Scholarlink

<p>Written educational materials have been found to be effective in the delivery of pre-operative information. However, creating personalized educational materials using large-language models (LLM) such as GPT-4 tailored to parental educational level has not yet been described in pediatric cardiology. A prospective single-center quality improvement study from December 2023 to March 2024 was completed with ChatGPT used to generate two pediatric cardiac catheterization information sheets at two different reading levels (6th and 10th grade) to improve procedural understanding by families. Surveys were distributed according to the highest level of education of the&#xa0;parent, along with a clinic satisfaction survey using a Likert scale. Twenty-six families were recruited. ChatGPT rapidly and accurately generated information sheets at the reading level requested. Mean pre- and post-education sheet Likert scale scores for “do you feel well-informed why your child’s cardiac catheterization is being done?” were 4.27 and 4.85, respectively, with a significant mean improvement of 0.57 (<i>p</i> &lt; 0.01). Families responded that the education sheet improved their overall satisfaction of the clinic visit with a mean survey score of 4.5 ± 0.8. 96% of families responded that the education sheet improved their understanding of the child’s procedure. Our study demonstrates that LLMs such as GPT-4 can be valuable tools to adjust medical education to a specific reading level and augment a pre-procedural visit with healthcare professionals by improving clinic satisfaction and understanding of an otherwise complex procedure.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Customized Education Sheets Generated by ChatGPT Improve Parental Visit Satisfaction and Procedural Knowledge Prior to Pediatric Cardiac Catheterization

  • 期刊论文

  • 发表日期：

    2025年10月28日

查看全文
  • Sarah Pradhan,
  • Kate McCormick,
  • Luc Mertens,
  • Faruq Pradhan

摘要

Written educational materials have been found to be effective in the delivery of pre-operative information. However, creating personalized educational materials using large-language models (LLM) such as GPT-4 tailored to parental educational level has not yet been described in pediatric cardiology. A prospective single-center quality improvement study from December 2023 to March 2024 was completed with ChatGPT used to generate two pediatric cardiac catheterization information sheets at two different reading levels (6th and 10th grade) to improve procedural understanding by families. Surveys were distributed according to the highest level of education of the parent, along with a clinic satisfaction survey using a Likert scale. Twenty-six families were recruited. ChatGPT rapidly and accurately generated information sheets at the reading level requested. Mean pre- and post-education sheet Likert scale scores for “do you feel well-informed why your child’s cardiac catheterization is being done?” were 4.27 and 4.85, respectively, with a significant mean improvement of 0.57 (p < 0.01). Families responded that the education sheet improved their overall satisfaction of the clinic visit with a mean survey score of 4.5 ± 0.8. 96% of families responded that the education sheet improved their understanding of the child’s procedure. Our study demonstrates that LLMs such as GPT-4 can be valuable tools to adjust medical education to a specific reading level and augment a pre-procedural visit with healthcare professionals by improving clinic satisfaction and understanding of an otherwise complex procedure.