Coronavirus Case Counts And Death Tolls Are Likely Undercounts : Goats and Soda

The U.S. has the most coronavirus deaths of any country in the world — on May 11, the death toll passed 80,000.

And that’s likely an undercount.

“Almost certainly it’s higher,” said Dr. Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, at a virtual Senate hearing on Tuesday. “There may have been people who died at home who were not counted as COVID because they never really got to the hospital.”

It’s not just the death toll that’s likely higher. Medical statisticians say we have been undercounting cases since the pandemic started — not only in the U.S. but also around the world. That’s because in countries rich and poor and in between, people who contract COVID-19 are not necessarily diagnosed — because of the shortage of tests and the difficulty of getting tested in some parts of the world, especially remote spots.

So you should keep in mind certain caveats as you consider the endless stream of numbers.

Here’s what we know about some of the most frequently discussed data points.

Total confirmed cases

What we want this to tell us: How many people have had COVID-19.

What it actually tells us: How many people have tested positive for the live virus (the most common way of being diagnosed) or, in some cases, been diagnosed based on symptoms. The actual number of people who have had COVID-19 is probably higher.

Caveats:

• Access to testing is limited in many parts of the world, so not everyone with COVID-19 symptoms is confirmed and counted.
• Official counts of “confirmed cases” miss people with no symptoms or mild symptoms who may not seek testing or medical care.
• The swab tests are considered to be generally reliable but aren’t fail-safe. There’s a margin of error from mistakes in sampling or lab processing techniques as well as from faulty tests.

Daily case counts

What we want this to tell us: Is the number of new cases of COVID-19 rising or falling over time?

What it tells us: How many cases are reported on each day and how that number compares with previous days. Only in areas where large numbers of people are tested regularly could the curve of daily case counts be considered a reliable marker of spread or containment.

Caveats:

• Testing delays and backlogs in many countries means that a case may be reported days or weeks after a sample was taken. It can also take a few days for a confirmed case to be reported to national and global health authorities. So the number of new cases reported on a specific date may reflect both new and recent cases rather than the new case count on one specific date. 

• Countries don’t always use the same definition for “confirmed cases,” and some have changed their criteria as the pandemic has evolved. On April 14, the U.S. Centers for Disease Control and Prevention began including people presumed to have COVID-19 based on symptoms in their daily tally of confirmed cases even if the individuals had not been tested. China has revised its case definitions for infected people multiple times, briefly including diagnosis by symptoms in one province, then returning to a requirement for a lab test to confirm. 

Total deaths

What we want this to tell us: How many people have died from COVID-19 from the start of the outbreak until now.

What it tells us: How many people have died with the official cause of death listed as “COVID-19” — excluding those who died of the disease but were not identified as COVID-19 casualties. These numbers generally represent a conservative count of deaths and will likely be revised upward on review. But they are probably more accurate than case counts. “The [under]reporting issue for death numbers is less severe than case numbers, but it still exists,” says Sen Pei, a public health research scientist at Columbia University.

Caveats:

• As with testing, issues with varying definitions and delays in reporting also apply to deaths. In Wuhan, China, for example, where the first COVID-19 cases were discovered, the death tally jumped by 50% overnight in mid-April. WHO says this is because a review of past records found 1,290 deaths had been missed in the official count because of patients dying at home, medical staff too busy to fill out paperwork at the time and incomplete reporting from field hospitals.

Death rate for coronavirus patients

What we want this to tell us: How likely are people who get COVID-19 to die from it?

What it tells us: The total number of reported deaths divided by the total number of confirmed cases — both of which could be undercounts.

Caveats:

• In the middle of an epidemic, the death rate changes day by day based on the number of reported cases and deaths. Death rates could decrease over time as more survivors are identified. 
• It’s difficult to meaningfully compare death rates across countries because of variable factors including the extent of testing and the demographics of the population (countries like Italy, which have a greater proportion of older people, have seen high death rates). 

Recovery rate for coronavirus patients

What we want this to tell us: How many people who have contracted COVID-19 recover.

What it tells us: How many people who have been officially diagnosed with COVID-19 have kicked the virus.

Caveats:

• The number of confirmed cases may be missing many people with mild cases, who tend to recover — so the total count of recovered people is probably higher. 
• The definition of a recovered case varies. The World Health Organization states there should be two negative tests at least 24 hours apart while the Centers for Disease Control and Prevention also offers a symptom-based definition in places where tests are not readily available. Neither definition takes into account long-term physical and mental health effects, such as reduced lung function from scarring or memory issues from treatment drugs, especially in people who have recovered from severe cases.