目的:估计北京市流感流行阈值和分级强度阈值，对2019—2020年流感流行季的流行水平进行分级预警，并对应用的方法进行评价。方法:应用北京市2014—2019年5个流感流行季的流感样病例数、流感病毒核酸检测阳性率和流感聚集性疫情数资料，经标准化处理整合为综合指数，结合移动流行区间法（moving epidemic method，MEM）估计流行阈值和分级强度阈值，并应用交叉验证法评价预警效果，评价指标为灵敏度、特异度、马修相关系数和约登指数。结果:将流感流行水平分为5个等级，预警2019—2020年流感流行季流行水平的流行阈值、中位阈值、高强度阈值和超高强度阈值分别为0.73、2.12、3.56和4.49。评价指标灵敏度、特异度、马修相关系数和约登指数分别为98%、68%、65%和66%。结论:本研究为2019—2020年流感流行季建立了流行水平分级预警标准，为流感预警防控工作提供了依据。综合指数结合MEM预警的灵敏度较高，准确性在可接受范围内，该方法可在北京市流感分级预警中应用。“,”Objective:To calculate influenza epidemic thresholds and intensity thresholds, and to provide tiered warning alert for 2019-2020 influenza season in Beijing, as well as to evaluate performance of the applied method.Methods:The data of the last 5 influenza seasons from 2014 to 2019, including weekly counts of influenza-like illness, influenza virus positive rates and influenza outbreaks were integrated into the form of synthetical index after standardization, and then combined with the moving epidemic method (MEM) to calculate the epidemic thresholds and intensity thresholds. Cross-validation procedure was used to evaluate the alert performance using Matthew correlation coefficient, Youden’s index, sensitivity and specificity as the evaluation indicators.Results:Influenza epidemic intensity was graded as five levels. The epidemic threshold, medium, high and very high intensity thresholds for 2019-2020 influenza season were 0.73、2.12、3.56 and 4.49, respectively. As for evaluation indicators, sensitivity, specificity, Matthew correlation coefficient, and Youden’s index of the epidemic threshold were 98%、68%、65% and 66%, respectively.Conclusions:This study established early warning epidemic thresholds for 2019-2020 influenza season in Beijing, which provided evidence for the prevention and control of influenza epidemic. MEM combined with synthetical index produced a sensitive signal for detecting influenza epidemics and the accuracy was acceptable. The method could be applied in practical influenza epidemic alert in Beijing.