Despite dire warnings this summer from public health experts, over a third of U.S. colleges and universities went full steam ahead with reopening, saying they had no choice due to financial or political pressures. The results, in some instances, have been catastrophic.

From August 26 to September 10, 2020, there were at least 62,000 new positive test results at U.S. colleges and universities. A recent preprint study reports that colleges that reopened for in-person instruction this fall probably contributed more than 3,000 cases daily to their counties. About half of the counties with colleges around the country reported their worst week for cases in August. Given these mishaps, is there any way that colleges can successfully reopen in-person instruction during this pandemic?

Protecting students from getting infected needs to remain a priority. Its true that COVID-19 rarely kills young adults, but they can get sick and around 10% of infected people at any age can develop a long-term illness. Infected students can also infect older, vulnerable adults, including instructors and university maintenance and service staff. A recent CDC study showed that those between 20-29 years of age accounted for 20% of the new cases from June to August and in regions where infections among youth were seen, spikes in cases among seniors appeared about 9 days after spikes in the young. So, the young are contributing to community transmission and campus outbreaks can drive infection rates in the communities surrounding a university.

Campuses, like nursing homes and jails, are congregate settings and it is really hard to avert outbreaks under these kinds of living situations. The best-laid plans for reopening can still go awry. Even universities that seemed to have robust plans (like the University of Colorado, Boulder, which called itself a COVID-19 ready campus before reopening) have had outbreaks. The truth is you can make a situation safer or less safe but there will be an unclear demarcating line between the two, its position driven by both human behavior and the arc of the pandemic.

We need to quickly learn lessons from what went wrong with the first attempt at reopening. We should also learn from the colleges that, in the words of Erica Pandey, a business reporter at Axios, are getting reopening rightMiddlebury College in Vermont, for example, had tested 6,735 students and staff by September 28 and has had only two infections.

We wish that universities had taken the opportunity of reopening to formally conduct large-scale, forward-looking research that could guide our knowledge of safer reopening. In the absence of such research, however, we can still make some reasonable assertions based on case studies and on scientific modeling of different scenarios. Some common strategies are emerging from these experiences that may help colleges navigate the next semester more successfully.

Unless youre a tiny campus remote from town and you ban students from leaving, as Amherst did to create a bubble, you just cant seal off your campus from its surroundingsso the first step in improving safety is waiting until levels of community transmission have been driven down before reopening. While this is no guarantee of successBoston College had an outbreak in the week of September 7-13 (73 new cases) even though Massachusetts was doing relatively well in controlling the virusit is a lot easier to keep campus rates low when theres little virus in the surrounding community. Particularly if you are in an area of high transmission, it is completely reasonable to plan for an online-only spring semester, especially given that opening and then shutting down within a few weeks (as the University of North Carolina and North Carolina State University did) is hugely disruptive to education.

Entry testing and then high frequency surveillance of all students (and potentially staff) is of utmost importance. One modeling study found that, assuming typical student behavior, testing all students every two to three days with a rapid, cheap, high specificity test (a test thats very good at identifying people with the virus), is the best strategy to avert outbreaks. The CDC, whose leadership has repeatedly relented to pressure from the Trump Administration, unwittingly gave universities license to take risks when they came out in June against entry testing, a stance that was not based on evidence. An August survey led by the California Institute of Technology found that only one third of U.S. colleges that responded had done entry testing at the time, and only 20% planned on doing ongoing surveillance.

All colleges and universities need robust quarantining and isolation facilities. Even if a campus is in a county or state with little community transmission, arriving students come from all over the country and overseas, so at the start of semester a 14-day quarantine on arrival can be valuable. During the semester, any student who tests positive must be isolated, and those exposed to an infected person need to be quarantined. Universities must not send infected or exposed students home across the country, as this risks seeding further outbreaks nationwide. Such quarantining is, of course, much harder for colleges with large number of commuting students.

