More than 100,000 U.S. COVID-19 Deaths May Be Uncounted, Many of Them of Blacks or Latinos

This story is co-published with Capital & Main

When Denny Gilliam heard from his agency, in April of last year, that New York City was facing urgent nursing staff shortages during a torrent of rising COVID-19 cases, he felt compelled to serve. The former Army and Air Force vet had watched the Berlin Wall fall from the front lines of Germany, and now—as a longtime acute-care travel nurse—decided to heed the call once again. Within days, the healthy 53-year-old left his family in Pelham, Tennessee, and ventured north. From a city in desperate need, he sent recordings of applause echoing from the rooftops. He felt "more alive than ever," Denny told his wife, Amanda.

After weeks of long hours at NewYork-Presbyterian Brooklyn Methodist Hospital, the mood of the messages shifted. Denny sent videos of the endless ambulance queues that greeted him at the beginning of his shift. He called the hospital a "war zone." He told Amanda that he couldn't wait to come home. Amanda couldn't wait either; she'd worried about the danger of the assignment from its beginning.

Amanda and Denny Gilliam with two of their three children.

She was also anxious to spend family time with him — after some ups and downs, they had "finally got [their relationship] right," she said. Among the activities on the itinerary: fishing for crappies and bluegill in Chickamauga Lake, hunting ginseng on Monteagle Mountain, blasting Simon and Garfunkel around the house, or piling into bed with their three young children for a movie night complete with chips and a jug of Heluva Good! French Onion Dip.

"We had so many plans," she said.

Then there was silence. Two nights passed without a word from Denny. Amanda began to panic. She called 35 hotels across Brooklyn. Finally, at 12:13 a.m. on a Thursday in May, her phone rang. It was an investigator with the medical examiner's office: Her husband had been found dead in his room. "I was in shock," she said.

As shock melted into confusion, grief and anger, Amanda began seeking answers. She didn't know how it had happened; the investigator had provided little detail. Denny did not receive a test for COVID-19, the investigator told her. Around the time of his death, even living patients were facing a "serious shortage" of tests, then reserved for hospitalized patients only.

A year later, Amanda has more questions than answers: Denny's original death certificate, which she received months after his passing, listed the cause of death as "pending further study." Later, his diagnosis was updated to "acute intoxication" without her knowledge, and even though many deaths have multiple causes listed on the certificate, COVID-19 was not mentioned on Denny's. Amanda received no further information as to the timing or rationale behind the pivot, and remains convinced COVID-19 played a role in his passing. He sounded "freaking horrible" when they spoke over the phone before losing contact, she said. "You could hear it in his throat." Multiple calls and emails to the medical examiner requesting additional comment went unanswered.

"[Denny] paid for a round trip ticket," Amanda said. "The least they could send back are some answers."

In Augusta, Georgia, Bruce Davis' family was also left without answers when his death certificate listed sepsis and renal failure, not COVID-19. Davis, a Pentacostal pastor and nursing assistant, had been caring for hospitalized prisoners when he became ill himself. Two weeks later, he died.

Across the country, tens of thousands of families whose loved ones died during the pandemic without a COVID-19 diagnosis may be asking similar questions.

Excess deaths far outweigh official COVID-19 toll
Since early in the pandemic, epidemiologists, concerned with the potential undercounting of virus cases in situations like Gilliam's and Davis' (especially in the wake of an early bungled rollout of COVID-19 testing), have tracked "excess death" as a rough approximation of the pandemic's impact on the country. The statistic, a widely accepted way of monitoring the emergence of mass casualty events such as infectious outbreaks, compares the overall death rate over the past 16 months to that of the prior eight years.
The spike since last year has been stunning. Between Feb. 1, 2020, and June 9, 2021, the U.S. Centers for Disease Control and Prevention tracked up to 713,873 excess deaths, of which nearly a quarter—up to 169,687—are not currently attributed to COVID-19. That many Americans would fill the New Orleans Superdome twice over. By June of last year, Americans' average life expectancy had fallen to 77.8 years, meaning Americans were expected to live a full year less, on average, than they had been expected to live in 2019.

While not all of the excess deaths during the pandemic are likely to have been caused directly by COVID-19, experts say the discrepancy points to the likely undercounting of COVID-19 cases because it is far higher than can be explained by historical patterns or official COVID-19 numbers. The World Health Organization (WHO) recently estimated that, globally, the true death burden from the pandemic is up to three times that of official statistics. Widely accepted scientific models estimating excess death—like the one used by the WHO—have been met with skepticism from some conservatives, who have derided them as efforts to inflate the counts. But if the models are accurate, it means that thousands of deaths resembling Davis' and Gilliam's—with official causes other than COVID-19, but reasons to suspect otherwise—could be going uncounted.

