What two testosterone videos actually claim, checked against the papers
Private deliverable. Enter the hub password to continue.
That's not right. Try again.
What two testosterone videos actually claim, checked against the papers
83 high-risk claims from 101 minutes of video. Both videos recommend blood levels above the female physiologic range, which the global consensus statement says not to do.
Why these two videos
They are sources, not output. Both sit inside the Testosterone for Women notebook as YouTube inputs, and both were indexed and then never checked. The existing audit on the hub grades 46 NotebookLM podcasts, which are the machine's summaries of material like this. Nobody had gone back to the material itself.
| Video | Channel | Length | Transcript |
|---|---|---|---|
| Optimal Testosterone (GSD Hormone) Levels for Women | Dr. Amie Hornaman | 49 min | local whisper |
| Why TESTOSTERONE Management Is Critical For Women's Health with Amie Hornaman | Ben Greenfield Life | 53 min | human subtitles |
Both feature the same clinician, once solo and once interviewed. That matters when reading the tally below: this is close to one person's position checked twice, not two independent sources agreeing.
The finding that matters
Strip out everything else and one thread runs through both videos: the blood levels being recommended sit above the female physiologic range.
| Said in the video | Where | What the library says |
|---|---|---|
| There is no too high because I have women in the hundreds and they love it | Hornaman solo, 34:06 | The global consensus statement explicitly recommends limiting testosterone dosing to the physiologic premenopausal range, contradicting the claim that there is no too-high level and that levels in the hundreds are acceptable. |
| I want to keep my total testosterone level at a 200. That's where I feel my best. | Greenfield interview, 35:35 | The claim of 200 ng/dL as desirable for women is far above the normal female range and even exceeds supraphysiologic targets used in studies, which the provided literature does not support as desirable. |
| it's a 15, that's too low for me and you have lack of muscle tone you can't build muscle... you can't burn fat | Hornaman solo, 40:17 | The passages describe a general association between low testosterone and reduced muscle mass and increased fat, but they do not specify a threshold of 15 or state that such a level directly prevents muscle building and fat burning, making the podcast's claim an overstatement. |
The reference range for total testosterone in women runs roughly 15 to 46 ng/dL. The 2019 global consensus position statement, which is in the library in full text, says dosing should be limited to concentrations in the physiologic premenopausal range. The videos recommend 50 and above, describe 200 as a personal target, and say there is no such thing as too high.
That is a real disagreement with the guideline, and it is the one thing from these videos that should not be repeated on camera without saying so plainly.
The tally
83 high-risk claims were checked. High-risk means a listener could act on it medically, or it names a number, or it claims prevention of a disease.
| Verdict | Count | Share |
|---|---|---|
| Supported | 8 | 10% |
| Overstated | 4 | 5% |
| Contradicted | 6 | 7% |
| Under Review | 3 | 4% |
| Not In Corpus | 62 | 75% |
75% is NOT_IN_CORPUS, and that is not a verdict against the videos. It means the library holds no paper that speaks to the claim. Much of what these two discuss (thyroid interaction, pellets, oxandrolone, how a patient feels at a given level) is clinical practice talk that the papers Annette holds simply do not address. A claim can be perfectly true and sit in this bucket.
Where the videos overshoot (4)
The finding underneath is real. The confidence is not.
