How Many People Have Really Died Because of Coronavirus?

Researchers are struggling to calculate mortality statistics as the pandemic intensifies. Here is the true toll of the novel-coronavirus-related COVID-19 pandemic and how they measure it!

In early March, Andrew Noymer, a demographer (population statistician) at the University of California, Irvine, began to have doubts. He was watching countries in Europe and North America record their first deaths caused by COVID-19 and knew there could be problems with these data.

Andrew Noymer believes that, even in an ordinary winter, some deaths from a well-known disease such as flu (influenza) are misclassified as bacterial pneumonia. Unreported COVID-19 deaths were therefore inevitable, and explaining this to people would be very difficult.

His suspicions were confirmed in March and April, when national statistical offices began publishing death counts: the outbreak was killing far more people than the COVID-19 figures alone suggested.

In times of upheaval such as wars, natural disasters and epidemics, researchers need to calculate deaths quickly and often turn to an imprecise but reliable measurement system: excess deaths!

Excess deaths are typically defined as the difference between the number of deaths observed during particular periods and the number expected during those same periods (excess deaths = observed – expected). This compares expected deaths with what actually happened and, for many scientists, is the most robust way of measuring the pandemic’s impact. It can help epidemiologists make comparisons between countries and, because it can be calculated quickly, can identify even deaths only distantly associated with COVID-19 that would otherwise go undetected. Data from more than 30 countries where excess-death estimates are available show that, between the start of the pandemic and the end of July, these countries experienced approximately 600,000 more deaths than would normally have been predicted (413,041 of them were officially attributed to COVID-19).

However, this calculation of excess deaths also has several shortcomings.

• It cannot distinguish people who died from COVID-19 from those who succumbed to other pandemic-related factors, such as not seeking medical help or disruptions to routine health services that could delay treatment.
• It depends on accurate and timely reporting of deaths, which can be limited by underdeveloped death-registration systems or even deliberately kept secret.
• And, as with so much else in the pandemic, statistics have become politicized, becoming a way for countries to claim superiority over one another.

Experts worry that simple excess-death reports are leading to premature or erroneous comparisons of countries’ responses to the pandemic and that, because of missing data, the situation in low- and middle-income countries is largely being overlooked.

There are more sophisticated ways of classifying mortality to determine how many people died directly from SARS-CoV-2 (the novel coronavirus) infection and how many died from other pandemic-related factors. Eventually, demographers and public health researchers will obtain the necessary information from death certificates. They will then be able to assess which interventions worked best to inform responses to future pandemics.

Many media organizations are already calculating the data and drawing such conclusions. Some statisticians argue that, as the first wave of the pandemic subsides in many places, comparisons between government policies can and should be made to see how they might have affected mortality. But many experts say it is still too early in the outbreak to do this rigorously. Jennifer Dowd, a demographer and epidemiologist at the University of Oxford in the United Kingdom, says this process can be distorted because some early outbreaks spread at random while others die out, making the analysis a complicated, long road until the pandemic has run its course.

A blunt knife

As deaths began spreading across Europe, Lasse Vestergaard was among the first to notice. Vestergaard, an epidemiologist at the Statens Serum Institut in Copenhagen, brings together weekly all-cause mortality data from 24 European countries or regions through the European Mortality Monitoring Project (EuroMOMO), which he leads. Between March and April, EuroMOMO’s tracker showed tens of thousands more deaths than expected (approximately 25% more than the official COVID-19 death figures). Infections were being missed because of a lack of testing and because different countries counted deaths in different ways, for example by excluding deaths in care homes. Getting a true sense of how countries were doing was almost impossible.

Researchers, journalists and politicians therefore turned to calculations of excess-death data. Rather than getting bogged down in the cause, this system compares all deaths in a particular week or month with the deaths statisticians estimate would have occurred without the pandemic, usually by averaging the previous five years. More sophisticated versions model how a population ages or changes as a result of immigration and emigration, but these additions can make countries harder to compare. Some excess-death analyses, such as a July 30 report published by the UK Office for National Statistics, standardize mortality rates to account for differences in population age structure between countries. Because officials can record a death relatively quickly if they do not record its cause at the time it occurs, these statistics can be compiled much faster than cause-specific data.

