Hierarchical meta-regression: a unified approach to model multiplicity of bias in combining randomized and non-randomized evidence in meta-analysis

本项目的目的是在Meta分析中结合随机化和非随机化的证据，为偏差建模提供一种综合的方法。将要开发的统计方法是分层的Meta回归模型。该模型允许结合不同研究类型的证据片段(例如随机对照试验、观察性队列研究等)。和不同的数据类型(来自出版物的聚合结果、单个患者数据等)。该模型明确区分了用于数据收集过程的参数(偏差建模)和用于临床问题的参数(例如，患者组的治疗效果或诊断准确性)。在这种方法的基础上，将实施一个R包。该项目应对统计证据合成和卫生技术评估作出重大贡献。

The aim of this project is to provide an integrated approach for bias modeling when randomized and non-randomized evidence is combined in meta-analysis. The statistical approach to be developed is a hierarchical meta-regression model. This model allows combining pieces of evidence of different study-types (e.g. RCTs, observational cohort studies, etc.) and different data types (aggregated results from publications, individual patient data, etc.). The model explicitly distinguishes parameters used for data collection processes (modeling of bias) and parameters used for clinical questions (e.g. treatment effects in group of patients or diagnostic accuracy). Based on this approach an R-package will be implemented. The project should make a significant contribution to statistical evidence synthesis and health technology assessment.