Uncle Ebenezer asked a good question yesterday:
What’s the best response when Conservatives play the Cases Are Only Rising Because of More Testing card?
There are two answers to this question.
The first is to look at positivity rates. The positivity rate is the number of new confirmed cases divided by the number of new tests.
If we are getting more confirmed cases in the context of more tests being conducted, we look at the positivity rate to gain some insight as to what is happening to the unknown and only indirectly observed state of the infection at a given point in time for a given geography. If the positivity rate is constant or decreasing, the new cases that are being captured by the explicitly observed data is mostly a function of more testing being conducted.
However if positivity rate is increasing, that that is a strong signal that the newly observed cases are at least partially brand new case growth.
We have to make a few behavioral assumptions about testing proclivities. I think there are two facially defensible assumption sets. The first is that testing is fairly random. The second (and in my opinion, more likely) assumption is that testing is non-random. Under this assumption, people who have reason to think that they are COVID exposed are more likely to spend the time, cost, stress and hassle of getting testing than people who personally assess their exposure risk to be low or nil. These two assumptions have slightly different interpretations if there were to be an extra 1,000 tests taken on a given day in the normal workflow of a state. If we assume randomness, an extra thousand people tested won’t change the positivity rate. We’ll be getting background infection rates. If we assume adverse selection, an extra 1,000 people tested will have a slighlty lower incremental positivity rate as the first people in line are more likely to be positive than the last person to get tested with the extra tests.
We can look at two different states and their testing experience over the past two weeks. New York state has increased their daily testing rate by about 3%. During this same time of basically flat testing, positivity rate decreased by 30% (1.5% to 1% of tests came back as COVID positive). We can contrast that experience with Arizona. Over the past two weeks, testing has increased by 40% (14 day rolling average). At the same time positivity rate has increased from 12.7% to 21.2%. This is a very strong hint that the increased case count in Arizona is a reflection of ongoing spread and not just a function of increased revelation of a steady state.
The other way of looking at community spread is to disregard testing entirely. There are a lot of assumptions built into extrapolating an unobserved from a biased observable. Instead we should look at age and demographic (including race/ethnicity) hospitalization rates. Hospitalization rates lag infections by a couple of weeks. Very few people actively like to go to the hospital. Instead it it is bad option that is just superior to the viable alternatives. A hospital admission is a costly, and thus a believable signal.
If hospitals are filling up with people who are infected by COVID, this gives us some insight on the state of spread a few weeks ago. We can get more insight by examining the age and risk distribution. If the marginal patient is old, that tells us one story of disease spread. However, if the marginal patient is comparatively young and otherwise low risk, that implies more spread as a lot more infections are needed to generate a large number of hospitalizations for people under the age of 50.
This approach is not perfect. We can probably assume that a marginal patient admitted when there is 50% capacity remaining is probably healthier than a patient admitted when a hospital has 3% remaining capacity. Comparisons over time are also problematic as we are still in a stage of rapid learning on what does and does not work well clinically. But new hospitalizations on a daily basis and total bed counts are at least directionally useful.
Identifying current community spread is challenging as it is an unobserved function and we must estimate it from known and limited observable information sources but we can make reasonable approximations as long as we acknowledge assumptions and data limitations.
David C
Some of the hospitalization rate data are hidden – at least in FL.
https://www.wtsp.com/article/news/health/coronavirus/florida-coronavirus-cases-hospitalizations-deaths/67-4bbd0c35-6742-4f51-a59a-ea1d101f54ea
Mike J
As things have started to ease in King County, WA, more non critical health care that didn’t happen over the past three months is now ramping up. Instead of just zoom calls with doctors, patients are going into medical buildings attached to hospitals and having normal tests for non covid conditions. But they need covid tests before being allowed in the hospital. My mom’s had two in the past two weeks.
She’s been militant about limiting her exposure, so she had a pretty low chance of exposure because she had a very high chance of it being bad if she did get it.