Emerging research published on May 20 in PLOS Medicine, an international medical journal, breaks these excess deaths down on a county by county basis for the first time. And like many consequences of the pandemic, the death burden isn't distributed evenly across the country.

Instead, the excess death toll brings the impact of socioeconomic and racial inequities into even sharper relief, with the estimated additional pandemic deaths higher than official numbers in communities that are rural, poorer, less educated, Southern, non-white (especially Black), or where more people face medical risk factors (such as diabetes, obesity and smoking). People of color suffered even greater drops in their life expectancy than white Americans. In Hispanic communities, it dropped 1.9 years. In Black communities, it dropped 2.7 years.

"Our analysis suggests that the substantial racial inequities observed in directly assigned COVID-19 death rates for the non-Hispanic Black population are even larger in excess death rates not assigned to COVID-19," the Boston University-led research team concluded, noting "a pattern related to structural racism."

If, as experts suspect, at least some of those 169,687 deaths are COVID-19 cases that have gone uncounted, the consequences could be significant.

View our Excess Death map below. Use the search bar on our interactive map to find your county.

On the one hand, there's the importance of a historical record: Without accurate data, we may never know the virus's true toll—who we've lost to it over the last 16 months. These excess death numbers suggest the pandemic's impact on the country is likely even greater than the official statistics have shown. Uncounted cases also mean that grieving families may never receive the support they need and deserve. For example, Politico reported that thousands of Americans have faced delays or denials in reimbursement from federal funeral assistance programs due to absence of a COVID-19 diagnosis on their loved one's death certificate.

There's also a more immediate question, as the country teeters on the edge of new surges as vaccination rates plateau: What does a failure to capture COVID-19 deaths mean for public health measures to protect the living? Prompt identification of the virus is a matter of life and death. Identifying what epidemiologists call a "sentinel" case, an outbreak's version of the canary in the coal mine, can determine whether a disease is controlled in a given community of people or if it surges and overwhelms them. Throughout the pandemic, these cases have been missed. Outbreaks, illness, hospitalization and death have followed.

So experts wonder: How many Americans would still be alive if communities had found and responded to the virus earlier? Would hundreds of Americans in El Paso, Colorado, and East Baton Rouge, Louisiana, still be alive? And how many more can be saved, as the pandemic persists, if we can get accurate, comprehensive information to epidemiologists and the medical community more quickly?

For families still waiting for answers, the question is simpler: Would their loved ones have made it home for movie night? Will they?

The death maze

Hundreds of thousands of American families like the Gilliams have turned to death certificates to help them understand their loved ones' deaths during the pandemic. The paperwork has carried heightened importance: Families rely on the documentation not only for emotional closure, but also for economic relief and legal recourse.

Death certificates also play an integral role in public health. By communicating essential details uncovered over the course of a patient's clinical care—the presence of a novel virus, a resistant bacterium or a potent narcotic—the documents allow health authorities to take measures, such as contact tracing, to protect communities.

However, amid uncertain science, variable protocols, strained budgets and a politicized virus, the process of certifying COVID-19 deaths has been less than straightforward.

Understanding why excess deaths during the pandemic might be so different from the official COVID-19 death counts, said Robert Anderson, chief of mortality statistics within the CDC's National Center for Health Statistics, requires some background on how the systems that track vital statistics came into being. Like many federalist heirlooms, the process of death certification was delegated to local jurisdictions. As a result, Anderson said, these systems are variable in structure and process, with standards that are idiosyncratic and inconsistent.

Vital statistics systems, like the one Anderson now directs, are the great-great-grandchildren of methods of death and disease documentation birthed in ancient Rome. The American incarnation formally began in 1632 with a Virginia law requiring church ministries to track deaths in the colony; by the early 1900s, death records were being used to monitor tuberculosis mortality in New York City, bubonic plague mortality in San Francisco and, later, Spanish flu mortality across the continent.

For patients receiving medical care, some counties limit certification powers to physicians, while others permit physician assistants, nurses, dentists, or midwives to certify deaths.

For patients who are not receiving care—or for those whose deaths appear "unnatural" or otherwise "reportable"—death certification falls under the purview of forensic experts, which might mean medical examiners, who are usually politically appointed physicians or forensics experts with years of graduate-level medical training, or coroners, political electees or appointees whose backgrounds may lie in funeral home services, criminal justice, prosecutorial law or none of the above.

Local decisions, major consequences

Relying on counties for death investigation has led to fundamental differences in how COVID-19 deaths have been certified during the pandemic.

