The unevenness of COVID-19 weekly rise in cumulative cases

[nitpicker]

Last edited Sat Nov 28, 2020, 09:01 PM - Edit history (1)

I do ((manually)) a rolling weekly count ((start with Monday reported data, finish as fast as I can, usually Friday data)) that now has expanded to cover about 2900 of the 3100+ of the counties/independent cities/equivalents in the 50 states and DC.

My criterion is if a place gets 200 or more cumulative cases, it gets counted. For some states, it goes down to 100, or even 50 for some states (like WY).

There are some states where every county/parish has logged 200 or more cases.

The latest weekly run:

Some counties "get it". Queens and Kings counties in NY had cumulative case weekly increases of roughly 3%. Using the "rule of 72", one could project that the time for cumulative cases to double would be 24 weeks, thus weekly detected cases doubling at the end of the 24 weeks.

Others... not so great. A lot of counties most recently had cumulative detected cases increase by 15-35% week to week. That gives detected case doubling time of 5 to 2 weeks.

The number of hospitalizations, however, depends on the population in which the new infections are detected. Is it an inmate population or food processing plant outbreak where mass testing ensues? Or college students and other under-30s doing youthful things? If so, the death rate per detected cases is much lower. Some states (such as NE) that had meatprocessing plant outbreaks, and prison cases, have a reported cumulative death rate of under 1%, while many states with wider community spread are running about 2%.

If a care home outbreak occurs, all bets are off. A care home outbreak in Norton County KS was reported to have sickened all 61 residents and killed 21. NY and NJ cumulative death percentages are relatively high (4-5%) because of the aftereffects of not knowing how best to care for care home cases early on.

Same for super-senior spread. Cumulative case data would suggest that 15 to 30+ percent of over-80s that get noticeably sick from coronavirus die, BUT again that is skewed by earlier cases.

So, will we go from 1500 deaths a day to 3000 in a month's time as some doubling data would suggest? Again, IMO it will depend on (1) the impact of Thanksgiving and other mass events, and the age/condition of those with detected cases.

[NRaleighLiberal]

1. just looking regularly at some of the sites, it looks like wave 3 has crested...or has it?

will the holiday idiocy make just a small dip, then a big wave 3 prime - or wave 4?

Amazing to see this mostly national experiment in ignorance happening before our very eyes.

I still see republicans as largely a suicide cult.

[nitpicker]

2. Covering the data reminds me to act like I'm a supersenior

Because I live with one.

[ProfessorGAC]

3. A Tip

The farther one gets from 8-10%, in either direction, the rule of 72 gets worse and worse.
Your 2 weeks of 35% isn't that close to doubling. It would only be around an 82% rise.
Likewise, 3% doesn't double in 24 weeks. It would go up to a bit under 2.04.
The rule of 72 is really convenient for estimations of thing like loan interest, which over the years has typically been 3-12%.
For current bank interest, or exponential disease spread, it gets less & less an accurate estimate.
It's an excellent post, but given your screen name, I figured you'd understand.

[nitpicker]

4. I meant this in glittering generalities

As an order of magnitude, not precision.

A few counties for this "week" have measured over 50% cumulative cases increase (presumably mostly congregate living/working situations and/or testing drives.)

It doesn't mean that recent "mask" orders won't have an impact, or conversely "covididoits" getting together in large groups for Thanksgiving, or other mass events (e.g. Black Friday) won't set a relatively "clean" county back on its heels.

The important question is how many of the new cases will require hospitalization.

And where.

In the spring, staff could stream in from all over the country, and if needed people could be shipped out to other facilities.

Many places are now reported bursting to the seams, with few or no relief valves.

Even drawing on health students, the retired, the military may not be enough.

IF the numbers of those needing hospitalization parallel or exceed in percentage detected new cases.

((This assumes everyone who wants (or their provider wants) a test can get one.))

But if it's "all" under-60s detected, this may be less bleak.

[ProfessorGAC]

5. I'm On Board!

I was just picking a nit.
I know you understand.
But, your concerns are valid.