key: cord-1040263-wnrpz0js
authors: Zu, Jian; Shen, Mingwang; Fairley, Christopher K.; Li, Miaolei; Li, Guoqiang; Rong, Libin; Xiao, Yanni; Zhuang, Guihua; Zhang, Lei; Li, Yan
title: Investigating the Relationship between Reopening the Economy and Implementing Control Measures during the COVID-19 Pandemic
date: 2021-09-11
journal: Public Health
DOI: 10.1016/j.puhe.2021.09.005
sha: 247104660770eb30ce16d3fa06c179d1c59c7b1a
doc_id: 1040263
cord_uid: wnrpz0js

Objectives The COVID-19 pandemic has resulted in an enormous burden on population health and the economy around the world. Although most cities in the United States have reopened their economies from previous lockdowns, it was not clear how the magnitude of different control measures—such as face mask use and social distancing—may affect the timing of reopening the economy for a local region. This study aims to investigate the relationship between reopening dates and control measures and identify the conditions under which a city can be reopened safely. Study Design A mathematical modeling study. Methods We developed a dynamic compartment model to capture the transmission dynamics of COVID-19 in New York City. We estimated model parameters from local COVID-19 data. We conducted three sets of policy simulations to investigate how different reopening dates and magnitudes of control measures would affect the COVID-19 epidemic. Results The model estimated that maintaining social contact at 80% of the pre-pandemic level and a 50% face mask usage would prevent a major surge of COVID-19 after reopening. If social distancing were completely relaxed after reopening, face mask usage would need to be maintained at nearly 80% to prevent a major surge. Conclusions Adherence to social distancing and increased face mask usage are key to prevent a major surge after a city reopens its economy. Findings from our study can help policymakers identify the conditions under which a city can be reopened safely.

The COVID-19 pandemic has resulted in tremendous health and economic losses in the United States (US) and globally. 1 Despite tremendous risks, most US states and cities have reopened their economies from previous lockdowns. The latest research has documented the effectiveness of public health control measures-such as face mask use and social distancing-in curbing the spread and mitigating the severity of the COVID-19 pandemic. [2] [3] [4] [5] [6] [7] [8] When city or state officials make the important decision of reopening the economy, they have to emphasize the importance of face mask use and social distancing as a way to prevent a future surge of COVID-19 infections. 4 In addition to enacting and ensuring public health control measures, the timing of reopening also plays an important role in preventing a future surge. Reopening a city or state too soon may not control the epidemic effectively even if control measures are in place. 9 On the other hand, a delay in reopening will result in a more severe economic toll, which will, in turn, affect both the mental and physical health of the residents. 10 Local government officials need more evidence to make wise reopening decisions that would slow the spread of the virus while sustaining the economy.

In this study, we use a dynamic model of the COVID-19 epidemic to investigate the relationship between reopening dates and control measures (i.e., face mask and social distancing) and identify the conditions under which a city can be reopened safely. We use New York City (NYC) as a case study because NYC was the epicenter of the pandemic in the US and has rich COVID-19 related data that can be used to develop and validate the model.

Mathematical models of COVID-19 have been widely used to inform policies and interventions to curb the pandemic. 11 Unlike most existing models, [12] [13] [14] infections and deaths. Findings from this study will provide important evidence for the government officials in NYC and other states and cities to make more informed, targeted decisions on strategies for reopening the economy and controlling the pandemic.

We obtained NYC COVID-19 data, including the number of daily and total confirmed cases and deaths, from the NYC Department of Health and Mental Hygiene (DOHMH). 15 We used data from February 29 to June 7, 2020, because NYC reopened its economy on June 8, 2020, and only data prior to the reopening date would allow us to capture the COVID-19 epidemic under city lockdown (see Table S1 in the Supplementary Document for data details). These data were also used to estimate and calibrate the unknown parameters of the mathematical model. Other disease progression parameters and behavior data related to COVID-19 were estimated based on the published literature (see Table S2 in the Supplementary Document for details about parameter estimation).

