Covid-19 Calculations: Innumeracy Revisited
Roger LeBlanc (author of The Punter’s Tale)
I’m not a mathematician. Or a scientist. But I’ve spent my life immersed in numbers and calculating probabilities. As a public service I’d like to offer this important message: a little math can take you a long way.
My first book, The Lazy Bettor’s Guide to the Kentucky Derby, appears to be held together by its paperback binding, but it’s the figures and formulas within the pages that keep the book from falling apart. After spending decades calculating fractional and final times of horse races and using them to objectively define “mediocrity” and “excellence,” I developed an appreciation for math’s ability to reveal truth.
And that brings me to Covid-19 and a months-long debate with a conservative friend. It started in early February, a few weeks after China quarantined Wuhan City (pop. 11 million). That number, 11 million, prompted me to start wearing protective gloves at work. And my actions prompted him to roll his eyeballs and proclaim that coronavirus was just another flu bug.
His main objection to the growing panic amounted to this:
“As soon as we have widespread testing in the U.S., the 5% death rate will drop to about half a percent, the same as the flu.”
He reasoned that, as testing became more widely available, comparatively healthy people outside of hospitals and nursing homes would test positive and not die. Although I agreed the death rate would drop, I doubted it would drop to flu-like levels. When the death toll in the U.S. reached 25,000 (about 6 weeks later) and he clung to his opinion, I worked through some back-of-the-envelope calculations with him. Just some rough math is all anyone should need to clarify the crisis we face.
Valuing Slapdash Math over Knee-Jerk Opinions
Putting aside the fact that my friend was contradicting his hero, President Trump, by suggesting in late March that widespread testing was not yet available, I focused on the number 25,000. I placed that number in my context of a high death rate and in his context of a low death rate:
- High death rate and contained spread. I assumed the severity of the virus was twice as bad as the flu and would kill about 1% of those afflicted with it. I also assumed, optimistically, each carrier would infect only two people before being quarantined. If quarantines, shutdowns and contact tracing were used to slow the spread, the math projected this result:
25,000 deaths = 1% of 2.5 million infected Americans
2x spread results in 34 million (about 10% of U.S. population) infected = 340,000 deaths
The spread to 10% of the population would leave us reeling from a death total about four times higher than the worst flu season in the past 100 years. According to this math coronavirus deserved to be met with immediate, unprecedented preventive measures.
- Low death rate but a wide spread. If my friend was right that the virus should just be allowed to spread (affecting, say, 25% of the population) and that we should not be worried about its low assumed death rate of .5%, the numbers would look like this:
85 million infected x .5 = 425,000 deaths
What my friend failed to realize was that his “Don’t Worry, Do Nothing” reasoning led to a more deadly result than I was warning of!
Assuming widespread shutdowns in the economy might improve my projections, I predicted the U.S. would surpass the 100,000 death total by July 4. (That number was surpassed in mid-June.) And I told him that 200,000 deaths by the end of 2020 was likely.
Now three months and 110,000 deaths after our talk, the hard mathematical truth still hasn’t penetrated his even harder head. My friend is taking verbal victory laps, convinced he was right. A growing number of medical experts admit the death rate is far lower than originally anticipated, and it might drop lower still. He sees this as confirmation that the dangers of Covid-19 were overhyped.
(As I wrote this article, news broke that President Trump wore a mask in public. So maybe my friend and the “Don’t Worry” crowd will soon pay attention to the math.)
Countering Claims of Inflated Death Totals
The “Don’t Worry” crowd dismisses mathematical models that show the catastrophe continuing. They claim the death total is overstated (allegedly by medical-industry folks attempting to increase access to emergency funds and research grants).
Here are death totals for the U.S. in the early months of this pandemic:
- January: zero (Zero cannot be an overcount!)
- February: one (One cannot be an overcount!)
- March: 5,200
- April: 59,000
As of March 8, there were still 12 U.S. states reporting zero cases of coronavirus. Were there really none, or were these states not yet set up to track them? Given the CDC’s delay (until the end of February) in distributing test kits, it’s tough to imagine any state had implemented fully functional testing before mid-March.
When counting was (perhaps) fully in place nationwide in April, the monthly death count rocketed to 59,000. That suggests thousands of Covid-19 deaths from January through March went unreported. So we can safely assume 134,000 deaths (as of early July) is an undercount.
Bolstering my undercount assumption is a study by the Yale School of Public Health showing 15,400 excess deaths in the U.S. from March 1 through April 4 (when comparing historical numbers for that period to 2020). Although not all the excess deaths were caused by the virus, that’s twice as many deaths as the 8,100 actually attributed to Covid-19. And again, January and February were left at essentially zero Covid-19 deaths.
Projecting the 2020 Death Total in the “Don’t Worry” Scenario
Let’s take the Trumpian approach of assuming a low (.5%) death rate. This means that 26,800,000 Americans (about 8% of the population) have been infected in the first half of 2020, with 134,000 of them (.5%) dying from the virus.
In the following example, I use an optimistic spread rate of just 2.5 people infected by each carrier. I bumped that up from the earlier 2.0 figure because shutdowns are ending in most states and resistance to preventive measures is growing.
Applying a low spread rate to the low death rate assumed by the “Don’t Worry” crowd, we can expect one-third of the U.S. population to be infected by year end:
26,800,000 x 2.5 = 67,000,000
67,000,000 + the original 26,800,000 = 93,800,000
Add to the 93,800,000 figure all those infected before year end by the additional 67,000,000 carriers. That brings us to well over 100 million Americans testing positive for Covid-19 by year end.
Here’s the year-end death toll in the “Don’t Worry, Do Nothing” scenario:
100 million x .5% = 500,000 Covid-19 deaths
So the scenario and course of action that Trump and other shutdown dissenters are pushing results in a half million Americans dying from Covid-19 in 2020. Even if we chop the death-rate assumption in half to .25%, deaths in their scenario will surpass 250,000 for the year.
Absent from these calculations are the premature passings of Covid-19 sufferers listed as “Recovered.” Many victims have been left with serious lung, heart and liver damage that will shorten their lives by a little or a lot. That figure, known as the morbidity rate, has yet to be calculated, which means the true death total remains significantly undercounted.
It saddens me to say the combination of Covid-19 lethality and American innumeracy adds up to prolonged catastrophe. Masks and social distancing can help solve this problem, but they’d cure it faster if everyone understood the math.
