What kind of study?

Here’s a quick primer on the two types of studies on diet and health you’re likely to hear about, and a few of their pros and cons.

Observational studies

Most are “prospective cohort” studies that ask thousands of people what they typically eat, then wait five or 10 years to see who gets heart disease, diabetes, cancer, etc. The best studies filter out the impact of smoking, exercise, and other potential “confounders”—that is, something else about the participants or what they do that accounts for the results.

Pro: You can see a decade’s-long impact on health.

Con: You can’t tell if unknown confounders explain the results.

Randomized clinical trials

Researchers randomly assign people to eat one of two (or more) diets. The best studies provide all the food. After a few months, the scientists see if the diets made a difference in a risk factor (like blood pressure). If trials want to look at diseases like diabetes, cancer, or heart disease, they typically have to enroll thousands, wait for years, and rely on participants to choose their own food.

Pro: If the study finds a difference in a disease or risk factor, you can be sure the diet caused it.

Con: If a diet has no effect, it’s possible that people didn’t stick to it, or that the trial was too short or too small.

 

Illustrations: Loel Barr. Photo: apinan/stock.adobe.com.


The information in this post first appeared in the December 2018 issue of Nutrition Action Healthletter.

Find this article interesting and useful?
Nutrition Action Healthletter subscribers regularly get sound, timely information about staying healthy with diet and exercise, delicious recipes, and detailed analyses of the healthy and unhealthy foods in supermarkets and restaurants. If you don’t already subscribe to the world’s most popular nutrition newsletter, click here to join hundreds of thousands of fellow health-minded consumers.

Have a comment, question, or idea?
Send us an email at comments@nutritionaction.com. While we can’t respond to every email, we’ll be sure to read your message.