We developed a dynamic compartment model to capture the transmission dynamics of COVID-19 in NYC. The model has been validated extensively in our previous studies and used to study pandemic control strategies such as face mask use and vaccination. 16 Andrew Cuomo announced the "New York on PAUSE" Executive Order, so we assumed that starting from March 20, the public contact rate would decrease gradually, described by a decreasing logistic function. However, the contact rate in households would gradually increase, described by an increasing logistic function. Note that contact rate refers to the number of social contacts that may lead to virus transmission (details about the logistic functions for capturing contact rates are provided in the Supplementary Document). We also assumed that the proportion of frequent handwashing among people in NYC would increase from 77% to 95%. 18 Based on a literature review, we estimated that the effectiveness of handwashing in preventing COVID-19 infection is 42% (13-62%), which means frequent handwashing would reduce the risk of COVID-19 infection by 42% among susceptible individuals. 19 The US Centers for Disease Control and Prevention (CDC) started recommending face mask use by the general public on April 3, and after two weeks, on April 17, an Executive Order was implemented that required all residents over age 2 in the State of New York to wear masks or face coverings when they are in public and social distancing is impossible.

Based on a systematic review and meta-analysis of the effectiveness of face mask use for preventing COVID-19, 20 we estimated that the effectiveness of face mask use in preventing COVID-19 transmission was 85%, which means individuals wearing a face mask would have J o u r n a l P r e -p r o o f an 85% less risk of contracting COVID-19 compared to those without a face mask. Table S2 in the Supplementary Document presents the values of all model parameters and their sources.

We calibrated and validated our model based on the NYC COVID-19 data. More specifically, we estimated the unknown parameters and initial values in the model by using a nonlinear least square method. 21 Based on these estimates, we further estimated the mean values and 95% confidence intervals (CI) for unknown parameters and initial values by using an Markov Chain Monte Carlo (MCMC) approach. 21 We used the Metropolis-Hastings (M-H) sampling algorithm to calculate 10,000 times and used the last 1,000 iterations to derive the mean values and 95% CI of unknown parameters (see Table S2 in the Supplementary Document). All simulation runs and analyses were performed on the MATLAB R2020a platform. We validated the model by comparing model estimated values with the daily reported data (see Figure S3 in the Supplementary Document for details about model validation).

We conducted three sets of policy simulations to investigate how different reopening dates and magnitudes of control measures would affect the COVID-19 epidemic. First, we varied the reopening date and assessed the risk of a major surge of COVID-19 under each scenario. We examined four alternative reopening dates, including June 8, July 8, August 8 and September 8, 2020. For each reopening date, we fixed the contact rate in public settings and varied the face mask usage from 55%, 65%, to 85%. We also fixed the face mask usage and varied the public contact rate from 70%, 85%, to 100%. Results from these simulated Second, we fixed the reopening date on June 8, which is the actual reopening date, and then varied the face mask usage and public contact rate simultaneously. We assessed what combinations of face mask usage and public contact rate would lead to a major surge in the future. If there were a major surge, we assessed if it was less or more severe than the first wave of the epidemic. Figure 4 presents the results from this set of policy scenarios.

Third, we assessed the effect of different face mask usage on the total number of infections and asymptomatic infections. Since the prevalence of face mask use has been changing dramatically in the US over the course of the pandemic and there is a large geographical variation, we conducted simulation analyses under three scenarios, including (1) the worst-case scenario (0% face mask use); (2) the status quo scenario (85% face mask use); and (3) the best-case scenario (100% face mask use). Figure 5 presents the results from this set of policy scenarios.

We performed sensitivity analyses to explore how the magnitudes of control measures could affect the epidemic under different reopening dates. Specifically, for different reopening dates, we examined the changes in the number of infections and deaths by varying face mask usage and contact rate (results are presented in Figure S6 in the Supplementary Document). Lastly, we examined how the uncertainty in the effectiveness of face masks could affect the epidemic (results are shown in Figure S7 in the Supplementary Document).