As you were saying, much depends on who is getting the new tests. If people are getting them for work or for doctor visits, it will actually look more like a random distribution, or may even be people less likely to be sick. In King County, we have a cumulative positive rate of 6.4%. Yesterday we had a 1% positive rate (39 out of 3100 given yesterday.)
Edited: corrected cume rate from 9.4% to 6.4%
https://www.kingcounty.gov/depts/health/covid-19/data/daily-summary.aspx
GSmith
Am I missing something? It seems to me like people are overlooking the most obvious flaw in the argument that rates are only increasing because of increased testing. That argument basically boils down to “it’s not getting worse, it’s really been far, far worse than we thought all along, and we’re only now realizing it”. That’s not very reassuring.
Roger Moore
The big problem with all this is that everything we’re looking at is a lagging indicator. Tests lag infections, hospitalizations lag tests, and ICU stays and deaths lag hospitalizations. So the more concrete the evidence, the longer the delay before you get it. If you wait for deaths as your indicator, you’re dealing with people who were infected as much as a month earlier.
As an example, we can be extremely confident the increased case count in Texas is because of real growth in infections rather than increased testing because their hospitals are so full that Texas Children’s Hospital is taking adult patients to take up the slack. Unfortunately, that concrete indicator is way too late. The people being hospitalized in Texas were infected at least a couple of weeks ago, so even if they completely shut the state down yesterday the case load would still grow beyond what their hospitals can handle for a few more weeks.
Steeplejack
Tangential, but “numbers.”
A friend of mine’s 90+-year-old mother is in an assisted living facility here in Arlington, VA. She (friend) found out last week that the guy across the hall from her mother died of coronavirus last week. No notification or information from management. I think she found out in a chat with a janitor.
In subsequently talking with the facility’s management, she discovered that they have been underreporting their numbers—both infections and deaths—to the state, with various excuses like “unconfirmed,” “not required to be reported,” etc. One set of their numbers doesn’t add up with another set they reported elsewhere.
This is not a story about my friend going all Karen because “It’s my mother, damn it!” But when I talked to her yesterday we both realized (again) that stuff like this makes you distrust any statistics about the pandemic. Too many different sources, using different criteria, different time periods, different levels of competence, accountability and “statistical rigor,” etc. It’s nuts.
SiubhanDuinne
@GSmith:
This.
Trump seems to be engaging in the ever-popular post hoc ergo propter hoc fallacy. It’s like he thinks testing actually causes the disease.
Roger Moore
@GSmith:
Not really. The things that are scariest about COVID-19 are concrete and easy to identify. We don’t have any trouble figuring out that people are getting very sick and dying. What we really don’t know is how many people there are with either very mild or no symptoms. If we find more cases through increased testing, they are likely to be among those less scary cases. Finding them makes things better because A) those people can now isolate and reduce the risk of infecting others and B) it makes us believe the average severity of the disease is lower.
Patricia Kayden
Ruckus
@SiubhanDuinne:
I don’t think it’s like he thinks, it is he thinks. He is very limited and he has limited himself to numbers. Notice that he’s not very good with numbers, especially what the individual numbers mean in any way, but he is a numbers guy. Bigger numbers on his side is good, smaller numbers against him is good. He has no regard for other humans, none, other than what they can do for him. So more people having this is bad for him. Fewer people having this is good for him. It really is that simple. We want to make it more complex because in reality it is. But in trump world one should never bet on complex at any level because complex on any level is way, way beyond him.
Amir Khalid
@SiubhanDuinne:
I agree completely about what Trump is implying. Thanks for giving the Latin name for it, which I could not remember.
SiubhanDuinne
@Ruckus:
Yes. I rarely indulge in the “it’s like…” formulation, but I was being lazy. You’re right; it is just how he thinks.
Er, “thinks.”
:-)
SiubhanDuinne
David, or any FP: the Congressional hearing with Fauci et al. is going to start within moments.
Any chance of a dedicated thread with a link to the video feed?
Thanks!
