Maximilian Haas, Veronika Saberi, Christopher Gossler, Anna Schmelzer, Anton Kravchuk, Johannes Breyer, Johannes Bründl, Simon Engelmann, Clemens Kirschner, Christian Gilfrich, Maximilian Burger, Dominik von Winning, Christian Wülfing, Hendrik Borgmann, Severin Rodler, Axel S Merseburger, Emily Rinderknecht, Matthias May
{"title":"[使前列腺癌研究更容易获得:chatGPT-4作为加强外行交流的工具]。","authors":"Maximilian Haas, Veronika Saberi, Christopher Gossler, Anna Schmelzer, Anton Kravchuk, Johannes Breyer, Johannes Bründl, Simon Engelmann, Clemens Kirschner, Christian Gilfrich, Maximilian Burger, Dominik von Winning, Christian Wülfing, Hendrik Borgmann, Severin Rodler, Axel S Merseburger, Emily Rinderknecht, Matthias May","doi":"10.1007/s00120-025-02558-w","DOIUrl":null,"url":null,"abstract":"<p><strong>Background and objective: </strong>The abundant potential arising from various applications of artificial intelligence is gradually influencing academic and scientific communication. This study examines the suitability of ChatGPT‑4 for generating layperson's summaries (LS) of scientific articles published in the four journals within the European Urology family and compares the quality of these newly generated LS with the original texts.</p><p><strong>Methods: </strong>A total of 327 articles on prostate cancer published between January 1, 2023, and June 30, 2024, were analyzed. ChatGPT‑4 generated patient summaries using both a basic and an advanced prompt, the latter specifically optimized for enhancing readability. Readability was assessed using established indices, while two blinded reviewers evaluated content quality on a 5-point Likert scale. Additionally, readability, content quality, and adherence to journal guidelines were combined into an overall scoring system.</p><p><strong>Results: </strong>The advanced prompt led to significantly improved readability compared to the basic prompt (p < 0.001) and the original LS (p < 0.001). Content quality was comparable between the two ChatGPT‑4 prompts (p = 0.665) but was higher than that of the original summaries (p = 0.001 and p = 0.002, respectively). Both prompts demonstrated superior adherence to journal guidelines (p < 0.001), with error-free LS rates of 29.4% (original), 76.1% (basic prompt), and 92% (advanced prompt) (p < 0.001).</p><p><strong>Conclusion: </strong>ChatGPT‑4 is a validated and effective tool for generating LS, offering superior readability and high compliance with editorial guidelines. It has the potential to assist researchers and scientific journals in enhancing the accessibility and comprehensibility of scientific content, thereby, improving patient engagement and understanding.</p>","PeriodicalId":29782,"journal":{"name":"Urologie","volume":" ","pages":""},"PeriodicalIF":0.5000,"publicationDate":"2025-04-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"[Making prostate cancer research accessible: chatGPT-4 as a tool to enhance lay communication].\",\"authors\":\"Maximilian Haas, Veronika Saberi, Christopher Gossler, Anna Schmelzer, Anton Kravchuk, Johannes Breyer, Johannes Bründl, Simon Engelmann, Clemens Kirschner, Christian Gilfrich, Maximilian Burger, Dominik von Winning, Christian Wülfing, Hendrik Borgmann, Severin Rodler, Axel S Merseburger, Emily Rinderknecht, Matthias May\",\"doi\":\"10.1007/s00120-025-02558-w\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Background and objective: </strong>The abundant potential arising from various applications of artificial intelligence is gradually influencing academic and scientific communication. This study examines the suitability of ChatGPT‑4 for generating layperson's summaries (LS) of scientific articles published in the four journals within the European Urology family and compares the quality of these newly generated LS with the original texts.</p><p><strong>Methods: </strong>A total of 327 articles on prostate cancer published between January 1, 2023, and June 30, 2024, were analyzed. ChatGPT‑4 generated patient summaries using both a basic and an advanced prompt, the latter specifically optimized for enhancing readability. Readability was assessed using established indices, while two blinded reviewers evaluated content quality on a 5-point Likert scale. Additionally, readability, content quality, and adherence to journal guidelines were combined into an overall scoring system.</p><p><strong>Results: </strong>The advanced prompt led to significantly improved readability compared to the basic prompt (p < 0.001) and the original LS (p < 0.001). Content quality was comparable between the two ChatGPT‑4 prompts (p = 0.665) but was higher than that of the original summaries (p = 0.001 and p = 0.002, respectively). Both prompts demonstrated superior adherence to journal guidelines (p < 0.001), with error-free LS rates of 29.4% (original), 76.1% (basic prompt), and 92% (advanced prompt) (p < 0.001).</p><p><strong>Conclusion: </strong>ChatGPT‑4 is a validated and effective tool for generating LS, offering superior readability and high compliance with editorial guidelines. It has the potential to assist researchers and scientific journals in enhancing the accessibility and comprehensibility of scientific content, thereby, improving patient engagement and understanding.</p>\",\"PeriodicalId\":29782,\"journal\":{\"name\":\"Urologie\",\"volume\":\" \",\"pages\":\"\"},\"PeriodicalIF\":0.5000,\"publicationDate\":\"2025-04-07\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Urologie\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1007/s00120-025-02558-w\",\"RegionNum\":4,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q4\",\"JCRName\":\"UROLOGY & NEPHROLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Urologie","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1007/s00120-025-02558-w","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"UROLOGY & NEPHROLOGY","Score":null,"Total":0}
[Making prostate cancer research accessible: chatGPT-4 as a tool to enhance lay communication].
Background and objective: The abundant potential arising from various applications of artificial intelligence is gradually influencing academic and scientific communication. This study examines the suitability of ChatGPT‑4 for generating layperson's summaries (LS) of scientific articles published in the four journals within the European Urology family and compares the quality of these newly generated LS with the original texts.
Methods: A total of 327 articles on prostate cancer published between January 1, 2023, and June 30, 2024, were analyzed. ChatGPT‑4 generated patient summaries using both a basic and an advanced prompt, the latter specifically optimized for enhancing readability. Readability was assessed using established indices, while two blinded reviewers evaluated content quality on a 5-point Likert scale. Additionally, readability, content quality, and adherence to journal guidelines were combined into an overall scoring system.
Results: The advanced prompt led to significantly improved readability compared to the basic prompt (p < 0.001) and the original LS (p < 0.001). Content quality was comparable between the two ChatGPT‑4 prompts (p = 0.665) but was higher than that of the original summaries (p = 0.001 and p = 0.002, respectively). Both prompts demonstrated superior adherence to journal guidelines (p < 0.001), with error-free LS rates of 29.4% (original), 76.1% (basic prompt), and 92% (advanced prompt) (p < 0.001).
Conclusion: ChatGPT‑4 is a validated and effective tool for generating LS, offering superior readability and high compliance with editorial guidelines. It has the potential to assist researchers and scientific journals in enhancing the accessibility and comprehensibility of scientific content, thereby, improving patient engagement and understanding.