Comments on: Arizona COVID-19 Update – 6/27/20

By: Tod Newman

Tod Newman — Wed, 01 Jul 2020 00:59:53 +0000

Alan, it’s a conspiracy between me and your former Economics instructor at Baylor to keep you in a perpetual state of “About to get a 13 on a test”! 🙂

The ordinances aren’t based on solid science like you’d hope. I would have thought the CDC would have been curious about the effectiveness of pandemic control measures, but it seems like they weren’t. There’s work in this space from the Asian CDC’s, but not much that’s very mature. Of course it’s hard to design experiments like this at a large scale, but even exploiting natural experiments (like a sick passenger on an airplane and how they spread the virus) is kind of lacking.

Take care. Hopefully you’ll only get sheared, not slaughtered! 🙂 (you’re kind of skinny, so maybe you have a chance)

By: Alan E. Gooding

Alan E. Gooding — Wed, 01 Jul 2020 00:26:45 +0000

All this makes me break out in a cold sweat like I just read a question on a statistics or economics exam that I don’t understand an will never be able to answer! But its very informative, thorough, and full of good analysis. The frustrating part to me are the ordinances that seem to be based on pure guesswork. Its throwing darts at a board with a blindfold. The mask thing seems to be just to make people feel better and I DO NOT like being told what to do just to make people feel better. Slippery slope from a civil liberties standpoint if you ask me. Maybe a mask is a small thing but what is next? I don’t like how easily the herd falls in line with this stuff. Sheep always get slaughtered.

By: Tod Newman

Tod Newman — Tue, 30 Jun 2020 21:47:26 +0000

In reply to Jim. Oh, and I just figured out that you're Jim Palmer. Good to hear from you, friend! :)

By: Tod Newman

Tod Newman — Tue, 30 Jun 2020 21:39:59 +0000

In reply to Jim.

Hi Jim. I tend to show tables from time to time with the Deaths per 1000 persons. Is that what you’re looking for? I’ve been trying not to contribute to COVID-overload by posting the tables each day, but maybe I ought to. In short, Death rates are way down. Case Growth rates are way down too from the surge in April/May. And the areas where surges are happening right now seem to be having a very large number of asymptomatic cases. In the west our hospitals are getting hit hard, but some component (large?) is actually due to surges of US citizens and legal residents coming to the US for COVID treatment. The case counts for the Mexican states below us are probably 5-10x higher than they’re reporting, just judging from the high number of deaths they have to a low number of reported cases.

I’ve heard the stories of overcounts, too, but I tend to give less weight to some peoples’ perception that this is the Gov’t manipulating numbers. If it’s true, it’s likely due to an economic cause… hospitals are getting hit hard in their cash flow. I expect any of this would be offset by the practice of undercounting that happens in nursing homes where (I’m told by a nursing home MD) they often don’t bother to give the COVID test because a person is so sick with whatever other thing ails them. In those cases they move straight to palliative care and never have a confirmed case. So maybe these things balance?

And yes, I’m a huge proponent of Jupyter Notebooks. Most of my standard visualizations are just straightforward Matplotlib. Everything is done in Python.

By: Tod Newman

Tod Newman — Tue, 30 Jun 2020 21:29:43 +0000

In reply to James Head.

Hey Jim, you’re right. I keep thinking about how the phased expiration could inform us better as to what causal activities for COVID infections actually are… unfortunately, the data’s a bit unclear. The Zip Code data I’ve been using lately helps a bit, though, because certain activities are more likely in some zip codes than others. AND the mask rollout being inconsistent gives us a really interesting natural experiment. I’m curious if we’ll see increased rates in a large, non-mask county (Pinal?) compared to Pima and Maricopa. I did an analysis of two neighboring border counties, Webb County, TX, (masks mandatory 4/3/2020) and Hidalgo County, TX, (no masks until mid June) but can’t see any real evidence that Webb County’s ordinance had an effect. Quite a bummer.

By: James Head

James Head — Tue, 30 Jun 2020 02:41:26 +0000

Hi Tod – a trigger event missing from Analysis (1) is the phased expiration of the stay at home / business closure orders. One expects a 2-3 week lag in case diagnosis due to the incubation period (in most cases up to 14 days) and the delay in getting a test and diagnosis. The increase in cases at Memorial Day aligns to this timing well, better than I thought it would, particularly since behavior (as you note) is more important than orders. A note on (5) – there is and has been no State-wide mask order. Those are implemented at the local level, when allowed by the State government. From that perspective your focus on Maricopa and Pima counties is spot on.

By: Jim

Jim — Tue, 30 Jun 2020 01:03:59 +0000

Two questions/thoughts…
1. I would really like to see the death rates… since we aren’t hearing anything more about the rate of death due to COVID-19, I’m assuming that that rate has rated relatively constant (I.e. the rate of confirmed cases rising exponentially has little affect on the death rate).
2. I have heard from several health care professionals that due to government incentives, just about anyone showing symptoms similar to COVID-19 symptoms were being reported as having COVID-19… seems like this practice would cause a big skew in the perceived reality of positive cases. Have you heard any similar reports?

P.S. Tod, you’re using Jupyter Notebook, aren’t you?