Conformity of ChatGPT recommendations with the AUA/SUFU guideline on postprostatectomy urinary incontinence

Pinto, Vicktor B. P.; de Azevedo, Matheus F.; Wroclawski, Marcelo L.; Gentile, Guilherme; Jesus, Vinicius L. M.; de Bessa Junior, Jose; Nahas, William C.; Sacomani, Carlos A. R.; Sandhu, Jaspreet S.; Gomes, Cristiano M.

doi:10.1002/nau.25442

Cited by 4 publications

References 19 publications

(80 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

ChatGPT and the AUA/SUFU guideline on postprostatectomy urinary incontinence

Daungsupawong,

Wiwanitkit

2024

Neurourology and Urodynamics

View full text Add to dashboard Cite

ChatGPT and the AUA/SUFU guideline on postprostatectomy urinary incontinence Dear Editor, We would like to discuss on "Conformity of ChatGPT recommendations with the AUA/SUFU guideline on postprostatectomy urinary incontinence." 1 Using ChatGPT versions 3.5 and 4, a series of questions based on the AUA/SUFU Guideline were assessed. According to the findings, ChatGPT 4 fared better than ChatGPT 3.5 overall and especially in the conceptual and case-based questions. ChatGPT 4 scored 90.0% accuracy, demonstrating an improvement in performance over ChatGPT 3.5, which scored 57.5% accuracy.Compared to ChatGPT 3.5, ChatGPT 4 did a better job of correctly answering both conceptual and case-based questions in terms of certain question kinds. Both versions, meanwhile, were found to have certain shortcomings, mainly in the areas of assessment, available treatments, surgical complications, and unique circumstances. These flaws imply that the model's capacity to comprehend intricate medical ideas may be restricted in some circumstances.Future iterations of the ChatGPT model could concentrate on strengthening its performance in the areas where it was shown to be weak, including assessment, available treatments, surgical complications, and unique circumstances. Furthermore, investigating methods to improve the model's comprehension of intricate medical standards and settings may result in improved general accuracy and dependability when delivering medical information. The model's performance in answering medical questions could be enhanced by working with medical experts to refine it and by adding real-time feedback methods. Finally, since human user input is the foundation for all ChatGPT output, human user behavior code is required. 2

show abstract

ChatGPT and the AUA/SUFU guideline on postprostatectomy urinary incontinence

Daungsupawong,

Wiwanitkit

2024

Neurourology and Urodynamics

View full text Add to dashboard Cite

show abstract

Artificial Intelligence in Hand Surgery – How Generative AI is Transforming the Hand Surgery Landscape

TAN,

TEO,

PUHAINDRAN

2024

J Hand Surg Asian-Pac Vol

View full text Add to dashboard Cite

Artificial intelligence (AI) has witnessed significant advancements, reshaping various industries, including healthcare. The introduction of ChatGPT by OpenAI in November 2022 marked a pivotal moment, showcasing the potential of generative AI in revolutionising patient care, diagnosis and treatment. Generative AI, unlike traditional AI systems, possesses the ability to generate new content by understanding patterns within datasets. This article explores the evolution of AI in healthcare, tracing its roots to the term coined by John McCarthy in 1955 and the contributions of pioneers like John Von Neumann and Alan Turing. Currently, generative AI, particularly Large Language Models, holds promise across three broad categories in healthcare: patient care, education and research. In patient care, it offers solutions in clinical document management, diagnostic support and operative planning. Notable advancements include Microsoft’s collaboration with Epic for integrating AI into electronic medical records (EMRs), enhancing clinical data management and patient care. Furthermore, generative AI aids in surgical decision-making, as demonstrated in plastic, orthopaedic and hepatobiliary surgeries. However, challenges such as bias, hallucination and integration with EMR systems necessitate caution and ongoing evaluation. The article also presents insights from the implementation of NUHS Russell-GPT, a generative AI chatbot, in a hand surgery department, showcasing its utility in administrative tasks but highlighting challenges in surgical planning and EMR integration. The survey showed unanimous support for incorporating AI into clinical settings, with all respondents being open to its use. In conclusion, generative AI is poised to enhance patient care and ease physician workloads, starting with automating administrative tasks and evolving to inform diagnoses, tailored treatment plans, as well as aid in surgical planning. As healthcare systems navigate the complexities of integrating AI, the potential benefits for both physicians and patients remain significant, offering a glimpse into a future where AI transforms healthcare delivery. Level of Evidence: Level V (Diagnostic)

show abstract