key: cord-0757318-r2mwzzaw
authors: Zhang, Zhechun
title: Prevent the resurgence of infectious disease with asymptomatic carriers
date: 2020-04-19
journal: nan
DOI: 10.1101/2020.04.16.20067652
sha: 86e2eac17a09f9d06fd2592d0d9b6848496b2b19
doc_id: 757318
cord_uid: r2mwzzaw

As many countries reached the peak of the COVID-19 outbreak, there is debate on how to reopen the economy without causing a significant resurgence. Here we show, using a microsimulation model, that how to reopen safely depends on what percentage of COVID-19 cases can be detected by testing. The higher the detection rate, the less restrictive the reopen plan needs to be. If 70% of cases can be detected, schools and businesses can reopen if 2-layer quarantine is imposed on each confirmed case. Our results suggest that increasing the detection rate is essential to prevent the resurgence of COVID-19.

To address this question, we developed a microsimulation model 2,3 assuming only a portion of cases is detected by testing. Disease carrier is infectious between days 5 and 14 post-infection 4, 5 ( Fig. S1 ). After day 7, some carriers are detected by testing (Fig. S1 ), which results in the quarantine of these carriers and their close contacts. We applied this model to a hypothetical region similar to Massachusetts. There are 6.8 million residents (0.4 million 4-and-under, 1.6 . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.16.20067652 doi: medRxiv preprint million K-12 and college students, 3.9 million employees, and 0.9 million retirees), 2.3 million households, and 210,000 employers (Fig. S2) . The model has a timeline similar to the COVID-19 outbreak in Massachusetts (Fig. S3) . Patient 0 was introduced on February 26 at a company conference; K-12 schools and colleges were closed on March 14; non-essential works were shut down on March 22. To test this model, we first compared the number of simulated cases to reported cases. Because the latter depends on testing availability and turnaround time, it sets a lower bound of actual cases approximately one week before the report date (Fig. S4 ). Reported death, in comparison, provides a more realistic estimate of actual cases 1-2 weeks before the report date (Fig. S5 ). Here we assumed the death rate is ~1% as reported in Korea, where extensive testing started early in the outbreak. Third, our model predicts a peak date of reported cases on April 24 (Fig. S6 ). In comparison, the University of Washington model (covid19.healthdata.org) predicts a peak date of resource use on April 28 (Fig. S7) . Fourth, our model predicts a total death of 9,106 by August 5th, 11% more than the University of Washington model (Fig. S7) . Fifth, our model predicts a serial interval of 5.04 +/-0.04 days, consistent with the reported interval of 4.4-7.5 days 4 . Last, our model predicts a reproduction number of 5.96 +/-0.18. This number is higher than SARS 6, 7 and the COVID-19 estimate 4 based on the early outbreak in Wuhan, where testing was limited.

Therefore, we modeled a disease that is more infectious than SARS. According to the model, the safe reopen plan depends on the detection rate. If 70% of cases can be detected, all restrictions can be lifted (Fig. 1a) as long as 2-layer quarantine is in place (Fig.   1b ). 2-layer means not only the confirmed carrier and its close contacts, but their workplace, . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.16.20067652 doi: medRxiv preprint 3 schools, homes, and communities, need to be quarantined (Fig. 1c) . This additional layer leads to 950 +/-69 quarantines per confirmed case. Nevertheless, at any given time, only a small percentage of the total population is under quarantine (Fig. S8 ). . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.16.20067652 doi: medRxiv preprint 2-layer quarantine is necessary to prevent a resurgence (Fig. 1b) . In general, the number of layers required is determined by two timescales: ti, the time when the patient becomes infectious, and tq, the time when quarantine starts. It may take 4-5 days (ti) for a patient to become infectious 4 and another 4-5 days for the test to come back positive and for public health officials to reach close contacts. At the time quarantine starts (tq), the disease could have spread from the first layer of contacts to the next layer (Fig. 1c) .

If 60% of cases can be detected, schools and businesses can reopen, but community contacts need to be restricted (Fig. 1a) . Community activities outside schools, work, and home account for ~30% of total daily contacts 8 (Fig. S3 ).

If 30% of cases can be detected, in addition to restricting community activities, schools and large businesses (>2000 employees in our model) need to remain in remote mode (Fig. 1a) . Due to their large size, schools and large businesses are hot spots for disease transmission. Compared to the rest of the population, students are 66% +/-12% more likely to transmit the disease.

Employees of large businesses, 94% +/-5%. According to the model, had schools, large businesses, and community activities been shut down on March 3, the outbreak in Massachusetts would be limited to 3220 +/-650 people and peaked by March 31 (Fig. S9) .

The resurgence of infectious disease has significant impacts on public health and the economy.

Our model shows for diseases with asymptomatic carriers, the reopen plan needs to be developed based on the estimated detection rate. At the initial stage of reopening, it is safer to assume only a small percentage of cases (~30% or less) are detected. In order to increase the detection rate, it . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.16.20067652 doi: medRxiv preprint 5 may be necessary to test all symptomatic patients and potential contacts of confirmed cases, and routinely screen high-risk sub-populations. 100,000 COVID-19 tests a day cost $5million 9 , which is only 0.3% of daily GDP in Massachusetts 10 . Therefore, dramatically expanding COVID-19 tests is a small investment with a high return.

. CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.16.20067652 doi: medRxiv preprint Figure S8 . The percentage of the Massachusetts population under quarantine after reopening.

Error bars denote S.E.M (n=20).

. CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.16.20067652 doi: medRxiv preprint Error bars denote S.E.M (n=20).

. CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.16.20067652 doi: medRxiv preprint . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.16.20067652 doi: medRxiv preprint

Percentage of asymptomatic cases

Modeling targeted layered containment of an influenza pandemic in the United States

Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand

How will country-based mitigation measures influence the course of the COVID-19 epidemic?

SARS-CoV-2 Viral Load in Upper Respiratory Specimens of Infected Patients

Epidemiological determinants of spread of causal agent of severe acute respiratory syndrome in Hong Kong

Transmission dynamics and control of severe acute respiratory syndrome

Social contacts and mixing patterns relevant to the spread of infectious diseases

Medicare Administrative Contractor (MAC) COVID-19 Test Pricing. 2020. 10. fred.stlouisfed.org. Total Gross Domestic Product for Massachusetts