Modelling the pulse population-wide nucleic acid screening in mitigating and stopping COVID-19 outbreaks in China

Abstract

Background: During 2021-2022, mainland China experienced multiple times of local COVID-19 outbreaks in several cities, including Yangzhou, Xi'an etc., and the Chinese government persistently adopted the zero-COVID policy in combating with the local outbreaks.

Methods: We develop a mathematical model with pulse population-wide nucleic acid screening, part of the zero-COVID policy, to reveal its role in controlling the spread of COVID-19. We calibrate the model by fitting the COVID-19 epidemic data of the local outbreaks in Yangzhou and Xi'an, China. Sensitivity analysis is conducted to investigate the impact of population-wide nucleic acid screening on controlling the outbreak of COVID-19.

Results: Without the screening, the cumulative number of confirmed cases increases by [Formula: see text] and [Formula: see text] in Yangzhou and Xi'an, respectively. Meanwhile, the screening program helps to shorten the lockdown period for more than one month when we aim at controlling the cases into zero. Considering its role in mitigating the epidemics, we observe a paradox phenomenon of the screening rate in avoiding the runs on medical resource. That is, the screening will aggravate the runs on medical resource when the screening rate is small, while it helps to relieve the runs on medical resource if the screening rate is high enough. We also conclude that the screening has limited effects on mitigating the epidemics if the outbreak is in a high epidemic level or there has already been runs on medical resources. Alternatively, a smaller screening population per time with a higher screening frequency may be a better program to avoid the runs on medical resources.

Conclusions: The population-wide nucleic acid screening strategy plays an important role in quickly controlling and stopping the local outbreaks under the zero-COVID policy. However, it has limited impacts and even increase the potential risk of the runs on medical resource for containing the large scale outbreaks.

Keywords: COVID-19; Mathematical model; Nucleic acid screening; Runs on medical resource; Screening paradox.

Fig. 1

Schematic diagram of the model for illustrating the COVID-19 infection dynamics. The infected individuals in class I can be diagnosed and isolated through opportunistic diagnosis and the population wide nucleic acid screening

Fig. 2

A Data of COVID-19 outbreak in Yangzhou from July 28, 2021 to August 26, 2021; B Data of COVID-19 outbreak in Xi’an from December 9, 2021 to January 9, 2022. The data include daily opportunistic confirmed cases, daily confirmed quarantined exposed cases and daily confirmed cases through nucleic acid screening

Fig. 3

Best model fitting results for the transmission dynamic model in Yangzhou and Xi’an. A, C and E are the daily opportunistic confirmed cases from I, the daily confirmed cases from $E_q$ and the cumulative confirmed cases in Yangzhou, respectively. Similarly, B, D and F are those in Xi’an. The yellow and green curves are the estimated curves in Yangzhou and Xi’an, respectively, with the shadow areas being the corresponding 95% confidence intervals. The circles are the corresponding observed data

Fig. 4

Comparing the impact of previous implemented screenings and implementing no screening on the COVID-19 epidemic in Yangzhou and Xi’an. The blue and green curves are the estimated curves corresponding to previous screening and no screening strategies, respectively, with the shadow areas being the corresponding 95% confidence intervals. The other parameter values are fixed as those listed in Table 1

Fig. 5

The impacts of various nucleic acid screening schemes on the COVID-19 epidemic in Yangzhou and Xi’an. The specific schemes are: Remain the first half screenings in Yangzhou (remain the screenings before 22/1/1 in Xi’an); Remain the screenings every other two days; Remain the last half screenings in Yangzhou (remain the screenings after 22/1/1 in Xi’an). The other parameter values are fixed as those listed in Table 1

Fig. 6

Solutions of model (2). Here, we assumed a screening frequency of each two days (i.e. $T_s=2$). $d_e=1$ and the other parameter values are fixed as the same as those estimated from the outbreak in Xi’an, as listed in Table 1

Fig. 7

A-D Solutions of model (2); E Relation curves of the peak values of confirmed and isolated population (H(t)) with respect to the screening rate. Here, we fix the screening period as $T_s=7$, and vary the screening rate and contact rate. The other parameters are fixed as the same as those estimated from the outbreak in Xi’an

Fig. 8

Solutions of model (2) by choosing different combinations of the screening rate and the screening period. Here $d_e=1$ and the other parameters are fixed as the same as those estimated from the COVID-19 outbreak in Xi’an