More Infections From an Online Semester?

Many universities have released statements about their intent to reopen. And every university leader ideally would like to invite students back to campus, since that's what students say they want (and will pay for).

Cornell University joined the chorus of reopening statements on Tuesday in announcing that its Ithaca, N.Y., campus will be open for in-person instruction in the fall.

But for Cornell, one additional piece of information was "very important" in making that decision, according to Martha Pollack, the university's president. That was the finding from Cornell researchers that holding the semester online potentially could result in more infections and more hospitalizations among students and staff members than holding the semester in person would.

A study by Cornell researchers concluded that with nominal parameters, an in-person semester would result in 3.6 percent of the campus population (1,254 people) becoming infected, and 0.047 percent (16 people) requiring hospitalization. An online semester, they concluded, would result in about 7,200 infections and more than 60 hospitalizations.

The conclusion rested on a few different assumptions. First, the study assumed about 9,000 Cornell students would return to Ithaca -- even if there is no in-person learning or physical campus life.

Researchers concluded that during an in-person semester, asymptomatic testing is crucial for containing an outbreak and keeping the total number of infections low. When students live and take classes on campus, the university can enforce such a testing program with a variety of methods. For example, students who don't get tested can lose access to residence halls or be locked out of their email accounts, said Peter Frazier, a data scientist and professor in Cornell's School of Operations Research and Information Engineering, who led the study.

But when instruction is online, the university loses much of that ability to encourage and enforce testing.

"If we have a residential, on-campus semester, then we have the authority to put all kinds of expectations and requirements on our students," Pollack said. "If we were only in an online basis, then it would be really difficult to impose regulations on students who happen to be living in Ithaca, as opposed to, say, happen to be living in Atlanta or San Francisco."

Frazier said the university still could choose to ask students where they are living and attempt to enforce asymptomatic testing for those living in Ithaca. But students could misrepresent where they are residing, and the spotty enforcement could result in outbreaks. The model assumes students in Ithaca are entirely outside the university's testing purview.

The assumption that 9,000 students will return to campus is based on student surveys and conversations with area landlords about their fall tenants.

In a recent survey of 10,365 Cornell undergraduates, 31 percent of respondents said they were "very likely" to return to Ithaca if instruction is online. Twenty-two percent said they were "somewhat likely" to return to the area for the semester. (Also noteworthy, only 32 percent of students said they were "very likely" to enroll at Cornell in the fall if instruction is entirely online. Twenty-three percent said they were "somewhat likely" to enroll.)

On social media, some students and instructors voiced concerns about using the survey data to come to the conclusion that 9,000 students will be in the Ithaca area. The survey was completed weeks ago, before the COVID spikes and travel bans that are now evident in a number of states. (New York now has quarantine requirements for anyone arriving from 16 different states, including Texas and California.) Students may not have consulted with their families before signaling their intent to return in the survey.

Additionally, 53 percent of Cornell's undergraduate population (the share likely return to campus) comes out to under 8,000, not 9,000 students. (The 9,000 number does include some graduate students.)

The ‘Break-Even’ Point

Frazier emphasized that interpreting the survey results must be done with care. The survey suggests many students will come to Ithaca, he said, though it's possible that prediction will not bear out.

But the "break-even" point for the data -- when on-campus instruction and online instruction result in the same number of infections -- occurs only when the number of students coming to Ithaca in an online scenario gets down to 2,000.

"The risk associated with the virtual instruction seems to be a lot higher because, even though it might result in fewer infections, it might result in way more infections because we would have so little control," Frazier said.

The uncertainty in how many students will choose to come to the area creates the risk of high infections. And that risk breeds danger.

Frazier said the applicability of the study to other colleges and campuses is not entirely clear. A college's setting and the propensity of students to return to a campus town even when there's no in-person instruction are both things to consider, he said. Cornell's New York City campus -- Cornell Tech -- will be doing online instruction.

"I would urge a university to at least survey their students," Frazier said.

More applicable to other institutions is the importance of asymptomatic testing, he said.

"It's really a fantastic tool that we have," he said. "If you have this ability, avail yourself of it."

‘Dealing With Misbehavior’

As for the actual path Cornell plans to take, the university follows a well-worn path in planning to end on-campus instruction by Thanksgiving, but it has a few other distinctive elements.

Students and potentially their parents will be asked to sign a behavioral expectation form, with potential penalties for noncompliance.

"We are hoping to have a series of escalations for dealing with misbehavior," Pollack said. "Look, people are going to make mistakes. Someone's going to forget their mask and we're going to tell them to put their mask back on, but we will be escalating if the misbehavior gets too serious."

The university is also looking to work with student leaders on bystander intervention training to prevent potentially risky contact.

Students also will be required to submit to testing and to report any symptoms daily.

Faculty will deliver instruction both in person and online, to accommodate students who can't or don't feel comfortable coming to campus or who are in quarantine.

Eligible international students will have the option to take part in residence programs in their home or nearby countries through a program called StudyAway.

"These international students will live and study at a local campus in their country or region while taking a mix of online and in-person classes. They will share co-curricular activities with their Cornell peers and have access to local facilities and services," Pollack wrote in a message to students.

Sites include China, Colombia, France, Ghana, Hong Kong, India, Italy, Korea and Vietnam.

For their testing regime, the university will be relying on pooled tests of the general population for surveillance and individual retesting of pools with positive results.

"Pooled testing can decrease the number of laboratory tests required by 10-fold or more. Absent this critical and longstanding method of surveillance testing, Cornell couldn’t test our 24,000 students at a high enough frequency," Pollack and Michael Kotlikoff, the university's provost, wrote in a Wall Street Journal op-ed.

Ultimately, Pollack said, they are relying on math and models.

"Even with all the limitations and uncertainties in any sort of modeling, we still think it makes sense to rely on the science."