PreprintPDF Available

Comparative cost-effectiveness of SARS-CoV-2 testing strategies

Authors:
Preprints and early-stage research may not have been peer reviewed yet.
ResearchGate Logo

This preprint is featured on the COVID-19 research community page

View COVID-19 community

Abstract and Figures

Alessandro Vespignani
Content may be subject to copyright.
1
Comparative cost-effectiveness of SARS-CoV-2 testing strategies
Zhanwei Du1+, Abhishek Pandey2+, Yuan Bai3+, Meagan C. Fitzpatrick2,4+, Matteo Chinazzi5, Ana
Pastore y Piontti5, Michael Lachmann6, Alessandro Vespignani5, Benjamin J. Cowling3, Alison
P. Galvani2, and Lauren Ancel Meyers1,6*
1. The University of Texas at Austin, Austin, Texas 78712, USA
2. Center for Infectious Disease Modeling and Analysis, Yale School of Public Health, New
Haven, CT 06510, USA
3. The University of Hong Kong, Hong Kong SAR, China
4. Center for Vaccine Development and Global Health, University of Maryland School of
Medicine, Baltimore, MD 21201, USA
5. Laboratory for the Modeling of Biological and Socio-technical Systems, Northeastern
University, Boston, MA, USA
6. Santa Fe Institute, Santa Fe, NM, 87501, USA
Corresponding author: Lauren Ancel Meyers
Corresponding author email: laurenmeyers@austin.utexas.edu
+ These first authors contributed equally to this article
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
2
Abstract
Background: To mitigate the coronavirus pandemic that emerged in 2019 (COVID-19),
countries worldwide have enacted unprecedented movement restrictions, social distancing
measures, and face mask requirements. Until safe and efficacious vaccines or antiviral drugs
become widely available, viral testing remains the primary mitigation measure for rapid
identification and isolation of infected cases.
Methods: We evaluate the economic tradeoffs of expanding and accelerating SARS-CoV-2
testing using a multi-scale model that incorporates SARS-CoV-2 transmission at the population
level and daily viral load dynamics at the individual level.
Findings: Assuming a willingness-to-pay of $100,000 per year of life lost (YLL) and a price of
$5 per test, the strategy most likely to be cost-effective under a rapid transmission scenario (Re
> 2) is daily testing followed by a one-week rather than two-week isolation period subsequent to
test confirmation. Under lower transmission scenarios, weekly testing of the population is
expected to be more cost effective. Expanded surveillance testing is expected to be cost
effective if the price per test is less than $400 across all transmission rates considered.
Interpretation: Extensive expansion of testing coupled with isolation of confirmed cases is
essential for mitigating the COVID-19 pandemic. Further, resources recouped from shortened
isolation duration could be cost-effectively allocated to more frequent testing.
Funding: US National Institutes of Health and US Centers for Disease Control and Prevention.
Keywords: COVID-19; SARS-CoV-2; Antigen test; Testing frequency; Epidemiological model;
Cost effectiveness
Research in context
Evidence before this study
We searched PubMed on Oct 6, 2020 for articles published after Aug 1, 2020 focusing on the
cost-effectiveness of expanding COVID-19 testing in United States using the search terms
("Economic")[Title/Abstract] AND ("Testing"[Title/Abstract] OR "Test")[Title/Abstract] AND
("SARS-CoV-2"[Title/Abstract] OR "COVID-19")[Title/Abstract] AND ("United
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
3
States"[Title/Abstract] OR "US"[Title/Abstract] OR "America"[Title/Abstract] OR
"U.S.")[Title/Abstract] with no language restrictions. We found only one article that addresses
the cost effectiveness of SARS-CoV-2 testing strategies in the US, specifically focused on using
PCR testing to bolster endoscopy practices. Four articles estimate the impact of social
distancing measures on the economic burden of COVID-19 in the US (e.g., city lockdown,
contact-tracing and household quarantine). However, no article provides estimates for the cost-
effectiveness of vastly expanding testing using the affordable 15-minute COVID-19 antigen test
approved by the U.S. Food and Drug Administration (FDA) on Aug 26, 2020.
Added value of this study
Using a data-driven model of SARS-CoV-2 transmission that incorporates daily viral load
dynamics of infected cases, we evaluate the economic tradeoffs of expanding and accelerating
SARS-CoV-2 testing. Our study is the first to identify strategies that are expected to be cost-
effective, depending on the local transmission rate of the virus, the cost of the test and the
societal willingness-to-pay for averting COVID-19 deaths. Given the epidemiological and
economic conditions in the US as of August 2020, the optimal strategy is staggered weekly
testing of the entire population followed by a one-week isolation period.
Implications of all the available evidence
Despite the intimidating up-front costs, mass testing coupled with strict but relatively short
isolation of confirmed cases is recommended to health authorities and local governments as a
cost effective strategy for mitigating the unprecedented threat of the COVID-19 pandemic,
before safe and efficacious vaccines and drugs become available.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
4
Introduction
A novel coronavirus (SARS-CoV-2) continues to threaten the health, economy and stability of
the world. The virus emerged from Wuhan China at the end of 2019 and was declared a
pandemic by the World Health Organization (WHO) on March 11, 2020 (1). As of October 2020,
the COVID-19 burden in the United States (US) has surpassed seven million confirmed cases
and 200,000 deaths. Globally, there have been over 33 million reported cases and one million
deaths (2). Moreover, the estimated economic costs have exceeded US$3.5 trillion (3).