Pre-COVID studies have found that coroners are less accurate than physicians when certifying causes of death. Clinicians' competency "ranges from OK to horrible to OMG," said Brian Peterson, chief medical examiner for Milwaukee, citing a lack of required training.

Despite the CDC's recommendation that certifiers include COVID-19 as a contributory cause in cases where they suspect the virus may have been present—regardless of whether they received a positive test—some counties have adopted different policies.

For example, in East Baton Rouge Parish, Louisiana, a county of 440,000 where the excess death rate is far higher than the COVID-19 rate, decedents have the virus on their death certificates only if they had both a positive test and known symptoms.

"Remember, without a positive swab, then (scientifically) you don't have COVID," William Clark, an emergency medicine physician and the parish's coroner, wrote in an email.

But testing may not be routinely performed or readily available in many counties, which was especially true early on in the pandemic, said Victor Weedn, a forensic pathologist, attorney and past president of the American Academy of Forensic Sciences. Nonmedical factors can affect testing and reporting, including concerns about stigma. In Arkansas and Georgia, for example, some individuals diagnosed with the virus say they were all but cut off by family, friends, neighbors and colleagues after testing positive.

Dashauna Ballard, a 29-year-old Alabama native, said she faced the derision of fellow churchgoers after she recovered from a bout with COVID-19 that landed her in the hospital. When people at her church implied that the infection was punishment for her sins, Ballard said, "I felt like I had a scarlet letter on my chest."

Financial concerns constitute another reason for diminished testing—despite CDC recommendations to test all suspected COVID-19 cases for the virus. Coroners on tight budgets may view kits for postmortem testing as "unnecessary expenses," Weedn said. Or they can be in short supply, as in Denny Gilliam's case, when testing simply wasn't available.

"Without testing, it's impossible to know what we don't know," Weedn said, "so we'll bury our mistakes—literally."

In counties where clear protocols do not exist, death documentation relies heavily on the subjective judgments, opinions and beliefs of certifiers.

That may be why people like Bruce Davis are going uncounted. Part of the issue for people like Davis is a dearth of lack of formal training in death certification for clinicians, said Dr. James Gill, Connecticut's chief medical examiner and president of the National Association of Medical Examiners. Accuracy and comprehensiveness remain major challenges, he added, especially for diseases that, like COVID-19, can be subtle in the absence of a medical workup or an autopsy when other conditions are present.

Certifiers "too often just chalk a death up to bronchopneumonia or dementia or cardiopulmonary arrest," he said, "none of which are causes of death, but any of which may co-occur with COVID-19."

Subjective factors also can influence whether a certifier codes for the virus, Weedn said. For example, clinicians may be pressured to leave a diagnosis off the certificate. In Colorado's El Paso County, coroner Dr. Leon Kelly said that families have demanded removal of COVID-19 from the documents, calling the virus a "conspiracy" and "hoax."

Additional forthcoming research by the Boston University excess death team—shared with Capital & Main in advance of publication—corroborates these claims, finding that Trump-voting counties, and those that are coroner-based, were more likely to have high rates of excess deaths that were never attributed to COVID-19. Andrew Stokes, an assistant professor of global health and lead author on the research, deemed this phenomenon a "MAGA modifier."

For example, in Florida, the politics of death have recently come under renewed scrutiny after months of accusations that Gov. Ron DeSantis (R) and his colleagues were suppressing or manipulating the state's numbers.

In May 2021, the Institute for Health Metrics and Evaluation (IHME), a nonpartisan population health research center associated with the University of Washington, released a model suggesting that nearly 14,000 deaths due to COVID-19 went uncounted in Florida since the beginning of the pandemic. (As of June 3, Florida Department of Health statistics counted 37,717 deaths compared to 51,496 in the IHME model. On June 4, Florida stopped releasing daily death counts and discontinued its online dashboard.)

The state's administration continues to rebut these accusations: A year ago, DeSantis called models like IHME's "totally unreasonable." In April, Shamarial Roberson, deputy secretary for health for the Florida Department of Health, referred to the most recent models as "excess analyzing."

"Florida is not undercounting deaths," she concluded.

More than paperwork: The consequences of missed cases

There's no doubt that some excess deaths during the pandemic had nothing to do with COVID-19. While the official CDC numbers have not yet been released, preliminary estimates of the rates of drug overdoses, homicides and accidental deaths, such as traffic accidents, are all up since March of 2020; the emotional and economic trauma caused by the COVID-19 have driven the so-called deaths of despair, along with homicides and accidents.