Our model projected that, with the current level of control measures, the cumulative (Figure 2a) . However, if face mask usage was reduced to 65%, then a moderate epidemic would occur (Figure 2b) . In this case, the final cumulative deaths would be 5-fold greater than the current situation ( Figure S5b In contrast, after reopening, if face mask usage were reduced to 45%, a major surge of the epidemic could be prevented if contact rate was less than 70% of the pre-pandemic level (Figure 3a) . However, if contact rate increased to 85% of the pre-pandemic level, then a moderate epidemic would occur (Figure 3b) . In this case, the final cumulative deaths would be five times greater than the current situation ( Figure S5e in the Supplementary Document).

A 100% contact rate would result in a major surge about three times greater than the first wave (Figure 3c) ; the final cumulative deaths would be 9-fold greater than the current situation ( Figure S5f in the Supplementary Document). Moreover, reducing face mask usage would not only increase the magnitude of a future surge but also lead to an earlier peak time (Figures 2c and 3c) . The figures showed that, regardless of the reopening date, a high contact rate would always need to be matched with a high face mask usage to avoid a future surge. After reopening, if the contact rate were maintained at 70% of the pre-pandemic level, a relatively low face mask usage (30%) would be sufficient to prevent a future surge. In contrast, if the contact rate was at 90% of the pre-pandemic level, 65% face mask usage was necessary to prevent a future surge. Furthermore, if the contact rate increased to the pre-pandemic level, at least 75% face mask usage was necessary to prevent a future surge. It is worth noting that if "revenge travel" occurred and the contact rate increased to 165% of the pre-pandemic level, a major surge of COVID-19 would always occur regardless of face mask usage.

Our model projected that if all the control measures were relaxed, a major surge of COVID-19 would occur and deaths due to COVID-19 would increase by nearly ten times.

Our results showed that it is important to maintain social contact rate lower than 80% of the pre-pandemic level but also ensure 50% of the population wore face masks. If the contact rate returned to 100% of the pre-pandemic level, face mask usage would need to be maintained at nearly 80% to prevent a major surge. Our results also showed that delaying reopening will only delay the peak of the surge but not reduce the magnitude of the pandemic if the other control measures stay constant. In other words, only if face mask usage and social distancing were strengthened can delaying reopening reduce the magnitude of the pandemic.

Maintaining a certain level of social distancing is necessary to prevent a future surge of COVID-19 after reopening the city. Recent studies based on mathematical modelling indicated that the government-initiated social distancing alone only had a short-term impact on the pandemic. 22, 23 It will unlikely eradicate the epidemic, but only enable a better preparation of the healthcare system for the major surge. Self-imposed social distancing by the population, in contrast, is more likely to be sustainable in reducing the transmission of the virus. Consistently, mass communication and social media to improve public awareness on COVID-19 prevention and control even after reopening should be promoted. If essential economic events, large group gatherings and business activities involving physical person-toperson contacts cannot be avoided, maintaining social distancing during these activities would help reduce viral transmission.

Face mask usage in public settings after reopening should remain in place. The latest J o u r n a l P r e -p r o o f evidence showed that face mask use could result in a large reduction in the risk of infection, even in non-health-care settings. 20 Our finding demonstrated that a higher contact rate after reopening required a higher face mask use correspondingly. For incidence, if the contact rate after reopening can be maintained at 70% of the pre-pandemic level, only 30% of face mask usage is sufficient to prevent a major surge of COVID-19. However, if the contact rate after reopening was 90% of the pre-pandemic level, face mask usage needs to be maintained at 65% to prevent a major surge. The required coverage needed to be 75% if the contact rate returned to the pre-pandemic level. With this consideration, the requirement of face mask use in public settings, such as public transportation, supermarkets, and shopping centers, is likely necessary.

