构建适合我国国情的医疗人工智能技术评估体系探讨  

Discussion on Building a Medical Artificial Intelligence Technology Assessment System Suitable for Chinese National Conditions

作　　者：汪茹 何达[1] 符雨嫣 罗雅双 邢倩 王海银[1] Wang Ru;He Da;Fu Yuyan(Shanghai Health Development Research Center(Shanghai Medical Information Center),Shanghai,201199,China;不详)

机构地区：[1]上海市卫生和健康发展研究中心(上海市医学科学技术情报研究所),上海201199

出　　处：《中国卫生经济》2024年第10期38-43,共6页Chinese Health Economics

基　　金：国家卫生健康委医院管理研究所项目(2023008AB)。

摘　　要：目的:探讨构建适合我国国情的医疗人工智能技术评估体系。方法:对国内外传统卫生技术评估体系进行总结,分析医疗人工智能技术的显著特性和新风险因素。结果:现有技术特性评价指标中最有可能根据可靠性的次级评价指标对人工智能技术进行评估,但是可靠性评估的侧重点无法针对算法、大数据、算力及存在的风险因素进行评估,现有卫生技术评估体系有待进一步完善。结论:传统卫生技术评估体系难以适用于医疗人工智能技术的评估,建议以传统卫生技术评估体系为基础框架,新增与医疗人工智能技术特征相关的评价指标,参考相关领域的评估指标和方法加以调整、完善医疗人工智能技术评估体系。Objective:To explore the construction of a medical Artificial Intelligence(AI)technology assessment system suitable for the national conditions in China.Methods:Summarize the domestic and international traditional health technology assessment system,and analyze the distinctive characteristics and novel risk factors of medical AI technology.Results:The existing evaluation indexes of technical characteristics are most likely to assess AI technology based on the sub-evaluation indexes of reliability,but the evaluation focus of reliability cannot be assessed for algorithms,big data,arithmetic power and the existence of risk factors,and the existing system of health technology assessment needs to be further improved.Conclusion:Traditional health technology assessment system is challenging to apply to the evaluation of medical AI technology.It is recommended to use the traditional health technology assessment framework as a foundational structure,incorporate evaluation indicators related to the characteristics of medical AI technology,and refer to evaluation indicators and methods from relevant fields to adjust and refine the medical AI technology assessment system.

关 键 词：人工智能 医疗 卫生技术评估 评估体系 

分 类 号：R1-9[医药卫生—公共卫生与预防医学] R19-0