PreprintPDF Available

Abstract and Figures

Background: As phases of COVID-19 vaccination are quickly rolling out, how to evaluate the vaccination effects and then make safe reopening plans has become a prime concern for local governments and school officials. Methods: We develop a contact network agent-based model (CN-ABM) to simulate on-campus disease transmission scenarios at the micro-scale. The CN-ABM establishes a contact network for each agent based on their daily activity pattern, evaluates the agent's health status change in different activity environments, and then simulates the epidemic curve on campus. Based on the developed model, we identify how different community risk levels, teaching modalities, and vaccination rates would shape the epidemic curve. Results: The results show that in scenarios where vaccination is not available, restricting on-campus students to under 50% can largely flatten the epi curve (peak value < 2%); and the best result (peak value < 1%) can be achieved by limiting on-campus students to less than 25%. In scenarios where vaccination is available, it is suggested to maintain a maximum of 75% on-campus students and a vaccination rate of at least 45% to suppress the curve (peak value < 2%); and the best result (peak value < 1%) can be achieved at a vaccination rate of 65%. The study also derives the transmission chain of infectious agents, which can be used to identify high-risk activity environments. Conclusions: The developed CN-ABM model can be employed to evaluate the health outcome of COVID-19 outbreaks on campus based on different disease transmission scenarios. It can assist local government and school officials with developing proactive intervention strategies to safely reopen schools.
Content may be subject to copyright.
Page 1/21
Reopen Schools Safely: Simulating COVID-19
Transmission on Campus With a Contact Network
Agent-based Model
Chuyao Liao
Sun Yat-sen University
Xiang Chen
University of Connecticut
Li Zhuo ( zhuoli@mail.sysu.edu.cn )
Sun Yat-sen University
Yuan Liu
Sun Yat-sen University
Haiyan Tao
Sun Yat-sen University
Research Article
Keywords: COVID-19, contact network, vaccination, agent-based modeling, school
DOI: https://doi.org/10.21203/rs.3.rs-627505/v1
License: This work is licensed under a Creative Commons Attribution 4.0 International License. 
Read Full License
Page 2/21
Abstract
Background: As phases of COVID-19 vaccination are quickly rolling out, how to evaluate the vaccination
effects and then make safe reopening plans has become a prime concern for local governments and
school ocials.
Methods: We develop a contact network agent-based model (CN-ABM) to simulate on-campus disease
transmission scenarios at the micro-scale. The CN-ABM establishes a contact network for each agent
based on their daily activity pattern, evaluates the agent's health status change in different activity
environments, and then simulates the epidemic curve on campus. Based on the developed model, we
identify how different community risk levels, teaching modalities, and vaccination rates would shape the
epidemic curve.
Results: The results show that in scenarios where vaccination is not available, restricting on-campus
students to under 50% can largely atten the epi curve (peak value < 2%); and the best result (peak value <
1%) can be achieved by limiting on-campus students to less than 25%. In scenarios where vaccination is
available, it is suggested to maintain a maximum of 75% on-campus students and a vaccination rate of
at least 45% to suppress the curve (peak value < 2%); and the best result (peak value < 1%) can be
achieved at a vaccination rate of 65%. The study also derives the transmission chain of infectious
agents, which can be used to identify high-risk activity environments.
Conclusions: The developed CN-ABM model can be employed to evaluate the health outcome of COVID-
19 outbreaks on campus based on different disease transmission scenarios. It can assist local
government and school ocials with developing proactive intervention strategies to safely reopen
schools.
1. Background
School environments are characterized by frequent social gatherings and extensive interpersonal
interactions, making them ideal breeding grounds for viruses [1, 2]. Due to the highly contagious nature of
the coronavirus disease 2019 (COVID-19), school closure and pedagogical transitions were widely
implemented to control the spread of the disease [3]. The World Health Organization (WHO) estimated
that about 90 percent of the world’s students, which were more than 1.5billion children and young adults,
were affected by school closure [4]. The cascading impacts of school closure included reduced work
eciency [5], lack of educational resources [6], and mental health issues caused by social isolation [7].
Burdened with these grave concerns and the political pressure to revive the economy, governments of
many countries laid out plans to reopen schools. However, these reopening plans and their
implementation often lacked scientic evidence to weigh the potential health consequences. For example,
a national survey of more than 1,500 universities and colleges in the United States revealed over 51,000
COVID-19 cases and 60 deaths as a result of school reopening in early September 2020 [8]. The sobering
Page 3/21
reality signies that it is of utmost signicance to employ rigorous measures to evaluate the health risk
for school reopening [9].
Existing studies have used process-based dynamic epidemic models, primarily the Susceptible-Exposed-
Infective-Recovered (SEIR) model and its variations, to simulate the potential for COVID-19 outbreaks
[10–13]. Although these models can project the epidemic development under different social distancing
scenarios, they cannot be easily implemented in a school environment without major modication. First,
existing studies are mainly focused on macro-scale assessments with aggregate administrative units
(e.g., counties, census tracts). They are unable to model the intricacy of the disease transmission process
at the individual scale, known as the micro-scale, which plays a vital role in determining the likelihood of
virus spread in a school environment. Second, existing studies mostly employ deterministic models
without considering the stochastic transmission process, such as the infection rates in different activity
environments (e.g., inside buildings, outdoor); and these uncertainties at the micro-scale can dictate the
likelihood of outbreaks when school reopens. To date, existing micro-scale predictive models, especially
those applicable to schools, are considerably lacking.
In this paper, we propose a contact network agent-based model (CN-ABM) to simulate the COVID-19
transmission process in a school environment. In the model, we dene on-campus students as agents
and construct a ne-grained dynamic contact network between agents based on their daily activity
patterns and movement trajectories. Then, we employ the model to simulate on-campus transmission
scenarios under WHO’s guidelines for school reopening [14]. Specically, we evaluate how the epidemic
curve (i.e., epi curve) is shaped by different community risk levels, teaching modalities, and vaccination
rates. It is hoped that the proposed model can help project timelines to safely reopen school amid the
ongoing COVID-19 vaccination practices.
