Paul L. Caron

Thursday, July 2, 2020

It Is Safer For Cornell And Other Schools To Reopen Than To Go Online

Wall Street Journal op-ed:  Why Cornell Will Reopen in the Fall, by Michael Kotlikoff (Provost, Cornell) & Martha Pollack (President, Cornell):

CornellColleges and universities in the U.S. face a difficult choice in the coming weeks. Some have decided that closing their campuses and offering online classes is the safest option. For others, the safer and more responsible alternative may be to allow students to return while putting in place a comprehensive virus-screening program that minimizes the risk of transmission. Contingent on New York state’s approval, we will be opening Cornell for residential instruction this fall.

Consider two scenarios. University A decides to reopen. For the health and safety of students, faculty and staff, it institutes a screening program to identify asymptomatic students infected with the novel coronavirus and prevent them from spreading it by repeated testing and isolation. The school also monitors symptoms daily, restricts group sizes, modifies classrooms and dorms, secures extensive quarantine capacity, restricts travel, and imposes requirements for masks and social distancing.

University B decides that this is too risky and chooses to play it safe. The school doesn’t reopen for residential instruction this fall and opts instead to teach all courses online. It takes cautious steps to open for selected professional or graduate programs and research efforts, but doesn’t implement the complex process of screening thousands of undergraduates and modifying the learning environment for social distancing.

Surprisingly, epidemiological modeling done by a group led by Cornell Prof. Peter Frazier [COVID-19 Mathematical Modeling for Cornell’s Fall Semester] suggests that despite playing it safe, sometime during the fall University B may well experience markedly worse health outcomes in its community, while University A will have more effectively safeguarded public health.

For many universities, closing the campus to undergraduates is probably not the safest option—notwithstanding concerns that college students may not adhere to public-health guidelines. That’s because at many colleges, students will gather on and around campuses whether classes are held in person or online. ...

As universities like Cornell make difficult decisions about the fall semester, it’s important to consider the risks of not reopening alongside the risks of opening. Epidemiological modeling suggests, perhaps counterintuitively, that if a university is prepared to put in place a comprehensive virus screening program followed up with supportive quarantine and isolation—in addition to other effective public health measures—reopening may be the more responsible option.

COVID-19 Mathematical Modeling for Cornell’s Fall Semester:

Initial modeling results suggest that a combination of contact tracing, asymptomatic surveillance, and low initial prevalence (supported through testing students prior to, and upon, returning to campus) can achieve meaningful control over outbreaks on Cornell’s Ithaca campus in the fall semester if  symptomatic surveillance is sufficiently frequent and if we have sufficient quarantine capacity. This would dovetail with a complementary effort at Cornell to reduce transmissions through housing policy, class organization, and regulations on social gatherings.

Cornell 1

We use our model to predict outcomes for a full return of students, faculty and staff in the fall semester over a 16 week time period, with cases imported from returning students and from Tompkins county, counterbalanced by aggressive asymptomatic surveillance where every member of the campus community is tested every 5 days. The course of the epidemic is random and we directly model that randomness. Accordingly, our model produces a range of potential futures. In the median random potential future, under our nominal set of parameters, 3.6% of the campus population (1254 people) become infected, and 0.047% of the campus population (16 people) require hospitalization. The 90% quantile rises to 4.02% infected and 0.051% requiring hospitalization. Of the 1254 infections in the median outcome, 570 are due to direct outside infections and ensuing additional infections prior to isolation, while 31 (0.09% of the campus population) are infected before arrival to campus but missed in the test-on-return protocol. There are an additional set of people infected before arrival, found through test-onreturn, and isolated in Ithaca (22 people) or at home prior to travel (180 people).

Legal Ed News, Legal Education | Permalink


I find the claim that COVID-19 will hit purely online programs more seriously than on-campus programs hard to swallow. The claim has to be that Cornell's safety measures will insulate its community better than they would be insulated at home. I would love to see the sensitivity analysis underlying its model.

In any event, time will tell.

Posted by: Theodore Seto | Jul 5, 2020 4:56:37 PM