Still, those causes can't account for all the excess deaths over the past year, experts say. While collecting the 2020 data is expected to take 18 months or more, it's unlikely that injury deaths could have increased 54% to account for all the excess deaths not attributed to COVID-19 in 2020. (They increased only 2.3 percent in the prior year, according to CDC data.) "I am sure we're missing [COVID-19] cases," Anderson said.

As such, county-level protocols, influences and pressures leading to undercounts may have contributed to undue illness and death from COVID-19. The public health implications of undercounting are "extraordinarily worrisome" to Gill—especially in small counties, where even a couple "silent" cases could ignite an outbreak. While vaccination can help insulate against the worst impacts, the communities that Stokes' team found have the highest risk of missing COVID-19 deaths also have lower rates of vaccination.

And because medical information is covered "lock and key" under privacy laws, Gill said that public health authorities rely on death certificates to "provide a critical red flag" for COVID-19 outbreaks. He's seen this up close, uncovering "dozens and dozens" of silent COVID-19 cases in his own region that health authorities subsequently acted upon.

"The death certification is the gold standard," Gill said. "It isn't, shouldn't, mustn't be a political piece of paper."

When data is a matter of life and death

Kelly, in El Paso, witnessed firsthand what can happen when cases go silent, even temporarily. In March of 2020, moments into his discussion with a deceased octogenarian's family, he realized his county was in trouble.

As loved ones shared the details—prior to her death, Grandma had played in a weeklong card tournament at the Colorado Springs Bridge Center with what, in hindsight, were active symptoms of COVID-19—Kelly remembers "the picture just kept getting worse and worse."

The old-timey card game pits paired combatants against one another at adjoining card tables, where they butt elbows as they circulate around the room. Within weeks, the Colorado Springs tournament had become a public health nightmare. Dozens fell ill. Four died.

"With the delays ... we were immediately overrun," Kelly said. "It was devastating: These were the socialites of Colorado Springs. Everybody knew these people."

In under-resourced communities where those seeking medical care for the virus have faced delays, as in Tennessee's majority-Black neighborhoods and the Navajo Nation, such outbreaks can cause particularly severe consequences.

Take East Baton Rouge parish, where William Clark decided not to attribute deaths to COVID-19 without a positive test. As of June 9, 14 percent of total U.S. deaths from COVID-19 were in Black communities, according to the CDC. In Louisiana, a state whose population is 33 percent Black, 38 percent of the state's COVID-19 death toll was borne by Black individuals, according to the Louisiana Department of Health. (Clark noted that most Louisiana coroners "follow [protocols] similar to mine.") In East Baton Rouge, that figure is higher still. In the parish, which is 46 percent Black, 51 percent of those who died from COVID-19 were Black. Hundreds more may have gone uncounted based on the many excess deaths not currently attributed to COVID-19.

"If you pile the cultural issues, the financial issues, the access issues and political issues on top of everything else about this disease—that creates even higher risk," Kelly said. "If you're behind the curve—because you're refusing to acknowledge it or because you simply don't have the resources—you're going to miss [cases]."

"And if you miss [cases], you will have deaths that otherwise wouldn't occur."

Living without answers: The legacy of pandemic deaths

As the pandemic persists, getting the numbers right isn't just a matter of arithmetic. It's about lives lost because a case slipped through the cracks. It's about communities decimated because an ember turned into a wildfire. And when the sun invariably, eventually, finally, sets on the pandemic, it's about historical memory: who died during the pandemic versus who died from it? In other words, who gets counted?

Back in Tennessee, Amanda maintains her belief that Denny died of COVID-19. Deaths commonly result from multiple causes: Amanda said that the context of his work—providing hands-on care in a hospital overwhelmed by patients with the virus—along with the presence of respiratory symptoms before his death have convinced her the virus is involved. But it's impossible to know whether Denny had it without a test.

Without a COVID-19 diagnosis, Amanda can't file for workers' compensation from his agency. (The agency, TruStaff, did not respond to multiple requests for comment.) Or apply for federal relief funds earmarked for the families of victims. Or begin the process of mourning in earnest, even after she paid out every last penny he made on his New York assignment to hold a funeral.

Without answers, "It's like my hands are tied," she said.

Then there's the lingering question of how things might have played out if suspected cases like Denny's were definitively diagnosed with the virus and if appropriate measures were taken as a result. How many Americans might still be alive, if the contact tracing, the quarantines, the lockdowns and the work-ups all happened more quickly? Would COVID-19 have flooded that hospital in Brooklyn way back when? Would he have traveled north at all?

In a different reality, maybe he and Amanda would be on a boat right now, catching crappies in a sparkling blue Chickamauga Lake. Home where my love lies waiting, they might be singing in the summer sun, silently for me.

Copyright 2021 Capital & Main