Dozens of candidate vaccines are in the development pipeline (4). Although none have
completed Phase III trials, some global leaders have projected that large vaccine campaigns
would be attainable by mid-2021 (4). There are also numerous SARS-CoV-2 antiviral drugs
under evaluation aimed at reducing the severity of COVID-19 and providing prophylaxis from
infection (5). Until such medical countermeasures become widely available, the world is
primarily combatting SARS-CoV-2 through unprecedented non-pharmaceutical interventions
including wearing of face masks, travel restrictions and strict social distancing measures that
can exact dire socioeconomic costs (6). The contribution of asymptomatic and pre-symptomatic
cases towards transmission makes control via non-pharmaceutical interventions challenging (7).
Recent studies suggest that at least 18% to 31% of infected people never develop symptoms (8,
9) and 50% of transmissions occur while the infected person is pre-symptomatic (10).
Therefore, symptom-based interventions alone may not be sufficient to curtail an outbreak (7).
Although mass diagnostic testing, contact tracing and isolation can substantially mitigate spread
(11), only a few countries have been able to scale such programs to levels required to contain
pandemic waves (6). In October 2020, the Chinese port city of Qingdao took the unprecedented
step of testing all 9.5 million of its residents (12).
The first and most widely used SARS-CoV-2 tests apply reverse transcription polymerase chain
reaction (PCR) to identify viral particles swabbed from the noses or throats of patients (12). In
the United States, limited availability and slow laboratory turnaround times have impeded the
use of testing to slow viral spread (13). As the pandemic continues, however, cheaper and
faster testing technologies are becoming increasingly common (13, 14). In August, the U.S.
Food and Drug Administration (FDA) approved both a rapid point-of-care saliva-based PCR test
providing results in three hours (15) and a 15-minute COVID-19 antigen test (16). The new PCR
tests, which require only a small sample of saliva, have a sensitivity of 94% and specificity of
100% at a price point of $1.21-$4.39 per test (17). Antigen tests, which detect viral surface
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
5
proteins, provide more rapid and accurate indication of active infection than PCR tests. They
cost roughly $5 per test and provide a sensitivity of 97.1% and specificity of 98.5% (13, 16). In
addition, serological tests which can identify SARS-CoV-2-specific antibodies starting a week
after infection and possibly for months or years after recovery, with an estimated sensitivity of
90% and specificity of 99% (13, 18). At roughly $50 per test (19), serological surveys are
providing retrospective insight into the spread of the pandemic throughout the US (20).
As affordable rapid SARS-CoV-2 tests become more widely available (16), testing and isolation
will become an increasingly economically viable strategy for slowing spread and averting large
pandemic waves. Using a network-based mathematical model of SARS-COV-2 transmission
that incorporates household-specific and age-stratified heterogeneities as well as temporal
changes in viral load that impact diagnostic sensitivity, we evaluated the cost-effectiveness of
cheap and fast antigen testing. We conducted cost-effectiveness analysis of a wide range of
testing strategies, taking into account the direct costs of testing, work lost during isolation, and
the economic burdens of COVID-19 illness. We identify the optimal testing and isolation
strategies under a range of transmission scenarios reflecting the heterogeneous implementation
of non-pharmaceutical interventions across the US and the globe.
Methods
Epidemic model
We simulate epidemic outbreaks for 150 days using a stochastic chain-binomial model in
contact networks with nodes representing individuals and edges representing epidemiologically-
relevant contacts between individuals (21). The degree of a node is the number of other nodes
connected to it via its edges.
We implement the stochastic individual-based model based on the parameters given in Table 1.
Each individual can be in one of 12 states that reflect both the progression of infection and
testing (Figure 1). Upon SARS-CoV-2 infection, an individual remains exposed for days,
1/𝜎
after which they enter either the asymptomatic or pre-symptomatic infectious state with
probabilities of and . Asymptomatic cases transition to recovered state after an
1
𝑝
sym
𝑝
sym
average asymptomatic infectious period of . Pre-symptomatic cases become symptomatic at
1/
𝛾
a rate and then recover at a rate . Recovered individuals are assumed to be permanently
𝛾
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
6
immunized against future infection. Relative to symptomatic individuals, the infectiousness of
asymptomatic and pre-symptomatic nodes is reduced by factors and . For each infected
𝜔
𝜔
individual, the model tracks the number of days since infection.
To calibrate the model to a specified effective reproduction number , we use an interior-point
𝑅
𝑒
algorithm to minimize the mean square error between the mean across 100 simulations and
𝑅
𝑒
the desired value. In each simulation, we simulate the first 100 infections assuming a status quo
strategy testing in which of symptomatic cases are tested and isolated for two weeks
𝜄
=
19%
beginning, on average, days after symptom onset. We estimate as the ratio between
𝜆
=
2
𝑅
𝑒
the number of infectees and infectors in our simulations.
Testing strategies
We model rapid antigen testing of the entire population at different frequencies, where
𝑑
test
indicates the interval between tests in days. For example, days means that each
𝑑
test
=
7
individual is tested once every seven days, with tests distributed equally across each seven-day
period. Test outcomes depend on the infection status of the tested individual as well as
sensitivity and specificity of the test. Susceptible and recovered individuals may test positive
based on the false positive rate of the test (16). Infected individuals test positive based on days
since infection and sensitivity of the test on that day (Table 2). Individuals that test positive
move into their corresponding isolation states where they are unable to infect others (i.e.,
susceptible-isolated, exposed-isolated, pre-symptomatic-isolated, asymptomatic-isolated,
symptomatic-isolated, or recovered-isolated). After a specified isolation period ( ), individuals
𝑑
iso
progress to a non-isolated state corresponding to their current infectious/non-infectious state. To
determine the public health benefits of each strategy, we also model a status quo strategy that
assumes a baseline level symptomatic testing without additional surveillance testing.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
7
Figure 1. A schematic of the individual-based SARS-CoV-2 infection dynamic model. Upon
infection, susceptible individuals (S) progress to exposed (E) where they are neither infectious nor
symptomatic. A fraction of cases become asymptomatic infectious (A) with lower infectiousness before
recovering (R); the remaining cases progress to pre-symptomatic (P) where they are moderately
infectious but not yet symptomatic, followed by symptomatic infectious (Y) and then recovered (R) state.
