Written and reviewed by physicians

Editorial standards

Our Methodology

Every article published on Petit Veinard passes through a multi-stage scientific pipeline designed to ensure accuracy, safety, and transparency.

📖
21
scientific journals monitored
🔬
6
anti-hallucination layers
📋
12
ESVS/AHA/ACC guidelines
100%
sources cited with PMID

1. Scientific sourcing

All articles are grounded in peer-reviewed literature retrieved from PubMed/MEDLINE, Cochrane, CrossRef, and EuropePMC. Each cited study is identified by its PMID (PubMed ID), allowing readers to verify sources directly.

Clinical practice guidelines from ESVS (European Society for Vascular Surgery), AHA/ACC (American Heart / College of Cardiology), SVS (Society for Vascular Surgery), and NICE (UK) are cross-referenced to provide an international perspective.

2. Six-layer anti-hallucination pipeline

PICO Screening
Each source is evaluated against Population–Intervention–Comparison–Outcome criteria. Off-topic studies and medical homonyms (e.g. amino-acid statins vs. statin drugs) are filtered out at this stage.
Bridge Consensus
Local guidelines (ESVS, AHA, ACC, SVS, NICE) are cross-referenced against recent studies. Conflicts are labelled CONFIRMED / EVOLVING / CONTRADICTED and disclosed in the article.
Skeleton Pre-Check
Before writing begins, every factual claim in the article outline is validated against its source. Unverified claims are removed or marked for additional sourcing.
Guided Writing (temp 0.35)
The medical writer follows the validated skeleton. Low temperature settings minimize invention while preserving clarity and readability.
Numeric Drift Detection
An algorithm compares every number in the article against source abstracts. Values diverging more than ±15% from any cited source trigger a correction pass.
Adversarial Fact-Check
A final verification step tests each statement for entailment (does the source support it?), attribution (is the right source cited?), and distortion (is context preserved?). Publication requires a score ≥ 0.85.

3. Publication logic

Publication requires both quality gates to pass simultaneously (AND logic).

Scores Outcome
Hallucination ≤ 3 AND adversarial ≥ 0.85 Auto-published
Hallucination 4–6 OR adversarial < 0.85 Submitted for human review (PR)
Hallucination ≥ 7 Blocked — article discarded

4. Monitored journals (21)

New England Journal of Medicine
The Lancet
JAMA
BMJ
Circulation
European Heart Journal
JACC
Lancet Neurology
EJVES (European Journal of Vascular & Endovascular Surgery)
Journal of Vascular Surgery
Thrombosis and Haemostasis
Thrombosis Research
JVS — Venous and Lymphatic Disorders
Vascular Medicine
Annals of Vascular Surgery
Phlebology
Cochrane Database of Systematic Reviews
Arteriosclerosis, Thrombosis, and Vascular Biology (ATVB)
Journal of Thrombosis and Haemostasis (JTH)
Mayo Clinic Proceedings
Vascular

5. Transparency & updates

Articles are updated when new major guidelines or meta-analyses supersede the cited evidence. The date of last update is displayed on each article.

No sponsored content. No pharmaceutical company funds or influences our editorial decisions. Petit Veinard is an independent media outlet.

Medical disclaimer. Content on this site is for informational purposes only and does not replace the advice of a qualified healthcare professional. Always consult a physician for personal medical decisions.

Newsletter

The best of vascular health, delivered weekly

Guides, expert insights, and medical news — written by physicians, for you. Free. No spam.

Subscribe for free