As I wrote my column last week, the outlook was mighty grim for Alabama. The University of Washington’s IHME COVID-19 model being used by the federal government, states and just about every media outlet in the land to predict how many will ultimately die from this pandemic, projected a horrifying April for the Yellowhammer State.
The death toll was supposed to skyrocket from less than 20 on April 1 to more than 7,500 by August. That would be an average of about 60 Alabamians dying every day from COVID-19 through most of the summer. It’s a terrifying thought.
In fact, those numbers had Alabama pegged to be one of the biggest tragedies of the pandemic. As al.com’s Ramsey Archibald reported just this past Saturday based on IHME, nearly 10,000 Alabamians might die by mid-May and our per capita and overall death rates would rank near the top when the dust all settled. Nationally IHME was still plotting anywhere between 100,000 and 240,000 deaths, with April seeing most of those deaths.
What a difference a week makes, though.
This past Sunday, UW drastically adjusted down its projections for Alabama to 923 — more than seven times fewer casualties. Were these current projections to prove closer to reality, that would still mean more than 800 Alabamians would die this month, which is tragic, but pales in comparison to the tale of carnage laid out before us a week ago when April was expected to see thousands upon thousands of deaths.
Nationally the story has brightened as well, if we’re living by the IHME projections — which we apparently are. Now 81,700 are projected to die nationwide, and the vast majority of that would happen by the end of April, meaning the daily death totals would start declining rapidly by then.
Also, projections of hospitals being overrun across the country and people dying without respirators have also been ratcheted down severely. IHME no longer projects Alabama running short on hospital beds, ICU beds or ventilators. In other words, according to projections, this state isn’t going to be overwhelmed. Nor is the vast majority of the country.
Projections are just that, and the fact the IHME numbers changed so wildly shouldn’t make any of us feel completely sure these new, less tragic numbers should be chiseled in stone either, but these changes do suggest a couple of things: 1. Social distancing, etc., has helped “flatten the curve.” 2. We should never again put so much faith in projections.
While the professors behind IHME are quick to point out their projections were never meant to be used to set policy, the fact is they absolutely were. Late last week, for example, political critics and the media were bludgeoning Gov. Kay Ivey for not uttering the magic words “stay at home.” Staring down the barrel of 7,500 deaths, Ivey blinked and finally issued the stay-at-home order Friday. Mayor Sandy Stimpson installed a curfew and other restrictions. The political pressure to do all of this was based upon the projections of the carnage to come.
On March 25, Lt. Gov. Will Ainsworth even released a highly critical letter attacking Ivey’s COVID-19 task force for not taking a “realistic view” of the “tsunami” of hospital patients likely to hit the state. Using “simple math and statistics,” Ainsworth projected 245,000 cases of COVID-19 in Alabama and a need for nearly 32,000 hospital beds. He was calling for tripling hospital capacity. Fourteen days later, his letter seems like the histrionics of a guy with a broken calculator.
Media outlets have also trumpeted the IHME numbers, and it’s cool slider technology definitely made the model easy to use. Those projections have been at the core of almost every discussion and political decision made regarding the COVID-19 outbreak, regardless of whether the UW professors ever intended that.
Even as these new, less-horrible projections hit the internet, the media and filter through the thick skulls of those in government, it seems more fair than ever to ask whether they’re any more accurate than those a week ago. The people behind IHME have explained that very little was initially known — which is understandable — and that a massive influx of new information was cause for the adjustments. All that sounds reasonable, and it’s being gobbled up in huge bites by just about everyone. But ….
Picking at IHME’s new numbers Monday morning, I found them still wildly out of sorts with what’s actually happening. For example, by Monday, according to projections, Alabama should have had 89 total deaths. In reality we had 53. That means IHME’s projection was 87 percent too high.
So I ran the numbers around the South. Not much better. Georgia was 40 percent high. Florida 33 percent high. Tennessee 25 percent high. Mississippi was actually 14 percent low and Louisiana was only 5 percent high, which is the closest they got to actual numbers. Even New York was 23 percent high. All of this seems especially troubling given the fact that Monday was just one day after they updated their info.
On top of that, even total death numbers for days prior to Monday — in which they should have had the official totals — were reported higher than what actually happened.
Reporter Jason Johnson talked to them about that specifically, and the IHME spokesman said it was just incorrect info still in the chart and would be corrected, but the wrong numbers had not been used in making projections. OK ….
But their daily death totals have also been extremely high. IHME projected 16 deaths on Monday; we had six. That’s 167 percent high. And their daily projections increase rapidly from here. Five days from the time this article hits the streets, they’re predicting 34 deaths. They project daily death rates in the 40s until things peak around April 24, then fall to two daily victims by May 9.
I fervently hope all these projections turn out wildly high, for obvious reasons. The numbers, after all, are human lives lost.
Still, it’s hard to understand why we are so quick to accept estimates and projections such as these as stone-cold facts when they are frequently so very far off base. In this case our leaders let projections lead us to doing something we’ve never done before — shutting down commerce in this country, derailing millions of businesses and thrusting tens of millions (so far) into unemployment.
Did it do something to help flatten the curve? Probably. Was the curve ever what we were warned it would be? I have serious doubts.
Those who embraced the projections as fact or made policy decisions because of them will certainly claim it was “shelter in place” or “stay at home” that made the difference. In Alabama’s case, less than 24 hours of “stay at home” saved thousands of lives according to IHME, which is pretty impressive.
I’m hopeful now another set of projections will be wildly high — projections of 30 percent unemployment, of another 5 million or 6 million filing for unemployment this week, of bankruptcies, foreclosures, increased homelessness, increased depression, suicides and civil unrest.
Perhaps our leaders can watch closely the numbers they’re using to make these choices over the next week and if what’s really happening isn’t meeting those projections, maybe it’s time to start up the economy again and do what we can to shorten the tsunami of misery that’s coming for tens of millions of Americans.
This page is available to subscribers. Click here to sign in or get access. During the month of December, give (or get) a one year subscription with TWO months FREE.