Recovered individuals remain permanently protected from future infection. The various testing strategies
assume that individuals are tested at a specified frequency, ranging from daily to monthly, according to an
evenly staggered testing schedule, regardless of their disease state. Those testing positive proceed to
isolate for the specified period (either 7 or 14 days). Susceptible and recovered individuals test positive
according to a false positive rate; infected individuals test positive based on the test sensitivity that
depends on the number of days since they were infected. Following isolation, individuals return to non-
isolated states corresponding to the current state of their health. The status quo strategy assumes that
19% of symptomatic cases are tested and isolated for two weeks.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
8
Table 1. Epidemiological and testing parameters for the individual-based SARS-CoV-2 infection
dynamic model. We model multiple strategies for expanding surveillance testing of asymptomatic
individuals via rapid antigen tests combined with a fixed baseline level of symptom-based testing and
isolation (i.e., status quo testing).
Values
Parameters
Faster transmission
scenario
Slower transmission
scenario
: effective reproduction number
𝑅
𝑒
2.2 (22)
1.3 (23)
: transmission rate
𝛽
Calibrated to Re
Calibrated to Re
Initial number of infections (in
exposed state)
10
: days of isolation following a
𝑑
iso
positive surveillance test
7 or 14
: testing frequency (days
𝑑
test
between tests)
1, 7, 10, 14, or 28
Probability of test detecting an
infection based on time since
infection
Table 2
Probability of false positive test
1.5% (16)
: transition rate out of exposed
𝜎
state
1/3 (24)
: recovery rate of symptomatic
𝛾
individuals
1/4 (11, 25)
: recovery rate of asymptomatic
𝛾
individuals
1/9
: transition rate from the pre-
𝜖
symptomatic to symptomatic stage
1/2 (24)
: relative infectiousness of pre-
𝜔
symptomatic cases
1.57 (11, 25)
: relative infectiousness of
𝜔
asymptomatic cases
0.5 (26)
: proportion of infections that
𝑝
sym
are symptomatic
75% (9, 11)
: COVID-19 mortality rate for age
𝛿
𝑎
group
𝑎
[0.0049%, 0.084%, 1%, 3.371%]
for [0-17y, 18-49y, 50-64y, >65y] (27)
: life expectancy (years) for age
𝜆
𝑎
group
𝑎
[70.2, 48.9, 27.5, 15.8]
for [0-17y, 18-49y, 50-64y, >65y] (28)
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
9
Status quo parameters: Baseline detection and isolation of symptomatic cases who are not
detected via surveillance testing
: proportion of symptomatic infections that are isolated through
𝜄
symptom-based testing
19% (29)
: daily isolation rate following symptom onset for such cases
𝜆
: days of isolation for such cases
𝜎
𝑖𝑠𝑜
14 (11)
Table 2. Assumed probability of detecting an infection using an antigen
test (16) by days post infection*.
Days since infection
Sensitivity
Source
1
26.3 %
2
30.0 %
3
36.3 %
4
68.8 %
5
97.1 %
extrapolated from
(30, 31), assuming
proportional to viral
load (32, 33)
6
100%
7
100%
8
100%
9
100%
10
95.7%
11
96.3%
12
97.1%
(31)
13−19
68.8 %
20-26
36.3 %
27−33
30.0 %
>33 days
26.3 %
(30)
* As of October 2020, we were able to obtain published estimates for SARS-
CoV-2 antigen test sensitivity only for days 6-12 post exposure (31). For
sensitivity beyond 12 days, we assume published estimates for SARS-CoV-2
RT-PCR tests under the assumption that antigen and PCR tests will have
similar sensitivity curves, increasing steadily following infection and peaking
around symptom onset (34). Prior to six days post exposure, we were unable
to find published sensitivity estimates for either type of test. To approximate,
we assume that sensitivity increases is proportional to viral load and use
values from the later in the infectious period with roughly comparable viral
levels (30, 31).
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
10
Individual-based network
The individual based SARS-CoV-2 infection dynamic model assumes that the virus spreads
through a fixed contact network consisting of 2019 individuals and 25,428 contacts between
those individuals. We populated our network by first constructing 1000 households. The size
and age composition of each household is based on a randomly sampled household from
among the 129,697 households included in 2017 National Household Travel Survey (35). We
assumed that households are fully connected (i.e., all nodes in the same household are linked
by edges). We construct random links between individuals in different households based on
reported age-specific contact rates in the US, stratified into age bins of 5-17, 18-49, 50-64, and
over 65 years (36). Specifically, to determine the number of contacts a node in age group has
𝑎
𝑖
with nodes in age group we draw a random deviate from the Poisson distribution centered at
𝑎
𝑗
the mean number of contacts between and . The resulting network includes 1000
𝑎
𝑖
𝑎
𝑗
households, 2019 nodes (people), and degrees (numbers of edges per node) that roughly follow
a gamma distribution with shape 3.69 and scale 3.41.
Estimating the Years of Life Lost (YLL) Averted
For each scenario, we run 1000 rounds of the 9 candidate testing strategies (including the
status quo). All parameters are identical except for those governing testing. For each round, we
determine the averted life loss (YLL) for each strategy , as follows:
𝜏
1. Calculate the difference in incidence by age group as , where and
Δ
𝑎,𝜏
=
𝐼
𝑎,0
𝐼
𝑎,𝜏
𝐼
𝑎,0
𝐼
𝑎,𝜏
are the total incidence of infection in age group produced by the status quo and
𝑎
strategy simulations, respectively.