The paper is organized as follows. Section 2 introduces the methods, including the conceptual framework
and the proposed simulation model. Section 3 applies the model to a school environment and derives the
results, including the epi curve and the transmission chain. Section 4 performs the sensitivity analysis of
the model by incorporating the vaccination rate, discusses its public health implications, and also
presents the limitations. Finally, Sect.5 concludes the study with future directions.
2. Methods
2.1 Conceptualizing contact network
The agent-based model (ABM) is a micro-scale model that simulates the synchronous interactions of
agents based on pre-dened rules [15]. The ABM has been integrated into network science to enhance our
understanding of human behaviors [16]. In a network-based ABM, each agent can be represented as a
node, which interacts with other agents under predened behavioral assumptions in an abstracted
interaction network [17]. Recently, the network-based ABM has received elevated attention in near real-
Page 4/21
time genome sequencing of the SARS-CoV-2 virus [18]. Its applications to the COVID-19 transmission in
communities (e.g., schools) are relatively lacking.
The transmission process of the SARS-CoV-2 virus can be regarded as the transition of an individual's
health status via interactions with infectious agents [19]. This process can be illustrated by Fig. 1, where
an agent’s health status could change if the agent’s contact network entails an infectious agent.
Additionally, the internal transmission process inside a school can be inuenced by the community
spread beyond the school environment [20]. We thus incorporate the community risk as an external node
into the contact network, meaning that each agent can become infectious at a given probability even
without contacting other internal agents (Fig. 1).
To construct the contact network, all agents’ daily activity patterns, including their movement trajectories,
locations of stay in an activity environment (e.g., dining hall, residential building), and periods of stay,
must be acquired. Deriving these activity patterns for all agents is a prerequisite for building the contact
network at each time point. This process in our case study is detailed in the Appendix.
2.2 Infection phases
Based on the classical SEIR model, an agent’s health status in an infection cycle can be divided into four
statuses: susceptible (S), exposed (E), infectious (I), and recovered (R). The transition of the health status,
reecting the disease progression, is key to simulating the spread of an epidemic. Based on the study of
modeling school reopening [21] and the successive vaccination stage [22], we introduce six heath
statuses to illustrate a complete infection cycle, including susceptible (S), exposed (E), pre-symptomatic
(Ip), infected (I), recovered/removed (R), and vaccinated (V). These six health statuses constitute ve
infection phases, as shown in Fig. 2.
The transition of an agent’s health status at each infection phase is articulated below.
Phase 1 (S E): the probability of an agent transitioning from S to E in an activity environment
j
at time
t
is
Pt,j
. It is dependent on both the internal infection probability
Φt,j
and the external (community) infection
probability
Ψ
, as shown in Eq. (1). An agent’s internal infection probability
Φt,j
in activity environment
j
at
time t is shown in Eq. (2). An agent’s external infection
Ψ
is a ratio of the community infection rate
p
c to
the agent’s health level
H
, as shown in Eq.(3)
    (1)
    (2)
    (3)
Notation:
Page 5/21
H
: agent’s health level;
K
: decay coecient for the infection rate of an
Ip
-status agent;
mt
: number of
Ip
-status neighbors in an agent’s contact network at time
t
;
nt
: number of
I
-status neighbors in an agent’s contact network at time
t
;
pc
: community infection rate;
pj
: infection rate in activity environment
j
;
Pt,j
: probability of an agent transitioning from S to E in activity environment
j
at time
t
;
Φt,j
: internal infection probability;
Ψ
: external infection probability.
Phase 2 (E Ip): an E-status agent transitions to the Ip-status after a latent period (ε−1). During this
phase (
t
through
t
 + 
ε
−1), the agent is not infectious.
Phase 3 (Ip I): an Ip-status agent transitions to the I status after the prodromal period (
µp
−1); and the
duration from E to I is the incubation period
σ
−1. During this phase (
t
 + 
ε
−1 through
t
 + 
σ
−1), an Ip-status
agent infects all S-status neighbors at an infection rate
Kp
j if they are within the same contact network.
Phase 4 (I R): an I-status agent transitions to R-status after the infection period (
γ
−1). During this phase
(
t
 + 
σ
−1 through
t
 + 
ε
−1 + 
γ
−1), the I-status agent infects S-status neighbors in its contact network at the
infection rate
p
j.
Phase 5 (S V): an S-status agent transitions to V-status, if it is vaccinated. A V-status agent is not
infectious and cannot be infected. The number of S-status agents is determined by the initial number of
S-status agents
S0
and the immunized agents
αvS0
, where
v
is the vaccination rate and
α
is the vaccine
ecacy [23], as is shown in Eq.(4).
   (4)
The parameters in Fig. 2 and Equations (1) through (4) are derived from existing epidemiological
parameters, as shown in Table 1.
Page 6/21
Table 1
Epidemiological parameters used in the simulation model.
Parameter Explanation Value Reference
σ
−1 Incubation period
σ
−1 ~ lognormal (1.6, 0.5)
σ
−1[2, 14]
days
[24]
µp
−1 Prodromal period 2 days [25]
ε
−1 Latent period
σ
−1 -
µp
−1 -
γ
−1 Infection period
γ
−1 ~ lognormal (2.05, 0.25) [26]
K
Decay coecient for the
infection rate of an
Ip
-status
agent
1/3 [27]
p
jInternal infection rate in an
activity environment
j
Dining hall: 3.03e− 4/min;
residential building: 1.74e− 4/min;
lecture hall (including library): 3.30e− 
5/min
[27]
p
cCommunity infection rate 9.5e− 8/min [20]
H
Agent’s health level 1/
H
 ~ normal(1, 0.1) [28]
α
Vaccine ecacy 80% [29]
The workow for the ABM simulation is given in Fig. 3. We call this proposed model the contact network
agent-based model (CN-ABM).