Eunicecycle
@Steeplejack: My husband’s uncle is at a very good facility in NE Ohio that has been sending out updates indicating a few staff Covid positive and no residents positive. He was having some medical issues not explained by his current conditions so his children asked for him to be tested. His son found an article from Kaiser Health that talked about unusual presentation in the very elderly, the very issues he was having. It took 2 weeks but they finally tested him and he is Covid positive. He doesn’t have a cough or fever but had some stomach issues, dizziness, and extreme fatigue. Then we find out they basically are not testing anybody. The nurse manager said the county gave her 10 tests and she has used 4. The other residents were negative. Gee, I wonder if their numbers are so low because they aren’t testing hardly anyone? With hundreds of elderly they could have many with a different presentation.
Steeplejack
@Eunicecycle:
And she was given 10 tests for hundreds of residents?! ?
SiubhanDuinne
@Amir Khalid:
Possibly I remembered it because I just finished bingeing the complete West Wing on Netflix, and the Latin phrase is the title of an early episode.
I had four different Latin teachers in high school, and two philosophy professors at university. I doubt that any of them would be pleased to know that one of their alumnae is more conversant with a TV drama than with the writings of the ancient Mediterraneans.
artem1s
the only response to this idiot argument is this – Your GOP representative and Trump are lying to you. They have been lying to you all along. They have always lied to you. This time, your life is at risk if you believe them, not just some random ‘Mexican’. Go ahead and be a chump so they can keep the stock market bubble going. Go ahead and risk your life when your employer orders you back to work because he and your Red State governor doesn’t want to pay you unemployment. Those of us who are not buying into their lies, will stay home or mask up when in public if we are ordered back to work. But we won’t be spending our money at those businesses that have bought into the lies and refuse to protect their customers and employees. The stock market is more likely to crash if we don’t know the real numbers. We will have to have more shut downs, if we don’t have real numbers. Real testing numbers help us know the real risks. If the paying customer doesn’t trust the numbers, they are never coming back, never spending a dime on airfare or at a casino or a hotel or a baseball game. The virus doesn’t care if Trump is lying about the number of people infected. If you were smart and not part of a cult of anti-science nutballs, you’d be protesting the lack of testing, not that your favorite hair salon is shut down. Please proceed (at your own risk) accordingly and stay at least 6 feet the fuck away from me and my family.
Humdog
Have we heard from MomSense? Wasn’t she worried she was experiencing symptoms and was getting a test?
WaterGirl
@Humdog: Yes on symptoms, yes on worries, no on having heard from MomSense since yesterday evening. I think on Saturday she said she would get the test “tomorrow” but she did not say anything about tests or test results yesterday, just that she is tired and has some other symptoms, I believe she called them mild at this point.
SiubhanDuinne
@Humdog: She has posted briefly in a couple of recent threads. I’ll see if I can track down her comments and link or copy them here.
(I really love the way we all look after each other. There are very few blogs that could say the same.)
ETA: WaterGirl (ETAA: and Steep!) [ETAAA: never the fuck mind!] did the heavy lifting. I have nothing to add, except to wish MomSense complete and speedy good health.
Sure Lurkalot
@artem1s: Oh this is mighty excellent.
Steeplejack
Redacted.
Ohio Mom
It’s been my experience that when Conservatives ask a gotcha question, they are never interested in the answer.
Sometimes you can see them start to follow your argument but you can also see in their faces when their mind snaps shut. They wanted to watch you sputter, they didn’t want to actually consider new information.
Too much work and it requires too much ego strength to reconsider even a small part of one’s belief system.
It’s enough for me to check tne NYT Ohio Covid map and the rt.live site every morning. I don’t need to know how those stats were arrived at. I can easily tell that my state’s lull is ending and it is time to get back to battening down the hatches.
Jinchi
It’s worse than the disease, testing illustrates just how badly he’s failing.
On the day of his Tulsa rally, we learned that 6 (now 8) of his staffer’s tested positive. That was embarassing to him ( no word of how they are doing). Yesterday, they reported that they will stop testing for symptoms at the White House.
