Evaluation of control measures for COVID-19 in Wuhan, China.

Dear Editor,

Recent article in this Journal has reported trend of the coronavirus disease (COVID-19) in China. The outbreak of COVID-19 was first reported from Wuhan, China, had spread to more than 100 countries. Recent studies had forecasted COVID-19 incidence in China and worldwide based on mathematical models, , but they did not take into account the control measures adopted by the Chinese government; the forecast results were appreciably higher than actual numbers. For instance, the article forecasted that the number of infections in Wuhan would reach 75,815 on January 25, 2020, but the actual number was simply 618. Therefore, it is indispensable to establish a new dynamic model of the epidemic to evaluate the effectiveness of control measures in Wuhan (Fig. 1 ).

Fig. 1

Fit of prediction value and actual value.

In our study, we used data released by the National Health Commission. Since the data were not updated from January 1 to January 17, we used data from January 18 to February 13 to establish our model. Epidemic prevention and control in Wuhan can be divided into three stages: the first stage was the natural occurrence and spread of the epidemic from January 18 to January 23; the second stage was the blockade of Wuhan and the advocacy of residents going out less from January 23 to February 5; and the third stage was cabin hospitals put into use from February 6 to February 13 to ensure that all cases are admitted to hospitals and close contacts are under intensive medical observation. We used the susceptible-exposed-infected-removed (SEIR) dynamic model to simulate the spread of the epidemic and Simulated Annealing (SA) algorithm to identify optimal parameters. The formula of SEIR model is as follows:

The susceptible, exposed, infected, and r removed populations at time t were set as S (t), E (t), I (t), and R (t), respectively, and the total population in Wuhan was set as N. The number of the susceptible infected by each infected person per unit time (d) was set as α1, and by each exposed person, α2. Moreover, according to the literature, the incubation period was 1/β = 4 d, and the infection period was 1/γ = 3.5 d. After adopting control measures, the number of susceptible infected by each infected person per unit time (d) was set as C*α1, and by each exposed person, C*α2. The optimal parameters of the model found out through SA algorithm.

The goodness of fit between the forecast value and the actual incidence was 81.98%, suggesting that our model can reflect the actual incidence and spread of the epidemic. Based on this model, we calculated the basic reproduction number in Wuhan, R0=3.31, the effective reproduction number of the second stage, RT=1.12, and the effective reproduction number of the second stage, RT=0.71 (RT< 1, the inflection point of the epidemic appeared and the incidence began to decline).

In this article, the effectiveness of control measures at different stages of COVID-19 in Wuhan was evaluated for the first time. The results showed that the effective reproduction number after blockade of the city was RT=1.12, showing that blockade reduced transmission of infection by 66.16%. Although the adopted prevention and control measureswere effective, the effective reproduction number remained greater than 1, indicating that the epidemic continues to spread. This could be attributable to medical limitations: not all patients could be admitted to hospitals, and close contacts were mainly isolated at home, resulting in family outbreaks. Since February 6, cabin hospitalshave been put into use to ensure that all cases are admitted to hospitals and close contacts are under intensive medical observation. Consequently, the effective reproduction number has decreased to less than 1 (RT = 0•71), indicating the inflection point of the epidemic; thus, the incidence of the disease will gradually decrease until its disappearance.

Declaration of Competing Interest

The authors declare that they have no competing interests.