Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.
Abstract 18238: ChatGPT Has the Potential to Increase Accessibility to Cardiology Research
0
Zitationen
5
Autoren
2023
Jahr
Abstract
Introduction: Generative language models like ChatGPT have experienced a novel rise in their implementation. This study explored ChatGPT’s potential in simplifying cardiology research abstracts published in specific journals within the family of Journals of American College of Medicine (JACC). If successful, this can expand the understanding of cardiology research among the public and those with low health literacy. Research Question: Can ChatGPT transform abstracts to be understandable by patient populations who cannot understand medical text without including extraneous or false information? Goals: We quantified the readability of abstracts published in JACC with clinically verified metrics. We assessed the readability of these same abstracts post-ChatGPT transformation. Methods: Review article abstracts published in the JACC from January to June 2023 were simplified using ChatGPT. The following prompt was used in ChatGPT to transform abstracts: “Please rewrite the following text to be readable at a 5th-grade level. Do not include information not contained in the original text, and do not exclude information contained within the original text.” Readability was quantified using Flesch-Kincaid Grade Level (FKGL) and Reading Ease (FKRE) scores. A board-certified cardiologist assessed content validity and ensured accuracy. Results: All 24 abstracts had significantly lower FKGL and higher FKRE scores (P<0.001) after transformation by ChatGPT, indicating improved readability. No extraneous information was found in 24/24 transformed abstracts. Conclusion: Medical literature exceeds the average reading and comprehension level of US adults, especially among those with low socioeconomic status. Our results indicate that ChatGPT effectively simplifies complex cardiology literature, maintaining information accuracy. This opens the door for further exploration of ChatGPT's role in improving accessibility to various cardiology topics.
Ähnliche Arbeiten
Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI
2019 · 8.214 Zit.
Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead
2019 · 8.071 Zit.
High-performance medicine: the convergence of human and artificial intelligence
2018 · 7.429 Zit.
Proceedings of the 19th International Joint Conference on Artificial Intelligence
2005 · 5.776 Zit.
Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI)
2018 · 5.418 Zit.