𝜏
2. Estimate the YLL prevented by the testing strategy as
𝜏
𝐵
𝜏
=
𝑎
(
𝜆
𝑎
𝑎
)
𝛿
𝑎
Δ
𝑎,𝜏
where denotes the life expectancy for individuals of age (28) and denotes the
𝜆
𝑎
𝑎
𝜆
𝑎
age-specific case fatality rate for COVID-19 (27).
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
11
Estimating the Cost-Effectiveness Acceptability Curve
For each simulation of a testing strategy, we calculated the total incidence ( ), the YLL averted
𝐼
𝜏
in comparison to a base case simulation ( , as described in the preceding subsection) and cost
𝐵
𝜏
as the sum of salary loss of adults during isolation and the total cost of testing ( ) (Table 3).
𝐶
𝜏
For a given willingness to pay for a YLL averted ( ), we calculated the net monetary benefit
𝜃
(NMB) as: .
NM
B
𝜏
=
𝜃
𝐵
𝜏
𝐶
𝜏
We determined the optimal strategy across a range of scenarios, each defined by the effective
reproduction number ( ), willingness to pay, and cost of a test. For each scenario, we ran 1000
𝑅
𝑒
rounds of parallel simulations of each of the 9 candidate testing strategies (including the status
quo). For each of the 1000 rounds of 9 simulations, we identify the strategy with the highest
NMB. We then estimate the probability that a particular strategy has the greatest net benefit of
all strategies by the proportion of simulation rounds in which it resulted in the highest NMB. For
a given scenario, the strategy with the highest probability of having the highest NMB is
considered optimal.
Using this approach, we first assume a price of US$5 per test and determine optimal strategies
across a range willingness-to-pay per YLL averted up to US$200,000. Then, we fix willingness-
to-pay per YLL averted to US$100,000 and determine optimal strategies for a range of testing
prices up to $800 per test, given that tests are widely available for under $200 as of October,
2020 (37).
Table 3. Cost parameters
Parameter
Value
Cost of administering a single test
$5 US (16)
Average weekly salary of adults 18-65y
$1002 US (38)
Sensitivity Analysis We determined the optimal strategy across a range of effective
reproduction numbers ranging from 1.3 to 3 (Table 4). Assuming a willingness-to-pay per
𝑅
𝑒
YLL averted of $100,000 and a test price of US$5, the best strategy across all transmission
scenarios includes testing period and the shortest isolation considered (one-week post testing).
We also determine a testing threshold price (to the nearest US$100) assuming a willingness-to-
pay per YLL averted of $100,000, that is, the price per test above which the status quo is
expected to be optimal.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
12
Results
Using an individual-based model of within-host and between-host SARS-CoV-2 infection
dynamics, we compare eight testing strategies in which all individuals are tested at a frequency
ranging from daily to monthly, coupled with either a one-week or two-week isolation period for
confirmed cases (Table 1). Outcomes are quantified in terms of Years of Life Lost (YLL) from
infection, costs of diagnostic testing, and salary lost while in isolation.
The cost-effectiveness of each testing strategy depends on the willingness to pay for averted
YLL, the cost of each test and the transmission rate for the virus ( ) (Figure 2). Intuitively,
𝑅
𝑒
costs increase with the frequency of testing and the length of the isolation period; the costliest
option considered is daily testing coupled with a two-week isolation of confirmed cases. We then
identified the optimal strategy, that is, the strategy most likely to provide the greatest net
monetary benefits (NMB) for a given testing price and willingness-to-pay for YLL. If each test
costs US$5 and the society is willing to pay at least $90,000 to avert one YLL, the optimal
strategy under high transmission scenarios is daily testing coupled with one-week isolation for
confirmed cases. Under lower transmission scenarios, weekly testing with one-week isolation is
optimal (Figure 3). At intermediate transmission scenarios (i.e., between 1.7 and 2.0),
𝑅
𝑒
weekly testing with the longer, two-week isolation period is preferred (Table 4). Assuming a
willingness to pay per YLL averted of $100,000 (39), the optimal testing strategy pivotally
depends on the cost of a test. At US$100 per test, the optimal strategies are testing the
population either every two weeks or every week under the low or high transmission scenarios,
respectively, coupled with a two-week isolation period. Above US$500 per test, the optimal
strategy is the status quo (Table 4).
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
13
Figure 2. Estimated costs of testing strategies, assuming US$5 tests and an
effective reproduction number of either (A) or (B) . Each point
𝑹
𝒆
=
𝟏.𝟑
𝑹
𝒆
=
𝟐.𝟐
corresponds to one of 1000 stochastic simulations for the specified testing strategy,
under parameters given in Table 1. Costs include both the cost of the test and
salary lost during isolation following a positive test result; YLL averted considers
morbidity and mortality due to COVID-19 disease. The costs and YLL averted are
all scaled assuming a US population of 328.2 million, as estimated in 2019 (40).
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
14
Figure 3. Cost effectiveness acceptability curves (CEAC) assuming an effective reproduction
number of either (A,B) or (C,D) . (A, C) Assuming each test costs US$5, the probability
𝑹
𝒆
=
𝟏.𝟑
𝑹
𝒆
=
𝟐.𝟐
that a candidate strategy has the greatest net benefit under a given willingness to pay (x-axis) is based on
1000 rounds of stochastic simulations. In each round, every strategy is simulated and the one resulting in
the largest net monetary benefit is deemed optimal. (B,D) Assuming a willingness-to-pay of US$100,000,
the same procedure is applied across a range of testing prices (x-axis). The graphs depict only the three
best strategies for each x-axis value.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
15
Table 4. Optimal strategies and price threshold for cost effective testing under a range of SARS-
CoV-2 transmission scenarios, assuming a willingness to pay per YLL averted as $100,000. For
effective reproduction numbers ( ) ranging from 1.3 to 3, the middle columns give the optimal strategy
𝑅
𝑒
assuming that each test costs US$5. The rightmost column gives a threshold price above which the
status quo is expected to be more cost effective than all nine testing strategies considered. The threshold
value was identified by evaluating all strategies across a range of costs per test from $100 to $800 at $25
increments.
