Our Approach

Every article in the Nutrient Interaction Atlas is built the same way. This page describes that process, so you can judge the method rather than take the conclusions on trust.

How we research an interaction

The order matters. We establish what authoritative resources already say before going to individual studies, because a single striking paper read in isolation is the fastest route to a wrong conclusion.

1. Question — We write down precisely what we are trying to answer. “Does calcium block iron?” is too loose to research. “Does calcium in a meal reduce iron absorption from that meal, and does that change iron status over time?” is two answerable questions.
2. Authoritative sources — We check what established nutrition resources already state, which sets the baseline and surfaces the accepted mechanism.
3. Scientific research — We look for reviews and primary studies, noting study design, population, dose, and what was actually measured.
4. Context — We identify the conditions the effect depends on: dose, nutrient form, preparation, timing, and the rest of the meal.
5. Evidence — We assess how strong, consistent, and relevant the findings are, and where they disagree.
6. Explanation — We write what a reader can reasonably conclude, including when the honest answer is “less than you would think.”

Our default research foundation

Four resources form our starting point for most topics:

• NIH Office of Dietary Supplements — nutrient functions, food sources, intake recommendations, and documented interactions
• USDA FoodData Central — food composition and nutrient data
• Dietary Guidelines for Americans — broader dietary patterns and context
• PubMed / National Library of Medicine — peer-reviewed biomedical literature

Other authoritative government, academic, professional, and peer-reviewed sources are used where a topic calls for them. Referencing an organization does not mean it endorses, reviews, or is affiliated with NutriFoodGuide. More detail on each: our sources.

Different studies answer different questions

Evidence is not a single quantity that a topic has more or less of. Study designs answer genuinely different questions, and confusing them is the most common failure in nutrition writing.

• Systematic reviews and meta-analyses gather many studies and ask what the body of evidence shows as a whole.
• Randomized trials test whether changing one thing causes a measurable difference.
• Controlled absorption studies measure how much of a nutrient was taken up, usually from a single meal under tightly controlled conditions.
• Observational studies track what happens in populations over time and can show associations.
• Mechanistic research explains how something could happen, often in cells or animals.

Two consequences follow, and we apply both consistently. A mechanistic finding explains a possible route; it does not establish that the effect is meaningful in people eating ordinary meals. And an observational association does not establish that one thing caused the other.

There is a third, specific to this subject. A single-meal absorption study is excellent at isolating one variable and poor at describing a diet — participants arrive fasted and eat a simplified meal, which is exactly what makes the measurement clean and exactly what makes it unlike dinner. Where both kinds of study exist, we report both.

How we describe certainty

Articles sort what is known into plain categories. These are editorial descriptions written for clarity, not a formal grading system such as GRADE, and we do not present them as one.

• Well established — consistent findings across multiple studies
• Supported but context-dependent — real, but the size or direction shifts with conditions
• Limited evidence — few studies, small studies, or narrow populations
• Mixed evidence — credible studies disagree
• Uncertain — not enough good evidence to answer

Every Atlas article also carries a section stating what the evidence does not establish. Naming the limits is part of the explanation, not a disclaimer bolted onto the end.

The claim check

Before publication, each substantive claim is checked against the specific source cited for it. The test is not whether the source is authoritative — it is whether that source actually supports that sentence.

This catches a failure mode that is easy to miss and common across the web: a respected reference is attached to a claim it does not quite make, and the citation lends authority the underlying source never granted. If a claim cannot survive the check, we rewrite the claim rather than change the source.

We also do not cite sources we have not read, and we do not reconstruct citations from memory.

We explain the evidence — not the hype

Nutrition interactions are unusually easy to turn into viral rules, because the mechanisms are real and the numbers from controlled studies sound enormous. “Tea reduces iron absorption by up to 94 percent” is a real finding from a real study. It describes a specific test meal, not your dinner.

Our commitment is to keep the qualifying conditions attached to the finding, even when that makes the headline duller. When the honest conclusion is that something matters less than its reputation, we publish that conclusion.

Corrections

We get things wrong sometimes, and evidence moves. If you find an error, a misread source, a broken reference, or an explanation that no longer reflects current research, tell us. Corrections are reviewed against the underlying sources before any change is made.