EUVIMEDEUVIMED entdecken
VergleichsstudieOpen Access · CC BY 4.0

Do large language model chatbots perform better than established patient information resources in answering patient questions? A comparative study on melanoma

Nadia C W Kamminga iD, June E C Kievits iD, Peter W Plaisier, Jako S Burgers, Astrid M van der Veldt, Jan A G J van den Brand, Mark Mulder, Marlies Wakkee iD, Marjolein Lugtenberg iD, Tamar Nijsten iD

British Journal of Dermatology · 2025

DOI: 10.1093/bjd/ljae377 ↗

Kurz gefasst

Vergleichsstudie in der Zeitschrift British Journal of Dermatology, 2025. Verfasst von Nadia C W Kamminga und 9 weiteren Autor:innen. Beteiligte Einrichtung: Department of Dermatology, Erasmus MC Cancer Institute, University Medical Center Rotterdam, the Netherlands. Hauptthemen laut MEDLINE: Melanoma, Skin Neoplasms, Patient Education as Topic. Die Arbeit wurde bisher 28-mal zitiert (Crossref) und nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Bristol Myers Squibb.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Large language models (LLMs) have a potential role in providing adequate patient information.

Objectives

To compare the quality of LLM responses with established Dutch patient information resources (PIRs) in answering patient questions regarding melanoma.

Methods

Responses from ChatGPT versions 3.5 and 4.0, Gemini, and three leading Dutch melanoma PIRs to 50 melanoma-specific questions were examined at baseline and for LLMs again after 8 months. Outcomes included (medical) accuracy, completeness, personalization, readability and, additionally, reproducibility for LLMs. Comparative analyses were performed within LLMs and PIRs using Friedman’s Anova, and between best-performing LLMs and gold-standard (GS) PIRs using the Wilcoxon signed-rank test.

Results

Within LLMs, ChatGPT-3.5 demonstrated the highest accuracy (P = 0.009). Gemini performed best in completeness (P < 0.001), personalization (P = 0.007) and readability (P < 0.001). PIRs were consistent in accuracy and completeness, with the general practitioner’s website excelling in personalization (P = 0.013) and readability (P < 0.001). The best-performing LLMs outperformed the GS-PIR on completeness and personalization, yet it was less accurate and less readable. Over time, response reproducibility decreased for all LLMs, showing variability across outcomes.

Conclusions

Although LLMs show potential in providing highly personalized and complete responses to patient questions regarding melanoma, improving and safeguarding accuracy, reproducibility and accessibility is crucial before they can replace or complement conventional PIRs.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

MelanomaSkin NeoplasmsPatient Education as TopicHumansComprehensionReproducibility of ResultsNetherlandsMaleFemaleSurveys and QuestionnairesMiddle AgedLarge Language ModelsGenerative Artificial Intelligence

Autor:innen und Institutionen

  • Department of Dermatology, Erasmus MC Cancer Institute, University Medical Center Rotterdam, the Netherlands.
  • Department of Surgery, Albert Schweitzer Hospital, Dordrecht, the Netherlands.
  • Dutch College of General Practitioners, PO Box 3231, Utrecht, the Netherlands.
  • Care and Public Health Research Institute, Department Family Medicine, Maastricht UMC+, Maastricht, the Netherlands.
  • Department of Radiology & Nuclear Medicine, Erasmus MC Cancer Institute, University Medical Center Rotterdam, the Netherlands.
  • Department of Medical Oncology, Erasmus MC Cancer Institute, University Medical Center Rotterdam, the Netherlands.

ORCID-Profile: Nadia C W Kamminga, June E C Kievits, Marlies Wakkee, Marjolein Lugtenberg, Tamar Nijsten

Förderung

  • Bristol Myers Squibb

Zitieren

APA

Kamminga, N. C. W., Kievits, J. E. C., Plaisier, P. W., Burgers, J. S., van der Veldt, A. M., van den Brand, J. A. G. J., Mulder, M., Wakkee, M., Lugtenberg, M., & Nijsten, T. (2025). Do large language model chatbots perform better than established patient information resources in answering patient questions? A comparative study on melanoma. British Journal of Dermatology. https://doi.org/10.1093/bjd/ljae377

Vancouver

Kamminga NCW, Kievits JEC, Plaisier PW, Burgers JS, van der Veldt AM, van den Brand JAGJ, et al. Do large language model chatbots perform better than established patient information resources in answering patient questions? A comparative study on melanoma. British Journal of Dermatology. 2025. doi:10.1093/bjd/ljae377

BibTeX
@article{Kamminga2025,
  author = {Kamminga, Nadia C W and Kievits, June E C and Plaisier, Peter W and Burgers, Jako S and van der Veldt, Astrid M and van den Brand, Jan A G J and Mulder, Mark and Wakkee, Marlies and Lugtenberg, Marjolein and Nijsten, Tamar},
  title = {Do large language model chatbots perform better than established patient information resources in answering patient questions? A comparative study on melanoma},
  journal = {British Journal of Dermatology},
  year = {2025},
  doi = {10.1093/bjd/ljae377}
}
Kostenlos starten →