Comment on "Online and Chatgpt-generated patient education materials regarding brain tumor prognosis fail to meet readability standards"


Sivri I., Ozden F. M., Gul G., Kaygin E., Colak T.

JOURNAL OF CLINICAL NEUROSCIENCE, cilt.142, 2025 (SCI-Expanded, Scopus) identifier identifier identifier

  • Yayın Türü: Makale / Kısa Makale
  • Cilt numarası: 142
  • Basım Tarihi: 2025
  • Doi Numarası: 10.1016/j.jocn.2025.111699
  • Dergi Adı: JOURNAL OF CLINICAL NEUROSCIENCE
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, EMBASE, MEDLINE
  • Kocaeli Üniversitesi Adresli: Evet

Özet

This correspondence comments on the recent article by Shukla and Sun (2025), which compared the readability of online and artificial intelligence-generated patient education materials in neuro-oncology. While the study provides meaningful insights, several methodological aspects deserve further consideration. The exclusive use of zero-shot prompts may have limited the model's adaptive capability, potentially contributing to low readability scores. Employing structured prompting strategies, such as one-shot or few-shot methods, and patient-centered instructions like "Explain in simple terms," could yield more accessible content. Moreover, since ChatGPT has gained internet browsing functionality as of February 2025, future studies integrating this feature may produce different outcomes regarding accuracy, readability, and clinical relevance. Together, these refinements could enhance the educational utility of Artificial Intelligence-generated health information.