基于动力学模型的COVID-19疫情预测与控制策略研究

Predicting the Trend and Evaluating Control Strategies of COVID-19 Based on a Dynamical Model

本文基于COVID-19的传播机理以及追踪隔离染病者的密切接触者、治疗等公共干预措施,建立了非自治的动力学模型,对湖北省疫情发展进行预测并评估相应控制策略的有效性.首先基于湖北省卫生健康委员会公布的数据,利用最小二乘法以及MCMC估计动力学模型中的待估计参数值.然后基于估计得到的参数值,验证模型预测的有效性,估计湖北省总病例数达峰时间,峰值规模以及流行时间.此外,研究发现湖北省每天的总病例数中隔离的潜伏者以及现存确诊者占比重较大,因此,加快对于隔离人群的筛查、对于确诊人群的治疗能够加快疫情的结束.最后对有效再生数进行敏感性分析,揭示了即使在疫情发展的后期,加强对密切接触者的追踪隔离依然是疫情控制最有效的措施.

In this paper,based on the transmission mechanism of COVID-19 and public health intervention measures such as follow-up isolation of close contacts and treatment,we develop a non-autonomous dynamical model to predict the development of epidemic in Hubei Province and evaluate the effectiveness of corresponding control strategies.Firstly,using the data published by Health Commission of Hubei Province,and utilizing the methods of least square and MCMC,we estimate the model parameter values.Then,based on the estimated parameter values,we verify model prediction and estimate peak timing,peak scale and total cases during the epidemic in the Hubei Province.In addition,we reveal that quarantined exposed and existing confirmed individuals can account for a large proportion of the total cases in Hubei Province.Hence,speeding up the screening of the quarantined population and the treatment of the confirmed population can contain epidemic more quickly.Finally,a sensitivity analysis of effective reproduction number reveals that even in the later stage of the epidemic,strengthening the follow-up isolation of close contacts is still the most effective measure.