As Omicron Hits, COVID-19 Case Counts Don’t Mean What They Used To

America has begun the gradual process of accepting that COVID-19 is going to be endemic—meaning it will always be present in the population to some degree—due to inherent properties of the virus (animal reservoirs, high transmissibility, long period of infectiousness, symptoms similar to other pathogens), and will remain so for the foreseeable future. However, the U.S. has an impressive suite of tools to deal with this reality. Vaccine eligibility is widening and boosters are available for all adults who want one. Two effective oral antiviral drugs that prevent hospitalizations and death in people newly infected with COVID-19 are about to be authorized. There are also monoclonal antibody treatments for people whose immune systems do not mount a robust response to infection. The CDC signaled this important and realistic shift by acknowledging that herd immunity is not achievable. With this admission from our most important public health agency, policy must shift also.

We have reached a point in the pandemic where policy should no longer be based around the idea that we cannot resume normal life until case numbers are below a particular (arbitrary) level. One reason is that those levels were set before vaccination, and have not been adjusted accordingly, even though a large proportion of cases, in part due to the growing proportion of cases that are breakthrough cases, are now mild. Another reason is that these metrics were set at a time where policy makers were scrambling to set thresholds to open and close social institutions in the absence of robust data. Setting thresholds for activities according to cases no longer makes sense, but U.S. states and counties are still reporting daily case numbers and fluctuations as though policies should revolve around these numbers.

In a recent article about Oregon’s COVID-19 restrictions, currently the strictest in the country, the state epidemiologist said that in order to ease restrictions, “We need to see cases going down and no resurgence of disease.” But hospitalizations have fallen 57 percent in Oregon since their peak in late August-early September 2021 and have been flat for the past three weeks. Oregon has a low daily new hospitalization admissions rate of about three per 100,000 people down from five per 100,000 people in August 2021, meaning the vast majority of cases are not resulting in serious outcomes. It no longer makes sense to assume “cases” equate to “disease” in areas of high vaccination coverage and this misguided equivalency can have negative consequences.

These negative consequences of mitigations is why Marin County, California, did not re-impose masks recently even though their cases had crept back up into the CDC’s recommended masking zone. The county simultaneously had zero Covid-19 hospitalizations. Marin County’s public health officer recently noted that hospitalizations were at a four-month low, and explained that going forward mask mandates would depend largely on hospitalization and vaccination metrics (instead of case counts) because in areas of high vaccination rates, hospitalizations became largely uncoupled from cases during the U.S. delta variant surge.

In light of the properties of the virus, vaccines and emerging therapeutics, the CDC should stop basing guidance on when to unmask on case counts. In November, Montgomery County, Maryland, dropped indoor mask mandates based on the CDC’s case count guidance for metrics of transmission, then re-instated them a week later. This kind of whip sawing decreases confidence in public health institutions, sows confusion, and does little to nothing to meaningfully prevent the transmission of COVID-19.

Relying on case numbers as the key metric to decide whether COVID-19 mitigations are needed will trap us forever in mitigations aimed at, well, controlling cases. These include long term mask mandates in places where they are not necessary due to high vaccination rates; quarantining healthy children; disrupting school to prevent tiny numbers of COVID-19 cases spreading to other healthy children; and sending college students to remote learning and shutting dining halls, as Middlebury College announced they were doing in early December due to 50 positive COVID-19 tests among their 2580 vaccinated students. Cornell University banned all in-person student social gatherings for the remainder of the semester, though all cases among students were mild. Meaningless mitigations such as these do not substantially slow the spread of COVID-19, may deter people from vaccination if they signal that vaccination does not mean a return to normal life, and take a toll on students’ mental and emotional health. Removing social connection from already struggling cohorts of vaccinated students who are extremely low risk for a severe COVID-19 outcome defies common sense.

