Thread by @jeremyfaust

There’s one metric that can help us know when it is safe to re-open.

This evening, @CarlosdelRio7 and I propose what that is in The @WashingtonPost.

It is tracking all deaths of ALL CAUSES.

washingtonpost.com/opinions/2020/…

Thread...

First, we want everyone to understand that the death counts from covid-19 are *real and unprecedented*.

These deaths are not mostly “borrowed” from people who would have died anyway.

Nor are they mostly from missed opportunity for care because people are afraid to seek care.

Yes, some of that is happening. But it's not driving the crisis.

Not by a long shot.

Again, the graphic made with the help of the Mass Department of Health and @WashingtonPost says it better than words.

These are ALL deaths.

Note how stable death counts usually are (below)

The number ALL deaths here in Massachusetts, like many states, far exceeds ANY SPIKE IN AMERICAN HISTORY as far back as we can look.

Makes 9/11 look like a blip on the radar.

And yes, the numbers of excess all-cause deaths more or less mirror reported covid-19 deaths now.

It's worth noting that this historic rise in all deaths preceded the time when ANY effects of shelter-in-place orders could have been seen. (i.e. people not seeking care, other downstream costs that unemployment will cause).

For example, in Massachusetts, 60% of all covid-19 deaths occurred in nursing homes.

Those numbers FAR exceed usual nursing home deaths which, it bears noting, are basically immune to the deniers might say *could* be causing deaths for others by having the economy closed.

Excess deaths are occurring in many states, nationwide.

That's never happened. Not in any flu season. And not at this magnitude.

We've seen all-cause mortality go up in some groups before.

Like young men in the 1980s and 1990s.

This is causing FAR more excess deaths.

And we also know that WHAT WE DO will determine how long it takes for the expected number of deaths to come back to normal.

This is where tracking excess deaths will not only help us not open TOO SOON, but also, interestingly enough, it will help us not open TOO LATE! Why? ...

Because we know we will continue to test for SARS-CoV-2.

We WILL find people dying with covid-19.

But if the number of excess deaths abates, we will be able to say that the effect it is having on society is no longer "truly extraordinary."

Which it is now.

Let's imagine that in a couple of months, we are back to baseline of "expected all deaths."

If we still have people testing positive and dying with covid-19, we may fail to notice that these are deaths which, while sad, were expected to occur anyway.

An example...

If we have NO excess deaths for a month...and someone dies of stage IV cancer (as sadly expected), and they test + for SARS=CoV-2, we will have to choose which disease "gets the blame."

Docs are bad at that. We get that.

That's the POWER of tracking all cause excess mortality.

If and when we get back to "baseline death rate," we will be can be a lot less reliant on "covid-19 death counts" which, over time, may be harder to see as reliable. When you don't have thousands of extra deaths, it's harder to be sure that what you're seeing is "real."

In sum, we want to safely re-open the economy.

Following excess deaths is the least "political" way to do this.

You can't fudge it.

More people died. Or they didn't.

Currently, and please hear this, SO MANY MORE THAN EVER BEFORE ARE DYING.

When excess mortality goes away, it'll be important to notice that.

And to *not* blame every single death on covid-19 (ie. people die WITH the virus, not OF it).

At that stage, deaths from cancer or heart disease or other usual things that kill us all the time, should get their full due.

If we don't, we will stay closed too long (!).

So the punchline:

Excess deaths tells us when it is too SOON to open (now).

And it's disappearance will help us know when it's actually TIME to open (later).

Read my @PostOpinions piece with @carlosdelrio for more detail. Thanks for reading. fin.

washingtonpost.com/opinions/2020/…