Will we ever know the real death toll of the pandemic?

The past year has been a stark reminder of global inequalities—including the resources needed to collect timely and accurate data on deaths. These innovators aim to fix that.

By Amruta Byatnal
Published 15 Jun 2021, 13:24 BST
death toll
A grave is prepared in New Delhi on April 23, 2021, for a man who died from COVID-19.
Photograph by Atul Loke, T​he New York Times, via Redux

On a sweltering April night in Ahmedabad, the largest city in the Indian state of Gujarat, Shayar Rawal rode his motorbike to a COVID-designated government hospital at 11:55 p.m. Over the next 24 hours, he had just one task: to count the number of dead bodies being brought to the mortuary.

Rawal—a reporter at the Gujarati daily newspaper Divya Bhaskar—counted four deaths in the first hour. Grieving relatives collected five more dead bodies in the next hour. When the number reached 100, he says, reality sunk in.

“I knew the government was hiding numbers, but this was way more than what I had expected,” Rawal says.

Officially, the city recorded just 15 COVID-19 deaths that day. But by the end of his vigil, Rawal had counted 112 bodies—and that was just the dead from one city hospital. Over the next few weeks, his colleagues also counted bodies at crematoria and burial grounds and accessed death certificates from districts across the state. Their final analysis revealed that the death toll over nine weeks in the state of Gujarat was 10 times higher than the official figure. In response to the Divya Bhaskar report, the state government gave multiple reasons for the discrepancy, from saying deaths with co-morbidities cannot be counted as COVID-19 deaths to claiming that duplicates of death certificates had been issued. National Geographic’s questions to the state government went unanswered.

“There are fundamental, inherent challenges,” says Samira Asma, assistant director general for data, analytics, and delivery for impact at the World Health Organisation. Even in wealthy nations, officials are grappling with incorrect diagnoses, irregularities in the data tracking, and other factors that can obscure the virus’s true impact. “So because of this, we don't have a complete understanding of the entire scope of the pandemic.”

But the past year has also been a stark reminder of inequalities throughout the world—including the resources needed to collect timely and accurate data on deaths. In an assessment conducted in 2019, the WHO found that about two-thirds of the countries in the world lack strong civil registration and vital statistics systems that keep a count of births and deaths.

This disparity is having dangerous consequences with COVID-19. The World Health Statistics Report released last month stated that there were 3 million deaths directly and indirectly attributed to the SARS-CoV-2 virus—that’s 1.2 million higher than the official figures reported by countries and then tallied by the WHO.

The lack of robust data in low- and middle-income countries such as India means grassroots efforts such as Rawal’s are crucial in determining the true death toll, which in turn affects our understanding of the pandemic’s global trajectory.

“To have an accurate understanding of historic mortality is key in knowing how effective different interventions have been, but also in helping us to more accurately forecast what may happen in the future of the pandemic,” says Oliver Watson, a postdoctoral researcher in infectious disease epidemiology at Imperial College London.

Deaths due to COVID-19

At the beginning of the pandemic, a lack of standardised methods to designate COVID-19 deaths across countries led to underreporting. In a bid to understand the true toll better, demographers, reporters, and economists developed alternate tracking methods, and each approached the problem from different angles.

For Ariel Karlinsky it started with a meme.

Karlinsky is a graduate student at Hebrew University in Jerusalem and an economist at Kohelet Policy Forum, a think tank. When Israel locked down in March 2020, a popular meme doing the rounds was that COVID-19 was claiming as many lives as the flu had in previous years. Curious to find out if that was true, Karlisnky started digging for data, but he didn’t find any. So he started collecting data on his own from individual countries—he emailed national and regional statistics offices, as well as researchers who were working on the issue in various countries.

In January 2021 he started collaborating with Dmitry Kobak, a research scientist at Tübingen University. Their effort has led to the creation of the World Mortality Dataset, which includes information from 95 countries and territories. The Economist has since used the World Mortality Dataset to make its own projections on global deaths due to COVID-19. Recognising Karlinsky's work, the WHO invited him to be part of a technical advisory group that aims to map the death toll of the pandemic across the world.

Karlinsky is surprised. “Usually there are organizations like the World Bank and the OECD [Organisation for Economic Cooperation and Development] that essentially do what I did. They take data from each individual country and they harmonise it. For some reason, they hadn’t done it, and they still haven’t done it,” he says, speaking over Zoom on a recent Sunday afternoon. “Although they are now taking my data, which is sort of strange because they are official organisations with a much higher budget than me and my laptop.”

Karlinsky is now also working on local mortality figures based on data for smaller regions, such as cities or states across the world. By collecting and standardising this data in one place, more people can use the information to make informed decisions about their local pandemic response.

“A lot of data exists, but that data is not being used, the data is not being shared—it's the same as the data not existing at all, right?” he says.

New sources, old problems

One source that made its way into the World Mortality Dataset comes from Indian data journalist Rukmini S, who like Karlinsky believes that the data exists in India’s civil registration system, which records births and deaths across the country. Knowing this, Rukmini scraped together mortality data for all causes in the city of Chennai, in South India.

