Skip to Main Content

LOS ANGELES — They are the invisible victims of Covid-19, marginalized not just in life, but also in death.

Despite the extraordinarily detailed statistics that parse the ages, races, and comorbidities of the nation’s more than 500,000 Covid deaths, no one seems to have any idea how many homeless people have died.


One attempt to track all U.S. Covid-19 homeless deaths through official records turned up just 373. “It’s absolutely a vast undercount,” said Katherine Cavanaugh, a consumer advocate with the National Health Care for the Homeless Council. “Housing status is not on any major Covid dashboard.”

That so few Covid deaths have been recorded among homeless populations — far fewer than had been expected early on — has led many to describe the innate social distancing of homelessness as a kind of perverse advantage against the virus. Headlines asked if homeless populations had “dodged a catastrophe.”

While the idea makes intuitive sense, it may not be true. Many outbreaks occurred in homeless shelters and unsheltered people have been infected as well. One study showed the mortality rate of homeless people in New York was 75% higher than the city’s rate. It’s just that an accounting of these deaths, like homeless people themselves, seems to have slipped through the cracks.


“I don’t totally buy this narrative that Covid hasn’t been bad for homeless people,” said Margot Kushel, a professor of medicine who directs the Center for Vulnerable Populations at the University of California, San Francisco.

Shelter outbreaks not only led to infections but also forced people onto streets as shelters restricted capacity, while the pandemic shuttered or cut back social services and harm reduction programs. And people without homes could hardly follow stay-at-home orders. “It was just a recipe for disaster,” Kushel said.

What cities here in California, and across the nation, do know is that homeless deaths rose sharply in March 2020, just as the pandemic arrived. They tripled in San Francisco this spring, and this year rose 32% in Los Angeles and 54% in Washington D.C. Yet many experts remain puzzled because few of these excess deaths were officially attributed to Covid-19.

“I don’t fully understand what’s going on,” said Randall Kuhn, a social demographer at UCLA’s Fielding School of Public Health who researches homelessness. “But something happened in March.”

Los Angeles County, which records about 100 deaths among its homeless population monthly, saw an additional 40 to 60 deaths per month beginning in March. But of the nearly 1,000 homeless deaths that occurred between January and July 2020, the period for which statistics have been analyzed, just 36 were attributed to Covid.

Los Angeles paramedics cover the body of a homeless woman who died of unknown causes in Skid Row in April 2020. APU GOMES/AFP via Getty Images

The same phenomenon played out in neighboring Orange County. “Officially, only five people who are homeless died of Covid, yet there were more than 100 extra deaths since the pandemic started,” said Father Dennis Kriz, a homeless advocate and pastor of St. Philip Benizi church in Fullerton, Calif.

Kriz suspects a severe undercount of Covid deaths among the homeless because many do not receive autopsies. “If you find someone who died on the street and he wasn’t hit on the head and there’s no drug paraphernalia, how would you know how they died?” he said.

Homeless advocates and researchers say the pandemic is exposing a longstanding problem: the haphazard and often nonexistent accounting of homeless deaths and their causes. They say such data is desperately needed to address and remedy the steep human toll of homelessness, which, like the coronavirus, disproportionately harms people of color. “The big question is where did we lose people? Where did we go wrong?” asked Kuhn.

Matt Fowle, a doctoral student at the University of Washington, started tracking homeless deaths on a website he created — — after he discovered no one was collecting such data on a national level. Fowle came to Seattle from the U.K. in 2015 to study poverty and public policy. He was so shocked by the multitudes of homeless people he saw each day on his commute to campus — sometimes 100 or more — that he shifted his research focus to homelessness. He was shocked anew when he discovered the lack of basic data. “I wanted to know how many people experiencing homelessness had died in a given year in the U.S. and there was no number,” he said. “It seemed strange to me.”

One major barrier to collecting such data is that housing status is not listed on most death certificates, hospital records, or other important forms. “We don’t ask about housing status, we don’t record it, we don’t use it. What that does is obscures the scope of the problem and hides it,” said UCSF’s Kushel. “It’s almost like we don’t want to know.”

So what’s behind the spike in homeless deaths since the pandemic started? Some deaths were likely due to Covid, but were not recorded as such — especially in the early chaos of the pandemic, when many medical examiners were overwhelmed, Covid tests were hard to come by, and pauper’s graves were overflowing in cities like New York.

“There definitely was an undercount. There wasn’t any testing, and if tests were available, they weren’t giving them to the M.E.,” said Sally Aiken, the chief medical examiner in Spokane County, Wash., and the chair of the board of the National Association of Medical Examiners.

Aiken described a case of a man that had died of a massive heart attack, which seemed unusual because he was only in his 30s. Then she learned he’d recently been incarcerated. “Well, you say, the jails had a large amount of Covid infections. You wonder if he got Covid and had a huge clot that was still there and caused a massive heart attack,” she said. “It gets hard to sort out.”

