Comment on “ChatGPT Answers Common Patient Questions About Colonoscopy”

I read with great interest the recent article in Gastroenterology by Lee et al1Lee T.-C. et al.Gastroenterology. 2023; 165: 509-511.e7Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar that aimed to evaluate the ability of ChatGPT to provide satisfactory answers to common patient questions about colonoscopy. As artificial intelligence systems like ChatGPT become more prevalent in health care, it is crucial that we critically evaluate their capabilities and limitations. The authors' findings that ChatGPT can provide responses comparable with traditional medical resources in understandability, adequacy, and satisfaction ratings are important initial validations. However, several limitations and perspectives deserve further discussion before integrating ChatGPT into clinical use. First, the study evaluated only 8 common questions retrieved from 3 hospital websites. Expanding the sample size and diversity of questions could provide more robust results. Because patient questions often focus on procedural expectations, quality of life impact, and cost concerns, evaluating ChatGPT's performance on these topics will offer additional clinically useful insights. Second, the responses generated by ChatGPT bypass the clinical reasoning involved in providing personalized recommendations, which considers patients' medical history and test results. Third, the lack of patient raters is an important limitation, as acknowledged by the authors. Including patient perspectives would more authentically assess the utility of artificial intelligence–generated medical information from the end-user standpoint. Qualitative feedback could reveal nuances not captured by quantitative ratings. Furthermore, ethical implications of reliance on ChatGPT for clinical use also warrant consideration.2Rahimzadeh V. et al.Am J Bioeth. 2023; 23: 1-11Google Scholar Despite the limitations, the insights gained from this study open up avenues for impactful work in the near future, which deserve more attention. Looking ahead, an exciting area for further exploration is training domain-specific models like 'MedGPT.' Retraining ChatGPT on a clinical corpus could potentially enhance the accuracy, readability, and bedside manner of responses. Collaborations with patients to co-design prompts and grading rubrics would uphold patient-centered principles. Moreover, evaluating the integration of ChatGPT into electronic health records could facilitate seamless clinical application. With the rapid advancement of generative artificial intelligence tools, such as ChatGPT, there is substantial promise in leveraging these tools to enhance patient education and foster collaborative decision-making, if implemented under strict clinical oversight and guidance. In this regard, the work by Lee et al1Lee T.-C. et al.Gastroenterology. 2023; 165: 509-511.e7Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar has opened an insightful avenue for future research. In summary, although Lee et al provide early data on an innovative concept, expanding the scope and methodology could offer more comprehensive insights to realize the full potential of ChatGPT in patient education and other promising clinical areas. ChatGPT Answers Common Patient Questions About ColonoscopyGastroenterologyVol. 165Issue 2PreviewChatGPT (OpenAI) is a 175 billion–parameter large language model (LLM) artificial intelligence (AI) that was released in November 2022. ChatGPT is developed based on the generative pretrained transformer (GPT) 3.5 natural language processing technology and provides a conversational text response to a given prompt.1 Full-Text PDF Open AccessReplyGastroenterologyVol. 166Issue 1PreviewWe appreciate the comments by Kleebayoon and Wiwanitkit,1 Wu,2 and Tariq et al3 about our article.4 We agree with these valuable comments, which point out the importance of physician oversight, patient involvement, integration of evidence-based medicine, and patient-specific information to the optimal design and use of conversational artificial intelligence (AI) as a tool for shared decision making and for communication between patients and health care providers. Full-Text PDF