基于医院的上市后高风险医疗器械快速评估体系的循证构建

医疗器械近年呈快速发展趋势，我国对器械按风险高低实施分类管理，对高风险器械严格管控，但器械上市管理远不如药品严格，存在已有证据质量较低、证据不足、证据缺乏问题。加之面对大量频繁更新、使用日益复杂，费用日益昂贵的医疗器械，医院层面亟需对医疗器械开展准入评估。医院卫生技术评估（HB-HTA）随之出现，我国通过部分医院试点引入。快速生产高质量证据是HB-HTA的核心，但器械领域存在广泛采用替代指标、缺乏证据分级体系，及如何在确保评估质量情况下实现快速评估的问题。本课题旨在基于信息论方法，建立高风险器械替代指标验证研究的方法学体系，形成替代指标可靠评估模型；基于GRADE工具，采用Delphi法及综合指数法，构建高风险器械证据分级体系；并基于Meta流行病学和机器学习方法，针对评价限速环节循证重构高风险器械评估流程。有效解决缺乏器械证据分级体系，难以获取远期、终点指标结果，及评估速度慢的问题。

Medical devices have been developing rapidly in recent years. Devices were managed using the classification method according to the level of risk in China. Consequently, devices at high risk were tightly controlled. However, the management of getting devices on the market is far less effective than that of getting drugs. The existing problems are insufficient evidence and lack of high-quality evidence. Additionally, medical devices are frequently updated, increasingly complex, and increasingly expensive. Therefore, it is urgent to carry out access assessments for medical devices at the hospital level. Hospital-based health technology assessment (HB-HTA) has emerged, and it was introduced in some hospitals in China. The core method of HB-HTA is how to generate high-quality evidence rapidly. But there are problems in the field of devices, such as widely using surrogate endpoints, lack of evidence grading system, and spending a lot of time on access assessments. This research aims to establish a methodological system about the reliability of surrogate endpoints for devices at high risk based on the information theory and to build a relevant model; to construct an evidence grading system for devices at high risk using Delphi and comprehensive index method based on GRADE tools; and to re-construct the assessment process for devices at high risk using evidence-based method based on Meta-epidemiology and machine learning, especially for the rate limiting steps. This research can effectively solve the problems of lack of evidence grading system, difficulty in obtaining long-term endpoints, and slow assessment of medical devices.