Trump doesn’t care who gets coronavirus. He only cares about his own PR.
Eunicecycle
@Steeplejack: Yes. So she only wants to use them on people who are coughing and have a temp. And his doctor and the medical director agrees with her thus the fight for the test. He is going to be retested because they think he is a false positive, although even the doctors agree that false positives are rare. And if the next test is negative, are we doing best 2 out of 3?
ET correct a spelling error
SiubhanDuinne
@Jinchi:
Kind of proving my point. If the WH doesn’t test, then no one at the WH has or can get the Rona.
That’s just logic.
Mel
@WaterGirl: Thank you for the update. Let us know if we can do anything for her?
Jinchi
These are the people who invented the ‘Gish gallop’. The point is to swamp you with nonsense. You’ll often find them contradicting themselves within the same sentence. You can’t answer them faster than they can create another bogus argument and as long as you’re trying, they’re winning.
rp
It’s not that he thinks testing causes the disease. It’s that he wants COVID-19 to go away and not bother him, and in his mind the best way to accomplish that is to remain ignorant. That’s the extent of his thinking. Trump is a pure narcissist, and doesn’t view other people as real humans. He’s like a kid who thinks everyone else in the world is a robot — he’s never grown out of that phase. The best approach to dealing with the disease for him is to not hear about it, so less testing is better.
Gvg
Trump is an advertising guy mentally. He only thinks in terms of spin and ratings, not underlying reality. All through his Presidency, I have been remembering “Death of a Saleman” and how insane it seemed to me that the salesman disn’t understand that ads weren’t real, or all true. Trump believes not just his own bullshit, but a lot of other people’s too, especially if it reinforces his favorite themes that he is great and a lot of minorities are bad and to blame.
a lot of the most aggravating conservative loudmouths have this terrible inability to not know TV isn’t real. Cheney quoting that Tv show to justifiy torture to find a bomb was also part of it. Lots of fearful people didn’t understand dramatic plot lines were made up because reality would be too boring to watch. So Trump is the worst, but plenty of others can’t think their way out of a wet paper bag.
Jinchi
Early in the epidemic West Virginia was being praised as ‘the only state without any cases of coronavirus’. At the time, I think they’d run fewer than 200 tests.
Mel
@Eunicecycle: That’s just nuts!! Just 10 tests?
The asymptomatic spreaders (including employees) or the patients with mild, atypical symptoms are the big risk early on in a facility situation, as they are less likely to be isolated or to self-isolate, and will (at least in the case of employees) have widespread contact with both a large number of patients and with a large number of surfaces that other employees and the more mobile residents will be touching multiple times in the course of a day: doorknobs, computer keyboards, food trays, tables, chairs, elevator buttons, bathroom surfaces, linens, magazines and mail for residents, etc.
If they’re only testing residents, and only when they have become very obviously ill, they’re not likely to effectively contain their outbreak.
Laura Koerber
I just say, “The cases are real. DOn’t you want to know about them?” That results in the asshole shutting up or changing the subject. They don’t understand any of the stuff you guys are trying to explain to them.
Brachiator
I don’t understand what the point of this conservative play here might be, other than to support Trump? They seem to be rejecting science as some offshoot of liberal ideology. But they do not offer any coherent alternative explanation or plan to deal with the virus. I could understand if this were the 15th century and they were saying that the pandemic and whether one might die from the virus was God’s will. But at best (worst) conservatives here become much like anti-vaxxers and have to say either that the virus is unimportant or that herd immunity will magically protect everyone (after culling a few million people).
zhena gogolia
@Patricia Kayden:
That is a magnificent ad.
Bluegirlfromwyo
@Ohio Mom: Absolutely. One of the biggest mistakes we make as liberals is assuming our conservative friends are asking questions in good faith because they’re friends.
My standard answer to this “question” is “So if I stop getting mammograms, I won’t get breast cancer?” The blank looks are pretty priceless.
rikyrah
The answer?
