In early January the state of Massachusetts added a new set of figures to its Covid-19 dashboard. Two years into the pandemic, it began to draw a distinction between people who were hospitalized because of the virus and people who were there for other reasons but also happened to be infected.
Nothing changed inside the hospitals’ walls—a Covid-positive patient there because of a car crash still had to be isolated. But the effect on the state’s numbers was dramatic. It cut them in half.
While cases have plunged and the death tally is slowing, the U.S. will in the next few weeks pass 1 million Covid fatalities, and more than half the country has been infected. There will almost certainly be more waves of illness, either from new variants or as a seasonal event. When those waves come, they will hopefully be less deadly, thanks to a wall of immunity from vaccinations and prior infections. And because of that, political leaders, health experts, and regular people across the country are adopting new attitudes toward risk and what costs they’re willing to pay to stop transmission.
But they’re making those choices with flawed, or at least outdated, information, thanks in part to the U.S.’s fractured public health system. It’s a deficit that’s made it harder to assess the consequences of the pandemic and what we’re willing to do to avoid those costs, and has helped create a vacuum that’s been filled by fatigue and distrust.
Even after billions of dollars in spending and a million dead, the way we measure the risk of the virus hasn’t improved much in the past two years. The question of how many people are hospitalized is crucial—new thresholds for public health rules from the Centers for Disease Control and Prevention depend on it. If the virus does return in another wave, it will be essential to know how much vulnerability exists in communities—but U.S. data systems make that impossible. And how will we spot that wave when more people either stop testing or shift to at-home tests that don’t get reported?
The pandemic has changed. The way the country measures it needs to change, too.
How Many Hospitalized?
Throughout the pandemic, tallying hospitalizations has been one of the best ways of measuring the virus’s consequences. Case numbers undercount the number of sick and lump in the barely symptomatic with the gravely ill. Deaths are a final reckoning but come weeks or even months too late to have any predictive value. Hospitalizations tally the strain on the health system and the financial costs, as well as the impact on those who spend weeks in an inpatient bed.
In the first year and a half of the pandemic, hospitalizations were a simple measure: Almost everybody infected with Covid who ended up in the hospital during the surges was there because of the virus. But during the wave of omicron-variant driven disease that started late last year, something changed: About half the people with Covid who entered the hospital were there for something else.
Only a handful of researchers and public health departments have looked at the issue. A team based at Harvard Medical School examined medical records to separate patients hospitalized “with” Covid from those hospitalized “for” Covid. A team at the University of California at San Francisco has done the same. And New York and Massachusetts early this year began breaking out “with” versus “for” hospitalizations in their data. All found the same thing: As Covid became more widespread and more people gained protection from vaccines or prior illness, the number of people admitted “with” Covid but not “for” Covid made up a substantial share of the more than 20,000 people a day the CDC was counting as new Covid inpatients.
“As we started looking into this, we realized this has huge implications in public health reporting,” says Jeffrey Klann, an assistant professor at Harvard Medical School, who conducted one of the studies. It means the country has failed to count the real cost of Covid as the pandemic has evolved. And because the CDC and many states use hospitalizations as a core measure of the risk of the pandemic, it means that they, and members of the public, have been using far-too-blunt data as well.
A Bloomberg review of state Covid data dashboards found that although Massachusetts, New Hampshire, and New York post the data, only a handful of other states have even done one-time surveys.
How did this happen? How are only a few states regularly tracking what seems like a crucial distinction in consequences, with vast implications for how society judges risk?
One reason may be that the question has been poisoned by the politics that have engulfed seemingly every aspect of Covid in the U.S. The “with Covid” versus “for Covid” distinction has been used by skeptics pushing the idea that the pandemic was never as severe as others claim.
Klann got sucked into that undercurrent when his group published its work. “People were talking about how we must be anti-vaxxers because we’re trying to minimize the problem of Covid, or from the other perspective that Covid is not really a problem, and we’re spending too much money on it,” he says. “Which is not what we’re trying to show at all.”
A Question of Values
Almost every public health decision is a trade-off. A work-from-home order will reduce transmission but might crush the economy. Closing nursing homes to visitors might save people from dying of Covid but will cut them off from family members they depend on. How many hospitalizations are enough that we should put restrictions like those back in place? What price are we willing to pay to not have to wear masks at the airport? What if there’s another variant? Or, in the future, another pandemic? If you’re going to make those choices, wouldn’t it be nice to have better data?
“These are fundamentally questions of values,” says Jay Varma, who helped lead New York City’s Covid response. He is now a professor at Weill Cornell Medicine, directing its new Center for Pandemic Prevention and Response. “Do you value keeping everybody healthy at all times? And what’s the cost you’re willing to bear to do that?”
