文献速递烧伤创面修复

现代烧伤患者教育:ChatGPT 与 Google 回答烧伤及处理常见问题性能的比较分析

Burn Patient Education in the Modern Age: A Comparative Analysis of ChatGPT and Google Performance Answering Common Questions on Burn Injury and Management.
PubMed 文献 · J Burn Care Res · 2025-01-01

本研究比较 ChatGPT 与 Google 对烧伤及处理常见问题的回答质量。Google 搜索「burn」取前 10 个常见问题,再由 ChatGPT 回答,3 名烧伤与创伤外科医生用全局质量评分量表(1 分差至 5 分优)评价。ChatGPT 平均 4.57±0.73 分,Google 平均 2.80±1.03 分;半数问题上医生一致偏好患者从 ChatGPT 获取信息。

要点

  • ChatGPT 回答平均 4.57±0.73 分,Google 回答平均 2.80±1.03 分。
  • 3 名烧伤与创伤外科医生用全局质量评分量表评分,范围 1 至 5 分。
  • 半数问题上医生一致偏好患者从 ChatGPT 获取信息。

核心结论

  • 基于 3 名经验丰富的烧伤外科医生评价,ChatGPT 在回答烧伤及处理常见问题方面优于 Google。
  • 结果突显 ChatGPT 作为患者教育来源的潜力。

研究设计与样本

  • 研究使用 Google 搜索「burn」,记录前 10 个常见问题及其答案,再将这些问题输入 ChatGPT,由 3 名烧伤与创伤外科医生用全局质量评分量表评价回答质量。
  • 采用 Wilcoxon 配对 t 检验评估 Google 与 ChatGPT 回答得分差异。

主要发现

  • Google 回答平均得分 2.80±1.03,提示部分信息存在但重要主题缺失。
  • ChatGPT 生成回答平均得分 4.57±0.73,提示回答质量优秀且对患者实用价值高。
  • 对于半数问题,外科医生一致偏好患者从 ChatGPT 获取信息。

关键数值与推荐

  • 全局质量评分量表评分范围为 1 分(质量差)至 5 分(质量优)。
  • Google 回答平均 2.80±1.03 分,ChatGPT 回答平均 4.57±0.73 分。

临床怎么用

  • 原文未据此给出推荐。

局限性

  • 本文为 Google 与 ChatGPT 回答常见烧伤问题的初步比较。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

Patients often use Google for their medical questions. With the emergence of artificial intelligence large language models, such as ChatGPT, patients may turn to such technologies as an alternative source of medical information. This study investigates the safety, accuracy, and comprehensiveness of medical responses provided by ChatGPT in comparison to Google for common questions about burn injuries and their management. A Google search was performed using the term "burn," and the top 10 frequently searched questions along with their answers were documented. These questions were then prompted into ChatGPT. The quality of responses from both Google and ChatGPT was evaluated by 3 burn and trauma surgeons using the Global Quality Score scale, rating from 1 (poor quality) to 5 (excellent quality). A Wilcoxon paired t-test evaluated the difference in scores between Google and ChatGPT answers. Google answers scored an average of 2.80 ± 1.03, indicating that some information was present, but important topics were missing. Conversely, ChatGPT-generated answers scored an average of 4.57 ± 0.73, indicating excellent quality responses with high utility to patients. For half of the questions, the surgeons unanimously preferred their patients receive information from ChatGPT. This study presents an initial comparison of Google and ChatGPT responses to commonly asked burn injury questions. ChatGPT outperforms Google in responding to commonly asked questions on burn injury and management based on the evaluations of 3 experienced burn surgeons. These results highlight the potential of ChatGPT as a source of patient education.

来源:J Burn Care Res,查看原文。版权归原作者所有。

引用

  1. Pandya S, Alessandri Bonetti M, Liu HY, et al. Burn Patient Education in the Modern Age: A Comparative Analysis of ChatGPT and Google Performance Answering Common Questions on Burn Injury and Management. J Burn Care Res. 2025. PMID: 39761346. doi:10.1093/jbcr/irae211

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。