We use essential cookies (authentication, your saved goals/stack) by default. With your permission we’ll also enable privacy-respecting analytics (Vercel Web Analytics, anonymous load-time metrics) and error-replay diagnostics (Sentry — DOM snapshots only when an error fires) so we can fix bugs faster. Learn more about cookies
Supplements are sold on the promise of science. We scored 583 of them on a transparent 0–10 evidence scale, built from 8,250 curated studies and 1,999 meta-analyses — and looked at what the evidence actually says.
The short version: the evidence is real, but heavily concentrated. A small core is well-supported and safe — most of the market is marketing more than science.
By Baher Al Hakim, Founder, SupStack · Published July 4, 2026 · Updated daily from the data
On this page
Scored 0–10, the 583 supplements break down like this — the evidence is concentrated in a small core:
Cross evidence with safety and the real problem appears: 27% (roughly 1 in 4) have both weak evidence and a safety caution.
42 supplements have 15+ studies and still don't reach 5/10. Popularity and study count aren't efficacy — some of the most-talked-about supplements land here:
Just 12 supplements clear both bars — strong evidence (8+) and a high safety rating:
Multivitamin is the most-studied product in the corpus (44 studies) yet scores only 6/10. Being studied a lot and being shown to work are not the same thing.
About 37% of the supplements we track (215) are experimental or research compounds. They average 4.4/10 — and 20% don't even have enough evidence to be scored. The marketing energy is aimed squarely at the tier where the evidence is thinnest.
Where the research is heading — the supplements with the most new studies in the last 3 years:
The best-evidenced supplements are the boring ones — vitamins and minerals lead; the trendy tiers (mushrooms, peptides) trail. Average evidence score by type:
Map supplements to the goals they target and a pattern emerges: sleep and women’s health are best-evidenced — longevity, the trendiest area, is weakest. Average evidence score by goal area:
58% of supplements (336) are backed by at least one meta-analysis — the gold standard. And the evidence is human, not hype: the scores rest on 2,803 RCTs and 1,999 meta-analyses, and 75% of cited studies are human clinical research — only 6% are animal or in-vitro. The base is current too: the median cited study is from 2019, 42% from the last five years.
194 supplements (33%) carry a "use with caution" safety rating — for drug interactions, contraindicated populations, or dose sensitivity. "Natural" is not a synonym for "harmless."
Beyond the individual ratings, the corpus carries 416 documented drug interactions — 27% of supplements have at least one. St. John's Wort alone interacts with 24 medications:
All 583 supplements — search, filter, and sort by evidence, studies, or safety.
583 of 583
Some do, but the evidence is concentrated. Of 583 supplements we scored, only 4% have strong evidence, while 55% are limited or insufficient. A small core is genuinely supported — most of the market is not.
By type, vitamins and minerals lead. The "evidence-backed dozen" (strong evidence plus a high safety rating) includes Creatine, Whey Protein, Omega-3, Probiotics, Sunscreen (SPF).
Not automatically. 33% carry a use-with-caution rating, and 27% have both weak evidence and a safety caution. St. John's Wort alone interacts with dozens of medications.
On a transparent 0–10 scale weighting study quantity, quality (RCTs and meta-analyses over observational data), consistency, and effect size. Scores below 3 are treated as "insufficient evidence." The full methodology is public.
The ones with the most new studies in the last 3 years include Omega-3, Sulforaphane, Beta-Alanine, Inulin, Cistanche.
Free to reuse with attribution — copy the code to drop a chart (with a link back) into your article or post.
SupStack scores each supplement on a transparent 0–10 scale weighting study quantity, study quality (RCTs and meta-analyses over observational data), effect consistency, and effect size. Scores below 3 are treated as "insufficient evidence." Every figure is computed live from 583 supplement records (snapshot July 4, 2026).
Related report
Significance Without Size: How Supplement Trials Report
Almost every trial reports a p-value — only one in six reports the effect size.
Use the data
Every score, study count, and safety rating is public — browsable and available via the free API for independent verification.
Cite this: Al Hakim, B. (2026). The State of Supplement Evidence [dataset]. SupStack. https://doi.org/10.5281/zenodo.21311542