Research & Evidence

Evidence snapshot

Counts of PubMed records that mention each food or term. A count shows how much has been published, not whether a food works. Many searches also catch non-health research (farm crops, food chemistry) and other forms such as juice, extract and supplements. [1]

Counted 2026-10-22 using NCBI E-utilities. Counts include studies of extracts, blends and non-health topics, and a trial can find no effect. A higher count means more human research has been indexed, not that something works.

What this means

• Large counts (soy, legumes, tomato, citrus) include crop science and lycopene or hesperidin research; do not read them as food-trial counts.
• Fermented foods have much smaller trial counts than the probiotic capsule literature: probiotics have over 3,000 randomized-trial records, yogurt about 600, kefir about 36, kimchi about 9, kombucha about 6. Many probiotic trials are of capsules, not food.
• Terms like "functional food" and "superfood" are themselves search topics: "superfood" has few randomized trials.

An eating-pattern trial

In the CORDIOPREV randomized trial, 1,002 people with coronary heart disease ate a Mediterranean or a low-fat diet; the primary endpoint occurred in 87 of 502 Mediterranean-diet participants and 111 of 500 low-fat participants, with hazard ratios from 0.719 to 0.753, mainly in men. [2] It tests a pattern, not a single food.

How to judge a food claim

1. Randomized trial or observational?
2. Marker or outcome?
3. Whole food, juice, powder or supplement?
4. How large, how long, how many people?
5. Who funded it? (We did not check funding for every study.)
6. Does the claim need a medical reviewer? Our pages do not have one yet.

See Claim checks and Comparisons.