Nature collected figures from various databases used by demographers and from trackers run by The Financial Times and The Economist, two of the most comprehensive datasets on excess deaths. Although not universal, the data list 32 countries, largely in Europe, and 4 major world cities, include many countries with large outbreaks, and account for approximately two-thirds of the official COVID-19 death toll up to the end of July.

Nature conducted an analysis showing very large differences in excess-death data between countries. In the United States and Spain, the two worst-affected countries so far, approximately 25% and 35% of excess deaths, respectively, are not reflected in official COVID-19 mortality statistics. The mismatch is much greater in places such as Peru, where 74% of excess deaths are not explained by reported COVID-19 deaths. And some countries, such as Bulgaria, have so far experienced negative excess deaths during the outbreak, meaning that fewer people have died this year than expected despite the virus.

Picking apart the data

For most demographers, using excess mortality as a blunt knife is the best tool available during the outbreak. As time passes, they will be able to use hindsight and more detailed data to better understand the damage caused by the pandemic.

Eventually, deaths could be divided into 3 categories:

1. Direct deaths for which COVID-19 is recorded as the cause
2. Deaths directly caused by the virus but not officially recorded
3. Indirect deaths due to other changes caused by the pandemic

Recorded deaths directly attributable to coronavirus are usually found in pandemic trackers showing case and death counts updated daily by local and national health authorities. But Maimuna Majumder, an epidemiologist at Harvard Medical School in Boston, Massachusetts, warns that even this number is not as precise as it seems.

It can be difficult to distinguish people who died from COVID-19 from those who were infected but died from other causes. Majumder asks, "This will be a very critical part of all of this. If you have two simultaneous conditions, what is that classified as?" Majumder says separating these deaths will require a mortality classification system that accounts for underlying conditions that increase the likelihood of COVID-19 killing someone. Such a system means waiting for cause-of-death data, which take about a year to compile fully.

Researchers are already looking back at the first 6 months of the outbreak and adding deaths that were misclassified at the time. Several major outbreaks, including those in Wuhan, China, and New York City, revised their death tolls upward in April to account for deaths suspected to have been incorrectly coded.

There are also deaths directly caused by coronavirus that were not counted because the individual presented with symptoms that were not recognized as COVID-19. Natalie Dean, a biostatistician at the University of Florida in Gainesville, says, "We are still trying to understand exactly how the disease presents". Strokes and pulmonary embolisms (the clinical condition caused by partial or complete blockage by a clot of an artery carrying blood from the heart to the lungs) are two potentially fatal complications of the virus that might initially have been overlooked, she says.

A small proportion of excess deaths are indirect, resulting from conditions created by the pandemic’s impact rather than by the virus itself. Some hospitals report that patients with cancer and chronic conditions are missing regular checkups, potentially endangering their own health. Reports of domestic violence have increased in some places, and mental health researchers are concerned about frontline workers and people living under lockdown, but it is not yet clear whether the number of deaths has increased as a result.

- Related topic: How Can We Cope with the ANXIETY Caused by Coronavirus?

According to a report from the US Centers for Disease Control and Prevention (US-CDC), visits to emergency departments in the United States fell by more than 40% in the early days of the pandemic. Majumder says hospitals are overstretched even when people need health care. "You died from something else, but the reason you died from something else is that the systems initially set up to care for you are no longer strong enough." Early, incomplete CDC data offer a glimpse of these indirect deaths: in April in the United States, deaths from diabetes were 20-45% higher than the 5-year average; deaths from ischemic heart disease were 6% to 29% higher than normal.