One way to get a handle on COVID-19 deaths when specific causes are not available is to look at what are known as excess deaths. This figure represents the gap between the number of deaths in an average year, after adjusting for population growth, and that in a year with an extenuating circumstance, like a pandemic.

Excess deaths could be COVID-19 deaths that went unreported, but they might also be due to indirect causes, such as an inability to access healthcare during lockdown, or due to unrelated diseases. While more analysis and data will be needed to link the past year’s excess deaths to COVID-19, researchers are increasingly relying on excess mortality figures to understand the direct and indirect consequences of the virus.

In Chennai, the data show more than 74,000 deaths in 2020, which is 12,000 more than the average of the five preceding years. That’s a 20 percent increase, despite the fact the city’s official number of reported COVID-19 deaths for that time period was just 4,000.

In other places, public health experts have found even more novel data sources. In Damascus, Syria, public pressure led to the mortuary office releasing data on deaths due to all causes for a limited period of eight days between July 25 and August 1, 2020. Watson, of Imperial College London, matched this information with data from a Facebook group that uploads obituaries.

“If data is a public good, it should be open... And that is the only way to hold each other accountable.”

Samira Asma

Using these data points, Watson and his colleagues calculated estimates for the entire duration of the pandemic, from February up until September 2020, to arrive at a staggering figure: Their model shows that only 1.25 percent of COVID-19 deaths in Damascus had been reported as of September 2, 2020, and over 4,380 deaths might have been missed through the official reporting system.

For Watson, the analysis was an eye-opener. “It completely changed how we understood the scale of the pandemic,” he said. ”It was just shocking that you could have that amount of death happen, that level of health system collapse, and how it didn’t get much media attention.”

At first, he says, it was just hard to convince people that the officially reported death toll was inaccurate. But he notes that he has found a sense of community in other researchers who are also willing to question the official data and try various methods to arrive at better estimates.

“Coming across other individuals who are coming at it with an understanding of actually no, we all are missing a lot of what’s actually happened and we need to be doing much better—actually going out there and finding that data—it’s definitely quite a bonding experience,” Watson says.

Government transparency

While acknowledging the benefits of these kinds of creative methods, Rukmini, the data journalist from Chennai, emphasises that the most accurate data need to be made public by governments.

“While I am all for novel sources, and everyone working on this is doing it out of the spirit of scientific inquiry and obviously working very hard for it, I am a bit concerned that enough democratic pressure is not building for official statistics to be released,” she says.

Chinmay Tumbe, an economist who has studied past pandemics, thinks that relying on government figures would mean that the actual toll would only be revealed in a few years, when census data comes in. But that would be too late for governments, businesses, and other policy-makers to use the data to influence the course of the pandemic.

Getting to the true death toll will eventually help in figuring out the true infection fatality rate, the proportion of all infected people who died because of COVID-19. This value was crucial in the early stage of the pandemic to understand how serious the implications might be if the public health restrictions were not put in place and the virus was allowed to spread freely in populations.

“However, it is very difficult to estimate correctly if we don't know the true number of people who have died,” Watson says.

He explained that the infection fatality rate for COVID-19 has been a topic of debate. Some people argue that it has been inflated, and they have used that position to discourage lockdowns and other interventions. They also point to lower reported mortality in lower-income settings as evidence that the pandemic is not as deadly as initially thought. “Those arguments are blatantly incorrect, given the scale of underreporting that has been observed,” he says.

That is why researchers need to continue to put pressure on governments to prioritise accurate and timely data collection, Tumbe says. “I think it can easily become a poll issue. And it might even become a manifesto issue—saying that, well, if you elect us, we will form a committee with which we will count the deaths.”

Need for accountability 

In some countries, the pressure on the government has worked. Recently, Peru revised its official death count, which now stands at 185,380—almost three times the original figure of 69,342. At 5,551 deaths per million people, Peru has the worst official death toll in the world.

Convincing countries to accept a higher death count than what was officially reported is not going to be an easy task. With the help of the advisory task force, the WHO has a plan, says Asma. By November 2021, the WHO aims to standardise methodology for counting excess deaths, decide what parameters should be used, and have revised estimates for all countries, which will be followed by consultations with representatives of governments.

“We are going to do it in a transparent way, so that is going to be very critical,” she says. That way countries can work with the WHO to flag inconsistencies and land on the most accurate estimates. She believes a dialogue will encourage countries to come to a consensus on the true impact of the pandemic and acknowledge and release better data.

“If data is a public good, it should be open,” she says. “And that is the only way to hold each other accountable.”


Explore Nat Geo

  • Animals
  • Environment
  • History & Culture
  • Science
  • Travel
  • Photography
  • Space
  • Adventure
  • Video

About us


  • Magazines
  • Disney+

Follow us

Copyright © 1996-2015 National Geographic Society. Copyright © 2015-2023 National Geographic Partners, LLC. All rights reserved