R(t) is the "real-world" number of new infections spread per infected person, with mitigations (vaccines, masks)
My doctor recommended a blog called “Your Local Epidemiologist” to me. The author describes herself thus:
My name is Dr. Katelyn Jetelina. I have a Masters in Public Health and PhD in Epidemiology and Biostatistics.
During the day, I’m an Assistant Professor at a School of Public Health where my research lab resides and where I teach. I have a secondary appointment at a medical school too.
Methinks the lady is entitled to an opinion.
The picture at the top of this blog is from her blog of December 17th, but the projection was calculated December 14th. Since she credits someone else’s Twitter account with the graphic, I assume it’s available for fair use.
Now here’s what’s absolutely false about that graph — the United States does not have the capacity to perform a million tests a day. We can’t make the test kits that fast, and more importantly the lab capacity does not exist. Depending on what R(t) turns out to be — right now it looks like R(t) = 4 is a good guess, so closer to the red dashed line than the yellow one — somewhere around Christmas day the United States will lose the capacity to know just how bad it is. We will no longer be able to keep up with the spread.
She estimates that the peak in the United States will be mid-January, and then hospitalizations lag infections by a week or two, and deaths lag hospitalizations by two or three weeks. If you’ve been reading the Anti-vax chronicles, you’ve seen how this plays out. If we’re going to hit a million cases a day (which we won’t accurately measure) around New Year’s day, the peak will be several million a day. Two, three, four — we’ll never know.
Let’s guess three million cases a day.
Dr. Jetelina also notes that South African data suggests that the rate of hospitalization is less with Omicron. She states without showing her work (*sigh*, but it’s a spare-time blog) that Omicron causes 45% of the hospitalizations of Delta. Per the CDC data, recent weeks have seen about 122,000 new cases (mainly Delta) a day, with about 7800 hospitalizations.
A little math, then (I will show my work):
(3,000,000/122,000) = 24.5. My guess is that at the peak we’ll have almost 25 times as many cases per day as we do now.
(122,000/7,800) = 0.0639. Right now, about 6.4% of cases result in hospitalization.
0.0639*0.45 = 0.0288. If she’s right about the 45%, only 2.8% of cases of Omicron will result in hospitalization.
3,000,000*0.0288=86,311 hospitalizations. Per day. For an average stay of 8.4 days if not admitted to ICU and 16.2 days if you are.
The total number of hospital beds in the United States is about 900,000. In normal times maybe 500,000 or 600,000 of those beds are full with people sick with normal everyday stuff. There’s no way to fit 8*86,000=688,000 Omicron patients into our health care system. When the system is overwhelmed, the case fatality rate goes up.
So mid-January, look for 3 million new cases a day (which will not be accurately measured), 86,000 hospitalizations a day trailing that by a week or two, and a whole lot of deaths coming 2 or 3 weeks after the hospitalization. The CDC tracker shows roughly 1/7 of hospitalizations resulting in death; that’d be ~12,000 deaths a day, probably in February. The peak will last a couple of weeks.
Admittedly every single number in my math is an educated guess — the 3M / day, the 2.8% hospitalization, that the hospitalized patients will have the same duration of stay, the case fatality rate being the same. But the thing about estimates is that things could be better than my guess, or they could be worse.
UK data shows Omicron having about the same severity as Delta. South African data shows greatly reduced severity. It’s too soon to know.
