Basic Susceptible-Exposed-Infectious-Removed (SEIR) models of COVID-19 dynamics tend to be excessively pessimistic due to high basic reproduction values, which result in overestimations of cases of infection and death. We propose an extended SEIR model and daily data of COVID-19 cases in the U.S. and the seven largest European countries to forecast possible pandemic dynamics by investigating the effects of infection vulnerability stratification and measures on preventing the spread of infection. We assume that (i) the number of cases would be underestimated at the beginning of a new virus pandemic due to the lack of effective diagnostic methods and (ii) people more susceptible to infection are more likely to become infected; whereas during the later stages, the chances of infection among others will be reduced, thereby potentially leading to pandemic cessation. Based on infection vulnerability stratification, we demonstrate effects brought by the fraction of infected persons in the population at the start of pandemic deceleration on the cumulative fraction of the infected population. We interestingly show that moderate and long-lasting preventive measures are more effective than more rigid measures, which tend to be eventually loosened or abandoned due to economic losses, delay the peak of infection and fail to reduce the total number of cases. Our calculations relate the pandemic’s second wave to high seasonal fluctuations and a low vulnerability stratification coefficient. Our characterisation of basic reproduction dynamics indicates that second wave of the pandemic is likely to first occur in Germany, Spain, France, and Italy, and a second wave is also possible in the U.K. and the U.S. Our findings show that even if the total elimination of the virus is impossible, the total number of infected people can be reduced during the deceleration stage.

由于基本繁殖值较高，COVID-19 动态的基本易感暴露感染去除 (SEIR) 模型往往过于悲观，导致感染和死亡病例的高估。我们提出了扩展的 SEIR 模型和美国和欧洲七个最大国家的 COVID-19 病例的每日数据，通过调查感染脆弱性分层的影响和防止感染传播的措施来预测可能的大流行动态。我们假设（i）由于缺乏有效的诊断方法，在新的病毒大流行开始时，病例数会被低估；（ii）更容易感染的人更有可能被感染；而在后期，感染等的机会将会减少，从而有可能导致大流行停止。基于感染脆弱性分层，我们证明了大流行减速开始时人口中感染者比例对感染人口累计比例带来的影响。有趣的是，我们发现，适度而持久的预防措施比更严格的预防措施更有效，后者往往会因经济损失而最终放松或放弃，延迟感染高峰并且无法减少病例总数。我们的计算将大流行的第二波与高季节性波动和低脆弱性分层系数联系起来。我们对基本繁殖动态的描述表明，第二波大流行很可能首先发生在德国、西班牙、法国和意大利，第二波也可能发生在英国和美国 我们的研究结果表明，即使不可能完全消除病毒，在减速阶段感染人数总数也可以减少。