We found that nearly two-thirds of infections were transmitted by the pre-symptomatic and asymptomatic individuals, and face mask usage does not seem to change this proportion. This is consistent with the recent findings that a large proportion of infected individuals did not show any symptoms when they were actively tested for COVID-19. As a result, the proportion of infection caused by pre-asymptomatic and symptomatic individuals may account for most of the infections, as demonstrated by our model. Interventions for symptomatic individuals are relatively straightforward; timely isolation and contact tracing are effective means to reduce the further spread of the virus. 24 In contrast, for presymptomatic and asymptomatic individuals, prevention of viral transmission would be more difficult. We argue that the use of a face mask would be an effective way to prevent presymptomatic and asymptomatic individuals from transmitting the virus to others and, thus, persistent use of face masks after reopening should be promoted. 16 Our study has several limitations. First, we assumed a homogeneous population, but human behaviors tend to vary substantially. It is evident that social contact and self-imposed prevention behaviors have varied substantially in the population. Second, the current J o u r n a l P r e -p r o o f understanding of the infectiousness of the pre-symptomatic and asymptomatic individuals are still limited. In our study, we assumed the transmission rate among pre-symptomatic and asymptomatic individuals were about 25% of those who are symptomatic, which needs to be supported by more evidence. Third, the model did not take into consideration population mobility within the city during the pandemic. Fourth, the study did not specifically investigate the resumption of intercity or international travel but assumed that residual cases in the city would essentially have the same effect as people traveling in and out. Fifth, we did not evaluate the economic implications of the lockdown and benefits from a controlled epidemic. Sixth, the confirmed number of COVID-19 cases reflects the level of testing available in a population. Since testing is not optimal in NYC, it is likely that we underestimated the number of cases in all scenarios. Finally, we estimated model parameters from both the literature and a model calibration process when the relevant literature is not available. Model parameters estimated based on the model calibration process tended to have narrow confidence intervals, which may not reflect the reality. Despite these limitations, our model is well-calibrated using local data and, thus, can be used to inform pandemic mitigation strategies. severe/critical symptoms (I2), diagnosed infections with mild/moderate (T1) and severe/critical symptoms (T2), recovered (R) and deceased (D) cases. The force of infection is denoted as , which involves two transmission patterns: public settings (e.g., public transportation, supermarkets, offices, etc.) and households. The model includes three control measures: handwashing, social distancing and face mask use. The average incubation period is 1/k1, and the probability that an individual is asymptomatic is . Infectious individuals with mild/moderate and severe/critical symptoms are diagnosed and treated at the rates there will be no second wave of COVID-19; blue dots: there will be a minor second wave that is less severe than the first wave; red dots: there will be a major second wave that is more severe than the first wave) 

The COVID-19 pandemic in the US: a clinical update

Rational use of face masks in the COVID-19 pandemic

Face masks for the public during the covid-19 crisis

Association of public health interventions with the epidemiology of the COVID-19 outbreak in Wuhan, China

Nonpharmaceutical interventions and epidemic intensity during the 2019 novel coronavirus disease pandemic

Assessing the effects of metropolitan-wide quarantine on the spread of COVID-19 in public space and households

Can self-imposed prevention measures mitigate the COVID-19 epidemic

What Is Required to Prevent a Second Major Outbreak of SARS-CoV-2 upon Lifting Quarantine in Wuhan City

Second waves, social distancing, and the spread of COVID-19 across the USA

The mental health consequences of COVID-19 and physical distancing: the need for prevention and early intervention

How simulation modelling can help reduce the impact of COVID-19

Nowcasting and forecasting the potential domestic and international spread of the 2019-nCoV outbreak originating in Wuhan, China: a modelling study

The effect of travel restrictions on the spread of the 2019 novel coronavirus (COVID-19) outbreak

An investigation of transmission control measures during the first 50 days of the COVID-19 epidemic in China

COVID-19: DATA

Projected COVID-19 epidemic in the United States in the context of the effectiveness of a potential vaccine and implications for social distancing and face mask use

Effects of New York's executive order on face mask use on COVID-19 infections and mortality: A modeling study

The impact of cross-cultural differences in handwashing patterns on the COVID-19 outbreak magnitude. Regul Gov

Effectiveness of handwashing in preventing SARS: a review

Physical distancing, face masks, and eye protection to prevent person-to-person transmission of SARS-CoV-2 and COVID-19: a systematic review and meta-analysis. The lancet

DRAM: efficient adaptive MCMC

Impact of self-imposed prevention measures and short-term government intervention on mitigating and delaying a COVID-19 epidemic

Early characteristics of the COVID-19 outbreak predict the subsequent epidemic scope

Bidirectional contact tracing could dramatically improve COVID-19 control

This work was supported, in part, by the National Natural Science Foundation of China