Many have described homeless populations as less affected by the coronavirus because of the relatively low positivity rates found in many shelter testing programs. But programs that test everyone in a group generally have lower positivity rates than community testing programs where people who have symptoms or known exposures go for testing. “It’s apples and oranges,” Kushel said.

In California, many Covid deaths in the homeless population may have been prevented because vulnerable people were moved to hotel rooms as part of Project Roomkey, a program that is now winding down. “The good thing about the pandemic was that we had to get all of the older and more fragile people into hotels,” said Susan Partovi, a street physician and medical director of Homeless Health Care LA who thinks such housing programs should continue post-pandemic. “Can we just get everyone inside please?”

It’s clear that not all excess deaths are due directly to Covid. Some evidence comes from Los Angeles County, where public health officials have better data because of an innovative program they started in 2019 to track not only deaths, but also causes of death, among the county’s estimated 66,000 homeless residents.

“Mortality and morbidity is our bread and butter, but we had not focused on the homeless populations,” said Will Nicholas, director of the county’s Center for Health Impact Evaluation. “When you get better data, you can focus on why people who are homeless are dying and find out how to intervene.”

While the homeless morbidity report for all of 2020 has not yet been released, the county is now reporting 180 homeless Covid-19 deaths to date, which tracks with the extreme Covid surge LA saw in December and January. “The surge got so bad, its tentacles reached to more parts of the county, including the homeless population,” Nicholas said.

To obtain their homeless mortality data, Nicholas’ team faced a laborious task. They augmented coroners’ records where housing status was not known by checking if addresses were listed as shelters or street intersections with high homeless populations. They also cross-checked coroner’s reports against a heartbreaking list of words — shopping cart, bus stop, train tracks, overpass, dumpster — that could indicate if a decedent was homeless.

The LA data, Nicholas said, shows many increased deaths are due to drug overdoses, often involving the powerful synthetic opioid fentanyl. Numerous cities are reporting higher rates of substance abuse and opioid overdose deaths linked to the pandemic.

Layer the isolation and economic impact of the pandemic over the boredom and stress of homelessness and you have “a virtual prescription for self medication,” said Kriz, who writes a monthly column listing the homeless Orange County residents who have died to acknowledge their deaths, and their lives. “They walked this Earth and they deserve to be remembered,” he said.

In addition, the pandemic may have accelerated the deaths of people who were already sick. There are high rates of chronic disease like diabetes and hypertension among the homeless, as well as wounds and infections; the body of a 50-year-old person living on the streets is physiologically more like that of a 70-year-old, susceptible to Covid-19 as well as a host of other diseases.

For those who were already sick, the pandemic made health care much harder to access. “We’re all patting ourselves on the back because of telehealth, but that doesn’t work if you’re homeless,” Kushel said, especially for people who can’t afford to buy minutes for burner phones or charge phones with libraries and coffee shops closed. As for seeking care for emergencies, said Kriz: “The homeless, like many of us, were simply too afraid to go to the hospital.”

Other leading causes of death include traffic-related accidents, homicide, and suicide. The pandemic disrupted the few lifelines that existed for those who are homeless. “There’s been so much despair,” Kushel said. “People died of Covid, they died with Covid, and they died as an impact of disruptions,” Kushel said. “I 100% consider these all as Covid deaths,” she said.

There could also be more homeless deaths simply because there are more homeless people now. San Francisco officials reported that tents in the Tenderloin neighborhood increased 285% this May. And homeless advocates are worried that the economic fallout of the pandemic will continue long after the medical crisis eases. Fowle, who interviews low-income tenants as part of his research, said he’s worried about what will happen as eviction moratoriums expire. “People say they have no backup or safety net,” he said. “If they lose their current housing, they’ll be homeless.”

Many municipalities don’t even know how large their homeless populations are because transient populations are so difficult to count. LA County officials were planning a massive counting effort in January but had to cancel the event; it was deemed too dangerous because coronavirus cases were surging at the time.

This lack of data on homeless deaths is a void a growing number of groups are trying to fill. Philadelphia led the way in 2009 by creating “death review panels” to examine causes of homeless deaths, though that program was suspended two years ago; medical examiners are discussing whether death certificates should contain a box to indicate homelessness; Fowle updates his website whenever he tracks down new data; and the National Health Care for the Homeless Council has developed a toolkit localities can use to better track homeless deaths.

Many say we may never know the true toll Covid-19 has taken on homeless populations, but hope the crisis is highlighting the need to better account for what are — even without a pandemic — likely tens of thousands of deaths each year.

“If we want to address homeless mortality, we need to know what is causing homeless mortality,” the homeless council’s Kavanaugh said. “We really need data.”