Policymakers need to be humble about what we don’t know, especially with Covid-19.
The first Covid-19 strategy document I wrote, in April 2020, for Bill de Blasio when he was New York’s mayor had multiple pages devoted to metrics for when to relax or tighten restrictions — such as mask mandates — for offices, restaurants, sporting events and more.
But I’m just as perplexed now as I was almost two years ago about the best metrics to use to monitor the pandemic and how to use them to trigger actions that slow the spread.
It has become clear to me that decisions about restrictions cannot be solely determined by any single metric or combination of them. Whether they are case rates, test positivity, I.C.U. bed utilization or other figures, these metrics must be considered along with many other, qualitative factors. No Covid-related number can speak for itself and work at all times, because critical factors keep changing. These include the virus itself; tools like vaccines, drugs and better tests; our evidence for what works to prevent spread; and public attitudes about pandemic control measures.
For me, the darkest moment of relying too much on metrics was when Mr. de Blasio temporarily closed in-person public school in November 2020 because he had publicly committed that he would do so when the citywide coronavirus test positivity rate rose above 3 percent. This was a metric I endorsed before the school year began but no longer did because we had collected enough data to show that measures to limit Covid-19 transmission in schools were working.
Decision makers continue to face similar challenges. Consider the question of when schools can stop requiring children to wear masks. Ideally, government or school officials would base this decision on how much Covid transmission is occurring inside schools. When there’s very little transmission, masks can come off. When transmission rises, masks go back on.
But this is incredibly difficult to measure. We do not yet have accurate, widely available technology to measure how much virus is in the air. Instead, most places simply report the number of cases that occur in anyone enrolled in school. But these case counts do not really show how much the coronavirus spreads in schools, because many kids and teachers get infected outside the school building. And in places where there is extensive in-school testing, that measurement also strongly reflects the intensity of testing, not where people got infected.
One option is to tie school policies to community transmission rates, but policymakers have learned that schools, like hospitals, can keep transmission low even when community rates are high.
Another option is to monitor the percentage of close contacts of teachers, students and other staff members who test positive for the virus, which can be a metric for how well the school’s practices prevent people with undiagnosed Covid from spreading it. But even if officials and parents make a judgment call on a safe target, masks are hardly the only policy intervention that affects transmission in schools. Health officials need to disentangle the individual benefit of masks from other Covid-19 mitigation measures — including vaccinating all kids and adults, having a robust test-to-stay plan for close contacts and improving ventilation — and decide if these measures are a package that must be provided together or they can be used separately.
Unfortunately, policymakers have to accept several inconvenient facts about masks at the same time. Masks on adults and children reduce coronavirus transmission indoors. But we don’t know precisely how much in different school settings with different populations at different levels of community transmission. During Covid-19 surges, many parents and educators may agree the harms of Covid-19 outweigh concerns about masks. Outside of surges, there is less agreement. Some people do not find it acceptable to wear masks continuously during the day. Some parents worry about potential speech and development delays from masks, even though that remains unproven.
Similar problems exist for other urgent Covid-19 policy questions, like whether employers should ever remove vaccine mandates and whether cities should remove vaccine verification for dining and other indoor activities when a certain percentage of the population is vaccinated.
The truth is that science doesn’t have an answer for what level of Covid-19 transmission is acceptable in schools before and after masks are removed or what level is acceptable in communities before and after vaccine verification. Someone has to decide, and that decision will involve subjective assessments of the risks people will tolerate.
Moreover, metrics that rely on public case numbers — such as the Centers for Disease Control and Prevention’s classification system for community transmission levels — may not be usable in the near future. Many states may eventually stop daily or weekly reporting of coronavirus case counts because so many people use at-home rapid tests, making public case counts inaccurate.
In the absence of reliable data and metrics, our elected officials will need to make considered judgment calls. For example, my ideal situation would be for masks to come off in school when there’s universal vaccination among kids and adults in schools, hospitalizations have been stable or declining for at least four weeks and rapid tests and high-quality masks are abundant and free for anyone who wants one in school. But since we are unlikely to reach these levels soon, the next best option is for governors and mayors to decide what to do, as governors in New Jersey, Virginia and elsewhere have.
Elected leaders should still consider data when making these decisions. They should start by asking public health officials: Is the overall burden of disease rising, and if so, could it overwhelm the health care system? Even if it’s not rising, are there gross inequities in disease burden by place or race that can be addressed, such as specific communities or facilities that require a targeted response?
To answer, health officials will need to use an approach similar to how they measure flu activity. For the flu, public health officials acknowledge they cannot accurately count all community and hospitalized cases, so they monitor samples of the population for flulike symptoms, flu strains circulating and vaccine effectiveness. They make a composite estimate of flu activity and compare it with historical data. We can shift to this approach for Covid-19 and complement it with other promising approaches, such as wastewater monitoring.
In most situations, however, elected officials need to operate in the far grayer area of interventions narrowly tailored to specific settings, like schools, nursing homes and jails. For each setting, officials need to consider whether the people in those places are vaccinated or are medically vulnerable, whether the building has optimized ventilation and whether there are high-quality masks and rapid tests available to all who want them — similar to the expectation that a building has abundant soap, water and paper towels for hand hygiene.
They also need to use their best judgment to estimate the level of public outrage if outbreaks occur in these settings — and balance that against continued dissatisfaction with mitigation measures. And then it’s up to all of us to hold these leaders accountable at the ballot box.
Jay K. Varma (@DrJayVarma) is a physician and epidemiologist focused on prevention and control of emerging infectious diseases, such as Covid-19, Ebola and others. He was the Covid-19 adviser to former Mayor Bill de Blasio and organized New York’s Covid testing, tracing and vaccination campaigns. He is a professor at Weill Cornell Medicine.
The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: letters@nytimes.com.
Follow The New York Times Opinion section on Facebook, Twitter (@NYTopinion) and Instagram.
