Math and More Questions

I’ve been working out math problems all day, trying to answer some questions. I’m not satisfied with the results. I think I’m getting more questions than answers. Some of this has to do with lack of accurate data—same problem everyone else has right now.

I’ll share some of the data anyway, and then just ask the questions.

This chart compares different populations and the number of deaths attributed to Covid-19. All I’ve done is divide the actual number of deaths by population in millions to find deaths per million people. 

Place	Population in millions	Covid-19 Deaths as of April 30, 2020	Covid-19 deaths Per million
Texas	30.54 M	722	23.64
Harris County	4.6 M	109	23.69
USA	328.88 M	58,356	177.44
World	7,597.78 M	233,014	30.67
Italy	60.46 M	27,967	462.57
New York State	19.44 M	18,274	940.0
New York City	8.75 M	18,069	2,065.0
Sweden	10.1 M	2,586	256.0
Norway	5.4 M	210	38.89

There's a lot of difference between New York and pretty much everywhere else.

This is a snapshot. It doesn’t tell you the death rate of Covid-19. For that you need to know how many people were exposed to and contracted the disease. Then you divide that number by how many deaths.

So, in the absence of that data, everybody’s trying to make guesses.

I took a good look at the video getting a lot of attention this week of two ER doctors in Kern County, California. They’re making the case for opening up. And their non-numerical arguments for that are persuasive—like how we lose immunity strength when we avoid all exposure to germs. And that the costs to society at large—and to overall healthcare and people’s well-being—because we are focusing on this single health issue are something we need to be talking about.

Dr. Dan Erickson (left) and Dr. Astin Massihi
screenshot found here

Drs. Dan Erickson and Astin Massihi are doing Covid-19 testing on everyone that comes in to their five ER centers, if I’m understanding them correctly. Not just for possible Covid-19 cases, but everyone who comes in for anything, which gives them their own raw data. Their numbers show that 6.5% of those they test (several thousand) test positive. I’m unclear whether that means people actually have it, people have antibodies because they had it, or a total including both. Then they point out how many have been tested in California. Dr. Ericksen says,

We have 33,865 Covid cases, out of a total of 280,900 total tested. That’s 12% of Californians were positive for Covid.

Then they extrapolate that 12% to the population at large. But that’s not accurate for a couple of reasons. Most tests, beyond their own sample, have been done on probable cases, not a random sample. If we were randomly testing populations, that information could be extrapolated to the community the sample represents. But most of our testing has been done on people with symptoms. That leaves out all people who are asymptomatic or have such mild symptoms they don’t get a test—or don’t qualify to get tested, in many places. Also, according to some, current tests tend to have a high false positive rate. 

They use this extrapolation as “the best we have” right now with the lack of data. But, while their arguments are valid, and a conversation worth having, we really can’t deduce death rates by this extrapolation. If you assume the sample covers everyone, then it looks like more have been infected than have been, and your death rate calculations will be far lower than reality—just as death rates appear far higher than reality when you only count verified cases. Too many assumptions.

As they say, the more testing, the better. We just don’t have the right sample sets yet.

We know the death rates are much lower than what the data shows. But how much lower? Because, if it’s in the range of a typical flu year, then we can let people take precautions and then take their chances—the same way we do for the flu.

What we also don’t know is, what will the contraction and death numbers be when we get through this season? By that I mean probably the year. Flu seasons tend to go from fall through the following spring, not the whole year. By the time it comes around again, it’s a different flu, which is why there’s an annual—different from the past year—flu shot. So we can count annual flu deaths and that data is relatively solid. Not all cases get reported, but deaths caused by flu would be.

We don’t know season length of this virus. We don’t know if it will mutate enough to make herd immunity a moot point or a proper goal. Either way, we’re nowhere near the 60-80% required for herd immunity, meaning enough people have had it that the virus has a hard time finding a new host.

Anyway, we don’t have a stopping point at which we can look at the different approaches and say one worked better than another.

Among all the questions, I’d also like to know a few more things:

·         If it’s true that urban areas are more likely to be hit hard, how do you explain the discrepancy between Houston and New York City? Houston’s shutdown was mid-March, close to New York’s. New York had a few more cases by then, but not that explosive a difference.

·         Does it occur to anyone else that public transportation has a lot to do with the spread of the virus? During the NYC shutdown, they have kept their subways running. (My germ senses are making me shudder.)

·         How much lower would the death data be if we took nursing homes out of the data? Less than half? If so, why have we shut down society instead of super-protecting nursing homes?

·         Sweden did some social distancing but no shutdown. Norway, by comparison, did a full shutdown. There are differences, but are they differences in timing only? How many inevitable cases have hit Sweden already but will take longer to eventually hit Norway? Do shutdowns stop cases from happening or simply postpone them?