But the U.S. is no longer measuring those costs using the right yardstick. And that means it’s not giving its policymakers or its citizens the best information about the choices they face.
There’s little sign that the country is ready to do better. The CDC’s new “Community Level” guidelines, which are the basis for recommending or scaling back measures like mask wearing, draw no data distinction between a hospitalization “with” or “for” Covid.
CDC Director Rochelle Walensky, in a press conference on Feb. 25, said the agency had decided to not ask hospitals for those details. Most places can’t or won’t report them, and a Covid-positive patient puts the same infection-control burden on hospitals, she said. (It does not, however, put the same strain on limited resources such as ICU beds, ventilators, and staff.)
Eventually, Walensky said, she expects U.S. hospitals will stop testing every patient for Covid. “When that happens, we won’t actually be able to differentiate,” she said.
That may be true, but it also fits a pattern at the agency, which has sometimes backed away from data collection that would have provided a clearer view. A few months into the vaccination effort, in 2021, the CDC decided to stop counting mild vaccine breakthrough infections, describing them as expected and a distraction. It was a decision that left the agency unable to see clearly when vaccine efficacy began to fade. (The CDC is taking steps to do better: It’s pushed for more authority to collect local data, and on April 19 the agency launched its new Center for Forecasting and Outbreak Analytics, promising that it would help modernize efforts to better understand and predict infectious diseases.)
Other countries do collect “with” versus “for” hospital data—the U.K. publishes regular updates, for example. But what other countries have done only highlights the U.S.’s deficiency: Lacking a national health records system, time and time again the U.S. has struggled to amass data—whether for hospitalizations, testing, vaccine efficacy, or other metrics—that could have provided a better picture of the pandemic.
This spring and summer, the U.S. is likely to go through another viral lull, one that is being accompanied by the ongoing relaxation of public health rules across the country. But Covid is raging in China and continues to transmit in the U.S. and everywhere else around the globe—it’s far from done. There will likely be another surge or another variant, perhaps one that’s better at evading our vaccines.
When that happens, will less, worse data really be the answer?
How Far Up Is Your Sandcastle?
The U.S. is already flying blind in at least one other respect.
Blood surveys show that more than 90% of Americans have some level of immune marker of protection. That includes the 66% of the population that’s been fully vaccinated (although fewer than half of those people have gotten a booster). Close to 60% of the country has been infected with SARS-CoV-2, according to the CDC’s surveys. Some people have been infected but not vaccinated, some vaccinated but not infected, some both, and some never exposed at all. All of those combinations carry different levels of protection.
And although the U.S. has relatively detailed vaccination records, there’s far less data on the infected, certainly not at the individual or community level. That makes it hard to tell where protection is the strongest, where it’s fading, and where there are already holes.
That ever-changing immunity can mean some places get hit harder than others and not always in straightforward ways. Last year, as vaccines rolled out across the U.S., researchers in Virginia began to follow how the state’s different levels of immunity functioned.
Northeast Virginia, with its liberal D.C. suburbs, vaccinated early and heavily. The southwest part of the state, rural and conservative, did not. When the delta-variant wave hit in late summer, Northeast Virginia was largely spared, says Bryan Lewis, a University of Virginia researcher who has tracked the virus and modeled how it might act.
A few months later, omicron hit and the result was very different. Cases surged in the northeastern part of the state but were far more mild in the southwest. The pattern had reversed. Vaccines had protected people from infection early on, then their efficacy had faded slightly over time. In the southwest, many people had gotten sick—and many had died—but those that were left had relatively robust protection from the virus when it came back.
So what happens later this year if a new wave of Covid hits the U.S. and immune protection has waned? Which parts of the U.S. will be most vulnerable? Which will be most protected?
“We don’t know what our current supply of immunity is,” Lewis says. “It’s really important to highlight this deficit.” He compares each community to a sandcastle as the tide is changing. “Think about waves hitting a beach. How high up is your sandcastle?”
Because the U.S. has never created any kind of national health records system or linked its vaccine records to its health records, the efforts to estimate what our immunity will look like in a few months, or next fall, are up to people like Lewis.
“We’re going to have this decline, maybe a bit more than we had last summer,” Lewis says.
“The big question is, what does this look like in the future?”
Spotting the Next Wave
It might seem hard to remember, but there was a time when people didn’t take a test every time they had respiratory symptoms. Three years ago, you were just “sick.”
While U.S. Covid testing has been a challenge since the start of the pandemic, it’s entering a new, more complex phase. Funding for Covid test reimbursement has expired, although at-home tests are plentiful. Many cases are mild, meaning a smaller proportion of cases show up at hospitals or other “official” points of care where they might be tallied. Daily case counts—always unreliable—have become even more so.
