We developed an optimal model ARIMA (2,1,2) (1, 1, 0)12 that fit historic varicella outbreak case data and used it to quantify the impact of COVID-19 non-pharmaceutical interventions on varicella outbreak cases. We showed that in 2019, prior to the pandemic, predicted outbreak cases were consistent with the actual outbreak cases, indicating that the model was could predict varicella outbreak cases. In the first half of 2020, however, the actual number of varicella outbreak cases was 93.37% lower than the predicted number of outbreak cases. No outbreak cases were reported in March 2020. The reduction of outbreak cases in the first half of 2020 was associated with implementation of a series of effective NPIs taken by China’s government to mitigate the spread of SARS-CoV-2 [22]. Many kindergartens and schools, which are common settings for varicella outbreaks, were closed, significantly reducing the opportunity for transmission among children. Keeping social distance in general populations reduced opportunity for airborne transmission – normally an important means of spreading varicella virus [9].
Varicella is often considered a benign disease with no sequelae, provided that the immune system is functional and efficient [23]. From the perspective of disease burden, however, varicella has one of the highest-incidence vaccine-preventable disease, causing huge economic losses and many temporary school closures [8]. Our study showed that reported cases of varicella outbreaks increased continuously since 2010 and began to rise rapidly and peaked from 2017 to 2019, indicating that varicella was not controlled before the COVID-19 epidemic. Lack of control of varicella is due to the low vaccination rate of varicella vaccine in China [24] and is associated with an increase of breakthrough cases[25].
Due to COVID-19-related NPIs, varicella outbreak cases declined dramatically in 2020. Although this is favorable for reducing the spread of varicella, we should be aware that due to the nearly ubiquitous nature of varicella, children who would normally have contracted varicella in 2020 instead experienced extended periods of low exposure to VZV and therefore remain susceptible. The buildup of susceptibles leads to a so-called immunity debt, which puts children at a greater risk of outbreaks and possible epidemics.[26] Without timely vaccination to “repay” the immunity debt and reduce the number of susceptibles, the average age of contracting varicella could increase, causing more severe and complex varicella cases. In a study by Ma Chao and colleagues that used a modified Delphi technique to obtain opinions of experts [27], varicella vaccine was considered the top priority vaccine to introduce into China’s Expanded Program on Immunization (EPI). Since protection provided by VarV is similar to that provided by natural infection, widely implementing varicella vaccine should be considered in order to avoid a rebound of varicella outbreaks[10].
Other studies have shown that transmission of respiratory infectious diseases has been interrupted by the NPIs that were used to reduce SARS-CoV-2 spread [28,29,30]. Several countries have experienced large outbreaks of respiratory infectious diseases following the relaxing or lifting of NPIs, for example England, Italy, the United States, and Australia [12, 25,31-32]. These outbreaks led to an increase in hospitalizations. In China, the stringent dynamic COVID-zero policy has continued for sustained containment of COVID-19. Several domestic studies have shown that the incidence of notifiable infectious diseases has been significantly affected by the SARS-CoV-2 NPIs [33,34]. However, these studies are of notifiable infectious diseases. As a non-notifiable infectious disease, varicella has not been included in these studies. Because varicella is one of the most common infectious diseases childhood, clinically significant burdens of disease and economic losses are unavoidable during outbreaks. For this reason, we explored changes in varicella epidemiology caused by COVID-19 NPIs. To our knowledge, our study is the first to use an ARIMA model to quantify changes in varicella outbreak cases associated with COVID-19 NPIs. The methods and results may provide reference for future studies.
A limitation of our study is possible underestimation of cases because the source of the data was the Emergency Reporting and Information Management System, which does not include sporadic cases of varicella. Therefore, the overall number of varicella cases is an underestimate that may affect the accuracy of our analysis but would not change the direction and qualitative findings of our study. Varicella outbreak cases are cluster events that are likely to attract public attention, so underreporting is less of a problem with outbreak cases compared with sporadic cases. For our study, this should mitigate somewhat the impact of underreporting. A second limitation is that decreased outbreak-related cases could be partly driven by seasonality of varicella and school holidays [35]. However, looking at the dynamics of the varicella in China in the past 15 years, we found that the decrease in 2020 was too large to be explained by the regular seasonality patterns of varicella epidemic36. As with other respiratory infectious diseases, the decline is almost certainly due to the COVID-19 NPIs. Continued evaluation of varicella surveillance data will help better understand the causal relations.