Thread by @SethAbramson

(THREAD) This is #FakeNews intended to understate the lethality of coronavirus and force an end to social distancing. It will get people killed.

Case Fatality Rate (CFR) is deaths divided by closed cases (active cases may still result in death).

The US CFR is 1080/ 2943—or 37%.

https://twitter.com/johncardillo/status/1243206904882528256

1/ If a man with *no morals* wanted to outrageously understate the CFR for coronavirus, what he'd do is divide deaths by *all* US coronavirus cases—thus fraudulently assuming *every active case* will result in full recovery.

Calculated that way, the US CFR is 1080/75066—or 1.44%

2/ If a man with *few morals* wanted to wildly understate the CFR for coronavirus, what he'd do is look at research suggesting 50% of coronavirus cases are asymptomatic, *then* assume all *estimated* cases will result in full recovery.

The US CFR would then be 1080/150132—0.72%.

3/ But that man with few morals—swayed by research as he pretends to be, despite using it to devise fake CFRs that include active cases—would *also* have to consider research suggesting 30% to 70% of all Americans will get sick.

A .72% CFR with 30% infection is *712,800 deaths*.

4/ So the question is this: when a putrid piece of human garbage like the poster below posts data so *off the charts fake* it wouldn't even be posted by a genocidal maniac hoping to kill Americans, what's the motive?

Simple: re-elect Donald Trump over a mountain of human bodies.

5/ Now some of you will say—quite rightly!—that CFRs can only *really* be calculated *after* a pandemic, when all cases are "closed" and a reasonable estimation of untested cases can be made. So—accepting the limited utility of a CFR *now*—how would we calculate it? Here's how.

6/ Because coronavirus has a *very* long "tail"—post-infection it takes a *very* long time for a positive case to become a *closed* case (i.e. we're sure that the infection has run its course)—one could *estimate* CFR by this equation: deaths divided by total cases *14 days ago*.

7/ 14 days ago, we had 1,630 confirmed cases of coronavirus. But testing was scarce—and again, some research says 50% of cases are asymptomatic and won't be tested. So we don't know what our denominator. 1,080 deaths divided by *what*? Even if we say *50,000*, the US CFR is 2.2%.

8/ As we don't have a denominator, we might instead look at *global* rates—feeling safe there, as the US has handled this crisis worse than almost any country, so global data should be *conservative* as a stand-in for US data. Right now the deaths/closed-case CFR globally is 16%.

9/ But again, say you had *no morals* and decided to fraudulently state *all active cases will recover*. In that case you'd divide global deaths (22,339) by cases worldwide (503,071)—a 4.44% CFR.

If 30% of America gets infected with a 4.44% CFR... 4.4 million Americans will die.

10/ The goal of this thread was to give us *resources* to fight back against genocidal maniacs like @johncardillo who care only about their stock portfolios—not *millions of innocents being killed*. Use whatever metric you like—I've gone through many here—to call such people out.

PS/ People falsely say that as more cases close, the global CFR goes down. Maybe eventually, but not yet. With *143,566* cases closed worldwide, the CFR is 16%—and has been *rising* for *weeks*, from about 4% originally. So please don't say that global CFR looks at too few cases.

PS2/ Folks *also* falsely say increased testing lowers CFR. Again, maybe eventually—but not yet. Even if you use *total cases*—which CFR can't/doesn't—the mortality rate in South Korea (one of the top 3 testing nations) has been *rising* daily. It was 0.7%; then 1.24%; now 1.39%.

PS3/ The simple fact is, the CFR for coronavirus—however you look at it—is *bad*.

It just is. It's *bad*.

Coronavirus is *many times more deadly* than the flu and *many times more contagious* than the flu and *there is no vaccine for it*—unlike the flu.

Don't get it twisted.

PS4/ So:

1⃣ Always point out CFR can't be known yet.
2⃣ Never allow someone to use active cases in a denominator.
3⃣ Always—for denominator—use case data from a period of time in the past matching the virus's "tail."
4⃣ Consider testing frequency.
5⃣ Consider asymptomatic cases.

PS5/ Moreover:

6⃣ Do not compare the U.S. to *any* country (e.g. Germany, South Korea) so far ahead of us in testing and tracking we *know* our CFR will be worse.
7⃣ Do not compare coronavirus to *any* public health threat for which there is a *vaccine* (e.g., the seasonal flu).

PS/ And also:

8⃣ Always consider *infection* rate.
9⃣ Always consider *total deaths*—not just mortality rate—to fully gauge the impact of the virus.
🔟 Understand—as to total deaths—that coronavirus will be an *annual* event until we have a vaccine. So don't just look at 1 year.

NOTE/ I know Twitter is a bad place for math-heavy threads that consider the "technical" definition of things like CFR, particularly at a time that rate can't be calculated in a useful way. But we must *try* to talk this way—or liars like John Cardillo win and endanger all of us.