“Cases, Cases, Cases! If we didn’t test so much and so successfully, we would have very few cases. If you test 40,000,000 people, you are going to have many cases that, without the testing (like other countries), would not show up every night on the Fake Evening News.”Tweet from the Commander in Chief, July 4, 2020.
And what is YAM, you ask? Yet another metric, of course!
I first heard about this on a TV news show, but I can’t remember the name of the doctor who was touting as a valuable metric. I started looking into it, and I agree with him. When you look at localities across the world, it sheds much light on the situation within those countries and their success or failure in containing the virus.
Specifically, it gives us an indication of whether a locality is just testing when they really have to, e.g., when someone is delivered to a hospital on a stretcher, or are they actively employing wide spread testing and case tracking to contain any and all outbreaks.
This update is a bit of me thinking aloud as it were, because I plan to incorporate this metric, where possible, into the Zorgi Score. Sources for the data are listed on zorgi.me.
TPC1 shows the localities I follow where both case and testing data is available.
Here’s what pops out at me as I look at this graph. First, it is kind of a naturally normalized metric. You don’t have to worry about dividing by population or millions or anything else, because testing in a locality generally scales to the population all by itself.
Next, there’s a very clear “outlier” on this chart: Denmark, with 86 tests per case. Is it any wonder than Denmark has less than 1 death a day and fewer than 10 cases a day?
Now look at the worst country on there – Brazil. The number isn’t actually zero; it’s 0.2. This reflects the policies of Bolsonaro, a populist-autocrat who doesn’t want to be bothered with COVID-19.
The US is better off than Brazil, but nothing to brag about at 13. Germany is more than twice that. Interestingly, LA, which has done a lot of testing, is also at 13. Of the three counties out of 10 providing testing data, only San Diego has a relatively decent number, at 23.
The state number provide some insight into AZ’s terrible situation. At 6, AZ reflects Sweden more than it does the US, and Sweden has one of the highest normalized cumulative death rates in the world. Interestingly, OK is at 23, just like SD County. California, at 18, is better than the US as a whole, but not where it should be. The rest of the states are not doing well at all with this metric.
I wondered, what would the countries with the lowest cases per 1M people look like? TPC2 shows that.
China skews the numbers pretty badly, but we still can formulate something of an ideal from the average of all 10 countries, i.e., 387 tests per case. There are 3 countries that aren’t very high on the scale – Nigeria, Japan, and Indonesia. Nigeria is adding almost 600 new cases per day. Indonesia is adding 1,200 per day. Japan, with 25 tests per case, is adding around 250 cases per day.
[ Project for all you stats nerds out there: Take the top 30 countries with more than 5,000 cases, compute the tests per case, and report back on the correlation coefficient for daily new cases per day and tests per case.]
So, how would I use this metric in the Zorgi Score. First, I’d set a level that would reduce the overall score by a couple of points. Second, I’d find a number that would give zero points. Third, I’d use Brazil as the absolute worst you could do and make that level the top of the range.
So here’s my initial thinking on this:
|Tests per Case||Zorgi Score||Example|
|0 to 1||5||Brazil|
|1 to 5||4||Mexico|
|5 to 10||3||Arizona|
|10 to 20||2||California|
|20 to 50||1||Germany|
|50 to 100||0||Denmark|
|100 to 500||-1||Australia|
Tell me what you think!
A bunch of you have offered to help out, and I want to take advantage of that. It involves finding the best source for hospital data for all 10 counties I cover. If you’re interested, please DM me, and I’ll give you the password and the url to a page on zorgi.me that has all the details.
That’s it for today. I’m including the underlying tables for the charts above, as well as the local charts and tables. Discussion on those tomorrow.
Comments from Readers & My Responses
The following are some of the comments from readers on Reddit and other social media platforms where I regularly post. Reader comments are in italics and color. My responses are in plain text. If there is more than one commenter without a response, they are separated by different colors.
Hey zorgi, always look forward to your updates!!
Where do I begin…. Yikes, very alarming to see about a 10% positivity rate today, I listened to a bit of the San Diego board meeting today and Dr. Wooten seemed very concerned that we will almost certainly see a uptick in more deaths per day and would most likely pass our daily high of 584 positive test. It’s extremely upsetting seeing people in my age demographic(20-29) not giving a damn and having parties on the weekends. All I can do is shake my head and hope there’s a light at the end of the tunnel.
Don’t beat up your generation too much. After all, it’s my generation that has bequeathed you a pile of s**t. And I sure did a ton of stupid things in my 20’s; hell, it’s a wonder I’m still alive.
The difference it, you and many like you know and care about things like positivity rates, BLM, global climate change and all the rest. The light at the end of the tunnel is if you mobilize your generation to get out there and vote. In 2018, the turnout rate for your generation went up by 79%, but that still meant that 64% of you didn’t vote.
You and your friends need to make sure there is a light at the end of the tunnel. You can do that by changing that turnout to at least 50% or higher. A few people in my generation are starting to catch on, but your generation needs to make sure mine shuts up and gets on the right side of history.
Actually, I’m very hopeful it can happen. We never saw protests like these in the 60’s. It was always a relatively small number of people, and the protests weren’t integrated the way these were. I think we’re in for a real historical change.
Does that metric discriminate unique tests versus repeated testing of same individual and would it matter? I haven’t thought much about this but a family member in another state mentioned something that caught my attention. They downplay the value of positive case metrics due to the same individuals being tested multiple times, either as confirmation of a positive or to track infectivity.
So just to see if it mattered, I made the following assumptions (for simplicity). 1 million tests, 100,000 positive, 5000 hospitalized. So this gets reported as 10% positive rate and 5% of affected people require hospitalization. Further I assumed that every one of the Hospitalized were tested a total of 5 times during the course of their stay (really have no idea about this one), and that 5% of the nonhospitalized Infected were retested once (again, no idea). That retesting, if not segregated in the data, lowers the actual positive test rate from 10% to 7.6%, which is significant but still high. But calculating it that way raises the actual hospitalisation rate from 5% to 7.4% which is even more alarming. not sure if that calc is right but do states account for this and does it matter?
Very astute observations there, and this is something I’ve been complaining about all along — not that my little voice will change anything. The CDC specifically allows all 3 types of testing data to be intermingled: diagnostic tests by unique individual, all diagnostic tests, and serological tests. So, as you point out, we have a complete mishmash in test reporting. That means if we see positivity rates go up (or it’s counterpart, tests per case), it is even more alarming.
Do states account for it? That’s the big question. Some do, some don’t. Some counties split it out, some only report total tests, and 7 out of 10 counties in southern CA don’t report their testing numbers at all.
Frankly, it’s an outrageous state of affairs, but this is what we get when we gut the budgets of our public health departments over a 40 year period. We’re happy to give millions and millions to police departments to do jobs that social workers should do, but when it comes to doling out $100k to public health departments, there’s suddenly no way to pay for it.