As colleges and universities reopened in the fall of 2020, many emerged as hot spots of Covid-19 transmission. The narrative that has emerged to explain this centers on irresponsible, party-going 20-somethings who refuse to listen to reason.
Some schools have expelled students who held gatherings in their dorms. A few governors have called out these students as drivers of the epidemic. With no national Covid-19 control plan in sight, it may be politically expedient to blame youths for the nation’s woes.
Is controlling Covid-19 on college and university campuses really as simple as ending parties? Or are other forces at play?
Data analysis and simulation modeling point to the latter. Understanding the dynamics of campus coronavirus transmission can help schools be better prepared for the spring of 2021.
Connecting the dots
Let’s start with the summer of 2020, when a second wave of Covid-19 flooded the U.S., with the daily case count exceeding 70,000 in July. The number of total active cases was many folds higher than the reported cases. Based on estimates from the Covid-19 Simulator we developed with other colleagues, Texas and Florida each had more than 100,000 active cases in mid-August, just as college students were preparing to return to school.
Those surges coincided with students leaving their homes and traveling to colleges in other states. This is an important point: Many of the outbreaks seen on college campuses started with cases that were imported from communities with high Covid-19 incidence and prevalence as students left those states and returned to campus.
Using enrollment data from more than 800 colleges and universities and active Covid-19 cases in each state, we estimated the incoming prevalence of Covid-19 cases on those campuses if all undergraduate students had attended college in person for the fall semester. This information helps connect the dots and identify the underlying source of outbreaks.
Source: The COVID-19 Simulator, developed by Massachusetts General Hospital, Harvard, Boston Medical Center and Georgia Tech
Estimated number of Covid-19 cases in each college in the beginning of the fall semester
Take, for example, Brigham Young University in Provo, Utah. When school began there, Utah had one of the lowest Covid-19 case counts in the U.S. Roughly a month into the school year, BYU had more than 1,200 cases of Covid-19, which likely stemmed from cases imported from students’ home states. This reflects the lack of control of community transmission in the states where the students had been living, in addition to the lack of control of transmission on confined college campuses.
Given such examples, Covid-19 appears to spread more quickly in colleges than it does in the community. We estimated the effective reproduction number, known as RE — the number of new infections caused in the real world by each infected person — on selected college campuses given available week-by-week trends.
For an epidemic to sustain itself, the RE needs to be at least 1.0. The goal of interventions like wearing face coverings and social distancing is to reduce the RE below 1.0, which leads to the epidemic extinguishing itself. On most college campuses, RE is much higher than 1.0. We estimated that the RE was 2.8 at Notre Dame University and 1.5 at the University of North Carolina at Chapel Hill within the first few weeks after colleges resumed classes.
These high RE values imply that the way that college students interact does lead to more transmission on campuses than in the community. Some of the high RE likely reflects the reality that college students live in congregate settings and cannot effectively social distance. Some may reflect behavioral choices, like going to parties and forgoing mask use. Whatever the reason for the high RE, it is essential to get it under control. The interactive map below also shows how cases grow exponentially in colleges in four weeks at different RE values.
Source: The COVID-19 Simulator, developed by Massachusetts General Hospital, Harvard, Boston Medical Center and Georgia Tech
Estimated number of Covid-19 cases in each college after 4 weeks
Together, these findings paint a picture of why we are seeing outbreaks on college campuses. Colleges are importing students from around the country, and many arrive from communities with uncontrolled Covid-19 epidemics. When students arrive on campus, they enter congregate living situations in which it is hard to isolate and some are making behavioral choices likely to drive even further transmission.
What actions do we need to take for spring 2021?
To safely open colleges and universities, Covid-19 cases must first be controlled in the communities students come from so infections aren’t being exported to campuses.
A clear first step would be to enforce masking mandates and limit the size of indoor gatherings. There is reason to believe that if we took those steps we could control community spread. If masking and social distancing are not effective, we should be open to the idea of a limited period of lockdown — strict stay-at-home orders. Based on modeling, we believe it can be worthwhile to lock down for a brief period to then enjoy a longer period of Covid-19 control.
Using our Covid-19 Simulator, we estimated that a lockdown of four to six weeks could substantially decrease Covid-19 in most states. Following such an intense period of sacrifice, it would be possible to reopen colleges and universities, and likely other schools and businesses, with firm masking and distancing policies to contain transmission.
Also needed are appropriate actions to stop Covid-19 from spreading across campuses and keeping RE below 1.0. Here too, masking and physical distancing are the foundational principles for reduction. In addition, frequent and routine testing of asymptomatic individuals can help catch otherwise undetected cases. Availability of point-of care tests can make this practically feasible and such tests could be performed conveniently at home or in dorms.
Reducing transmission on campuses also requires more public awareness and education for all stakeholders — students, parents, administrators, and staff. It is disingenuous and unfair to blame Covid-19 transmission on party-going college students when political leaders in many of the states these students come from refuse to implement even the most basic Covid-19 control policies. At the same time, it is unwise to ignore the reality that RE will likely remain high on many campuses.
Putting the focus on penalizing adolescents and emerging adults for engaging in what is typical social behavior in any setting other than this pandemic is not helpful because it fails to address the root cause of the problem. It is important, of course, that college students understand and accept their social responsibilities in this time. But their institutions also have responsibilities, especially thinking creatively about how to support students in making safer choices about socializing on campus.
Jagpreet Chhatwal is a data scientist at Massachusetts General Hospital and an assistant professor of radiology at Harvard Medical School. Trisha Dwivedi is a senior at Carnegie Mellon University studying artificial intelligence. Benjamin Linas is an infectious disease physician at Boston Medical Center. Other collaborators who contributed to this work include Turgay Ayer and Yingying Xiao at Georgia Tech, Ozden Dalgic at Value Analytics Labs, and Sarah Shirley at Harvard University.
