but I don’t think the Paleo plan is practical for the majority of Americans.

For those who missed it, the Paleo Diet advocates eating only foods available to cavemen: grass-fed meat, pastured poultry, fish, fruits, vegetables, nuts and seeds. It excludes milk, yogurt and cheese and all grains such as corn, wheat, oats and rice.

While there are dozens of articles on the Paleo Diet, only four are clinical trials, which are the gold standard of scientific research. I feel four small trials is inadequate evidence to recommend it for all people.

Several readers wrote to tell me how much weight they lost on the diet. While I congratulate them, that does not prove that the Paleo Diet is the best diet in the world, or one I should recommend at this point.

Research comparing several different diets has shown that low-carb versus low-fat versus whatever other diet doesn’t matter as much as the total calories a person consumes. In other words, you can lose weight and improve your cholesterol level and blood pressure (at least in the short-term) on almost any diet if you eat fewer calories.

While individual stories of losing weight and better blood pressure are inspiring, they do not guarantee results for everyone. Nor do they guarantee that people will maintain their weight loss or other improvements.

Scientific studies are the best way to test diets. For example, to prepare my article on the Paleo Diet, I reviewed the latest scientific literature. Readers can do the same. I recommend PubMed, a free online database of the U.S. National Library of Medicine, which comprises more than 21 million scientific articles.

I rely on scientific articles rather than Internet searches alone, because scientific articles are peer-reviewed, meaning a team of experts reads the articles before they are published. This does not guarantee the articles are unbiased, but they are less biased than random Web sites, which tend to be selling something.

When writing a column, I often review more than 100 recent scientific abstracts and articles on a single topic. That way, I can learn about discoveries that have occurred since I earned my degree in nutrition from Cornell University. It’s important to keep current.


But even keeping current is not enough. That’s because not all scientific studies are equal. And it’s easy to lie with statistics. Some statistics show just a link, not cause and effect.

For example, there’s a humorous article circulating on the Internet stating that 90 percent of felonies were committed within 24 hours of eating bread; that bread can choke newborn babies; and other funny examples insinuating that bread is the source of all evil.

These  facts  don’t prove anything. Eating bread no more causes a person to commit a felony than a rooster crowing causes the sun to rise. They coincidentally happened at the same time.

To prove cause and effect, you must go beyond statistics to experiments. The best are randomized prospective clinical trials. For example, in studying the Paleo Diet, a randomized study might randomly assign two groups of volunteers to either follow the Paleo Diet or a heart-healthy diet.

It’s important that the volunteers are assigned randomly—you want equal numbers of volunteers in both groups, and you want roughly equal numbers in each group in demographics such as old, young, rich, poor, women, men, and more. You wouldn’t want all the people on the Paleo Diet to be young women and all the people on the heart-healthy diet to be old men, because then you’re comparing apples to oranges.

A prospective trial means you actually measure things before and after the new diet. You test blood pressure, cholesterol, weight, waist circumference, blood sugar and more. You also measure what people truly eat, either with food diaries or weekly surveys or ideally by actually providing them with meals.

One reader wrote about a well-designed trial showing that the Paleo Diet makes folks healthier regardless of their weight: “Dr. Frassetto and colleagues controlled for weight loss and found that the Paleo Diet ‘improves [blood pressure] and glucose tolerance, decreases insulin secretion, increases insulin sensitivity and improves lipid profiles without weight loss in healthy sedentary humans.’”

That’s true, but it was a small study involving only nine people; hardly enough for me to make blanket recommendations to the nation about the Paleo Diet.

I’ve seen research on the consistent-carbohydrate diet and the Mediterranean diet, the DASH diet and even the vegan diet that shows similar results. I’m not saying the Paleo Diet is bad—just that we need larger, longer-term studies before I’m convinced it’s best for everyone.

At this point, I feel the weight of the evidence—studies of thousands of people—supports a Mediterranean eating plan, which emphasizes eating lots of fruits and vegetables, fish, and healthy fats such as olive oil.

But this may change. Historically, as evidence accumulates, physicians and dietitians change their recommendations. Just 20 years ago, experts said that all fat was bad for the heart. Now we know more details: While some fats clog arteries, other fats, such as olive oil, are very healthy.

I’m sure we’ll know even more in the future. Maybe someday the evidence will change my tune about the Paleo Diet—who knows what the future holds?

In the meantime, I keep sifting through the latest evidence, looking for healthful, tasty things to eat.

Jennifer Motl welcomes reader questions via her website,, or by email at