Mark Ramsey, my SREC Chairman, compared Sweden and the USA graphically this week on Facebook, with this explanation:

Graph of the day. USA and Sweden. As identical as possible, except for Y-axis, which is different mostly due to extreme difference in size of the population, and increased testing in the US. Based on the SHAPE of the two bar-graphs, has turning off a VAST PART of the US economy been significantly better than Sweden, who simply had recommended practices and a VERY FEW closures? Has it been "worth it"? The initial models were very very wrong, and we now have millions of tests to calibrate the risk with. The answer needs to change accordingly. (data from WorldoMeters.info at 13:00 CDT, 4-28-2020)

Comparison graphs of Sweden and USA
from Mark Ramsey on Facebook

Then there are some political questions. I hope this virus hasn’t actually become political. Maybe it’s more of an urban/suburban/rural difference—which, again, doesn’t explain Houston vs. New York. Is it the prevalence of cars instead of mass transit in Houston? By the way, California’s deaths per million is about 31. I didn’t have today’s data, so I didn’t include California on the chart, above. But I’m guessing driving instead of mass transit is a reasonable explanation there too. Also sunshine.

Facebook friend Leslie May has been looking at the politics and the data the last couple of days. She said this yesterday on Facebook:

CHANCES OF CONTRACTING OR DYING OF COVID-19.... overall, Democratic-governed states are more urban with an average population density of 123 people per square mile....

average population density of Republican-governed state: 71 people per square mile

for all Democratic-governed states, the average rate of COVID-19 cases reported per 100,000 residents is 414

for R-governed states, the average is only 180 (less than half)

the average mortality per 100,000 in D-governed states is 23

the average mortality per 100,000 in R-governed states is 7, less than 1/3

by the way, Harris County, Texas, rates are 130 reported cases per 100,000 residents, and 2 deaths per 100,000 residents, lower than the average STATE....Texas as a state is 93 reported cases per 100,000 residents with 3 deaths per 100,000 residents.

Today she adds a map and more questions and data:

Wondering why certain states have no stay home orders (labeled 1 on the map below), others just "recommendations" (2 on the map), and some STILL have them in place (3 on the map)? Here's what our country looks like as with the "order" status / party of the governor superimposed on the actual cases per 100,000 people in that state. Also looked at where cases are increasing, decreasing or about the same.

7 states, all with R governors, are under NO stay home orders: SD, ND, OK, WY, NE, IA, AR. SD, WY, OK are in the lowest category on the map for number of cases/100,000. The others are, pardon the pun, "all over the map."

4 states are under stay home recommendations, including TX. The others are KY, UT and CT (kinda another outlier in the northeast, although cases decreasing may explain it). 2 with D governors, 2 with R.

39 states are still under stay at home, 22 D governors, 17 R. Oregon, Montana, Minnesota, NC, WV and ME are low in cases/100,000. Only WV's governor is R, although if memory serves, I think that is a relatively recent development.

In terms of increasing, decreasing or about the same number of cases, I could see little rhyme or reason to how that influenced these decisions -- for those in the middle with recommendations, TX and KY about the same, UT increasing, CT decreasing.

Finally, Harris County, TX, has 6161 cases, or 134 cases /100,000 residents, with 109 deaths or 2/100,000 -- about the same size geographically as Rhode Island with 8247 cases, 778/100,000 people, 251 deaths, or 24/100,000 residents. RI has twice the death rate of Harris County per capita.

Graphic from Leslie Joan May, from Facebook

Texas starts opening up tomorrow. There’s a mixture of “hurray”s and “no no no”s. I lean toward the hurrays. But I’m not rushing to a theater any time soon. And I’m still wearing a mask at the grocery store, and carrying hand santizer, wipes, and gloves for use as needed.

If we can be sensible while getting people back to work, I think that’s a win.

I don’t think it was ever our intention to hold everyone housebound until there was no more risk. We were told 2-3 weeks, and maybe a little longer (it’s been 4 weeks tacked on to the original 3 already) to “flatten the curve.” That never meant fewer people would contract the illness; it meant fewer would get it at the same time. The number that would eventually get it was expected to remain the same. See my question about Sweden and Norway above. So my question is, what's the rationale for making it take longer but not limiting the actual number who get it? Is that goal, whatever it is, worth economic collapse, with associated famine, hunger, poverty, and hopelessness?

One advantage of time has been the possibility of finding treatments that work. The new drug remdesivir was announced this week as a successful treatment. Because it’s new, it will be more expensive than hydroxycholoquine plus zinc plus Z-pac (one meme I saw showed $1,000 instead of $20). Still, it’s great news. Enough for a stock market bump upward.

Others treatments are coming. Even a vaccine may be coming soon, which would be in record time. So this could mean that the pause was of some value beyond guaranteeing our hospitals wouldn’t be overwhelmed, which it turned out was not the case even in NYC.

This morning Ben Shapiro talked about a new data piece on his show (haven’t tracked it down elsewhere yet) that there’s not a single case in which a child has spread the illness to an adult. I don’t know how they know that, but that’s great news. If true, then schools and play dates will start looking safe again. And maybe we’re in time for summer sports leagues.

At the very least, I think we need an absolute rational explanation for every infringement on our freedom from this point forward. “For your safety” won’t do. Neither will “The public can’t be trusted.” We need something like, “Here's the data that shows this (X requirement) limits spread of the disease by (X)%, and it is imperative that you not spread this disease in (X location) before (X time).” 

Failure to provide that information might just lead us to believe that fear and control are a goal when what we really want are freedom and innovation.