Comparing Personal Responses to Foods: A Careful Approach

The idea

People differ in how they respond to the same food. Researchers have shown this with blood glucose, and it has made "personal food response" testing popular. This page explains what research shows about those differences, and how a careful person could compare ordinary foods using plain observation. It gives no diet plan, no protocol with exact amounts, and no advice to cut out a food group.

Evidence labels used here: promising but limited: encouraging results from small or early human studies; insufficient evidence: too little to judge; marketing claim: a statement used to sell. A change in a number after a meal is a biomarker change, not a health outcome.

What research shows about differences between people

• In the PREDICT 1 study of 1,002 adults in the United Kingdom, researchers found large variability between people in blood triglyceride (103%), glucose (68%) and insulin (59%) responses after identical meals. Person-specific factors such as the gut microbiome influenced some responses more than the nutrient mix of the meal, and a machine-learning model predicted glucose responses with a correlation of 0.77. [1] This is a research result in a study setting, not a promise that a consumer product will predict your response.
• In a continuous-glucose-monitoring study, glucose patterns after identical nutrient challenges varied widely between people. Even people considered normoglycemic by standard measures had glucose reach prediabetic and diabetic ranges 15% and 2% of the time. [2] This shows that CGM numbers can look unusual in healthy people; it does not mean those people are ill.
• In a series of n-of-1 trials with 176 healthy Chinese adults eating standardized meals, there was large variation between individuals but high consistency within the same person over repeated tests. A personalized score from that work was reproducible over two years in the research setting. [3]

So the differences are real. What is unproven is whether acting on a personal response result improves long-term health. Label: promising but limited for the existence of individual differences, and insufficient evidence for benefits of personalized food advice from consumer tests. Proper evidence not found, in the sources I checked, that consumer food-response reports improve health outcomes in people without a medical condition.

What "n-of-1" means

An n-of-1 trial studies one person across periods, comparing exposures. See N-of-1 Research: Meaning and Limits and Tracking vs Experimenting. A formal n-of-1 trial needs a fair comparison; casual tests usually do not have one.

A careful approach: benign foods only

The only example this library gives is an ordinary-food comparison with no health goal beyond curiosity.

1. Keep the question small and harmless. For instance, noticing whether plain oatmeal or plain toast with the same breakfast leaves you more or less hungry before lunch. It should not involve fasting, extreme foods, supplements or cutting out whole food groups.
2. Change one thing. Keep the rest of the day similar. See Changing One Low-Risk Variable at a Time.
3. Record plainly. Write down what you ate and how you felt in your own words. Compare repeated days rather than a single day, because one reading can mislead. See Baseline Variation and Normal Fluctuation and Establishing a Baseline.
4. Watch for expectation. Knowing what you are testing can shape how you feel. See Placebo and Expectancy Effects in Wellness Tracking.
5. Do not label a food "bad" from one result. See Why One Reading Does Not Prove a Food Is Harmful.
6. Stop if it stops being useful. If recording makes you anxious or preoccupied with food, stop. See Tracking Anxiety and Compulsive Measurement.

Direct, estimated and calculated

A symptom diary is a direct record of what you report. A glucose monitor gives an estimate of glucose, and an app "response score" is a calculation by the company. The studies above used research-grade methods, so their results should not be copied onto consumer scores. See Nutrition Tracking and Continuous Glucose Monitors and Glucose Food Responses and Individual Variation.