GPT-4 improves sex-specificity in cardiovascular patient education but may perpetuate gender biases: A mixed-methods audit
GPT-4 verbessert die geschlechtsspezifische Anpassung von Patient:inneninformationen zu kardiovaskulärer Prävention, kann aber Gender-Bias fortschreiben: Mixed-Methods-Audit
PLOS Digital Health · 10.09.2026 · 5 (9) · e0001122
DOI: 10.1371/journal.pdig.0001122 ↗Auf einen Blick
GPT-4 verbesserte die Sex-Inklusivität von 7 kardiovaskulären Präventionshandouts deutlich und hielt die Lesbarkeit aufrecht. Qualitativ zeigten sich aber genderstereotype und teils inkonsistente Formulierungen.
Fragestellung
Kann GPT-4 kardiovaskuläre Patient:inneninformationen sex-spezifisch verbessern, ohne Gender-Bias zu verstärken?
Methode
Mixed-Methods-Audit von 7 englischsprachigen CVD-Präventionshandouts; GPT-4-Revisions für 55-jährige männliche und weibliche Zielgruppen; Bewertung mit Flesch-Kincaid, 10-Punkte-Checkliste und thematischer Analyse.
Ergebnisse
Sex-Inklusivität stieg: Original median 3,0/10 (IQR 0,0–3,0), männlich angepasst 7,0 (IQR 7,0–8,5), weiblich angepasst 10,0 (IQR 10,0–10,0). Lesbarkeit blieb erhalten; qualitative Analyse zeigte Bias-Muster.
Schlussfolgerung der Autor:innen
LLMs können die Sex-spezifität von Patient:inneninformationen schnell verbessern, können aber Gender-Stereotype und sprachlich verzerrte Rahmungen reproduzieren. Die Nutzung erfordert kritische menschliche Kontrolle.
Kernaussagen
- Sex-Inklusivität stieg von median 3,0/10 auf 7,0 bzw. 10,0.
- Lesbarkeit der Handouts blieb laut Autor:innen erhalten.
- Qualitativ traten Gender-Stereotype und Bias in beiden Versionen auf.
Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.
Kurz gefasst
Mixed-Methods-Studie in der Zeitschrift PLOS Digital Health, online erschienen am 10.09.2026. Verfasst von Samah Khan und 1 weiteren Autor:innen. Beteiligte Einrichtung: School of Global Public Health, New York University, New York, New York, United States of America. Die Arbeit nennt 23 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Patient education materials for cardiovascular disease (CVD) prevention frequently omit clinically important differences in disease manifestation, risk, and prevention between sexes. Socially constructed gender norms further shape how health information is communicated and received. Large Language Models (LLMs) like GPT-4 offer a potential tool for personalizing health communication, but their capacity to incorporate evidence-based sex differences without introducing or perpetuating gender biases is unknown. We identified seven publicly available English-language CVD prevention handouts from major health organizations. Using the GPT-4 API in August 2025, we generated sex-specific revisions for a 55-year-old male and female audience via standardized prompts. We provided structured prompts instructing the model to include evidence-based, sex-specific risk factors and symptoms. Original and revised materials were evaluated using Flesch-Kincaid Reading Ease, a novel 10-point sex-inclusivity checklist, and qualitative thematic analysis using a reflexive approach informed by thematic analysis principles. GPT-4 revisions substantially improved sex-inclusivity scores (Original median: 3.0/10, IQR = 0.0–3.0; Male-tailored median: 7.0, IQR = 7.0–8.5; Female-tailored median: 10.0, IQR = 10.0–10.0). Readability was maintained. However, qualitative analysis identified instances where outputs introduced content inconsistent with an equity-oriented framework: female-tailored content occasionally framed female symptoms as deviations from a male norm, male-tailored content frequently omitted erectile dysfunction as a CVD risk marker despite explicit prompting, and both versions at times reflected gendered social stereotypes (e.g., “bottling up emotions”) rather than clinical evidence. LLMs can rapidly improve sex-specificity in patient education materials but may also reproduce gender stereotypes and linguistically biased framings present in their training data. Their integration into clinical content development requires critical human oversight to support, rather than undermine, health equity goals.
Autor:innen und Institutionen
- School of Global Public Health, New York University, New York, New York, United States of America.
- Herbert Wertheim College of Medicine, Florida International University, Miami, Florida, United States of America.
ORCID-Profile: Samah Khan
Zitieren
APA
Khan, S., & Vaidean, G. (2026). GPT-4 improves sex-specificity in cardiovascular patient education but may perpetuate gender biases: A mixed-methods audit. PLOS Digital Health, 5(9), e0001122. https://doi.org/10.1371/journal.pdig.0001122
Vancouver
Khan S, Vaidean G. GPT-4 improves sex-specificity in cardiovascular patient education but may perpetuate gender biases: A mixed-methods audit. PLOS Digital Health. 2026;5(9):e0001122. doi:10.1371/journal.pdig.0001122
BibTeX
@article{Khan2026,
author = {Khan, Samah and Vaidean, Georgeta},
title = {GPT-4 improves sex-specificity in cardiovascular patient education but may perpetuate gender biases: A mixed-methods audit},
journal = {PLOS Digital Health},
year = {2026},
volume = {5},
number = {9},
pages = {e0001122},
doi = {10.1371/journal.pdig.0001122}
}
EUVIMED