| Claim | Where | Why it overshoots |
|---|---|---|
| Vaginal dryness is caused by low estrogen; low libido is caused by low testosterone. | Hornaman solo, 8:54 | While the claim that vaginal dryness is caused by low estrogen is supported (hypoestrogenism causes atrophy), the claim that low libido is directly caused by low testosterone overstates the evidence, which describes testosterone as playing a role in libido and its decline may contribute, not that it is the sole cause. |
| Women with low testosterone are very apathetic and more than likely depressed. | Hornaman solo, 15:59 | The claim presents low testosterone as causing strong apathy and high likelihood of depression, but the literature only suggests a likely contribution to depressed mood, not a definitive causal relationship. |
| A testosterone level of 15 is too low and associated with lack of muscle tone, inability to build muscle, and inability to burn fat. | Hornaman solo, 40:17 | The passages describe a general association between low testosterone and reduced muscle mass and increased fat, but they do not specify a threshold of 15 or state that such a level directly prevents muscle building and fat burning, making the podcast's claim an overstatement. |
| Oxandrolone can boost testosterone and can shut down natural testosterone production via negative feedback. | Greenfield interview, 31:33 | The claim correctly notes that oxandrolone suppresses natural testosterone via negative feedback, but the assertion that it 'boosts testosterone' is not supported by the passages, which describe AAS as negative modulators that suppress, not boost, endogenous testosterone. |
Where a paper says the opposite (6)
| Claim | Where | What the library says |
|---|---|---|
| Higher testosterone levels are safe if there are no androgenic side effects like facial hair or acne. | Hornaman solo, 8:07 | The passages state that even without androgenic side effects, supraphysiological testosterone levels are unsafe and require dose reduction, contradicting the claim that higher levels are safe if no side effects are present. |
| There is no universally too-high testosterone level for women; some women feel good with levels in the hundreds. | Hornaman solo, 34:06 | The global consensus statement explicitly recommends limiting testosterone dosing to the physiologic premenopausal range, contradicting the claim that there is no too-high level and that levels in the hundreds are acceptable. |
| In women taking testosterone, the goal is to mitigate androgenic effects while promoting anabolic effects, muscle growth, and strength. | Greenfield interview, 0:15 | The consensus statement on testosterone therapy in women finds no significant effect on muscle strength or lean body mass, contradicting the claim that testosterone promotes these anabolic effects in women. |
| For men, total testosterone should not be below 800 ng/dL, with an optimal range of 1000-1500 ng/dL. | Greenfield interview, 10:57 | The podcast claim that total testosterone should not be below 800 ng/dL and optimal is 1000-1500 ng/dL is contradicted by the literature, which defines the normal male range as 300-800 ng/dL. Reviewer note: Stands, but it is a claim about MEN and sits outside the subject of this site. Kept for completeness, excluded from the women's dosing picture. |
| For women, total testosterone should be at least 50 ng/dL. | Greenfield interview, 11:44 | The passage states that normal total testosterone for women is 15-46 ng/dL, which contradicts the podcast's claim that it should be at least 50 ng/dL. |
| A total testosterone level of 200 ng/dL is desirable for some women. | Greenfield interview, 35:35 | The claim of 200 ng/dL as desirable for women is far above the normal female range and even exceeds supraphysiologic targets used in studies, which the provided literature does not support as desirable. |
Verdicts withdrawn on review (3)
The checker returned an adverse verdict on these and a human pass overturned it. They are listed rather than deleted, because a silently dropped row is how an audit flatters itself.
| Claim | Where | Machine said | Why it was withdrawn |
|---|---|---|---|
| Testosterone protects against autoimmune disease. | Hornaman solo, 1:03 | OVERSTATED | Direction is plausible but the evidence does not carry it. The retrieved passage is an animal model of multiple sclerosis under electromagnetic field exposure, which is not the right paper for a general claim about autoimmunity in women. |
| Without sex hormones, you have no metabolism. | Hornaman solo, 13:04 | CONTRADICTED | Contradicted only on a literal reading. 'Without sex hormones you have no metabolism' is hyperbole in context, and grading hyperbole as a factual error is the same overreach this audit exists to catch. |
| Testosterone cream applied clitorally prior to sex can enhance orgasm intensity. | Greenfield interview, 36:40 | OVERSTATED | Extraction artifact rather than a claim. The words actually spoken were 'there is a little bit to that', which asserts almost nothing. Do not count this either way. |
What this audit cannot tell you
Claim counts are not stable. The Greenfield video was transcribed twice, once by local whisper and once from the human subtitles YouTube already publishes. The same extractor at the same settings pulled 105 claims from one and 80 from the other. The published run uses the human subtitles, which are the better source, but the gap means any headline count here is softer than a round number looks.
The library is partial. Nothing in NOT_IN_CORPUS has been shown false, and 157 of the papers in the library still have no full text at all.
A machine wrote the first verdict on every row. Fourteen adverse verdicts were read by hand before publishing and 4 did not survive: three were withdrawn outright and one was overturned to supported (the checker rejected a claim about oxandrolone toxicity because the passage said "17 alpha-alkylated androgens" instead of naming the drug, and oxandrolone is one).
One person, two appearances. Treat the tally as a reading of a position, not a survey of the field.
This is not a verdict on a clinician. It is a check of specific on-camera statements against specific papers, done so Annette knows which of them she can build on. Nothing here speaks to anyone's practice or patients.
How it was produced
Transcript (published subtitles where they exist, local whisper large-v3 where
they do not), claim extraction and per-claim judgement on OpenRouter, retrieval
by BM25 over 20,387 paragraph-level passages from the research library notes
that actually hold text. Code and raw data in
research/youtube-sources/ in the testosteroneinwomen repo.
Published to Annette's hub. Rebuilt from the source markdown, so edit the source and rerun rather than editing this page.