Optimal strategy
(assuming US$5 per test)
Re
Testing frequency
(days between tests)*
Isolation period
Testing threshold
(Cost per test in US$)
1.3
7
one week
400
1.4
7
one week
425
1.5
7
one week
450
1.6
7
one week
450
1.7
7
two weeks
425
1.8
7
two weeks
425
1.9
7
two weeks
500
2.0
7
two weeks
500
2.1
1
one week
475
2.2
1
one week
450
2.5
1
one week
450
3
1
one week
400
*A seven-day testing frequency means that individuals are tested once every seven days, on a rotating
basis.
Discussion
Aggressive and sustainable SARS-CoV-2 testing programs can substantially mitigate the threat
of COVID-19 in vulnerable communities and ensure the integrity of our healthcare systems,
while bolstering our societies and economies (41). In the US, where the willingness-to-pay to
avert one YLL is roughly $100,000 (39) and the cost of testing for SARS-CoV-2 is rapidly
declining, the optimal strategy depends on the transmission rate of the virus and the cost of the
test. In communities where non-pharmaceutical measures have substantially curtailed spread,
testing everyone weekly coupled with a one-week isolation period following a positive test result
is expected to be optimal. However, if the virus is spreading rapidly, daily testing is advisable.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
16
Given the estimated 7-10 day infectious period (25), many governments and public health
agencies have recommended isolating confirmed cases for 10-20 days after symptom onset
(42). In evaluating different isolation periods, we start the isolation clock from the date of testing,
which may occur before or after symptom onset. Notably, in many cases ( over 2.0 or under
𝑅
𝑒
1.7), we find that a one-week isolation period is expected to be more efficient than a two-week
isolation period. Roughly speaking, loss of salary in the second week outweighs the costs of
infections that occur during the second week following testing.
Our economic calculations have a limited scope, as we consider only the expense of testing and
the loss of salary during isolation. Testing may lead to additional expense, as asymptomatic or
paucisymptomatic individuals testing positive may seek healthcare when they would otherwise
not have suspected themselves to be infected. Conversely, preventing infectious spread averts
the costs associated with unrealized symptomatic COVID cases. Likewise, we quantify the
economic benefits of averting mortality, but do not consider the prevention of non-fatal morbidity
caused by SARS-CoV-2 infection, which is substantial (43), or the indirect health and mental
health consequences of the pandemic (10). Finally, we evaluate testing strategies across a
range of reproduction numbers, but do not account for the district or indirect costs of the non-
pharmaceutical interventions enacted to slow transmission in the milder scenarios.
Despite the intimidating up-front costs, our results demonstrate that ramping up mass
asymptomatic testing for SARS-CoV-2 across the US is a cost-effective and impactful strategy
for mitigating the unprecedented threat of the COVID-19 pandemic. When coupled with an
expansion of contact tracing programs, testing can be instrumental in averting pandemic waves
and allowing the relaxation of costly travel restrictions and social distancing measures (6, 44).
As we look towards the 2020-2021 influenza season, simultaneous surveillance testing for
SARS-CoV-2 and influenza may provide additional public health and economic benefits (7).
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
17
Author Contributions
Zhanwei Du, Abhishek Pandey, Yuan Bai, Meagan C. Fitzpatrick, Michael Lachmann,
Alessandro Vespignani, Benjamin J. Cowling, Alison P. Galvani, and Lauren Ancel Meyers:
conceived the study, designed statistical methods, conducted analyses, interpreted results,
wrote and revised the manuscript. Matteo Chinazzi and Ana Pastore y Piontti: data analysis,
generation of age-stratified contact patterns of the US and revised the manuscript.
Acknowledgments
Financial support was provided by US National Institutes of Health (grant no. U01 GM087719).
Declaration of interests
We declare no competing interests.
Role of the funding source
The funders had no role in the design, analysis, write-up or decision to submit for publication.
Data sharing
Data collected within the framework of this study are accessible to interested parties by
contacting the corresponding author.
Ethics committee approval
Not applicable.
References
1. W. Wan, WHO declares a pandemic of coronavirus disease covid-19. The Washington Post
(2020) (March 25, 2020).
2. COVID-19 Dashboard by the Center for Systems Science and Engineering (CSSE) at
Johns Hopkins University (JHU) (September 28, 2020).
3. D. Strauss, Pandemic knocks a tenth off incomes of workers worldwide. Financial Times
(2020) (September 28, 2020).
4. J. Corum, S.-L. Wee, C. Zimmer, Coronavirus Vaccine Tracker. The New York Times
(2020) (September 25, 2020).
5. September 4 coronavirus news. CNN (2020) (September 28, 2020).
6. E. Han, et al., Lessons learnt from easing COVID-19 restrictions: an analysis of countries
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
18
and regions in Asia Pacific and Europe. Lancet (2020) https:/doi.org/10.1016/S0140-
6736(20)32007-9.
7. S. M. Moghadas, et al., The implications of silent transmission for the control of COVID-19
outbreaks. Proc. Natl. Acad. Sci. U. S. A. 117, 17513–17515 (2020).
8. K. Mizumoto, K. Kagaya, A. Zarebski, G. Chowell, Estimating the asymptomatic proportion
of coronavirus disease 2019 (COVID-19) cases on board the Diamond Princess cruise ship,
Yokohama, Japan, 2020. Euro Surveill. 25 (2020).