3. Results
Following the aforementioned method, we built the proposed CN-ABM model in MATLAB R2019a. The
code of the model can be access in GitHub (https://github.com/xic19022/cnabm). Our computational
environment was as follows: CPU Intel(R) Xeon(R) E5-2620 (8-core/2.10GHz) and RAM of 64GB. We then
applied the model in a real-world school environment with different environmental settings.
3.1 Model initialization
Our study area was a university campus located in Southern China. We derived the building footprints
and the road networks on campus using the Baidu Map application programming interface (API)
(https://lbsyun.baidu.com/), as shown in Fig. 4. The shortest network distance between every two
buildings was also derived by the API and was converted to walking time (in minutes) based on the
Page 7/21
average human walking speed (i.e., 5 km per hour). To initialize the simulation, we generated 10,000
agents in the starting scenario (
t
 = 0) and introduced 1 pre-symptomatic cases among them. We
allocated these agents into the residential buildings proportional to the building capacities. Then, each
agent was assumed to follow a similar activity pattern but select the activity location of the same type
randomly. The details about generating the agents’ activity patterns are given in the Appendix.
In the rst set of analyses, we consider twelve school reopening scenarios based on different teaching
modalities (i.e., the composition of on-campus and distance-learning students) and community risk levels
(i.e., the community infection rate), as shown in Table 2. We consider that the distance-learning students
were not present on campus and were thus excluded from the simulation. We did not consider the
vaccination phase (i.e., Phase 5 in Fig. 2) in the initial set of analyses, as it is discussed in the follow-up
section. For each scenario, we performed the simulation for twenty-ve weeks at a one-minute timestamp;
we also repeated each simulation ve times to account for the stochastic nature of the disease
transmission.
Table 2
Twelve school reopening scenarios based on different community risk levels and teaching modalities.
Student composition (% of distance-learning)
Community risk level
(community infection
rate)
In-class
(0%) Mostly in-class
(25%) Hybrid
(50%) Mostly online
(75%)
No risk (0) N-S0 N-S25 N-S50 N-S75
Low risk (
p
c)*L-S0 L-S25 L-S50 L-S75
High risk (5x
p
c) H-S0 H-S25 H-S50 H-S75
*The community infection rate
p
c is 9.5e− 8/min [20].
3.2 Simulated epi curves
The simulation results in terms of the epi curves (i.e., the percentage of active infectious agents over
time) under the twelve school reopening scenarios are given in Fig. 5. Figure 5 reveals two important
epidemic patterns. First, the teaching modality largely affects the epi curve. For example, on a low
community risk level when all students are on campus (L-S0), the peak value of the epi curve or the
maximum percentage of infections is 22.38% (the blue curve, Fig. 5b). Then, reducing the on-campus
students by 25% (L-S25), 50% (L-S50), and 75% (L-S75) will considerably atten the epi curve, where the
peak values are 12.40%, 1.64%, and 0.40%, respectively (Fig. 5b). Second, the inuence of the community
risk on the transmission is moderate, and a high community risk advances the emergence of the peak
time. For example, under the same modality of “mostly in-class (red curves in Fig. 5)” but different
Page 8/21
community risk levels, the peaks of the epi curves emerge on the 95th day (N-S25, no risk, Fig. 5a), the
83rd day (L-S25, low risk, Fig. 5b), and the 64th day (H-S25, high risk, Fig. 5c), respectively. The statistics
of the epi curves, including the peak height and the peak day, are given in Table 3.
Table 3
Statistics of the simulation results under school reopening scenarios in Table 2.
Scenario Peak height (min, max)*Peak reduction rate** Peak day (min, max)*
N-S0 21.49% (21.30%, 22.02%) - 74.49 (66.54, 74.66)
N-S25 12.03% (11.29%, 12.40%) 44.02% 94.61 (83.81, 96.24)
N-S50 0.14% (0.12%, 0.36%) 99.35% 10.53 (8.07, 159.98)***
N-S75 0.12% (0.12%, 0.16%) 99.44% 6.38 (5.27, 8.35)***
L-S0 22.38% (21.18%, 23.77%) - 64.91 (60.44, 67.21)
L-S25 12.40% (11.60%, 13.23%) 44.59% 82.93 (78.30, 86.56)
L-S50 1.64% (1.56%, 2.02%) 92.67% 124.93 (121.58, 144.74)
L-S75 0.40% (0.32%, 0.48%) 98.21% 146.17 (111.33, 168.63)
H-S0 23.38% (22.66%, 24.46%) - 53.26 (52.40, 55.18)
H-S25 13.52% (13.41%, 14.11%) 42.17% 63.64 (62.96, 69.87)
H-S50 3.64% (3.10%, 4.00%) 84.43% 73.54 (63.37, 90.77)
H-S75 1.12% (0.92%, 1.16%) 95.21% 151.72 (100.54, 165.57)
*The values of the peak height and peak day are derived from the median of the ve repeated
simulations for each scenario.
**Dened as the current peak height relative to the peak height in the corresponding baseline scenario
(S0).
***In this scenario, the peak emerges early, as the outbreak is not prominent.
3.3 Transmission chain
The proposed CN-ABM model is capable to project the COVID-19 epi curve for a potential outbreak
scenario. More importantly, it can be used to infer the timeline of the disease progression, such as the
transmission chain. Figure 6 illustrates an example of the inferred transmission chain for an infectious
agent (i.e., the initial case) over the rst 35 days of the disease progression. Specically, the gure shows
that under the N-S0 scenario, the initial case can affect 104 other agents in its contact network, while
most of the infections (58.65%) occur in a residential building (i.e., 61 agents in the residential building,
30 agents in the dining hall, 12 agents in the lecture hall, and 1 agent in the library). Constructing this
Page 9/21
transmission chain can help track where and when the transmission takes place and can provide
evidence to formulate proactive intervention strategies in high-risk activity environments [30].
