Coronavirus Mortality: What's the Denominator?
David L. Katz, MD, MPH
CMO, Tangelo. Founder: Diet ID; True Health Initiative. Founding Director, Yale-Griffin PRC (1998-2019). Health Journalist. COVID Curmudgeon
How lethal is the coronavirus, compared to, say, the flu?
The only way to know a rate- any rate- is to know both the numerator (i.e., those who suffer the outcome), and the denominator (i.e., the population at risk/exposed).
As I noted in my first column about COVID-19 (or whatever we have started calling it today...), we were very unlikely to overlook deaths from the virus, but very likely to overlook mild cases.
The global epidemiologic data now suggest exactly that. Look at the just-updated tabulation of world data above (or visit https://www.worldometers.info/coronavirus/ to get it at the source), and something pretty dramatic jumps off the page from the top three rows. Namely, Italy and South Korea have total cases in much the same ballpark (~9200 in Italy, 7500 in South Korea)- but Italy has ten times as many deaths.
Does this mean the coronavirus is ten times more lethal to Italians than Koreans? ALMOST CERTAINLY NOT. There is no plausible explanation for any such discrepancy.
What it almost certainly means is that Italy actually already has ten times as many cases as they've detected.
Why the difference? Because South Korea is thus far the only country to implement a rigorous surveillance program, looking for cases in the total population. They are finding a much higher percentage of total cases than any other country.
And their data are importantly reassuring. Rather than a fatality rate above 3%, the data from South Korea suggest a fatality rate of 0.6%, one fifth as much. Compared to some of the high-end mortality statistics I've seen cited, it's an order of magnitude less. And, it puts the fatality rate of COVID-19 much closer to that of...the flu. A condition we take for granted. There are millions and millions of flu cases around the world right now, and many tens of thousands of deaths- both orders of magnitude greater than the stats attached to the coronavirus. But how many major events have been canceled, how many schools closed for flu? To the best of my knowledge, none.
I am not saying the coronavirus pandemic isn't serious; obviously, it is. But our actions reflect fear not only of getting the infection, but dying from it. The world's best data suggest that many people, perhaps even most, get the infection without even knowing they have it.
Yes, the disease can kill, clearly- and the elderly, frail, and chronically ill are at special risk. But if the South Korea data generalize, as they probably do- then the likelihood of recovery should be something like 99.4%; for flu, it is better, at 99.9%. But many more people have the flu, and are not quarantined- so they are busy spreading it around. The odds are very long indeed that the death toll from the coronavirus will come anywhere near this year's death toll from flu- which, tragically, and unlike coronavirus, includes children.
Trained in epidemiology, I always ask: what's the denominator? For coronavirus, the answer, courtesy of a robust public health system in South Korea- is a source of comfort in a time of woe.
-fin
Dr. David L. Katz is a board-certified specialist in Preventive Medicine/Public Health
Head of Operations and Talent, Diviner
4 年David L. Katz, MD, MPH, FACPM, FACP, FACLM, thanks again for your thoughtful post. However, things may be more severe than you stated above (on March 9 2020): "The odds are very long indeed that the death toll from the coronavirus will come anywhere near this year's death toll from flu." The CDC estimates 50,000 flu deaths this season. Dr. Anthony Fauci today (March 29 2020) stated that for COVID-19, "Looking at what we're seeing now, I would say between 100,000 and 200,000 cases... excuse me, deaths." Have your expectations changed over the last month to be consistent with his, or are you still confident that COVID-19 will be no more lethal than the flu?
BD Drug Discovery Consulting
4 年Are you suggesting that all people who get the flu are known documented but those with Coronavirus are not? What percentage of people with flu actually seek medical attention?
Physician at UC Davis Volunteer Clinical Faculty
4 年Dr. Katz's emphasis on the underestimation of denominator in COVID-19 reported case fatality rates is right on, but the degree of underestimation is even greater than he suggests. The test that is presently being done to detect COVID-19 infection is one which is only positive if live COVID-19 virus is positive in nasal secretions. If individuals have recovered from COVID-19 and are no longer infectious, the test will be negative. The denominator in calculations of COVID-19 case fatality rates should include both cases in which virus is detected in nasal secretions and cases in which people are no longer secreting the virus, but have formed antibodies. The latter category of patients probably vastly outnumbers the former.? The generally accepted case fatality rate for influenza in the United States is 0.1%, but the denominator in this calculation includes an estimate of the total number of influenza cases, not just those who have tested positive. Another way to compare the lethality of different infectious agents, and one that is commonly used in epidemiology, is to compare mortality rates expressed as the number of deaths attributable to different agents per 100,000 population. Approximately 34,000 people have died of influenza in the United States this flu season, which is about average as compared with past years. About 3,200 deaths due to COVID-19 have been reported in China, where the epidemic seems to have ended. The mortality rate for the flu in our country of 330 million people, therefore, is 10.3 deaths per 100,000 population, which is 45 times higher than the COVID-19 mortality rate of 0.23 deaths per 100,000 population in China, a country of about 1.4 billion people. Even if it’s assumed that all 3,200 Chinese COVID-19 deaths occurred in China’s Hubei province, where the epidemic began, the mortality rate in this province of 58.5 million people is 5.5 per 100,000 population, which is about half the U.S. influenza mortality rate.
Chief Technology Officer
4 年Since you are an M.D, I was wondering if you could explain to me how the medical community computes the denominator for the flu. We don't test everyone for that either do we?
Occupational & Environmental Medicine Physician
4 年There is hope that Plaquanil, an anti malaria drug, may help effectively fight the virus