Well, you know, pregnancies are on the rise because of pregnancy tests.
Breast cancer is on the rise because of those darn mammograms??
kindness
I’m very impressed with you all as you are trying to talk to people who are not convinced this virus is a special case and explain to them why it is. I’m honestly having a hard time with that one. Not so much that I couldn’t discuss the underlying biology and risk avoidance but because the people who ask us to convince them of the need are going out of their way to not believe or listen to what physicians say and prefer to parrot Fox News. It isn’t a serious conversation in my mind. I’m feeling done with having to convince people of something they fundamentally don’t want to hear or believe.
You have better zen than I do apparently.
Doc Sardonic
People are spending too much time ascribing a thought process to Nube de Pedo Naranje. His entire universe consists of a simple binary system good for me/bad for me, there is no other in depth process. Paying the full bill for a contractor, bad for me stiff them(good for me), virus numbers up bad, slow testing down numbers down good, the fact that 100,000 plus Americans have died is not even on his radar. As has been said by Fran Leibovitz “You don’t know anyone as stupid as Donald Trump. You just don’t”
rikyrah
@David C:
a lot is hidden in Florida???
rikyrah
I don’t think that I trust this process in America.
Matt McIrvin
I think that much of the rise in California may be due to increased testing (because their test positivity is stable around 5% and testing volume has been increasing more or less linearly). Definitely not in Arizona, Florida, or Texas.
Just Chuck
Anyone not convinced by now is beyond fucking hope. The only thing I would say to those people is “goodbye”.
Brachiator
@Ruckus:
I think you have nailed Trump’s psyche here. It’s not just that he is stupid and dishonest, although he is both. It is his simple reduction of whether something is good for him.
The tragedy for America is that his base continues to prop him up because they absolutely believe that he cares about them, no matter what he actually does.
Hoodie
@rp: That’s a subconscious motivation behind it, but I think what he believes is that (1) COVID is not as bad as everyone makes it out to be and (2) increased testing is just revealing people who are not that sick, but this makes the problem look worse because BIG NUMBER. There is probably some truth to that, but it’s irrelevant because the reason you test is to give information for tracing so we can eventually open up. Increased testing would turn the debate to how he has done zilch as far as tracing. Countries that are doing effective tracing also have things like quarantine hotels for mild and asymptomatic cases. We have nothing like that. He made a decision several weeks ago that we’re going to take the punch. If you’ve made that decision, why bother to waste money on tracing? That’s ok by him because he thinks he is not likely to get infected and he doesn’t give a shit about anyone but himself.
The danger for those fighting Trump’s sociopathy is getting hung up on his announcement about tests takes attention away from his utter failure to do anything to create a workable system for opening up, which would necessarily involve a tracing regimen and associated things like payments for quarantine, etc. Trump wants to talk about testing because he can bullshit about that and not have to deal with real issues. It’s another example of his technique of using controversy to create distractions.
Brachiator
@SiubhanDuinne:
West Wing clip. Pretty funny, actually.
Just Chuck
@Brachiator: His base knows deep down that T doesn’t give an inkling of a shit about them. They just like that he’s hurting the people they hate. Spite is all these monsters are made of. Spite and deliberate proud stupidity.
Just Chuck
@Hoodie: Maybe, but I think we still need to frame it as “T says more testing causes more disease” and make HIM explain it. Like LBJ would say, make him deny he’s a pigfucker.
The dirty fighting is just part of it. The other part is that MAGAts don’t understand nuance.
Dev Null
Uncle Ebenezer asked yesterday about a source for hospitalization data.
I don’t recall seeing a response other than “CDC”, so here’s another that looks like it might be more serious about hospitalization data:
https://covidtracking.com/data
I haven’t looked at their data download format, so no idea how easy to use nor how comprehensive their data is.
YMMV.
Also too, as people have noted, some states don’t report hospitalization data, some obfuscate, some outright lie, and what two different states report as hospitalization data might be very different.
Youse pays your moneys and youse takes your chances.