4. Discussion
4.1 How does vaccination shape the epi curve?
While phases of vaccination are rolling out in most world regions, there has been a lack of understanding
about how the vaccination shapes the epi curve at the community scale [23]. In the second set of
analyses, we evaluated the health outcomes under fteen school reopening scenarios based on
combinations of community risk levels and vaccination rates (Table 4). The vaccination phase (i.e.,
Phase 5 in Fig. 2) presumes that when a portion of S-status agents convert to V-status, they become
immune to the virus and are excluded from the infection cycle. To facilitate the bi-variate analysis, we
assumed all agents were in-class (0% distance-learning) in the model initialization. The simulation results
in terms of the epi curves are given in Fig. 7.
Table 4
Fifteen school reopening scenarios based on different community risk levels
and vaccination rates.
Community risk level
(community infection rate)
Vaccination rate (% of vaccinated agents)
5% 25% 45% 65% 85%
No risk (0) N-V5 N-V25 N-V45 N-V65 N-V85
Low risk (
p
c)*L-V5 L-V25 L-V45 L-V65 L-V85
High risk (5x
p
c) H-V5 H-V25 H-V45 H-V65 H-V85
*The community infection rate
p
c is 9.5e− 8/min [20].
Figure 7 reveals two important epidemic patterns. First, the vaccination rate is the primary parameter
shaping the epi curve. For example, on a low community risk level, high vaccination rates will largely
atten the curve—the vaccination rates of 5% (L-V5), 25% (L-V25), 45% (L-V45), 65% (L-V65), and 85% (L-
85) yield a peak value of 19.62%, 11.73%, 5.23%, 1.14%, and 0.20%, respectively (Fig.7b). Second, similar
to the results in Sect.3.2, the inuence of the community risk is moderate, and a high community risk
slightly advances the emergence of the peak. For example, at the vaccination rate of 45% (the green
curves in Fig.7), the peak of the epi curve emerges on the 102nd day (N-V45, no risk, Fig.7a), the 90th
day (L-V45, low-risk, Fig.7b), and the 71st day (H-V45, high risk, Fig.7c), respectively.
Our third set of analyses simulates the health outcome as a result of different teaching modalities and
vaccination rates. Since the distance-learning modality and vaccination are both preventive measures to
Page 10/21
contain the disease transmission, it is expected that combining these two measures will mitigate the
likelihood of outbreaks to the largest extent. Thus, we designed twenty school reopening scenarios by
combining different teaching modalities and vaccination rates (Table5), whereas all scenarios are
predicated on a low community risk level (
p
c = 9.5e− 8/min). The simulation results in terms of the epi
curves are given in Fig.8.
Figure 8 shows that when 50% (Fig. 8c) or 75% (Fig. 8d) of the students are off-campus, the virus spread
will be well controlled. While in reality, it is not likely to maintain a high rate of distancing learning
modality over an elongated period, the result further suggests that a modest vaccination rate will also
help to curb the infection when most students return to campus. Specically, when 75% of the students
are on campus, a vaccination rate of 45% or above can effectively atten the curve by limiting its peak
value to under 2% (the green curve in Fig. 8b); and the best result (peak value < 1%) can be achieved when
the vaccination rate is above 65% (the purple curve in Fig. 8b). This result indicates that social distancing
interventions, such as hybrid teaching modalities, can be cautiously relieved when a critical threshold of
the vaccination rate is reached.
The statistics of the epi curves in Fig. 7 and Fig. 8 are given in the Appendix.
Table 5
Twenty school reopening scenarios based on different teaching modalities and vaccination rates.
Student composition (% of distance-
learning) Vaccination rate
5% 25% 45% 65% 85%
In-class (0%) S0-V5 S0-V25 S0-V45 S0-V65 S0-V85
Mostly in-class (25%) S25-
V5 S25-
V25 S25-
V45 S25-
V65 S25-
V85
Hybrid (50%) S50-
V5 S50-
V25 S50-
V45 S50-
V65 S50-
V85
Mostly online (75%) S75-
V5 S75-
V25 S75-
V45 S75-
V65 S75-
V85
4.2 Public health implications
The proposed CN-ABM model and its application for school reopening scenarios are among the rst to
model the COVID-19 infection at the micro-scale. By running the model under different community risk
levels, teaching modalities, and vaccination rates, we can glean practical implications for school’s
reopening plans in response to the evolving epidemic situations.
First, micro-scale modeling and simulations are of critical signicance for uncovering the pandemic’s
development mechanism and potential health outcomes. Existing COVID-19 modeling work largely
utilizes macro-scale models to simulate variables and trends of disease development, such as daily
Page 11/21
cases of infection (e.g.[10] ). These studies were mostly implemented for a large region (e.g., country,
state) with the smallest analysis unit being an administrative unit (e.g., county, town). Such macro-scale
modeling is not able to account for the uncertainties in the transmission process, such as individual
attributes, activity patterns, activity environments, and interpersonal interactions. The proposed CN-ABM
model, on the contrary, takes a bottom-up approach to construct the transmission mechanism on the
individual level at rened spatiotemporal scales. This attempt is necessary as the implementation of
preventive measures, such as social distancing, is mostly oriented towards individuals [31]. In addition,
the CN-ABM model can generate the transmission chain of an infectious agent, which helps to identify
high-risk activity environments for proactive interventions.
Second, as the CN-ABM model is built on three guiding metrics (i.e., community risk level, teaching
modality, and vaccination rate), it can provide quantiable assessment results to evaluate the
effectiveness and consequences of different school reopening plans. Specically, our three sets of
analyses suggest practical school reopening strategies to minimize potential health adversities in a
general scenario (i.e., low community risk): (1) in scenarios where vaccination is not available, it is
suggested to restrict on-campus students to under 50%, as it can largely atten the epi curve (peak value 
< 2%, the green curve in Fig.5b); and the best result (peak value < 1%) can be achieved by limiting on-
campus students to < 25% (the purple curve in Fig.5b). (2) In scenarios where vaccination is available, it
is suggested to maintain < 75% on-campus students and a vaccination rate of > 45% to suppress the epi
curve (peak value < 2%, the green curve in Fig.8b); and the best result (peak value < 1%) can be achieved
at a vaccination rate of > 65% (the purple curve in Fig.8b).
