How we evaluate evidence

Every clinical claim on this site is graded before it is published. This page describes the exact hierarchy we use, the five questions each condition page must answer, and what we refuse to treat as evidence of benefit.

Maintained by the StemCell Longevita editorial team · Medically reviewed by Prof. Dr. Erdal Karaöz Team, Professor of Histology and Embryology · Last reviewed: August 23, 2026

The source hierarchy

Sources are ranked in a fixed order. A lower tier can never override a higher one, and a page may not describe a benefit that only appears in the bottom two tiers.

Tier	Source type	How it may be used
1	Systematic reviews and meta-analyses	May support a statement about clinical effect.
2	Randomised controlled trials	May support a statement about clinical effect, with the population and endpoint named.
3	Regulatory documents (FDA, EMA, MHRA, Türkiye Ministry of Health)	May support statements about legal status, approval and warnings — never about efficacy.
4	Professional-society guidance (e.g. ISSCR, ISCT)	May support statements about accepted practice and cautions.
5	Non-randomised human studies and registries	May describe what has been observed and its limitations.
6	Case series, preprints, conference abstracts, company material	May describe only what is being studied. Never used to claim benefit.
7	Laboratory and animal studies	May describe mechanism only, and must be labelled as non-clinical.

The five questions on every condition page

Each condition page carries an "Evidence at a glance" block answering the same five questions, so a reader can compare indications without reading the prose:

Is there human evidence? — and if so, in what population and at what size.
Are there randomised trials? — stated separately from uncontrolled human data.
What outcomes were measured? — named endpoints, not general impressions.
Is there long-term follow-up? — with the longest reported follow-up window.
Is there regulatory approval? — by jurisdiction, including where treatment is lawful but unapproved.

What we do not accept as evidence

Lawful availability in a country. Being permitted is not being proven.
Patient testimonials, before-and-after imagery, or unverifiable success percentages.
Laboratory or animal results presented as if they were clinical outcomes.
Cell counts, viability figures or dose numbers used as a proxy for benefit.
Citations to a paper we have not read in full, or to a source that cannot be linked.

Wording rules

Where evidence is uncontrolled or early, pages use language that matches its strength: studied in, reported in early-phase work, investigational. Words such as cure, guaranteed, proven and reverses are prohibited on clinical pages unless a Tier 1 or Tier 2 source supports the exact claim, and a periodic automated audit flags them for correction.

Review cadence and corrections

High-risk pages — serious, progressive or weak-evidence indications — are reviewed at least every six months; all other medical pages at least annually. Pages are also reviewed out of cycle when a trial reports, a regulator acts, or a reader tells us something is wrong. Review dates and the reviewer's name are printed on the page itself. Corrections are made in place and move the review date forward.

Frequently asked questions

Do you rate every treatment the same way?

Yes. The same five questions and the same source hierarchy are applied to every indication, including the ones we are asked about most often commercially.

What happens when the evidence is negative?

The page says so. Failed or inconclusive trials are reported on the condition page alongside the rest of the evidence rather than omitted.

Who signs off the rating?

Ratings are compiled by the editorial team and reviewed by Prof. Dr. Erdal Karaöz, whose publication record and affiliations are published in full.