Garcia Valencia Oscar A, Thongprayoon Charat, Jadlowiec Caroline C, Mao Shennen A, Leeaphorn Napat, Budhiraja Pooja, Khoury Nadeen, Pham Justin H, Craici Iasmina M, Gonzalez Suarez Maria L, Cheungpasitporn Wisit
Division of Nephrology and Hypertension, Department of Medicine, Mayo Clinic, Rochester, MN, United States.
Division of Transplant Surgery, Department of Surgery, Mayo Clinic, Phoenix, AZ, United States.
Front Public Health. 2025 Jan 27;13:1484790. doi: 10.3389/fpubh.2025.1484790. eCollection 2025.
Health equity and access to essential medical information remain significant challenges, especially for the Spanish-speaking Hispanic population, which faces barriers in accessing living kidney donation opportunities. ChatGPT, an AI language model with sophisticated natural language processing capabilities, has been identified as a promising tool for translating critical health information into Spanish. This study aims to assess ChatGPT's translation efficacy to ensure the information provided is accurate and culturally relevant.
his study utilized ChatGPT versions 3.5 and 4.0 to translate 27 frequently asked questions (FAQs) from English to Spanish, sourced from Donate Life America's website. The translated content was reviewed by native Spanish-speaking nephrologists using a standard rubric scale (1-5). The assessment focused on linguistic accuracy and cultural sensitivity, emphasizing retention of the original message, appropriate vocabulary and grammar, and cultural relevance.
The mean linguistic accuracy scores were 4.89 ± 0.32 for GPT-3.5 and 5.00 ± 0.00 for GPT-4.0 ( = 0.08). The percentage of excellent-quality translations (score = 5) in linguistic accuracy was 89% for GPT-3.5 and 100% for GPT-4.0 ( = 0.24). The mean cultural sensitivity scores were 4.89 ± 0.32 for both GPT-3.5 and GPT-4.0 ( = 1.00). Similarly, excellent-quality translations in cultural sensitivity were achieved in 89% of cases for both versions ( = 1.00).
ChatGPT 4.0 demonstrates strong potential to enhance health equity by improving Spanish-speaking Hispanic patients' access to LKD information through accurate and culturally sensitive translations. These findings highlight the role of AI in mitigating healthcare disparities and underscore the need for integrating AI-driven tools into healthcare systems. Future efforts should focus on developing accessible platforms and establishing guidelines to maximize AI's impact on equitable healthcare delivery and patient education.
健康公平和获取基本医疗信息仍然是重大挑战，尤其是对于讲西班牙语的西班牙裔人群而言，他们在获取活体肾捐赠机会方面面临障碍。ChatGPT是一种具有先进自然语言处理能力的人工智能语言模型，已被视为将关键健康信息翻译成西班牙语的有前景的工具。本研究旨在评估ChatGPT的翻译效果，以确保所提供的信息准确且具有文化相关性。
本研究使用ChatGPT 3.5版和4.0版将来自美国捐赠生命网站的27个常见问题（FAQ）从英语翻译成西班牙语。由以西班牙语为母语的肾病学家使用标准评分量表（1 - 5分）对翻译内容进行审核。评估重点在于语言准确性和文化敏感性，强调保留原文信息、恰当的词汇和语法以及文化相关性。
GPT - 3.5的平均语言准确性得分为4.89 ± 0.32，GPT - 4.0为5.00 ± 0.00（P = 0.08）。在语言准确性方面，GPT - 3.5的优质翻译（得分 = 5）百分比为89%，GPT - 4.0为100%（P = 0.24）。GPT - 3.5和GPT - 4.0的平均文化敏感性得分均为4.89 ± 0.32（P = 1.00）。同样，两个版本在89%的案例中都实现了文化敏感性方面的优质翻译（P = 1.00）。
ChatGPT 4.0通过准确且具有文化敏感性的翻译，提高讲西班牙语的西班牙裔患者获取活体肾捐赠信息的机会，显示出增强健康公平性的强大潜力。这些发现凸显了人工智能在减轻医疗保健差距方面的作用，并强调了将人工智能驱动的工具整合到医疗保健系统中的必要性。未来的努力应集中在开发可访问的平台并制定指南，以最大限度地发挥人工智能对公平医疗服务提供和患者教育的影响。