The study is not without limitations. First, some of the epidemiological parameters used in the model,
such as the infection rate and vaccine effectiveness, were solicited from the literature. In reality, tiers of
uncertainties, such as the room capacity and disinfection levels, can dictate the infection rate. Also, in
many countries, such as the United States, the active infection rate is not published due to the
enforcement of health data regulations, such as the Health Insurance Portability and Accountability Act
[32]. Future research should investigate the variation and temporality of these epidemiological
parameters with clinical evidence in a study area [33]. Second, our model initialization was based on
predened activity patterns by applying a stochastic selection approach. Thus, the simulation results
provide only general criteria for school reopening and may not accommodate the needs of every
educational institution, where the student enrollment, teaching modalities, course schedules, and social
distancing policies vary considerably. Future implementation of the model should incorporate real-world
mobility data (e.g., travel diaries, GPS trajectories) and local social distancing policies to improve the
applicability of the model.
5. Conclusions
While phases of COVID-19 vaccination are rolling out, amid the need to revive the economy, it is essential
to reopen schools safely while minimizing potential health adversities. Although levels of policy guidance
to initiate the school reopening are underway, many of these recommendations lack scenario-based
Page 12/21
evaluations. Under this context, the proposed CN-ABM model can provide scientic evidence to
corroborate the preventive measures and public policies for school reopening. The scenario-based
assessments can help government stakeholders and school administrators understand the joint effects
of the community risk level, teaching modality, and vaccination rate on shaping the epi curve. Findings in
this paper can further contribute to strategic decision-making that weighs the timing of school reopening
against the projected health consequences. It also complements the lack of micro-scale study on COVID-
19 and can be extended to other community settings, such as residential neighborhoods, workplaces, and
shopping malls, and can eventually contribute to developing proactive health intervention strategies in
these settings.
Abbreviations
ABM: agent-based model;
API: application programming interface;
CN-ABM: contact network agent-based model;
COVID-19: coronavirus disease 2019;
SEIR: Susceptible-Exposed-Infective-Recovered;
WHO: World Health Organization
Declarations
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Availability of data and materials
The data and codes for this study can be accessed in GitHub (https://github.com/xic19022/cnabm).
Competing interests
The authors declare that they have no competing interests.
Funding
Page 13/21
This research was supported by the National Natural Science Foundation of China (Grant No. 41971372),
in part by the Natural Science Foundation of Guangdong Province (Grant No. 2020A1515010680).
Authors’ contributions
CL contributed towards conceptualization, methodology, and writing the initial draft. XC and LZ
contributed towards conceptualization and revising the draft. YL and HT contributed towards revising the
draft. All authors read and approved the manuscript.
Acknowledgements
Not applicable.
References
1. Chao DL, Halloran ME, Longini IJ: School opening dates predict pandemic inuenza A(H1N1)
outbreaks in the United States.
J INFECT DIS
2010, 202(6):877-880.
2. Gog JR, Ballesteros S, Viboud C, Simonsen L, Bjornstad ON, Shaman J, Chao DL, Khan F, Grenfell BT:
Spatial Transmission of 2009 Pandemic Inuenza in the US.
PLOS COMPUT BIOL
2014,
10(6):e1003635.
3. Viner RM, Russell SJ, Croker H, Packer J, Ward J, Stanseld C, Mytton O, Bonell C, Booy R: School
closure and management practices during coronavirus outbreaks including COVID-19: a rapid
systematic review.
Lancet Child Adolesc Health
2020, 4(5):397-404.
4. WHO: Coronavirus disease (COVID-19). Situation report - 77. In
.
; 2020. [cited April 6, 2020].
https://www.who.int/docs/default-source/coronaviruse/situation-reports/20200406-sitrep-77-covid-
19.pdf?sfvrsn=21d1e632_2
5. Li X, Xu W, Dozier M, He Y, Kirolos A, Theodoratou E: The role of children in transmission of SARS-
CoV-2: A rapid review.
J GLOB HEALTH
2020, 10(1):11101.
. UNESCO: Educational applications and platforms to mitigate impact of COVID-19 school closures.
In
.
; 2020. [cited March 9, 2020]. https://en.unesco.org/themes/education-emergencies/coronavirus-
school-closures/solutions
7. Brooks SK, Webster RK, Smith LE, Woodland L, Wessely S, Greenberg N, Rubin GJ: The psychological
impact of quarantine and how to reduce it: rapid review of the evidence.
The Lancet
2020,
395(10227):912-920.
. Cai W: Tracking coronavirus cases at US colleges and universities.
The New York Times
2020, 25.
[cited September 3, 2020]. https://www.nytimes.com/interactive/2020/us/covid-college-cases-
tracker.html
9. Mallapaty S: How schools can reopen safely during the pandemic.
NATURE
2020, 584(7822):503-
504.
Page 14/21
10. Chen X, Zhang A, Wang H, Gallaher A, Zhu X: Compliance and containment in social distancing:
mathematical modeling of COVID-19 across townships.
International journal of geographical
information science: IJGIS
2021, 35(3):446-465.
11. Read JM, Bridgen JR, Cummings DA, Ho A, Jewell CP: Novel coronavirus 2019-nCoV: early
estimation of epidemiological parameters and epidemic predictions.
MedRxiv
2020.
12. Wu JT, Leung K, Leung GM: Nowcasting and forecasting the potential domestic and international
spread of the 2019-nCoV outbreak originating in Wuhan, China: a modelling study.
The Lancet
2020,
395(10225):689-697.
13. Zeng T, Zhang Y, Li Z, Liu X, Qiu B: Predictions of 2019-nCoV Transmission Ending via
Comprehensive Methods.
arXiv preprint arXiv:2002.04945
2020.
