论文部分内容阅读
常用的疫苗不良反应信号检测方法存在阈值确定比较武断、多重性等问题,导致信号检测的准确性降低。本文介绍了多重性问题对信号检测准确性的影响,揭示了基于错误发现率(False Discovery Rate,FDR)控制方法可解决信号检测中的多重性问题。基于FDR的信号检测方法,在提高了检测信号生成数量同时控制了生成信号中的错误数量。用R语言实现了基于FDR控制的疑似预防接种异常反应(Adverse Events Following Immunization,AEFI)的信号检测。
Commonly used vaccine adverse reaction signal detection methods exist threshold determination is arbitrary, multiple and other issues, resulting in lower accuracy of signal detection. This paper describes the impact of multiplicity on signal detection accuracy and reveals that the False Discovery Rate (FDR) control method solves the problem of multiplicity in signal detection. The FDR-based signal detection method increases the number of detection signals generated while controlling the number of errors in the generated signal. The signal detection of Adverse Events Following Immunization (AEFI) based on FDR control was implemented in R language.