Asal Maha Gamal Ramadan, Alsenany Samira Ahmed, Badoman Talal Emad Ahmed, El-Sayed Ahmed Abdelwahab Ibrahim
Nursing Department, College of Pharmacy and Applied Medical Sciences, Dar Al Uloom University, Riyadh, Saudi Arabia.
Public Health Department, Faculty of Nursing, King Abdulaziz University, Jeddah, Saudi Arabia.
J Eval Clin Pract. 2025 Aug;31(5):e70241. doi: 10.1111/jep.70241.
The integration of large language models (LLMs) into healthcare offers transformative potential but raises significant ethical challenges. Addressing these challenges requires a comprehensive framework to assess healthcare professionals' ethical awareness of LLMs usage.
To develop and validate a scale designed to evaluate healthcare professionals' ethical awareness regarding the integration of LLMs into clinical practice.
This two-phase methodological study was conducted in 2024 across nine healthcare institutions-five in Egypt and four in Saudi Arabia. In Phase I, a comprehensive literature review and semi-structured interviews with healthcare professionals guided the development of the initial scale and item pool. In Phase II, the psychometric properties of the scale were evaluated using data collected from 658 healthcare professionals. Construct validity was assessed through exploratory and confirmatory factor analyses, while internal consistency reliability was examined using Cronbach's alpha (α) coefficients and item-total correlation metrics.
An initial pool of 36 items was refined to 21 items across 6 dimensions: data privacy and confidentiality, consent and autonomy, transparency and accountability, bias and equity, safety and professional integrity, and education and sustainability. EFA identified a six-factor structure accounting for 71.5% of the variance. CFA confirmed the scale's structure with strong model fit indices for first-order analysis (CMIN/DF = 1.798, CFI = 0.967, RMSEA = 0.050) and second-order analysis (CMIN/DF = 2.862, CFI = 0.927, RMSEA = 0.058). The scale demonstrated excellent internal consistency (overall Cronbach's α = 0.90; dimensions ranging from 0.780 to 0.964) and achieved satisfactory composite reliability, convergent validity and discriminant validity. Moderate statistically significant inter-factor correlations confirmed the multidimensional nature of the scale.
The developed scale is a valid and reliable tool for assessing healthcare professionals' ethical awareness in the use of LLMs in healthcare. It provides a comprehensive framework for evaluating and enhancing ethical considerations, promoting the responsible and informed use of LLMs technologies in clinical practice.
将大语言模型（LLMs）整合到医疗保健领域具有变革潜力，但也带来了重大的伦理挑战。应对这些挑战需要一个全面的框架来评估医疗保健专业人员对LLMs使用的伦理意识。
开发并验证一个量表，旨在评估医疗保健专业人员对将LLMs整合到临床实践中的伦理意识。
这项两阶段的方法学研究于2024年在九个医疗保健机构进行，其中五个在埃及，四个在沙特阿拉伯。在第一阶段，对医疗保健专业人员进行全面的文献综述和半结构化访谈，指导初始量表和项目池的开发。在第二阶段，使用从658名医疗保健专业人员收集的数据评估量表的心理测量特性。通过探索性和验证性因素分析评估结构效度，同时使用克朗巴哈α（α）系数和项目-总分相关指标检查内部一致性可靠性。
最初的36个项目池被提炼为21个项目，涵盖6个维度：数据隐私和保密性、同意和自主权、透明度和问责制、偏差和公平性、安全性和专业诚信，以及教育和可持续性。探索性因素分析确定了一个六因素结构，解释了71.5%的方差。验证性因素分析通过一阶分析（CMIN/DF = 1.798，CFI = 0.967，RMSEA = 0.050）和二阶分析（CMIN/DF = 2.862，CFI = 0.927，RMSEA = 0.058）的强模型拟合指数确认了量表的结构。该量表显示出优异的内部一致性（总体克朗巴哈α = 0.90；各维度范围从0.780到0.964），并实现了令人满意的组合可靠性、收敛效度和区分效度。适度的具有统计学意义的因素间相关性证实了该量表的多维性质。
所开发的量表是评估医疗保健专业人员在医疗保健中使用LLMs时伦理意识的有效且可靠的工具。它为评估和加强伦理考量提供了一个全面的框架，促进在临床实践中负责任且明智地使用LLMs技术。
