Gettin’ sick (of this image).

Please, oh please, stop publishing predictions for when this will “peak”.

They’re not useful.

Every time the local paper talks to the county health administrator, they cite the same prediction that the county will have “up to 1,500 patients” in the hospitals with COVID-19 “in about two weeks”.


That’s just insane.

If you check the county’s dashboard, you’ll see we’ve had a steady 85-ish active cases for two weeks already, with a 15% hospitalization rate so that means we have 10-15 patients in hospital now and have had roughly the same number for the past two weeks.

There’s been no growth in test-confirmed cases and the hospitalization rate has actually been going down; it would have gone up if we’d started missing lots of un-hospitalized cases, so this is likely trustworthy.

There’s no hint in the published data that would support expecting a massive surge in infections.

So, clearly they think something is going to change if they think that soon we’re gonna have 100 times that number.

But what?

Dunno; she never says and the reporters don’t ask.

(And there’s no hint in the data that this county is in anything other than a steady state.)

That’s not useful.

Nor are the “models” that are using normal curve fits to predict a peak.

There is no reason to think the dropoff of cases will be a “normal” curve, symetrical to the growth curve.

Fitting an arbitrary curve onto the known growth rate and claiming that’s a predictive model isn’t useful either.

So, just stop it.

I understand that you’re all antsy to have some idea when and how this will progress, but badly-supported guesses masquerading as predictions just make things worse.

We need to trust that officials and the media are serious people doing their best, and we can’t when you’re dealing in unsupported numbers.

Leave a Reply