Mayo Clinic Proceedings just dropped this bombshell: Most nutritional studies are deeply flawed because human memory is notoriously unreliable, and also because people lie about what they eat.
Dr. Gregory House may only have been a TV doctor, but it looks like he was right—in this case, at least.
Doctors Edward Archer, Gregory Pavela, and Carl J. Lavie have concluded that dietary guidelines are largely based on information provided by memory-based dietary assessment methods (which they call M-BMs). Their paper (currently an in-press corrected proof) presents the case that decades of evidence show that actual nutrient and energy consumption bears little or no resemblance to what subjects self-report.
Furthermore, the authors assert that since the majority of evidence upon which these guidelines are based is subjectively (rather than objectively) measured, all such guidelines and studies that rely on that information are fundamentally flawed.
Say you’re participating in a nutrition study. Chances are excellent that researchers will ask you to either recall what you’ve eaten and drunk in the past 24 hours, or fill out a food frequency questionnaire. Both these methods rely solely on both your ability and willingness to accurately report what you’ve consumed. We may not be scientists, but we think the problems inherent in relying only on this information are obvious.
If that’s not enough, the doctors who wrote this paper cite a recent editorial in the British Medical Journal stating directly that data gathered in this way from the National Health and Nutritional Examination Survey is “incompatible with life.” We humbly suggest that NHANES applies water to that burn.
Seriously, if there’s such a discrepancy between what’s reported and what’s actually happening, how can these nutritional studies possibly help the people who need it most? That’s what these researchers hope to uncover.
Men And Women Both Misreport
While no one can say for sure if it’s intentional, the researchers found that both men and women have misreported their actual energy intake on both types of reports. In general, men underreported by 12 to 14 percent on two 24 hour food surveys, and by 31 to 36 percent on food-frequency questionnaires. Meanwhile, women underreported by 16 to 20 percent on the 24 hour surveys, and by 34 to 38 percent on the food-frequency questionnaires.
The paper lays it out very clearly here:
In other words, because you always eat a bowl of oatmeal with berries for breakfast, you assume definitely ate one yesterday morning—even though in reality, you were running late and grabbed a much less healthy bacon, egg, and cheese on the way to work instead.
If you want to read the entire systematic takedown of nutritional studies, the full text is available for free online. There’s no sass like scientific sass.