How frequently should students in a college community get tested in order to prevent large COVID-19 outbreaks?

Reopening School Testing and Contact Tracing

A: Every 2 days, according to one recently-published model.

Every 2 days!? Whaaaaah???? With hundreds of thousands of students around the country headed back to college campuses this month, COVID testing is something every college needs to plan for: Who, how, and how often?

The “how often” question was answered last week in a study published by researchers from Yale, Harvard, and Massachusetts General Hospital. Using a hypothetical cohort of 5,000 students that included 10 undetected positive cases, they found that screening every 2 days was needed to control outbreaks. The researchers could find no feasible scenario in which testing only symptomatic cases was sufficient. The model assumed a test with low sensitivity (lots of false negatives) and high specificity (few false positives), which matches the cost and supply constraints most colleges will face.

This news caused some raised eyebrows (and emergency meetings) at the many colleges and universities with no plans and no capacity to test at anywhere near that frequency. A more typical testing plan asks students to test before and/or on arrival to campus. After that, testing is reserved for athletes, anyone developing symptoms, and either campus-wide testing every few weeks, or a small random sample (e.g., 3%) of students more often. Given the new model, however, many university officials, faculty, and parents are wondering, “Is that enough?”

Dr. Elizabeth Bradley, the president of Vassar College and first author of an editorial comment in the same issue of the journal that published the Yale/Harvard/MGH model, says it is. She and her co-authors ran the same model for Vassar (a self-contained campus with a small student body) changing two key assumptions: 1. Fewer new cases would be brought onto campus; and 2. compliance with masking and social distancing guidelines would be high enough to limit transmission.

Their results? Testing every 4 weeks would be sufficient to keep number of new infections low, and would not exceed the capacity of planned isolation facilities.

So, is the right answer every 2 days or every 4 weeks? Unfortunately, this new model doesn’t give one “right” answer. Like all good models, though, it helps us see which factors make the biggest difference in determining the answer for a given situation. In this case, those factors include how many new cases “slip through” onto campus at the start of term; how well transmission is stopped by masking, social distancing, and hygiene; and test characteristics. The model also highlights critical constraints: cost, testing logistics and turnaround time for results; and isolation and contact tracing capacity on campus.

Stay tuned as campuses ramp up for the fall semester and put their testing plans in place. This is a test that everyone wants to pass.

New model
Bradley et al. editorial comment

News stories on testing and the new model:
The Scientist
Inside Higher Ed

Link to original FB post