9. H. Nishiura, et al., Estimation of the asymptomatic ratio of novel coronavirus infections
(COVID-19) https:/doi.org/10.1101/2020.02.03.20020248.
10. CDC, Coronavirus Disease 2019 (COVID-19) (2020) (September 28, 2020).
11. A. Aleta, et al., Modelling the impact of testing, contact tracing and household quarantine on
second waves of COVID-19. Nat Hum Behav (2020) https:/doi.org/10.1038/s41562-020-
0931-9.
12. N. H. S. England, N. H. S. Improvement, COVID-19 virus testing in NHS laboratories.
13. J. Ducharme, What to Know About COVID-19 Tests, From PCR to Antigen to Antibody.
Time (2020) (September 22, 2020).
14. Center for Devices, Radiological Health, In Vitro Diagnostics EUAs (2020) (September 28,
2020).
15. The Washington Times http://www. washingtontimes.com, T. Howell Jr, FDA gives
emergency approval to “game changer” COVID-19 sa (September 22, 2020).
16. A. Park, For $5 and in 15 Minutes You Can Learn If You Have COVID-19. Time (2020)
(September 9, 2020).
17. M. Greenwood, SalivaDirect: What You Need to Know About the New COVID-19 Test
(2020) (September 29, 2020).
18. CDC, Interim Guidelines for COVID-19 Antibody Testing (2020) (September 29, 2020).
19. U.S. health insurers may balk at paying for coronavirus antibody testing. Reuters (2020)
(September 29, 2020).
20. CDC, Commercial Laboratory Seroprevalence Surveys (2020) (October 8, 2020).
21. M. J. Ferrari, S. Bansal, L. A. Meyers, O. N. Bjørnstad, Network frailty and the geometry of
herd immunity. Proceedings of the Royal Society B: Biological Sciences 273, 2743–2748
(2006).
22. Q. Li, et al., Early Transmission Dynamics in Wuhan, China, of Novel Coronavirus–Infected
Pneumonia. N. Engl. J. Med. 382, 1199–1207 (2020).
23. P. J. Turk, et al., Modeling COVID-19 Latent Prevalence to Assess a Public Health
Intervention at a State and Regional Scale: Retrospective Cohort Study. JMIR Public Health
and Surveillance 6, e19353 (2020).
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
19
24. J. A. Backer, D. Klinkenberg, J. Wallinga, Incubation period of 2019 novel coronavirus
(2019-nCoV) infections among travellers from Wuhan, China, 20-28 January 2020. Euro
Surveill. 25, 2000062 (2020).
25. X. He, et al., Temporal dynamics in viral shedding and transmissibility of COVID-19. Nat.
Med. 26, 672–675 (2020).
26. D. Mc Evoy, et al., The relative infectiousness of asymptomatic SARS-CoV-2 infected
persons compared with symptomatic individuals: A rapid scoping review. Infectious
Diseases (except HIV/AIDS) (2020) https:/doi.org/10.1101/2020.07.30.20165084.
27. R. Verity, et al., Estimates of the severity of coronavirus disease 2019: a model-based
analysis. Lancet Infect. Dis. 20, 669–677 (2020).
28. E. Arias, M. Heron, J. Xu, United States Life Tables, 2014. Natl. Vital Stat. Rep. 66, 1–64
(2017).
29. Z. Wu, J. M. McGoogan, Characteristics of and Important Lessons From the Coronavirus
Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72 314 Cases From
the Chinese Center for Disease Control and Prevention. JAMA (2020)
https:/doi.org/10.1001/jama.2020.2648.
30. A. T. Xiao, et al., Dynamic profile of RT-PCR findings from 301 COVID-19 patients in
Wuhan, China: A descriptive study. J. Clin. Virol. 127, 104346 (2020).
31. Abbott Diagnostics Scarborough, Inc., BinaxNOW COVID-19 Ag CARD. Food and Drug
Administration (2020) (September 18, 2020).
32. N. Sethuraman, S. S. Jeremiah, A. Ryo, Interpreting Diagnostic Tests for SARS-CoV-2.
JAMA (2020) https:/doi.org/10.1001/jama.2020.8259 (May 6, 2020).
33. Y. Hirotsu, et al., Comparison of automated SARS-CoV-2 antigen test for COVID-19
infection with quantitative RT-PCR using 313 nasopharyngeal swabs, including from seven
serially followed patients. International Journal of Infectious Diseases 99, 397–402 (2020).
34. G. Guglielmi, Fast coronavirus tests: what they can and can’t do. Nature 585, 496–498
(2020).
35. U.S. Federal Highway Administration, 2017 National Household Travel Survey (June 16,
2020).
36. D. Mistry, et al., Inferring high-resolution human mixing patterns for disease modeling. arXiv
[q-bio.PE] (2020).
37. S. Kliff, Most Coronavirus Tests Cost About 100. Why Did One Cost 2,315? The New York
Times (2020) (October 6, 2020).
38. Bureau of Labor Statistics, USUAL WEEKLY EARNINGS OF WAGE AND SALARY
WORKERS SECOND QUARTER 2020 (2020) (September 4, 2020).
39. P. J. Neumann, J. T. Cohen, M. C. Weinstein, Updating cost-effectiveness--the curious
resilience of the $50,000-per-QALY threshold. N. Engl. J. Med. 371, 796–797 (2014).
40. US Census Bureau, National Population by Characteristics: 2010-2019 (October 1, 2020).
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
20
41. European Centre for Disease Prevention and Control, Testing strategies for SARS-CoV-2
(September 24, 2020).
42. CDC, Duration of Isolation and Precautions for Adults with COVID-19 (2020) (September
28, 2020).