14. WHO: Considerations for school-related public health measures in the context of COVID-19. In
.
; 2020.
[cited September 14,2020]. https://www.who.int/publications/i/item/considerations-for-school-
related-public-health-measures-in-the-context-of-covid-19
15. Wooldridge M: An introduction to multiagent systems: John wiley & sons; 2009.
1. Chen X, Kwan M, Li Q, Chen J: A model for evacuation risk assessment with consideration of pre- and
post-disaster factors.
Computers, environment and urban systems
2012, 36(3):207-217.
17. Cardinot M, O Riordan C, Grith J, Perc M: Evoplex: A platform for agent-based modeling on
networks.
SoftwareX
2019, 9:199-204.
1. Rockett RJ, Arnott A, Lam C, Sadsad R, Timms V, Gray K, Eden J: Revealing COVID-19 transmission in
Australia by SARS-CoV-2 genome sequencing and agent-based modeling.
NAT MED
2020,
26(9):1398-1404.
19. Chang SL, Harding N, Zachreson C, Cliff OM, Prokopenko M: Modelling transmission and control of
the COVID-19 pandemic in Australia.
NAT COMMUN
2020, 11(1):5710.
20. Gemmetto V, Barrat A, Cattuto C: Mitigation of infectious disease at school: targeted class closure vs
school closure.
BMC INFECT DIS
2014, 14:695.
21. Di Domenico L, Pullano G, Sabbatini CE, Boëlle P, Colizza V: Modelling safe protocols for reopening
schools during the COVID-19 pandemic in France.
NAT COMMUN
2021, 12(1):1073.
22. Huang B, Wang J, Cai J, Yao S, Chan PKS, Tam TH, Hong Y, Ruktanonchai CW, Carioli A, Floyd JR
et
al
: Integrated vaccination and physical distancing interventions to prevent future COVID-19 waves in
Chinese cities.
Nature human behaviour
2021.
23. Kim JH, Marks F, Clemens JD: Looking beyond COVID-19 vaccine phase 3 trials.
NAT MED
2021,
27(2):205-211.
24. McAloon C, Collins A, Hunt K, Barber A, Byrne AW, Butler F, Casey M, Grin J, Lane E, McEvoy D
et al
:
Incubation period of COVID-19: a rapid systematic review and meta-analysis of observational
research.
BMJ OPEN
2020, 10(8):e39652.
25. Kim SE, Jeong HS, Yu Y, Shin SU, Kim S, Oh TH, Kim UJ, Kang SJ, Jang HC, Jung SI
et al
: Viral
kinetics of SARS-CoV-2 in asymptomatic carriers and presymptomatic patients.
INT J INFECT DIS
Page 15/21
2020, 95:441-443.
2. Wolfel R, Corman VM, Guggemos W, Seilmaier M, Zange S, Muller MA, Niemeyer D, Jones TC, Vollmar
P, Rothe C
et al
: Virological assessment of hospitalized patients with COVID-2019 (vol 581, pg 465,
2020).
NATURE
2020, 588(7839):E35.
27. Chen Y, Wang A, Yi B, Ding K, Wang H, Wang J, Shi H, Wang S, Xu G: The epidemiological
characteristics of infection in close contacts of COVID-19 in Ningbo city.
Chin J Epidemiol
2020,
41(5):668-672.
2. Yang Y, Atkinson PM: Individual Space-Time Activity-Based Model: A Model for the Simulation of
Airborne Infectious-Disease Transmission by Activity-Bundle Simulation.
Environment and Planning
B: Planning and Design
2008, 35(1):80-99.
29. WHO: WHO lists additional COVID-19 vaccine for emergency use and issues interim policy
recommendations. In
.
; 2021. [cited May 7, 2021]. https://www.who.int/news/item/07-05-2021-who-
lists-additional-covid-19-vaccine-for-emergency-use-and-issues-interim-policy-recommendations
30. Kelvin AA, Halperin S: COVID-19 in children: the link in the transmission chain.
The Lancet infectious
diseases
2020, 20(6):633-634.
31. Gostin LO, Wiley LF: Governmental Public Health Powers During the COVID-19 Pandemic: Stay-at-
home Orders, Business Closures, and Travel Restrictions.
JAMA : the journal of the American Medical
Association
2020, 323(21):2137-2138.
32. Edemekong P, Annamaraju P, Haydel M: Health Insurance Portability and Accountability Act.
StatPearls
2020.
33. Corum J, Grady D, Wee S, Zimmer C: Coronavirus vaccine tracker.
The New York Times
2020, 5. [cited
February 5, 2021]. https://www.nytimes.com/interactive/2020/science/coronavirus-vaccine-
tracker.html
Figures
Page 16/21
Figure 1
Schematic illustration of the contact network. Red nodes are infectious agents; lines are interactions
between agents via their close contact at a time point. An external node is introduced to represent the
community risk.
Page 17/21
Figure 2
Schematic illustration of an infection cycle with the change of an agent’s health status (E = exposed, Ip =
pre-symptomatic, I = infected, R = recovered, and V = vaccinated).
Page 18/21
Figure 3
Workow of the CN-ABM model for simulating COVID-19 transmission in a school environment.
Figure 4
Building footprints of the campus in the study area. *The library is considered a lecture hall in the model
simulation; other buildings (e.g., administrative buildings) are excluded from the simulation.
Page 19/21
Figure 5
Simulation results under different school reopening scenarios at different community risks: (a) no risk, (b)
low risk, and (c) high risk. Curves in different colors represent different teaching modalities; the vertical
axis (Y) is the simulated active infectious cases as a percentage of the total agents (%); the horizontal
axis (X) is the day of simulation.
Figure 6
The transmission chain of an agent in the rst 35 days of the disease progression under the N-S0
scenario.
Figure 7
Simulation results under different school reopening scenarios at different community risks: (a) no risk, (b)
low risk, and (c) high risk. Curves in different colors represent different vaccination rates; the vertical axis
(Y) is the simulated active infectious cases as a percentage of the total agents (%); the horizontal axis (X)
is the day of simulation.
