Death tolls are often bullshit. They're also indispensable.

Welcome back to The Long Version, a newsletter by Jonathan Myerson Katz.

Over the last forty-eight hours or so, there has been a noticeable uptick in reporting about the rising cost in human lives of the coronavirus pandemic. Each story is greeted with broad alarm and an undertow of skepticism. “These death tolls are HIGHLY inflated,” someone invariably replies. “Are you out there counting the bodies?????”

In short: Aren’t the numbers bullshit?

In a way, the groundwork has been laid for weeks by conservative commentators. In early April, Britt Hume went on Tucker Carlson Tonight to complain that COVID-19 totals included people who had the disease but perhaps died from some other cause. “And there may be reasons that people seek an inaccurate death count but we can address that later,” Carlson responded, cryptically. The memes followed.

This isn’t really new, though. As a foreign correspondent, I often found myself calculating—and sometimes managing by default—death tolls in crises from epidemics and major disasters to social unrest. Each time, people made the same claims as Hume and the trolls, in various languages, at various volumes of anger. I thought it might be useful this week to bring my experience to bear.

My answer is … Yes. Death tolls on a massive scale always contain a high quotient of bullshit.

They are also extremely important.

How do we count?

The basic problem is that attributing a bunch of deaths to a single cause or event is by definition an abstraction. “Earthquake” and “war” are not causes of death. The tolls from each will contain people who “actually” died of heart attacks and car crashes. There are always “co-morbidities”—people who get shot only to die of asthma on the operating table. Until roughly World War II, the vast majority of soldiers who died in wars “actually” died of disease.

You’d think a medical pandemic would make things easier, but it doesn’t. At the height of the Haiti cholera epidemic, in late 2010, so many people were getting sick that the underfunded, overworked doctors could not test everyone for the bacteria. If the doctor decided they had cholera-like symptoms, they were a cholera case. If they died, they were recorded as a cholera death. (Sorry, Britt.)

Similarly, doctors now believe that the first COVID-19 death in the United States, 57-year-old Patricia Dowd of San Jose, Calif., occurred when the virus infected her heart and caused it to rupture. It was originally diagnosed as a heart attack—which, in a way, it was.

At the last count by the Haitian health ministry, in 2018, the death toll from the cholera outbreak stood at 10,191. The World Health Organization’s tally stood at an equally specific sounding 9,007. I could speculate on the reasons for the discrepancy, but that’s the point: At the end of the day, both are guesses. A Doctors Without Borders study in 2016 estimated the real number of deaths from cholera in Haiti could be as much as ten times greater in some areas. Some experts believe the “real” toll is 30,000, or more.

Who counts?

The Doctors Without Borders study was based on a different metric, a piece of statistical modeling called “excess deaths.” It compares how many people are recorded to have died in a given area at a given time to how many people should have died based on past experience.

“Excess deaths” were the basis of two widely shared articles this week by the Washington Post and New York Times. Both found that far more people have likely died since the start of the COVID-19 epidemic in the United States than has been officially reported.

As should be clear by now, this is in effect a guess compared to another guess based on past guesses. Yet “excess deaths” is often the best way to calculate a toll. It captures the people institutions miss: the poor who die in their homes; those in areas rescuers can’t reach; or those, often the very old or very young, who die before anyone knows what was wrong with them.

It can also correct for official coverups. If we ever get a good picture of how many people really died in Wuhan in the first months of the COVID-19 pandemic—a number the Chinese government assuredly wants no one to calculate—it will probably be through an “excess death” study.

If that method also counts those who die for other reasons—because the health system was too overburdened to help them, or they went hungry because the family breadwinner was gone—that is not necessarily inaccurate. It is not so different from counting a child who dies in a wartime famine as a casualty of war.

What counts?

Sometimes, people want to emphasize a high death toll. High numbers can convince donors and aid groups to give more money. (Something I’ve long thought ironic, since fewer dead can mean more survivors in need of help.)

At other times, governments want to keep the numbers low—especially those in power who need to cover up their guilt, incompetence, or both.

Those competing needs were on display in Puerto Rico after Hurricane Maria. President Trump was aware of the (accurate) perception that he had let Puerto Ricans suffer while focusing attention on the red states of Florida and Texas that same hurricane season. At a press conference in San Juan, he bragged that thanks to his and Gov. Ricardo Roselló’s efforts, only sixteen people had died—far less than “a real catastrophe like Katrina.