Combine waning surveillance with fading immunity, and what comes out is growing vulnerability, but fewer ways to spot the danger.
Some U.S. states and cities are trying to change that, with new surveillance methods that can spot infections without having to depend on that most unreliable of data sources: people. They’ve built networks that sample sewage for the virus. People with Covid shed the virus in their stool, often days before they test positive, and sampling wastewater can spot a wave before it shows up in tests.
“The advantage wastewater surveillance has is that it’s not dependent on human behavior, beyond using the bathroom,” says Amy Kirby, the head of the CDC’s wastewater program. “As the dynamics of the pandemic change, it remains an accurate measure.”
The CDC also plans to rely more on syndrome surveillance—tracking visits to hospitals and clinics for things that look like Covid. It’s been done before: In early March 2020 there was little to no testing and few other ways to broadly identify Covid cases. But a U.S. public health network that monitors emergency rooms and clinics for people with flu-like illness picked up a surge in patients who had similar symptoms but whose lab samples hadn’t tested positive for flu.
Fractured Ideas of Risk
If you put it all together, what you start to see is a bigger change. Our response to Covid is evolving—or being pushed—from a public-health problem to an individual one, from a shared risk to a personal one. For many people, a combination of vaccination, more widely available therapies and prior infection means their risk has fallen substantially from where it was a year ago. But not for everyone. The country is still recording more than 300 deaths a day. Many of the most vulnerable are the same people who were at risk at the start of the pandemic: the old, the frail, those with underlying medical conditions.
“This question of ‘we’re all in this together’ versus ‘everybody is on their own’ is the debate we’re going through right now,” says Varma.
Maybe that transition is inevitable. The flip side of being the land of opportunity has long been that America is also a land of inequality. And when we can’t agree on a common set of facts because we haven’t created the data to understand where we are, then by default individuality starts to win out. Why should somebody sacrifice more when they can’t even agree what they’re sacrificing for?
“People who are very vocal from the public health community and historically marginalized communities are saying, ‘You’re leaving us behind and our lives are not important,’” Varma says. “And you have people on the other side saying, ‘Ah, you’re stopping contact tracing, it was never useful.’ Actually, it was really useful for a while! It saved thousands of lives.”
Making those transitions means giving people who are at higher risk the tools to protect themselves, Varma says. He compares it to a public restroom—nobody walks in expecting they’ll need their own water, soap, and towel. Should we really expect anyone who might need a mask to have one at all times, whether to protect themselves or others?
“You have to move to an individual approach, but we have to make that handover the way it should be,” Varma says.
Bill Hanage, an associate professor of epidemiology at the Harvard T.H. Chan School of Public Health, is even more forceful. “When you say vulnerable people have the tools, do they? Do they really?” he says. “If you struggle with access to health care, how are you going to get your rapid test? If you do get a rapid test, how are you going to turn it into a prescription?”
The Biden administration has launched a “test to treat” program that would allow people with Covid to quickly get access to antiviral pills. At the same time, there’s still a lack of free access to masks, emergency money for vaccines, drugs and tests have run out, and the government has purchased only a portion of the pills it might need in a new surge.
A Broken Conversation
Hanage, who is from the U.K., has dreams about the pandemic. When things were going badly, he dreamed he had taken off his mask to step outside, and when he began to go back indoors he found himself staring at a huge pile of masks, unable to tell which was his. As the pandemic got better, he dreamed he was on the tube in London, riding along as things used to be.
Faced with a sea of Covid metrics, each with its own flaws, Hanage’s dreams may seem just as useful as any other method in understanding where the virus is heading and what do to about it. The reality of the response to Covid has sometimes been influenced as much by feeling as fact. It has come down to values, and the country has never agreed on what those values are—not before the pandemic, not after it.
Would better data create an easier conversation about values? Maybe. But look too deeply, and what reflects back isn’t a new, perfect view of Covid, it’s the unfairness of the U.S. health-care system, which was endemic long before the virus. Death data turns into people who didn’t get a cancer screening because hospitals were closed. Low vaccination rates kaleidoscope into the fact that one-fifth of Americans don’t have a primary care physician to talk with about their hesitations. Hospital capacity levels reveal not just how many Covid patients were sick but how deeply the health-care workforce has been broken.
In that imperfect view, the endemic phase of Covid becomes one more weight tilting the scales against the people who have always lost out in health care, either because of bad luck, bad genes, the ravages of time or poverty or circumstance—all the things that make us sicker and eventually pull us away from this world, some sooner than others. —With Andre Tartar, Madison Mills, Robert Langreth, and John LauermanRead next: Why Nasal Sprays Are Poised to Be the Next Weapon for Fighting Covid
©2022 Bloomberg L.P.
