真实世界数据适用性评价指标体系研究

Study on applicability evaluation index system of real-world data

目的:建立科学的真实世界数据适用性评价指标体系,为真实世界数据转化为真实世界证据提供定量分析依据。方法:通过研读国内外相关指南、文件及文献,对影响真实世界数据适用性因素进行分层次提取分析,综合运用层次分析法和群组决策特征根法对选取指标建立递阶层级结构,构建专家打分矩阵,利用MATLAB软件为评价指标体系确定权重。结果:从相关性和可靠性两方面构建了真实世界数据适应性评价指标。相关性下设包括数据来源与样本、暴露/干预和事件结局定义、多源数据的整合、监管要求、伦理要求5个子准则层指标及26项具体评价指标;可靠性下设数据的有效性、质量控制、质量保证3个子准则层指标及27项具体评价指标。结论:从建立的评价指标体系与所赋权重来看,采集和应用真实世界数据时,应重点关注数据的来源、有效性及质量控制手段,以保证数据与需求密切相关且真实可靠,提高真实世界数据转化为真实世界证据的适用性。

Objective:To establish a scientific applicability evaluation index system of real-world data(RWD),and provide a quantitative analysis basis for the transformation of RWD into real-world evidence(RWE).Methods:Hierarchical extraction analysis of factors affecting the applicability of RWD was carried out by studying domestic and international guides,documents and literature.Analytic hierarchy process(AHP)and group eigenvalue method(GEM)were used in combination to establish the hierarchical structures of the selected indicators,expert scoring matrix was constructed,and MATLAB software for index system was used to compute the weights.Results:Applicability evaluation index was constructed from two aspects:relevance and reliability.Relevance included five sub-criteria level indexes and 26 specific evaluation indicators:data source and sample,exposure/intervention and event outcome definition,integration of multi-source data,regulatory requirements,ethics.Reliability consisted of three sub-criteria level indexes and 27 specific evaluation indicators:data validity,quality control,and quality assurance.Conclusion:When gathering and applying RWD,attention should be paid to the data source,validity,and quality control,to ensure that the data is closely related to the demand and is reliable,and then to upgrade the applicability from RWD to RWE.