Focusing on case counts also creates the misperception that vaccination is not effective, because the proportion of breakthrough “cases” is going to grow as the number of vaccinated people increases. However, the vast majority of breakthrough cases do not result in hospitalizations. Oregon’s most recent breakthrough data from December 9th indicate that, to date, “4.4% of all vaccine breakthrough cases have been hospitalized and 1.2% have died. The average age of vaccinated people who died was 81.” Oregon data also show that, for younger adults, unvaccinated people are hospitalized with COVID-19 at 15-20 times the rate of vaccinated people. For those over 80, vaccination is four times as protective

Another reason it’s important to rely less and less on the metric of cases is that measurement of case positivity is becoming increasingly inaccurate. At-home COVID-19 tests, which the Biden Administration said are about to become much more available, mean that more case data (both positive and negative tests) will not be recorded by health departments. This means that reported case numbers are going to diverge even more widely from actual case numbers. Additionally, as mitigations drag on, the extreme consequence of reporting a positive test result, both for the person tested and for close contacts– which can include exclusion from school, sports, and social events for lengthy periods of time—are becoming a large disincentive to test at all. If case numbers are highly inaccurate, what is the point of using them to determine whether masks should come off, or whether to cancel elective surgeries, as New York did this month?

Florida is currently reporting half as many new daily cases as California (7 per 100,000 compared to 13 per 100,000). Is this lower case count real and due to higher natural immunity (the two states’ overall vaccination rates are very similar), or are lower cases an artifact of less testing, lower test reporting, less breakthroughs due to later vaccination, or other unmeasured factors? No one knows. This uncertainty argues strongly for looking at the more reliable and much more important metrics of COVID-19 hospitalizations and deaths, not cases, even if both are artificially inflated by people hospitalized who happen to have COVID-19 on routine screening (COVID-19 nasal swabs are generally performed for all hospitalized patients for infection control purposes). Hospitalizations for versus with COVID can be inflated by up to 25% to 40%. The relationship between hospitalization and deaths has changed also. With treatment, people who are hospitalized with COVID-19 are less likely to die from it than they were at the start of the pandemic.

So, what should we do instead?

The U.S. needs to transition to dealing with COVID-19 the way we do other endemic viruses such as the flu.

First, health departments should monitor COVID-19 cases the way they do influenza cases. Flu monitoring happens at two levels; by the CDC’s influenza surveillance network, keeping track of percent of positive tests, peaks during certain times of the year and, importantly, variants of concern, and by state and county health departments, who are largely trying to make sure influenza hotspots are not developing and are reporting potential concerns to the CDC. In addition, at certain times of year when county health officials know that flu is circulating in a particular area, they will make sure antiviral treatments such as oseltamivir are available to clinics.

Counties and states are not publicly reporting daily case counts of influenza, nor do they pivot what they are doing on a dime when cases reach some arbitrary number. One of the major elements of influenza tracking is monitoring for variants because of the potential of flu to switch from endemic to pandemic. With COVID-19, we are going the opposite direction—from pandemic to endemic—and we need to get to where influenza monitoring starts, which is very different from what we are doing with COVID-19 cases, where every uptick is seen as a cause for panic.

Even though the Omicron variant will likely drive up cases, initial evidence is that it may cause milder symptoms than delta and so far it is not causing higher hospitalization rates, though it is at this point not the dominant strain in the U.S.

It’s time for the U.S. to stop worrying so much about cases and redirect our energies to reaching the most high-risk unvaccinated people, a number we conservatively estimate at 10 million people, but which could be as high as 20 million. These extremely high risk people are unvaccinated retirees, people 55-70 employed in small companies that are exempt from mandates, and people on disability, half of whom do not work and who often have multiple chronic conditions. Finally, in addition to not targeting our most important metric, hospitalizations, these restrictions have a cost. The most recent Surgeon General’s report on the mental health crisis among young people tells a warning tale of the price of ongoing restrictions in their lives.

In light of the transformed landscape COVID-19 vaccines and treatments bring, shifting metrics to guide COVID-19 mitigations from cases to hospitalizations is not only good public policy, but good public health.