Page 20/21
Figure 8
Simulation results under different school reopening scenarios with different teaching modalities: (a) in-
class, (b) mostly in-class, (c) hybrid, and (d) mostly online. Curves in different colors represent different
vaccination rates; the vertical axis (Y) is the simulated active infectious cases as a percentage of the total
agents (%); the horizontal axis (X) is the day of simulation.
Supplementary Files
Page 21/21
This is a list of supplementary les associated with this preprint. Click to download.
SupplementaryMaterialv2.docx
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
The coronavirus disease 2019 (COVID-19) pandemic has posed substantial challenges to the formulation of preventive interventions, particularly since the effects of physical distancing measures and upcoming vaccines on reducing susceptible social contacts and eventually halting transmission remain unclear. Here, using anonymized mobile geolocation data in China, we devise a mobility-associated social contact index to quantify the impact of both physical distancing and vaccination measures in a unified way. Building on this index, our epidemiological model reveals that vaccination combined with physical distancing can contain resurgences without relying on stay-at-home restrictions, whereas a gradual vaccination process alone cannot achieve this. Further, for cities with medium population density, vaccination can reduce the duration of physical distancing by 36% to 78%, whereas for cities with high population density, infection numbers can be well-controlled through moderate physical distancing. These findings improve our understanding of the joint effects of vaccination and physical distancing with respect to a city’s population density and social contact patterns.
Article
Full-text available
As countries in Europe implement strategies to control the COVID-19 pandemic, different options are chosen regarding schools. Through a stochastic age-structured transmission model calibrated to the observed epidemic in Île-de-France in the first wave, we explored scenarios of partial, progressive, or full school reopening. Given the uncertainty on children’s role, we found that reopening schools after lockdown may increase COVID-19 cases, yet protocols exist to keep the epidemic controlled. Under a scenario with stable epidemic activity if schools were closed, reopening pre-schools and primary schools would lead to up to 76% [67, 84]% occupation of ICU beds if no other school level reopened, or if middle and high schools reopened later. Immediately reopening all school levels may overwhelm the ICU system. Priority should be given to pre- and primary schools allowing younger children to resume learning and development, whereas full attendance in middle and high schools is not recommended for stable or increasing epidemic activity. Large-scale test and trace is required to keep the epidemic under control. Ex-post assessment shows that progressive reopening of schools, limited attendance, and strong adoption of preventive measures contributed to a decreasing epidemic after lifting the first lockdown. The role of children in the spread of COVID-19 is not fully understood, and the circumstances under which schools should be opened are therefore debated. Here, the authors demonstrate protocols by which schools in France can be safely opened without overwhelming the healthcare system.
Article
Full-text available
In the early development of COVID-19, large-scale preventive measures, such as border control and air travel restrictions, were implemented to slow international and domestic transmissions. When these measures were in full effect, new cases of infection would be primarily induced by community spread, such as the human interaction within and between neighboring cities and towns, which is generally known as the meso-scale. Existing studies of COVID-19 using mathematical models are unable to accommodate the need for meso-scale modeling, because of the unavailability of COVID-19 data at this scale and the different timings of local intervention policies. In this respect, we propose a meso-scale mathematical model of COVID-19, named the meso-scale Susceptible, Exposed, Infectious, Recovered (MSEIR) model, using town-level infection data in the state of Connecticut. We consider the spatial interaction in terms of the inter-town travel in the model. Based on the developed model, we evaluated how different strengths of social distancing policy enforcement may impact epi curves based on two evaluative metrics: compliance and containment. The developed model and the simulation results help to establish the foundation for community-level assessment and better preparedness for COVID-19.
Article
Full-text available
After the recent announcement of COVID-19 vaccine efficacy in clinical trials by several manufacturers for protection against severe disease, a comprehensive post-efficacy strategy for the next steps to ensure vaccination of the global population is now required. These considerations should include how to manufacture billions of doses of high-quality vaccines, support for vaccine purchase, coordination of supply, the equitable distribution of vaccines and the logistics of global vaccine delivery, all of which are a prelude to a massive vaccination campaign targeting people of all ages. Furthermore, additional scientific questions about the vaccines remain that should be answered to improve vaccine efficacy, including questions regarding the optimization of vaccination regimens, booster doses, the correlates of protection, vaccine effectiveness, safety and enhanced surveillance. The timely and coordinated execution of these post-efficacy tasks will bring the pandemic to an effective, and efficient, close. A comprehensive strategy for the next steps to ensure vaccination of the global population against SARS-CoV-2 is now required, and key steps and challenges are detailed in this Perspective.
Article
Full-text available
There is a continuing debate on relative benefits of various mitigation and suppression strategies aimed to control the spread of COVID-19. Here we report the results of agent-based modelling using a fine-grained computational simulation of the ongoing COVID-19 pandemic in Australia. This model is calibrated to match key characteristics of COVID-19 transmission. An important calibration outcome is the age-dependent fraction of symptomatic cases, with this fraction for children found to be one-fifth of such fraction for adults. We apply the model to compare several intervention strategies, including restrictions on international air travel, case isolation, home quarantine, social distancing with varying levels of compliance, and school closures. School closures are not found to bring decisive benefits unless coupled with high level of social distancing compliance. We report several trade-offs, and an important transition across the levels of social distancing compliance, in the range between 70% and 80% levels, with compliance at the 90% level found to control the disease within 13–14 weeks, when coupled with effective case isolation and international travel restrictions.
