Two Centers for Disease Control and Prevention (CDC) studies published today in Morbidity and Mortality Weekly Report track COVID-19 spread on US college campuses, one showing that 101 counties with large universities offering in-person instruction saw a 56% jump in coronavirus cases after classes began, the other finding that 91% of gatherings at an Arkansas university were tied to fraternity and sorority activities.
Remote learning linked to drop in cases
The first study showed that counties with large, nonprofit colleges and universities conducting only remote learning experienced a 17.9% drop, on average, in new COVID-19 cases in the 21 days before and after the start of fall classes. In contrast, counties with institutions that featured in-person instruction observed a 56.2% increase, while those without large colleges or universities witnessed a 5.9% decline in new cases.
Percentages of positive COVID-19 test results and hot spot status followed the same pattern of rises in counties with colleges conducting in-person classes and declines in those with remote-only teaching.
Among all 101 university counties (3.2% of all US counties, accounting for 29.4 of the country's population), 22 (22%) offered remote-only instruction, and 79 (78%) offered in-person classes. University counties were already conducting more COVID-19 tests than nonuniversity counties before classes began.
Mean testing rates increased 4.2% during the first 21 days of classes among remote-learning instruction counties, versus 14.1% in in-person counties. Testing rates decreased 1.0% in counties without universities.
In that same timeframe, mean test positivity fell among remote-instruction university counties (absolute change, –1.8%) and nonuniversity counties (–0.6%) but rose among in-person learning university counties by 1.1%.
The percentage of counties identified at least once as a hot spot for COVID-19 activity climbed among all three groups, with in-person–learning counties having the highest absolute increase (30.4%), followed by remote-learning counties (9.1%) and nonuniversity counties (1.5%).
Compared with nonuniversity counties, in-person–learning counties saw a higher relative change in coronavirus testing rates (18.8% vs –5.6%), higher absolute change in test positivity (1.6% vs –0.8%), higher relative change in incidence (78.3% vs –19.5%), and higher absolute change in percentage identified as hot spots (33.8% vs 1.5%).
The authors concluded that colleges and universities should work with local authorities and public health officials to shore up community mitigation strategies and slow the spread of coronavirus on campuses.
"Some university counties might have one or more concerning factors, such as higher levels of older adult populations, high rates of obesity and cardiovascular disease, or strained health care resources," the researchers wrote. "These counties might need to consider the implications of in-person instruction on spread of COVID-19 among a student population that might have interactions with persons at higher risk in the community."
Greek life gatherings implicated in college outbreak
The second study identified a rapid rise in COVID-19 cases at one Arkansas university, with 956 cases with symptom onsets of Aug 20 to Sep 5.
Of the 54 student gatherings just before the outbreak, 49 (91%) were fraternity and sorority recruitment events, which were held for 1 week each in late August. Those activities accounted for 42 (72%) of links among campus gatherings.
The school year began on Aug 24 with both in-class and remote instruction for its roughly 20,000 students; most opted for remote learning. On Sep 4, the university banned gatherings of 10 or more people, and fraternity bid day, when selections for membership are announced, was held online.
Of the 956 students infected with COVID-19, 48 (5%) had in-person classes, 292 (31%) participated in fraternity or sorority gatherings, and 149 (15%) lived in a fraternity or sorority house.
An analysis traced 565 (59%) of cases to 16 dorms, 20 apartments and houses, 20 fraternity or sorority houses, and their case-to-event interactions. Researchers identified one large associated case cluster of 471 (83%) (86% in women, 14% in men) and eight smaller, unlinked gatherings involving 94 cases (49% in women, 51% in men), with cluster size ranging from 4 to 12 cases.
Among attendees at 44 (81%) fraternity and sorority gatherings, at least one person regularly attended in-person classes, and at least one at each of the 49 Greek gatherings (91%) reported participating in these events. On average, gatherings produced 20.3 cases each (median, 21; range, 5 to 44). Among 58 links between gatherings, 42 (72%) were associated with fraternity or sorority activities, 11 (19%) with on-campus dormitories, and five (9%) with off-campus apartments and houses.
While women made up 54% of students at the university, they accounted for 70% of coronavirus cases and 86% among linked gatherings, which the researchers attribute at least partially to gender-specific activities such as in-person sorority rush week activities and bid day.
The authors said that because most infected students said they attended classes only virtually, COVID-19 spread likely occurred largely outside of the classroom—a finding backed by the very small proportion of infected faculty and staff (0.5%).
"The rapid increase in COVID-19 cases was likely facilitated by on- and off-campus congregate living settings and activities, and health departments should work together with student organizations and university leadership to ensure compliance with mitigation measures," the authors said.
