Objective: To infer the start time of the resurgent COVID-19 epidemic in Xinfadi wholesale market in Beijing in June 2020 and evaluate the effect of comprehensive prevention and control measures in this epidemic. Methods: SEIR dynamics model was used to fit daily onset infections to search the start date of this resurgent COVID-19 epidemic in Beijing. The number of cumulative infections from June 12 to July 1 in Beijing were fitted considering different levels of control strength. Results: The current reemerged COVID-19 epidemic in Beijing probably started between May 22 and May 28 (cumulative probability: 95%), with the highest probability on May 25 (23%). The R(0) of the current reemerged COVID-19 epidemic was 4.22 (95%CI: 2.88-7.02). Dynamic model fitting suggested that by June 11, the cumulative number of COVID-19 cases would reached 99 (95%CI: 77-121), which was in line with the actual situation, and without control, by July 1, the cumulative number of COVID-19 cases would reach 65 090 (95%CI: 39 068-105 037). Since June 12, comprehensive prevention and control measures have been implemented in Beijing, as of July 1, compared with uncontrolled situation, the number of infections had been reduced by 99%, similar to the fitting result of a 95% reduction of the transmission rate. The sensitivity analysis showed consistent results. Conclusions: For the emergent outbreak of COVID-19, the dynamics model can be used to infer the start time of the transmission and help tracing the source of epidemic. The comprehensive prevention and control measures taken in Beijing have quickly blocked over 95% of the transmission routes and reduced 99% of the infections, containing the sudden epidemic timely and effectively, which have value in guiding the prevention and control of the epidemic in the future.
目的： 推测2020年6月北京市新发地新型冠状病毒肺炎（COVID-19）疫情首例感染的传播时间起点，辅助传染病溯源，评价当前综合防控效果。 方法： 根据北京市卫生健康委员会官方报告统计每日发病人数，建立SEIR传染病动力学模型，基于每日发病人数拟合动力学模型，并搜寻本次疫情的传播时间起点；考虑不同的防控效果而拟合6月12日至7月1日的累计发病人数，以评估当前综合防控措施效果。 结果： 北京市新发地疫情传播首例感染应起始于5月22日至5月28日之间（累计概率为95%），起始于5月25日的概率最大（23%）。本次疫情R(0)为4.22（95%CI：2.88~7.02）。模型拟合结果提示，截至6月11日，累计发病为99例（95%CI：77~121），符合实际情况。若不加控制，则截至7月1日累计发病估计将达到65 090例（95%CI：39 068~105 037）。截至7月1日，较之无防控措施的理论情况，实际感染人数减少了99%。自6月12日起，北京市采取了强有力的综合防控措施，疫情实际走势接近于传播率降低95%的推演结果，敏感性分析支持这一结果。 结论： 针对突发性疫情，传染病动力学模型可用来辅助推演传染病传播起始时间，辅助疫情溯源。北京市针对本次突发疫情所及时采取的综合防控措施迅速控制了95%以上的传播途径，减少了99%的感染人数，快速遏制了疫情，对于未来疫情防控具有重要的指导意义。.
Keywords: COVID-19; Epidemic tracing; Transmission dynamics model.