EUVIMEDEUVIMED entdecken
VergleichsstudieOpen Access

Patient-Friendly Simplifications of Outpatient Reports Using Large Language Models via Few-Shot Prompting: Blinded Comparative Evaluation

Timo Elias Reimann, Vanessa Haug, Filippo Maria Verri, Christoph Leinert, Tim Fleiner, Michael Denkinger, Thomas Derya Kocar

Journal of Medical Internet Research · 06.11.2026 · 28 · e101882-e101882

DOI: 10.2196/101882 ↗

Kurz gefasst

Vergleichsstudie in der Zeitschrift Journal of Medical Internet Research, online erschienen am 06.11.2026. Verfasst von Timo Elias Reimann und 6 weiteren Autor:innen. Beteiligte Einrichtung: Institute for Geriatric Research Ulm, Ulm University Medical Center, Zollernring 26, Ulm, 89073, Germany, 49 7311870. Hauptthemen laut MEDLINE: Outpatients. Die Arbeit nennt 26 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Effective communication between clinicians and patients plays an important role in the successful delivery of health care. In geriatric medicine, where multimorbidity, polypharmacy, and functional limitations often make treatment plans complex, AI simplifications of clinical reports into patient-friendly language offer a promising way to improve communication.

Objective

This study investigated whether AI could generate patient-friendly simplifications of outpatient reports that were comparable to those authored by human experts.

Methods

Gemma-3-27B-it, an instruction fine-tuned open-source large language model (LLM) by Google, was used, and few-shot prompting was applied to enable the generation of patient-friendly simplifications of 64 deidentified geriatric outpatient reports written in German while preserving clinical meaning. LLM-generated simplifications were compared with those produced by 2 medical experts in a blinded evaluation. Quality was assessed using three approaches: (1) embedding-based semantic similarity to the original report, (2) blinded expert review by 2 clinicians, and (3) review by patient representatives, with the latter 2 using a standardized Likert-scale questionnaire.

Results

AI-generated simplifications were comparable to geriatrician-authored simplifications regarding semantic similarity ( P =.02) and were preferred by patient representatives ( P

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

OutpatientsHumansLarge Language ModelsArtificial IntelligenceAged

Stichwörter der Autor:innen

AIartificial intelligencefew-shot promptinglarge language modelsmedical reportoutpatient clinic reportpatient communicationsimplification

Autor:innen und Institutionen

  • Institute for Geriatric Research Ulm, Ulm University Medical Center, Zollernring 26, Ulm, 89073, Germany, 49 7311870.
  • Department of Physiotherapy, Institute of Medical Engineering and Mechatronics, Ulm University of Applied Sciences, Ulm, Germany.

Zitieren

APA

Reimann, T. E., Haug, V., Verri, F. M., Leinert, C., Fleiner, T., Denkinger, M., & Kocar, T. D. (2026). Patient-Friendly Simplifications of Outpatient Reports Using Large Language Models via Few-Shot Prompting: Blinded Comparative Evaluation. Journal of Medical Internet Research, 28, e101882-e101882. https://doi.org/10.2196/101882

Vancouver

Reimann TE, Haug V, Verri FM, Leinert C, Fleiner T, Denkinger M, et al. Patient-Friendly Simplifications of Outpatient Reports Using Large Language Models via Few-Shot Prompting: Blinded Comparative Evaluation. Journal of Medical Internet Research. 2026;28:e101882-e101882. doi:10.2196/101882

BibTeX
@article{Reimann2026,
  author = {Reimann, Timo Elias and Haug, Vanessa and Verri, Filippo Maria and Leinert, Christoph and Fleiner, Tim and Denkinger, Michael and Kocar, Thomas Derya},
  title = {Patient-Friendly Simplifications of Outpatient Reports Using Large Language Models via Few-Shot Prompting: Blinded Comparative Evaluation},
  journal = {Journal of Medical Internet Research},
  year = {2026},
  volume = {28},
  pages = {e101882--e101882},
  doi = {10.2196/101882}
}
Kostenlos starten →