34%
of the 400 graded interventions covered here rest on Strong or Very Strong human evidence (136 of 400).
Original analysis
Every entry in this database is scored with the same published rubric — 45% Evidence, 15% Benefit, 40% Safety. Scoring a deliberately mixed set that way makes some patterns visible that no single review does. These are descriptive statistics about the database, updated whenever the data is.
34%
of the 400 graded interventions covered here rest on Strong or Very Strong human evidence (136 of 400).
19.5%
carry a high-risk evidence profile — a Safety score of 3.5 or below (78 of 400).
39%
involve a prescription form in at least one jurisdiction or route (156 of 400).
6/10
is the median Evidence score. Median Safety is 7/10 and the median composite is 6.2/10.
The rubric is not curved. Grades fall where the evidence, benefit, and safety inputs put them, which is why the distribution clusters in the B and C bands rather than spreading evenly.
| Grade | Entries | Share | |
|---|---|---|---|
| S | 8 | 2% | |
| A+ | 15 | 3.8% | |
| A | 13 | 3.3% | |
| A- | 14 | 3.5% | |
| B+ | 59 | 14.8% | |
| B | 66 | 16.5% | |
| B- | 49 | 12.3% | |
| C+ | 69 | 17.3% | |
| C | 37 | 9.3% | |
| C- | 15 | 3.8% | |
| D+ | 14 | 3.5% | |
| D | 34 | 8.5% | |
| D- | 3 | 0.8% | |
| F | 4 | 1% |
Grouping by what a thing is shows a wide spread. Groups smaller than eight entries are left out because a median over a handful of rows is not worth reporting.
| Class | Entries | Median composite | Median evidence | Median safety |
|---|---|---|---|---|
| Vitamin | 11 | 7.5 | 8 | 9 |
| Protocol | 23 | 7.1 | 6 | 7 |
| Supplement | 90 | 6.5 | 6 | 9 |
| Hormone | 15 | 6.4 | 8 | 5 |
| Mineral | 10 | 6.3 | 8 | 6 |
| Prescription | 133 | 6.3 | 8 | 5 |
| Herbal | 32 | 6.2 | 4.5 | 9 |
| Peptide | 20 | 5.3 | 2.8 | 9 |
| Anabolic | 33 | 4.9 | 6 | 3.5 |
| Research chemical | 28 | 4.5 | 3 | 4 |
Entries are cross-listed, so a compound studied for two goals counts in both columns. Totals across categories therefore exceed 400.
| Category | Entries | Median composite | Median evidence | Median safety |
|---|---|---|---|---|
| Essentials | 37 | 7.1 | 8 | 9 |
| Mitochondria & Cellular Energy | 28 | 7.1 | 6 | 9 |
| Immune & Inflammation | 53 | 6.8 | 6 | 9 |
| Mood, Anxiety & Stress | 52 | 6.5 | 6 | 6 |
| Metabolic Health | 88 | 6.4 | 6 | 5 |
| Sleep & Recovery | 69 | 6.3 | 6 | 9 |
| Longevity | 59 | 6.3 | 6 | 9 |
| Hormones & Endocrine | 51 | 6.2 | 6 | 5 |
| Cognitive | 77 | 5.7 | 6 | 7 |
| Muscle & Strength | 58 | 5.4 | 6 | 4.5 |
| Detox | 16 | 5.4 | 3 | 9 |
400 of 400 graded entries (100%) were reviewed within twelve months of the most recent review in the dataset. Review dates run from Jun 7, 2026 to Jul 27, 2026. Every material change is dated in the public archive.
These describe the entries covered in this database, not health interventions in general. The entry list is curated toward what readers of a longevity and health-optimisation publication ask about, which deliberately oversamples frontier and experimental topics relative to routine clinical practice. A figure like the Strong-evidence share is therefore a fact about this database's coverage; reading it as an estimate of the wider literature would misuse a convenience sample.
Scores summarise published evidence in studied populations. They do not predict an individual's response, and a grade does not establish that an intervention is appropriate, effective, or safe for any particular person. The full rubric, its weights, and its limitations are on the methodology page.
These figures are free to quote with attribution to Read Off Label and a link to this page.
Please carry the denominator with the number — "34% of the 400
graded interventions in the Read Off Label database" rather than "34% of
supplements". The whole dataset is available as
a plain-text index, and every entry has a Markdown representation
at /compound/<slug>.md.
A concise record of what changed, why it changed, and what is still uncertain.
Signup is handled by Beehiiv. See our privacy notice.