With the help of a technique that measures people’s energy expenditure, researchers came up with an equation to assess the accuracy of responses in dietary surveys. They found that more than half the records in large, widely used nutritional survey databases such as the U.S. National Health and Nutrition Examination Survey (NHANES) are likely wrong because of people underreporting what they consumed. The results call into question the thousands of studies that have used these data sets to link particular diets to human health, the authors claim.
Although the findings are not surprising to many experts, the paper represents the best attempt yet to quantify the problem, several tell Science. It also underscores the need for better ways to measure what people really eat. “If you want to try and set policy around food based on this type of data, then obviously your policy is fundamentally flawed to some extent,” says Gary Frost, a nutritionist and dietician at Imperial College London. “We’ve got to try and find new methodologies to actually understand what people consume.”