MedBot vs RealDoc: efficacy of large language modeling in physician-patient communication for rare diseases.

Journal: Journal of the American Medical Informatics Association : JAMIA
PMID: 39998911

Abstract

OBJECTIVES: This study assesses the abilities of 2 large language models (LLMs), GPT-4 and BioMistral 7B, in responding to patient queries, particularly concerning rare diseases, and compares their performance with that of physicians.

Authors

  • Magdalena T Weber
    Institute of Medical Informatics, University Medicine Frankfurt, Goethe University Frankfurt, Frankfurt 60590, Germany.
  • Richard Noll
    Institute of Medical Informatics, University Medicine Frankfurt, Goethe University Frankfurt, Frankfurt 60590, Germany.
  • Alexandra Marchl
    Institute of Medical Informatics, University Medicine Frankfurt, Goethe University Frankfurt, Frankfurt 60590, Germany.
  • Carlo Facchinello
    Santagostino Medical Center, Bologna 40138, Italy.
  • Achim Grünewaldt
    Department of Respiratory Medicine and Allergology, University Medicine Frankfurt, Goethe University Frankfurt, Frankfurt 60590, Germany.
  • Christian Hügel
    HELIOS Dr Horst Schmidt Kliniken Wiesbaden, Klinik für Pneumologie, Wiesbaden 65199, Germany.
  • Khader Musleh
    Department of Respiratory Medicine and Allergology, University Medicine Frankfurt, Goethe University Frankfurt, Frankfurt 60590, Germany.
  • Thomas O F Wagner
    European Reference Network for Rare Respiratory Diseases (ERN-LUNG), University Medicine Frankfurt, Frankfurt 60590, Germany.
  • Holger Storf
    Institute of Medical Informatics, University Medicine Frankfurt, Goethe University Frankfurt, Frankfurt 60590, Germany.
  • Jannik Schaaf
    Institute of Medical Informatics, University Medicine Frankfurt, Goethe University Frankfurt, Frankfurt 60590, Germany.

Keywords

External Resources

Popular Topics
Recent Journals