43. S. M. Moghadas, et al., Projecting hospital utilization during the COVID-19 outbreaks in the
United States. Proc. Natl. Acad. Sci. U. S. A. 117, 9122–9126 (2020).
44. European Centre for Disease Prevention and Control, Population-wide testing of SARS-
CoV-2: country experiences and potential approaches in the EU/EEA and the United
Kingdom (2020) (September 24, 2020).
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3714642
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Mathematical and computational modeling approaches are increasingly used as quantitative tools in the analysis and forecasting of infectious disease epidemics. The growing need for realism in addressing complex public health questions is, however, calling for accurate models of the human contact patterns that govern the disease transmission processes. Here we present a data-driven approach to generate effective population-level contact matrices by using highly detailed macro (census) and micro (survey) data on key socio-demographic features. We produce age-stratified contact matrices for 35 countries, including 277 sub-national administratvie regions of 8 of those countries, covering approximately 3.5 billion people and reflecting the high degree of cultural and societal diversity of the focus countries. We use the derived contact matrices to model the spread of airborne infectious diseases and show that sub-national heterogeneities in human mixing patterns have a marked impact on epidemic indicators such as the reproduction number and overall attack rate of epidemics of the same etiology. The contact patterns derived here are made publicly available as a modeling tool to study the impact of socio-economic differences and demographic heterogeneities across populations on the epidemiology of infectious diseases.
Article
Full-text available
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection is determined by reverse-transcription PCR (RT-PCR) in routine clinical practice. In the current pandemic situation, a more rapid and high-throughput method is in growing demand. Here, we validated the performance of a new antigen test (LUMIPULSE) based on the chemiluminescence enzyme immunoassay. A total of 313 nasopharyngeal swabs (82 serial samples from 7 infected patients, 231 individual samples from 4 infected patients and 215 non-infected individuals) were analyzed for SARS-CoV-2 by quantitative RT-PCR (RT-qPCR) and then subjected to LUMIPULSE. We determined the cutoff value for antigen detection using receiver operating characteristic curve analysis and compared the antigen test performance with that of RT-qPCR. Further, we compared the viral loads and antigen levels in serial samples from seven infected patients. When using RT-qPCR as the reference, the antigen test exhibited 55.2% sensitivity and 99.6% specificity with a 91.4% overall agreement rate (286/313). In specimens with > 100 viral copies and between 10 and 100 copies, the antigen test showed 100% and 85% concordance with RT-qPCR, respectively. This concordance declined with lower viral loads. In the serially followed patients, the antigen levels showed a steady decline along with viral clearance. This gradual decline was in contrast with the abrupt “positive-to-negative” and “negative-to-positive” status changes observed with RT-qPCR, particularly in the late phase of infection. In summary, the LUMIPULSE antigen test can rapidly identify SARS-CoV-2-infected individuals with moderate to high viral loads and may be helpful for monitoring viral clearance in hospitalized patients.
Article
Full-text available
While severe social-distancing measures have proven effective in slowing the coronavirus disease 2019 (COVID-19) pandemic, second-wave scenarios are likely to emerge as restrictions are lifted. Here we integrate anonymized, geolocalized mobility data with census and demographic data to build a detailed agent-based model of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) transmission in the Boston metropolitan area. We find that a period of strict social distancing followed by a robust level of testing, contact-tracing and household quarantine could keep the disease within the capacity of the healthcare system while enabling the reopening of economic activities. Our results show that a response system based on enhanced testing and contact tracing can have a major role in relaxing social-distancing interventions in the absence of herd immunity against SARS-CoV-2. An agent-based model of SARS-CoV-2 transmission shows that testing, contact tracing and household quarantine could keep new COVID-19 waves under control while allowing the reopening of the economy with minimal social-distancing interventions.
Preprint
Full-text available
Objectives: The aim of this study was to conduct a scoping review of estimates of the relative infectiousness of asymptomatic persons infected with SARS-CoV-2 compared with symptomatic individuals. Design: Rapid scoping review of literature available until 8th April 2020. Setting: International studies on the infectiousness of individuals infected with SARS-CoV-2 Participants: Studies were selected for inclusion if they defined asymptomatics as a separate cohort distinct from pre-symptomatics and if they provided a quantitative measure of the infectiousness of asymptomatics relative to symptomatics. Primary outcome measures: The relative number of secondary cases produced by an average primary case, the relative probability of transmitting infection upon contact, and the degree of viral shedding. Results: Very few studies reported estimates of relative infectiousness of asymptomatic compared with symptomatic individuals. Significant differences exist in the definition of infectiousness. Viral shedding studies in general show no difference in shedding levels between symptomatic and asymptomatic individuals but are likely to be impacted by insufficient statistical power. Two contact tracing studies provided estimates of 0.7 and 1.0, but differences in approach and definition preclude comparison across the two studies. Finally, two modelling studies suggest a relative infectiousness of around 0.5 but one of these was more reflective of the infectiousness of undocumented rather than asymptomatic cases. Importantly, one contact tracing study showing a very low level of infectiousness of asymptomatic was not included in the analysis at this point due difficulties interpreting the reported findings. Conclusions: The present study highlights the need for additional studies in this area as a matter of urgency. For the purpose of epidemiological modelling, we cautiously suggest that at present, asymptomatics could be considered to have a degree of infectiousness which is about 0.40-0.70 that of symptomatics. However, it must be stressed that this suggestion comes from a very low evidence base and that estimates exist that are close to zero and close to 1.