Students will socialize no matter what, so if your strategy to prevent outbreaks is just to urge students not to party or ask them to sign a personal responsibility compact, then you effectively have no strategy. Instead, universities need to offer safer alternatives to unmasked, indoor fraternity parties, such as outdoor and masked silent discos, movies, or yoga. As Julia Marcus, an infectious disease epidemiologist at Harvard says, What might actually stand a chance of working is giving students an opportunity to stay socially connected and have fun that are lower risknot necessarily zero riskbut lower risk than crowded indoor house parties. Universities have quickly identified the tension between taking a punitive approach to students breaking social distancing rules and then depending on those same students to truthfully share information for contact tracing purposes. It also remains too early to know if on-campus contact tracing is successful strategy in general, given concerns with privacy.

Universities must adopt the full range of so-called non-pharmaceutical measures to prevent the spread of the virus, including reducing campus density and enforcing universal masking, environmental cleansing, and distancing in common spaces such as dining halls. We now know the novel coronavirus can spread via aerosol transmission, which means that being inside around others in a poorly ventilated classroom is risky. This is why many universities are still leveraging digital learning, and pursuing a hybrid model even if they brought students on campus. If face-to-face teaching is absolutely essential, the risk of transmission can be reduced by adopting small class sizes, distancing, short class times, and fastidious attention to ventilation and air filtration.

If colleges and universities are putting lives on the line by reopening, they need to be fully transparent and provide a daily dashboard showing the number of tests, infections, and hospitalizations among staff and students both on and off campus. This type of transparency may help buy some community trust, and it also helps everyone learn which campuses are doing well in outbreak prevention. A team of independent public health experts recently started a website that rates university dashboards for these reasons. Seventeen out of the 175 universities rated so far received an A, the highest rating, based on criteria such as whether they state how soon test results come back and whether they report on city/county data (which acknowledges the potential effects on and from surrounding communities).

The challenge is that it looks as if you need to institute all of these strategies together to maximize chances of success. Focusing on just one or a few, rather than adopting a comprehensive integrated approach, has led to outbreaks. University of Colorado, Boulder, for example, had one of the best protocols in the country for ensuring classrooms were ventilated, and the University of Notre Dame conducted high frequency COVID-19 testingbut their plans were upended when off-campus parties became super-spreading events. Several large universities (such as Harvard, Boston and Tufts universities) in Massachusetts have adopted a fairly similar collection of interventions mentioned above including quarantine, arrival and surveillance testing and dashboard transparency with differing levels of in person learning. For now, these universities and others pursuing more comprehensive measures in areas of low prevalence appear to have the best chances of averting outbreaks as we enter the fall.

The interventions weve described take resources and, as with K-12 schools, this pandemic is going to worsen inequality in learning at the college level, as private universities can leverage more resources and wealthier students have a much larger range of options. Public institutions of higher learning need government support, not just to stay afloat but to ensure that they can offer high-quality digital education to students of all income levels. We are already seeing economically disadvantaged minority and rural students dropping out of college at higher rates. Private universities should consider partnering with public universities in their cities during these difficult times to share resources such as testing (large university labs at private universities could run some extra batches of tests at little cost). And we urgently need more information on best practices and logistic hurdles in digital learning so schools that cannot open can at least adopt these practices. Data show that entering into this pandemic, most instructors in the U.S. did not have experience with digital resources.

Universities will also face even bigger challenges ahead because the virus isnt going away any time soonin fact, cases are currently rising in many states. And with the onset of colder weather, which drives greater indoor gathering, and flu and respiratory virus season, the fall and winter are likely to herald nationwide increases in COVID-19 cases. Schools also probably need to start thinking of this as a multi-year challenge. Even in a best-case scenario, vaccine-generated herd immunity in the U.S. wont be reached until after late 2021when a candidate vaccine is hopefully expected to be widely availableso we have at least another pandemic school year ahead.

It will take a serious investment of resources for American universities to reopen safely. But this is an investment worth making, as this pandemic will be with us at least until the end of 2021, and maybe longer.

For your security, we've sent a confirmation email to the address you entered. Click the link to confirm your subscription and begin receiving our newsletters. If you don't get the confirmation within 10 minutes, please check your spam folder.

Contact us at letters@time.com.

More:
This Falls College Reopenings Were a MessHere's How to Avoid COVID-19 Outbreaks Next Semester - TIME