Modeling COVID-19 spread in small colleges - PubMed Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Multicenter Study
. 2021 Aug 18;16(8):e0255654.
doi: 10.1371/journal.pone.0255654. eCollection 2021.

Modeling COVID-19 spread in small colleges

Riti Bahl  1 , Nicole Eikmeier  2 , Alexandra Fraser  3 , Matthew Junge  4 , Felicia Keesing  5 , Kukai Nakahata  4 , Lily Reeves  6
Affiliations
Multicenter Study

Modeling COVID-19 spread in small colleges

Riti Bahl et al. PLoS One. .

Abstract

We develop an agent-based model on a network meant to capture features unique to COVID-19 spread through a small residential college. We find that a safe reopening requires strong policy from administrators combined with cautious behavior from students. Strong policy includes weekly screening tests with quick turnaround and halving the campus population. Cautious behavior from students means wearing facemasks, socializing less, and showing up for COVID-19 testing. We also find that comprehensive testing and facemasks are the most effective single interventions, building closures can lead to infection spikes in other areas depending on student behavior, and faster return of test results significantly reduces total infections.

PubMed Disclaimer

Conflict of interest statement

The authors have declared that no competing interests exist.

Figures

Fig 1
Fig 1
(A) The base model with single interventions applied. Note that the reduction in infections from “fewer students” is smaller than it appears since there are 50% fewer people on campus in that intervention. (B) The impact of testing latency on a campus with 25% fewer students and testing and quarantine in effect.
Fig 2
Fig 2. Total infections by room type in the base model and with the gym, library, and dining hall closed.
In an “austere closure”, students spend any extra free time alone. In a “social closure”, students spend half of their free time socializing.
Fig 3
Fig 3. The total infection counts colored by size for different policy and adherence intensities.
Fig 4
Fig 4. Schematic of the network.
Fig 5
Fig 5. Exposure profiles for 100 agents are arranged in decreasing order then averaged.
A 95% confidence interval is included around the curve. Panel A shows the exposure profile for off-campus students. The larger panel of Panel B shows the exposure profile for on-campus students with the maximum entry (corresponding to a dorm roommate) removed. The smaller subpanel in Panel B shows the exposure profile when the roommate is included. Panel C shows the ordered average exposure profile for 100 faculty.
Fig 6
Fig 6. Agent states over 100 days in the base model.
Panel A shows a 95% confidence around the mean behavior from 40 trials. Panel B shows the number of active infections over time for each trial.
Fig 7
Fig 7. The total number of cases (numeric) and the coefficient of variation (standard deviation/mean; colorbars) for different policy and adherence intensity levels.
Fig 8
Fig 8. Empirical measurements of R0(s) computed as in (4) with different initial seed sizes s of the on-campus student population infected.
The results from 100 runs are shown for each R0(s).
Fig 9
Fig 9. The average number of days (y-axis) to go from x/2 to at least x infections.
We omit x = 20 since we initially seed 10 agents in the exposed state and there is latency for infections to begin. We omit x > 320 since for such large x-value the doubling time slows significantly from a herd-immunity effect.
Fig 10
Fig 10. A sensitivity analysis of the tuning parameter, p.
We fix the student adherence to be medium, and show the total number of cases for each of the three administrative policies.
Fig 11
Fig 11. A sensitivity analysis of the off-campus multiplier.
We fix the student adherence to be medium, and show the total number of cases for each of the three administrative policies.
Fig 12
Fig 12. A sensitivity analysis of facemask effectiveness.
Displayed are total number of infections after a semester with f=1 (perfect facemask compliance), but no other intervention.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Long D, Graves D, Burton J, Kim C, Watson C, Mullen L, et al. COVID-19 Planning Guide and Self-Assessment for Higher Education. 2020.
    1. Wrighton MS, Lawrence SJ. Reopening Colleges and Universities During the COVID-19 Pandemic. Annals of Internal Medicine. 2020. doi: 10.7326/M20-4752 - DOI - PMC - PubMed
    1. Cheng SY, Wang CJ, Shen ACT, Chang SC. How to safely reopen colleges and universities during COVID-19: experiences from Taiwan. Annals of Internal Medicine. 2020. doi: 10.7326/M20-2927 - DOI - PMC - PubMed
    1. Guskiewicz K, Blouin R. Message from University leaders on important changes for Fall 2020 Roadmap;. Available from: https://www.unc.edu/posts/2020/08/17/shift-to-remote-cm/.
    1. News NDD. Notre Dame enacts two weeks of remote instruction;. Available from: https://news.nd.edu/news/notre-dame-enacts-two-weeks-of-remote-instruction/.

Publication types

Grants and funding

This research was supported by NSF RAPID Grants No. 2028892 and No. 2028880 to NE, FK, and MJ, and NSF Grant No. 1953141 to MJ. The research collaboration was initiated during the 2019 AMS Mathematical Research Community in Stochastic Spatial Systems.