State of Supplement Evidence 2026: What the Current Library Can and Can't Show

TL;DR

The 2026 Evidence Report now measures something narrower and more defensible than the original version: the distribution of evidence-grade labels across the herb and compound records currently rendered by the site.

It does not claim that every individual study in the library has been assigned one of those grades. It does not use a fixed study total, a hand-written percentage split, or a static list of “winning” supplements. The live distribution is calculated from the current runtime records and can change as the library changes.

See the current numbers in the Supplement Evidence Report.

---

Why the Report Changed

An evidence dashboard should not look dynamic while its headline statistics are actually constants in a page component.

The earlier version of this report used fixed totals and fixed percentages. That created a drift problem: the underlying herb and compound libraries could change while the report continued to display the same numbers. It also blurred two different units of analysis — profiles and studies.

That is now corrected.

The live report asks a simpler question:

Among the reference records the site can currently render, what evidence-grade labels are explicitly attached to those records?

That question can be answered from the runtime data itself.

---

What the Live Report Measures

The Evidence Report currently calculates:

• the number of renderable herb and compound reference records;
• how many of those records expose an explicit evidence_grade label;
• how many are currently unclassified at that field; and
• the distribution of the evidence-grade labels that are actually present.

The calculation happens when the site is built, so the report follows the current runtime library instead of relying on manually maintained percentages.

The unit is the profile

This distinction matters.

A profile can contain several claims, outcomes, mechanisms, preparations, populations, and citations. Those pieces may not all have the same evidence strength. A profile-level label is therefore a summary signal, not a substitute for reading the claim-level evidence and limitations.

---

What the Report Does Not Measure

The report should not be read as:

• a count of all peer-reviewed studies about supplements;
• a claim that every cited paper has been individually assigned the profile's grade;
• a ranking of the “best” supplements;
• proof that a higher-graded ingredient is effective for every outcome;
• proof that commercial products are equivalent to the preparations used in research; or
• a safety clearance for a particular person, medication list, dose, or medical condition.

Those are separate questions.

---

How to Read an Evidence Grade

An evidence label is most useful when it compresses a much larger appraisal without hiding the uncertainty behind it.

When evaluating a specific claim, the important questions still include:

1. Directness: Was the outcome studied in humans, in the population and context relevant to the claim?
2. Study design: Was the comparison appropriate, and were important sources of bias addressed?
3. Consistency: Do independent studies point in the same direction?
4. Precision: Are the effect estimates precise enough to rule out clinically important alternatives?
5. Replication: Has the finding been reproduced beyond one team, one branded extract, or one small trial?
6. Formulation: Is the studied preparation meaningfully comparable with the ingredient or product being discussed?
7. Safety: Do contraindications, interactions, population exclusions, or uncertainty change the practical conclusion?

A letter or tier is a starting point for those questions, not the final answer.

---

The Most Important Gaps Are Often Not “Does It Work?”

Supplement evidence can be weak in more than one way.

Product comparability

A study can be internally sound while still applying only to a specific extract, formulation, constituent profile, or dose. A different commercial product may not reproduce the same exposure.

Independent replication

A promising result is more convincing when independent researchers reproduce it under comparable conditions. Repetition by the same sponsor, extract, or research group can still be informative, but it does not answer every independence question.

Long-term safety

Short trials are usually designed to answer short-term questions. They generally cannot establish the frequency of uncommon harms or the safety of years of continuous use.

Outcome specificity

Evidence for one outcome does not automatically transfer to another. An ingredient can have credible evidence for a narrow endpoint and little evidence for broader marketing claims built around the same mechanism.

Missing or unstructured evidence labels

An unclassified profile does not mean “ineffective.” It means the current runtime record does not expose an explicit evidence-grade label at the field the report measures. That is a data-quality and review-status signal, not a clinical verdict.

---

What This Means for Readers

Use the report to understand the shape of the library, then move from the profile-level label to the underlying evidence.

For a specific herb or compound:

1. open the profile;
2. identify the exact outcome being discussed;
3. check whether the evidence is human, preclinical, traditional, or mechanistic;
4. read the limitations and safety section;
5. inspect the cited sources when the decision matters; and
6. avoid treating an evidence grade as personalized medical advice or as proof that a retail product matches the studied preparation.

For medication combinations or higher-risk contexts, a site-level evidence label is not enough. Use the Safety Interaction Checker as an educational screen for possible caution flags and review the exact combination with a clinician or pharmacist when appropriate.

---

Methodology and Reproducibility

The live Evidence Report is intentionally generated from the current runtime records rather than copied from this article.

• View the current Evidence Report
• Read the evidence methodology
• Use the Evidence Lookup
• Learn how to read scientific studies

The report should be treated as a snapshot of the site's current structured evidence labels. When the data changes, the distribution should change with it.

---

Last reviewed: August 12, 2026.