Scholarlink

Purpose <p>Patients increasingly rely on online resources to better understand their health conditions. ChatGPT could satisfy the demand for reliable and accessible online health education resources, yet few studies have applied this to pediatric orthopaedic counseling. This study quantifies the accuracy and comprehensibility of ChatGPT responses to frequently asked questions (FAQs) regarding scoliosis.</p> Methods <p>Twelve FAQs regarding scoliosis were compiled following a literature review, and ChatGPT Version 3.5 was utilized to answer them. The responses were analyzed for accuracy and clarity using the Mika et al. scoring system and modified DISCERN score in collaboration with two fellowship-trained pediatric orthopaedic surgeons. Readability was assessed using several published educational-level indices.</p> Results <p>The ChatGPT responses received a Mika et al. average of 2.4 (satisfactory requiring minimal to moderate clarification) and an averaged mean DISCERN score of 45.9. The estimated reading level necessary for comprehension ranged from 11<sup>th</sup> grade to college graduate.</p> Conclusions <p>When prompted with 12 scoliosis FAQs, ChatGPT produces responses of satisfactory accuracy but require further clarification and are written at an inappropriately high reading level for the scoliosis patient population. Future research should explore strategies to verify the reliability of AI services for counseling on other pediatric orthopaedic conditions.</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

Patient education strategies in pediatric orthopaedics: using ChatGPT to answer frequently asked questions on scoliosis

  • 期刊论文

  • 发表日期：

    2025年4月5日

查看全文
  • Brigitte Lieu,
  • Ethan Crawford,
  • Logan Laubach,
  • Teja Yeramosu,
  • Chester Sharps,
  • Joanna Horstmann,
  • Victoria Kuester

摘要

Purpose

Patients increasingly rely on online resources to better understand their health conditions. ChatGPT could satisfy the demand for reliable and accessible online health education resources, yet few studies have applied this to pediatric orthopaedic counseling. This study quantifies the accuracy and comprehensibility of ChatGPT responses to frequently asked questions (FAQs) regarding scoliosis.

Methods

Twelve FAQs regarding scoliosis were compiled following a literature review, and ChatGPT Version 3.5 was utilized to answer them. The responses were analyzed for accuracy and clarity using the Mika et al. scoring system and modified DISCERN score in collaboration with two fellowship-trained pediatric orthopaedic surgeons. Readability was assessed using several published educational-level indices.

Results

The ChatGPT responses received a Mika et al. average of 2.4 (satisfactory requiring minimal to moderate clarification) and an averaged mean DISCERN score of 45.9. The estimated reading level necessary for comprehension ranged from 11th grade to college graduate.

Conclusions

When prompted with 12 scoliosis FAQs, ChatGPT produces responses of satisfactory accuracy but require further clarification and are written at an inappropriately high reading level for the scoliosis patient population. Future research should explore strategies to verify the reliability of AI services for counseling on other pediatric orthopaedic conditions.