Article
Full-text available
Since the emergence of coronavirus disease 2019 (COVID-19), unprecedented movement restrictions and social distancing measures have been implemented worldwide. The socioeconomic repercussions have fueled calls to lift these measures. In the absence of population-wide restrictions, isolation of infected individuals is key to curtailing transmission. However, the effectiveness of symptom-based isolation in preventing a resurgence depends on the extent of presymptomatic and asymptomatic transmission. We evaluate the contribution of presymptomatic and asymptomatic transmission based on recent individual-level data regarding infectiousness prior to symptom onset and the asymptomatic proportion among all infections. We found that the majority of incidences may be attributable to silent transmission from a combination of the presymptomatic stage and asymptomatic infections. Consequently, even if all symptomatic cases are isolated, a vast outbreak may nonetheless unfold. We further quantified the effect of isolating silent infections in addition to symptomatic cases, finding that over one-third of silent infections must be isolated to suppress a future outbreak below 1% of the population. Our results indicate that symptom-based isolation must be supplemented by rapid contact tracing and testing that identifies asymptomatic and presymptomatic cases, in order to safely lift current restrictions and minimize the risk of resurgence.
Article
Full-text available
The pandemic of coronavirus disease 2019 (COVID-19) continues to affect much of the world. Knowledge of diagnostic tests for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is still evolving, and a clear understanding of the nature of the tests and interpretation of their findings is important. This Viewpoint describes how to interpret 2 types of diagnostic tests commonly in use for SARS-CoV-2 infections—reverse transcriptase–polymerase chain reaction (RT-PCR) and IgM and IgG enzyme-linked immunosorbent assay (ELISA)—and how the results may vary over time
Article
Full-text available
We report temporal patterns of viral shedding in 94 patients with laboratory-confirmed COVID-19 and modeled COVID-19 infectiousness profiles from a separate sample of 77 infector–infectee transmission pairs. We observed the highest viral load in throat swabs at the time of symptom onset, and inferred that infectiousness peaked on or before symptom onset. We estimated that 44% (95% confidence interval, 25–69%) of secondary cases were infected during the index cases’ presymptomatic stage, in settings with substantial household clustering, active case finding and quarantine outside the home. Disease control measures should be adjusted to account for probable substantial presymptomatic transmission. Presymptomatic transmission of SARS-CoV-2 is estimated to account for a substantial proportion of COVID-19 cases.
Article
The COVID-19 pandemic is an unprecedented global crisis. Many countries have implemented restrictions on population movement to slow the spread of severe acute respiratory syndrome coronavirus 2 and prevent health systems from becoming overwhelmed; some have instituted full or partial lockdowns. However, lockdowns and other extreme restrictions cannot be sustained for the long term in the hope that there will be an effective vaccine or treatment for COVID-19. Governments worldwide now face the common challenge of easing lockdowns and restrictions while balancing various health, social, and economic concerns. To facilitate cross-country learning, this Health Policy paper uses an adapted framework to examine the approaches taken by nine high-income countries and regions that have started to ease COVID-19 restrictions: five in the Asia Pacific region (ie, Hong Kong [Special Administrative Region], Japan, New Zealand, Singapore, and South Korea) and four in Europe (ie, Germany, Norway, Spain, and the UK). This comparative analysis presents important lessons to be learnt from the experiences of these countries and regions. Although the future of the virus is unknown at present, countries should continue to share their experiences, shield populations who are at risk, and suppress transmission to save lives.
Article
Rapid antigen tests are designed to tell in a few minutes whether someone is infectious. Will they be game changers? Rapid antigen tests are designed to tell in a few minutes whether someone is infectious. Will they be game changers?
Article
Background: Emergence of COVID-19 caught the world off-guard and unprepared, initiating a global pandemic. In the absence of evidence, individual communities had to take timely action to reduce the rate of disease spread and avoid overburdening their healthcare systems. Although a few predictive models have been published to guide these decisions, most have not taken into account spatial differences and have included assumptions that do not match the local realities. Access to reliable information that is adapted to local context is critical for policymakers to make informed decisions during a rapidly evolving pandemic. Objective: The goal of this study was to develop an adapted susceptible-infected-removed (SIR) model to predict the trajectory of the COVID-19 pandemic in North Carolina (NC) and the Charlotte metropolitan region and to incorporate the effect of a public health intervention to reduce disease spread, while accounting for unique regional features and imperfect detection. Methods: Using the software package R, three SIR models were fit to infection prevalence data from the state and the greater Charlotte region and then rigorously compared. One of these models (SIR-Int) accounted for a stay-at-home intervention and imperfect detection of COVID-19 cases. We computed longitudinal total estimates of the susceptible, infected, and removed compartments of both populations, along with other pandemic characteristics (e.g., basic reproduction number). Results: Prior to March 26, disease spread was rapid at the pandemic onset with the Charlotte region doubling time of 2.56 days (95% CI: (2.11, 3.25)) and in NC 2.94 days (95% CI: (2.33, 4.00)). Subsequently, disease spread significantly slowed with doubling times increased in the Charlotte region to 4.70 days (95% CI: (3.77, 6.22)) and in NC to 4.01 days (95% CI: (3.43, 4.83)). Reflecting spatial differences, this deceleration favored the greater Charlotte region compared to NC as a whole. A comparison of the efficacy of intervention, defined as 1 - the hazard ratio of infection, gave 0.25 for NC and 0.43 for the Charlotte region. Also, early in the pandemic, the initial basic SIR model had good fit to the data; however, as the pandemic and local conditions evolved, the SIR-Int model emerged as the model with better fit. Conclusions: Using local data and continuous attention to model adaptation, our findings have enabled policymakers, public health officials and health systems to proactively plan capacity and evaluate the impact of a public health intervention. Our SIR-Int model for estimated latent prevalence was reasonably flexible, highly accurate, and demonstrated the efficacy of a stay-at-home order at both the state and regional level. Our results highlight the importance of incorporating local context into pandemic forecast modeling, as well as the need to remain vigilant and informed by the data as we enter into a critical period of the outbreak. Clinicaltrial: