Did you know…

The dietary assessment methods that nutritionists and clinicians rely on to estimate a patient’s nutrient intake under-report actual intake by roughly one quarter. 24-hour recalls under-report energy intake by 22 to 25% when validated against doubly labelled water, the biological gold standard. Food diaries are similarly biased. No known self-report instrument is unbiased. Yet these instruments guide dietary policy, population surveillance, and individual clinical assessment.

The claim, in the reference of measurement-error research

Freedman and colleagues at the National Cancer Institute have anchored the field of dietary measurement-error research across multiple publications from 2011 onward. Their conclusion is direct:

“Evidence has accumulated that neither 24-hour recalls nor multiple-day food records are unbiased instruments, and no known self-report instrument is truly unbiased.”

— Freedman et al., National Cancer Institute methodological reviews (2011-2015)

In a validation study of Danish adults comparing multiple recalls against doubly labelled water:

“Compared with TEE, participants under-reported EI by 25% in the first recall, 22% for average of first two, and 25% for first three recalls.”

That under-report is not a marginal correction. It is 500 to 700 kcal per day systematically absent from the record.

What the instrument was originally designed for

The 24-hour recall was designed for large population-level surveillance of gross intake patterns, where systematic bias could be corrected for at the aggregate level. It was never designed for individual-patient assessment of micronutrient adequacy, and the technical literature has been clear about this limitation for decades. Yet in ordinary clinical nutrition, a patient reporting a three-day food diary receives a nutrient adequacy assessment based on that diary, without a correction factor, and without any acknowledgment that the record almost certainly reflects a subset of what the patient actually ate.

What the biological alternatives would give

Red-cell magnesium reflects intracellular status over months. Omega-3 index reflects membrane composition over the previous 90 days. Holotranscobalamin captures functional B12 status. Ferritin reflects iron stores. Each of these bypasses the memory and honesty of the patient and reads what the tissue has actually integrated. They are not part of standard nutritional assessment. The instrument that is biased is the instrument that gets used.

This entry is a companion to The Hunger We Don’t See and to arguments on reference ranges calibrated on the wrong questions.