Situations in which mortality rates declined

On the positive side, staying at home and behavioral changes such as mask-wearing and handwashing may have prevented deaths from other causes, particularly other infectious diseases such as flu. And with so many people around the world staying home, deaths from traffic accidents and interpersonal violence are likely to have fallen. These declines may be concealing some of the increase in deaths caused by COVID-19.

Some of these effects have already begun to appear in the data. The global surveillance system FluNet found that this year’s flu season was shortened by more than a month, probably because of staying at home and increased hygiene practices. In South Africa, a death-tracking system introduced at the height of the country’s AIDS epidemic allows epidemiologists to distinguish deaths from natural causes such as disease from those due to unnatural causes such as interpersonal violence. A team led by Debbie Bradshaw at the South African Medical Research Council in Cape Town showed that unnatural deaths fell to half their normal number following lockdown measures in late March. When lockdown rules began to ease in late May, these deaths returned to their expected levels.

Noymer says demographers will probably never know the outbreak’s final toll with certainty. "You cannot scratch a lottery ticket and find the real values underneath the gray plastic coating." Once the pandemic subsides, untangling the three types of deaths and determining how many would have occurred without the virus will take months or even years. "We have not even settled on how many people died in the great Spanish Flu Pandemic of 1918, and we have had 100 years to work out the numbers." he says.

Taking stock

For now, statistics on excess deaths are helping chart the course of the outbreak in different places. In the future, with complete cause-of-death data, researchers will be able to analyze the effects of lockdown and other interventions by looking at countries’ levels of direct and indirect mortality. Noymer says doing this now, while the outbreak continues and the final toll is unknown, is risky. "We do not yet have a realistic perspective because we are still on the rising wave. It is like trying to estimate the total rainfall in the middle of a hurricane." he says.

Even after the storm, these analyses will be possible only in high-income countries with robust systems for recording births and deaths known as civil registration and vital statistics (CRVS). Irina Dincu, an expert in such a program at the International Development Research Centre in Ottawa, says counting deaths in low- and middle-income countries is much easier. "Worldwide, approximately 50% of deaths that occur in a given year are registered. The remaining 50% do not exist; they are invisible." she says.

CRVS consultant Gloria Mathenge can think of many reasons why these deaths remain hidden. In her role at the Pacific Community, an international development organization in Nouméa, New Caledonia, Mathenge helps strengthen data systems in Pacific island countries such as Kiribati and Tonga. Even as the situation improves, an average of 20% of deaths in the region remain unregistered.

Many existing systems date back to their countries’ colonial pasts. Mathenge says they consequently do not reflect contemporary cultural and social standards, such as the fact that many people in low- and middle-income countries do not die in hospitals. Thus, in addition to missing COVID-19 deaths, there is no reliable way to establish a baseline for calculating the excess.

Stéphane Helleringer, a demographer at New York University in Abu Dhabi, says that to estimate death tolls in these countries, demographers will have to rely on less precise methods such as door-to-door surveys. "But these happen infrequently, and by the time we do them, they are already out of date." he says.

For some demographers, it does not matter whether someone died from the disease itself or because the health care system was pushed beyond its capacity; in their view, all deaths can in some way be attributed to the pandemic. Noymer says, "At some point, you have to say, ‘This is somehow related to COVID, because COVID disrupted the health system.’" He considers the question of what 2020 would have looked like if this had never happened to be a thought experiment.

Other researchers are keen to separate deaths directly caused by the virus from those due to other outbreak-related circumstances so that they can build an accurate picture of how lethal the virus itself really is.

But there is little time for such analyses at present. The pandemic has greatly increased the burden on death-registration systems and the scrutiny they face. Politicians, the media and the public demand daily or weekly statistics that would normally take months to compile. This makes some demographers uneasy. Noymer says that, because of the time and labor needed to compile and analyze death certificates, "We understand mortality best by looking back. But everyone wants to know it in real time."

*

- All Our Covid-19 Articles -

*

Manage preferences to view external content

Manage preferences to view external content