Why Biomarkers?
Objective measures for a subjective world
Nutrition science faces a fundamental problem: we cannot directly observe what people eat. Almost all large-scale studies rely on self-reported dietary data — food diaries, 24-hour recalls, food frequency questionnaires — and this data is systematically flawed.
The problem with self-report
When people record what they eat, they:
- Under-report total energy intake (often by 20–40%), particularly for socially undesirable foods
- Misremember portion sizes and preparation methods
- Change their behaviour in response to being observed
- Report what they think they should eat, not what they do
These biases are not random noise. They are systematic, and they differ by age, sex, body weight, and socioeconomic status — meaning that diet–disease associations estimated from self-report can be distorted, attenuated, or reversed.
What biomarkers can do
A dietary biomarker is a biological measurement — typically in blood or urine — that reflects intake of a specific food or nutrient, independently of what a person reports. Examples:
| Biomarker | What it reflects |
|---|---|
| Urinary sucrose and fructose | Added sugar intake |
| Urinary nitrogen | Total protein intake |
| Serum/plasma flavanols | Cocoa and tea consumption |
| Stable isotope ratios (δ¹³C, δ¹⁵N) | Animal protein and sugar intake over months |
| Urinary phytoestrogens | Soy and legume consumption |
Because these measurements are made from a biological sample, they are not subject to the reporting biases that affect questionnaire data.
Limitations
Biomarkers are not perfect. They may reflect:
- Short-term intake only — urinary recovery biomarkers are sensitive to the previous day or two
- Metabolism, not just intake — individual differences in absorption and excretion affect results
- Confounded exposures — a biomarker for one food may correlate with intake of others
My research addresses these limitations by validating biomarkers across diverse populations, combining biomarker and self-report data to reduce total measurement error, and pairing long-term markers (such as stable isotope ratios) with short-term urinary measures.
Why this matters for policy
Dietary guidelines are only as reliable as the evidence behind them. If that evidence rests on systematically biased self-report data, the recommendations may be wrong — sometimes in ways that harm health. Biomarker-based approaches help to:
- Validate or challenge findings from questionnaire-based epidemiology
- Detect and correct for measurement error
- Provide reference measurements for calibration studies
The goal is nutrition science robust enough to underpin effective policy.