Article
Full-text available
Objectives The aim of this study was to conduct a rapid systematic review and meta-analysis of estimates of the incubation period of COVID-19. Design Rapid systematic review and meta-analysis of observational research. Setting International studies on incubation period of COVID-19. Participants Searches were carried out in PubMed, Google Scholar, Embase, Cochrane Library as well as the preprint servers MedRxiv and BioRxiv. Studies were selected for meta-analysis if they reported either the parameters and CIs of the distributions fit to the data, or sufficient information to facilitate calculation of those values. After initial eligibility screening, 24 studies were selected for initial review, nine of these were shortlisted for meta-analysis. Final estimates are from meta-analysis of eight studies. Primary outcome measures Parameters of a lognormal distribution of incubation periods. Results The incubation period distribution may be modelled with a lognormal distribution with pooled mu and sigma parameters (95% CIs) of 1.63 (95% CI 1.51 to 1.75) and 0.50 (95% CI 0.46 to 0.55), respectively. The corresponding mean (95% CIs) was 5.8 (95% CI 5.0 to 6.7) days. It should be noted that uncertainty increases towards the tail of the distribution: the pooled parameter estimates (95% CIs) resulted in a median incubation period of 5.1 (95% CI 4.5 to 5.8) days, whereas the 95th percentile was 11.7 (95% CI 9.7 to 14.2) days. Conclusions The choice of which parameter values are adopted will depend on how the information is used, the associated risks and the perceived consequences of decisions to be taken. These recommendations will need to be revisited once further relevant information becomes available. Accordingly, we present an R Shiny app that facilitates updating these estimates as new data become available.
Article
Full-text available
In January 2020, a novel betacoronavirus (family Coronaviridae), named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), was identified as the etiological agent of a cluster of pneumonia cases occurring in Wuhan City, Hubei Province, China1,2. The disease arising from SARS-CoV-2 infection, coronavirus disease 2019 (COVID-19), subsequently spread rapidly causing a worldwide pandemic. Here we examine the added value of near real-time genome sequencing of SARS-CoV-2 in a subpopulation of infected patients during the first 10 weeks of COVID-19 containment in Australia and compare findings from genomic surveillance with predictions of a computational agent-based model (ABM). Using the Australian census data, the ABM generates over 24 million software agents representing the population of Australia, each with demographic attributes of an anonymous individual. It then simulates transmission of the disease over time, spreading from specific infection sources, using contact rates of individuals within different social contexts. We report that the prospective sequencing of SARS-CoV-2 clarified the probable source of infection in cases where epidemiological links could not be determined, significantly decreased the proportion of COVID-19 cases with contentious links, documented genomically similar cases associated with concurrent transmission in several institutions and identified previously unsuspected links. Only a quarter of sequenced cases appeared to be locally acquired and were concordant with predictions from the ABM. These high-resolution genomic data are crucial to track cases with locally acquired COVID-19 and for timely recognition of independent importations once border restrictions are lifted and trade and travel resume.
Article
Full-text available
Background: Understanding the role of children in the transmission of SARS-CoV-2 is urgently required given its policy implications in relation to the reopening of schools and intergenerational contacts. Methods: We conducted a rapid review of studies that investigated the role of children in the transmission of SARS-CoV-2. We synthesized evidence for four categories: 1) studies reporting documented cases of SARS-CoV-2 transmission by infected children; 2) studies presenting indirect evidence on the potential of SARS-CoV-2 transmission by (both symptomatic and asymptomatic) children; 3) studies reporting cluster outbreaks of COVID-19 in schools; 4) studies estimating the proportions of children infected by SARS-CoV-2, and reported results narratively. Results: A total of 16 unique studies were included for narrative synthesis. There is limited evidence detailing transmission of SARS-CoV-2 from infected children. We found two studies that reported a 3-month-old whose parents developed symptomatic COVID-19 seven days after caring for the infant and two children who may have contracted COVID-19 from the initial cases at a school in New South Wales. In addition, we identified six studies presenting indirect evidence on the potential for SARS-CoV-2 transmission by children, three of which found prolonged virus shedding in stools. There is little data on the transmission of SARS-CoV-2 in schools. We identified only two studies reporting outbreaks of COVID-19 in school settings and one case report of a child attending classes but not infecting any other pupils or staff. Lastly, we identified six studies estimating the proportion of children infected; data from population-based studies in Iceland, Italy, South Korea, Netherlands, California and a hospital-based study in the UK suggest children may be less likely to be infected. Conclusions: Preliminary results from population-based and school-based studies suggest that children may be less frequently infected or infect others, however current evidence is limited. Prolonged faecal shedding observed in studies highlights the potentially increased risk of faeco-oral transmission in children. Further seroprevalence studies (powered adequately for the paediatric population) are urgently required to establish whether children are in fact less likely to be infected compared to adults. Note: We plan to update this rapid review as new data becomes available. These updates are available at https://www.ed.ac.uk/usher/uncover/completed-uncover-reviews.
Article
Masks, class sizes and hygiene are important, but low community spread is key. Masks, class sizes and hygiene are important, but low community spread is key.
Article
Objective: To estimate the infection rate of close contacts of COVID-19 cases, and to evaluate the risk of COVID-19 under different exposure conditions. Methods: A prospective study was used to conduct continuous quarantine medical observations of close contacts of people infected with COVID-19, collect epidemiological, clinical manifestations, and laboratory test data to estimate the infection rate of close contacts under different exposures. Results: The epidemiological curve of COVID-19 in Ningbo showed persistent human-to-human characteristics. A total of 2 147 close contacts were tracked and investigated. The total infection rate was 6.15%. The infection rates of confirmed cases and positive contacts were 6.30% and 4.11%, respectively. The difference was not statistically significant (P>0.05). Among close contacts of different relationships, friends/pilgrims (22.31%), family members (18.01%), and relatives (4.73%) have a higher infection rate, and close contacts of medical staff were not infected. Differences in infection rates among the close contacts were statistically significant (P<0.005). Living with the case (13.26%), taking the same transportation (11.91%), and dining together (7.18%) are high risk factors for infection. Cross-infection in the hospital should not be ignored (1.94%). The median of incubation period is 5 days. Conclusion: The infection rate of close contacts of COVID-19 cases is high, and isolation medical observation measures should be implemented in strict accordance with the close contact management plan.