Less than a year later, with the electricity still out across the island—and people dying for want of basic medicine and other needs—it seemed clear that far more than a handful of people had died as a result of the storm. A team of Harvard researchers estimated the “excess deaths” from Sept. 20 to Dec. 31, 2017, ranged from 800 to 8,000. They settled on a midpoint of 4,645.

That number has become iconic in Puerto Rico. It was hoisted at protests that forced Roselló from office in 2019.

(Angel Valentin/The Guardian)

Yet that number is admittedly estimate, and a somewhat random one at that. As far as I can tell, the Harvard study ended its count arbitrarily on Dec. 31, probably just to use a nice, clean end of the calendar year. The blackout continued well into 2018. Deaths may continue to accumulate for a long time.

Under the pressure of that study, the Puerto Rican government came up with a revised estimate of its own: 2,975. That in turn was also the midpoint of yet another study, which topped out at 3,290.

Trump, of course, has angrily refused to acknowledge any deaths in Puerto Rico other than the ones he already bragged about.

We are already seeing even more blatant shenanigans with COVID-19. The Tampa Bay Times reported that Florida state officials have been suppressing statistics from the state medical examiner’s office, as pro-Trump governor Ron DeSantis tries to coax residents into going back to work.

… compared to what?

I was among the journalists who noted this week when the official COVID-19 death toll in the United States surpassed the official U.S. death toll for the Vietnam War.

Some complained about the comparison. As Aaron Blake wrote in the Washington Post: “Outbreaks, by and large, target people, who haven’t volunteered or been conscripted into a situation they know to be dangerous.”

(If you’re wondering which kind of death Blake thinks is more valid, he also said: “War deaths carry valor … that is not applicable to virus outbreaks.” That every conflict death is “valorous” is debatable. And there is quite a lot of valor happening in ICUs and homes across the world right now. But, again, that’s the point: Death tolls are subjective.)

I think the comparison is useful. No one knows how to feel about the number 61,187—the last estimate I saw of U.S. COVID-19 deaths from the Johns Hopkins Coronavirus Resource Center. But many know what it is like confronted with the names etched into the black granite walls of Maya Lin’s Vietnam Veterans Memorial—a feeling so overwhelming that many pro-war conservatives opposed the memorial’s construction in the first place.

Others use context to downplay death tolls—hence the persistent appeals to compare COVID-19 to opioid overdoses or the seasonal flu. Never mind that these are bad examples: U.S. COVID-19 deaths in two months have already exceeded the death toll of any year of the U.S. opioid epidemic. And, as Dr. Jeremy Faust of Harvard Medical School wrote, seasonal flu deaths may themselves be overestimated. Still, the trolls’ insight is clear. Any number looks small next to a bigger number. Those trying to wave away the deaths caused by the pandemic will inevitably find ways to do so.

Home on the range

Those of us who have been in the trenches tend to be circumspect. Whenever possible, I use a range of the most defensible estimates. In my book about the 2010 Haiti earthquake, I don’t even address the death toll until the middle of the fourth chapter, in a paragraph that begins, “Nobody knows how many died as a result of the earthquake.”

What I’ve really learned is that death tolls are flawed but necessary attempts at describing the indescribable. They are a way of finding the connections between the most individual of tragedies: Bodies decaying in a U-Haul truck in a Brooklyn alley, a beloved grandmother dying alone in a Minnesota nursing home, black families burying parents and most of their siblings at once in Louisiana.

Trying to add up a death toll in the middle of a crisis is always fraught, and always at the mercy of those with power who want to influence the numbers this way or that. But most people who calculate death tolls are trying their best. And without at least trying to figure them out, we can’t even guess where we are, or where we might be headed. They’re a road marker we build as we go.

Thanks for reading. You can help keep independent journalism alive by opting for a paid subscription to this newsletter. More details below. Stay safe.


Jonathan Myerson Katz is a journalist and the author of The Big Truck That Went By: How the World Came to Save Haiti and Left Behind a Disaster. His next book, Gangsters of Capitalism, explores the life of Gen. Smedley Butler and the rise and fall of American empire. Follow him on Twitter @KatzOnEarth.

Top photo: Leo Correa/AP