#242 Designing a clinical trial with chat generative pre-trained transformer (ChatGPT) on the role of hyperthermic intraperitoneal chemotherapy (HIPEC) in ovarian cancer

Abstract

<h3>Introduction/Background</h3> We live in an area of artificial intelligence(AI), currently used in research and medical practice. One of the most recent developments is Chat Generative Pre-trained Transformer(ChatGPT), launched on November 30, 2022 by OpenAI, an AI laboratory. The ChatGPT faced serious critiques from the scientific community particularly its use on manuscript writing. However, soon its potential use and aid in medical education and scientific writing have been appreciated by many researchers and publishing groups. We aimed to investigate the ChatGPT’s performance on designing a new randomised clinical trial(RCT) on the role of hyperthermic intraperitoneal chemotherapy(HIPEC) in ovarian cancer. <h3>Methodology</h3> We commend following to ChatGPT ‘’Design a new randomized controlled clinical trial on the role of HIPEC in ovarian cancer’ on May 1 and 7 2023. We assess the changes one week period. <h3>Results</h3> The first query involved the following subsections (n=12): title, objective, study design, participants, sample size, randomization, blinding, interventions, outcome measures, statistical analysis, ethical considerations, conclusion. It consisted of 482 words. The second query involved four main subsections: title, introduction, study design (randomization, blinding, sample size calculation, eligibility criteria, intervention, outcome measures, statistical analysis, ethical considerations), and conclusion. It consisted of 474 words. The second query was superior to the first one: i)concise, ii)higher level of details, and iii) better structure. Both queries lacked the dose, chemotherapeutic agent, duration, and temperature of the HIPEC. Both accepted ovarian cancer as stage III or IV. We did not detect any false information. both showed a highly acceptable level of scientific writing for the publication purposes. ChatGPT achieved to create a new RCT. <h3>Conclusion</h3> Given the ChatGPT’s shortly release date, our comparative results are supportive of its quickly evolving performance. In conclusion, we believe ChatGPT is a promising tool in medical education on focus of scientific writing and stimulating ideas for initial RCT design. <h3>Disclosures</h3> None

Ähnliche Arbeiten