Dev Null
@SiubhanDuinne: … proving my point …
I dunno, the story I saw said that “taking temperatures takes more time than we want to spend preventing staff infections.” For once they might not be lying.
I mean: if the Big Cheeses in the White House don’t care about the staff – it’s a good guess that they don’t – and if the Big Cheeses are thoroughly and continuously tested, what’s not to like (if you’re a Big Cheese)? Who cares what happens to the staff?
Interchangeable parts, and all that.
Dev Null
This has been making the rounds for a while, and has probably been posted here at BJ previously, but I think it’s insightful enough to post for anyone who missed it:
The Psychology of White Supremacy is similar to the psychology of narcissism
Relevant to your second para in the quote.
SW
I think the press is missing the story regarding Trump’s confession regarding slow rolling testing. To understand it, you have to revisit the very early stages of the pandemic. You have to go back to when the The WHO offered their test to the world and we refused. Why? This is when Trump stepped in. The CDC says it was because the test was unreliable and they had a better one. That is likely the cover story they came up with when Trump refused to let them use the WHO test because “testing is a two edged sword”. Then, when the CDC started to gear up their own tests I will bet dollars to donuts that Trump once again interfered and they had to come up with another lie. “The tests were contaminated”. I call bullshit. It was Trump trying to keep his numbers down.
There are certain things that no matter how many lies the press catches Trump telling they just don’t want to believe he is capable of doing. Like slow rolling the availability of testing during the early stages of the pandemic just to keep his numbers down, when this could lead to the needless deaths of tens perhaps hundreds of thousands of people. But refusing to believe that Trump wouldn’t do something because it is just too heinous is a suckers game that has proven to be stupid time and time again. He is telling us that this is what he did. It is just that no one wants to believe it. Believe it. That is exactly what he did. He knows that it will come out eventually and he is already making excuses.
Ruckus
@Brachiator:
The DSM is the go to manual about psychiatric disease. trump would be the perfect example of narcissism.
He is nth stage, n being the most narcissistic. That is his base disorder, everything else is layered on top of that. He sees himself as the center of the universe. He has never paid anyone to tell him what they know, only to give him positive news. He doesn’t hire or listen to experts because they tell him what he doesn’t want to hear, that he’s not the center of the universe and what they tell him has restrictions about him being/staying the center of the universe
Given one of my careers, I’ve met some people with money. Some people think that money makes them better humans and people with less are less human. shitforbrains is one of these people. Not everyone with money is narcissistic but having money and having done little to ever earn it spoils a lot of people. It’s a measure not of wealth but of value to them. Their value. shitforbrains is just one of them, abet farther progressed in the narcissism.
Ruckus
@Dev Null:
I believe that they are rather much the same disease, that the person with the disease has a misconception about their value, especially as it relates to others or other issues. Things that “prove” that the person is better are usually incorrectly asserted as true, “white people are superior,” “more money makes me a better person,” note that trump holds both those views. The base underlying disease is narcissism, a self centered view of the person with the disease, the racism or wealth issues are the measurement tools they use. The degree with which the views disrupt their lives and the lives of those around them is the degree of the disease.
Also note that people can have more than one disease concurrently.
Dev Null
@Ruckus: I get your point, but isn’t narcissism usually considered a character trait – meaning a characteristic of an individual? And isn’t “white supremacism” a group ideology?
Similar, yes, but the value I saw in the link I (thought I) provided is the discussion of the similarities. However it looks like I screwed up and failed to post the URL, so trying again:
https://twitter.com/_EthanGrey/status/1261824759404343298
A couple of Grey’s points reminded me of WJ Cash’s Mind of the South.
(Might be a copy here:
https://docuri.com/download/the-mind-of-the-south-w-j-cash_59b8e677f581717b5b8784fb_pdf
but I had trouble navigating the site so YMMV.)
Incidentally, IIRC Altemeyer’s The Authoritarians, which is a good read in its own right (and free!), makes a somewhat similar argument linking authoritarianism and narcissism.