Bolding added.
I seriously doubt that a medical professional said that.
Most doctors and nurses are not trained scientists, and are susceptible to cognitive flaws like confirmation bias just the same as anyone.
Although it’s implausible that a medical professional would have gone straight to COVID-19 without first ruling out Lupus.
The proper scientific way to approach any test is through Baysian statistics. If a person takes a test and it comes back negative, there are two possibilities: The person really is negative, or they’re positive and the test failed. Which is more likely? Well, the answer to that will depend on what you already knew. Before you even took the test, what was the prior estimate for the probability? Someone who has a fever and trouble breathing is more likely to have COVID-19 than someone who’s showing no symptoms. I don’t know the exact figures for this case, but it’s quite possible that the prior probability that Dallas Jones’ son had COVID-19 was high enough that, even with a negative test result, his probability of having it is still over 50%. And of course, even if he doesn’t have COVID-19 specifically, it’s extremely likely that his symptoms are due to some communicable disease or another, in which case the proper response would be mostly the same.
One might, of course, ask why one would bother with the test, when one’s response is going to be the same no matter what the test says. But even if the results aren’t relevant for his case, it still provides more data for studying the epidemic as a whole.
Playing devil’s advocate here, but how certain are you in making this assertion?
I think you need to at least know the population you’re dealing with, and a specific definition of “have COVID-19”.
What conceivable circumstances could there be in which showing symptoms consistent with a disease would not increase the probability of having the disease?
Well, okay, I can come up with a convoluted scenario. The disease is extremely rare, and there’s another much more common disease with identical symptoms, and [for reasons] it’s impossible to have both diseases at the same time.
But I can’t think of any realistic scenario.
“[H]ave COVID-19”, does this include asymptomatic carriers?
The prison my son works in early on had a cell block with many individuals with respiratory signs and fever. When tested they were negative for COVID19, but positive for plain old influenza. Throw in a high percentage of asymptomatic carriers, and things get confusing.
We’re trying to define probabilities, so we don’t have to be sure. We know that sometimes people with Covid-19 have no symptoms, we also know that people who do NOT have Covid-19 also don’t have symptoms. So, here is a person without any symptoms, do they have Covid-19? We have a pre-existing belief that MOST people without symptoms do not have Covid-19, so when somebody does not have symptoms, we say they probably do not have Covid-19.
Maybe based on other (made up as an example) data, we believe that 10% of the population as a whole DOES have Covid-19, but no symptoms. Then when we see somebody with no symptoms, we might say we are 90% sure they do not have Covid-19. The person with no-symptoms also gets a negative result on a Covid-19 test which we think has a 1% false negative rate. So now we can say that we are 99.9% sure that the person does not have Covid-19 (0.10 * 0.01 = 0.001). This is where lots of people get very upset. 99.9% sure seems like a big number, but if we test 10,000 people without symptoms, we will expect to be wrong 10 times saying somebody does not have Covid-19.
It works the same way for somebody with symptoms. If somebody shows classic Covid-19 symptoms, and is in an area where the disease is spreading (which is almost everywhere), then we think they probably have Covid-19. The test would add even more evidence that they do actually have Covid-19. So if the test gives a false negative even 1% of the time, then maybe that patient is in that 1%. If lots of patients have Covid-19 symptoms but are coming back negative on the test, then our assumptions are probably wrong. Perhaps the false negative is not 1%, but 25%. Perhaps it is not Covid-19, but Covid-20, and the virus is just enough different to not be detected by the Covid-19 test, but otherwise has the same symptoms.
It is always possible they instead have the flu and an opportunistic pneumonia. So, instead of the 1% of time that the test is wrong, this is the 1% of time that the nurse or doctor is wrong. I’m not that kind of doctor, but I’d think it would be worth it give somebody like that a flu test, too. Covid-19 like symptoms, but negative on the Covid-19 test and positive on a flu test, then it is probably the flu. Still could be wrong about that, though.
What if somebody is an asymptomatic carrier of Covid-19, but has the flu? It’s probabilities the entire way down.
It is appropriate to have a pedantic aside here.
COVID-19 is coronavirus disease associated with infection with SARS-CoV-2, which emerged in 2019. If an individual is infected with the virus but has no disease (i.e. no symptoms) they have been infected with SARS-CoV but they do not have the disease; they do not have COVID-19.
From a practical POV there is little difference between those who have infections with SARS-CoV-2 that never become symptomatic, and those who have had such mild disease that they never realized it, unless such translates into different degrees of infectivity … which is not currently clear.
But in a discussion like this precision is important.
every medical test has false positives. A while back I got a positive test for Hepatitis C but they did a better test and it came back negative.
Except that such is not really a cognitive flaw. They may actually understand Bayes theorem. Something sounding like a horse while surrounded by horses has high priors of being a horse. Even a test with a fairly low false negative rate is going to call many true positives not horse in that context. And most calls of “not horse” would be wrong.
Clinically if COVID-19 is frequent in the clinical population, very little causing those symptoms is, and someone has those symptoms, a good clinician would not trust a single negative result too much.
NM - hampsters.
They take revenge when posters can’t spell their species correctly.
The staffer is Katie Miller, the wife of real life Gorgomel, Stephen Miller.
I wonder if they will follow advice to quarantine them both?
I’m sure there is space in the immigrant concentration camps for Stephen.
IME, nurses don’t make diagnoses themselves. Also, no professional would make a diagnosis without the relevant testing.
Er… nurses routinely have opinions as to what is ailing their patients. They just aren’t allowed to share those opinions with their patients. And every professional has some a priori guesses as to what is the problem, and uses those guesses both to recommend tests. Frequently, they also recommend a course of action prior to the actual testing. For instance, my doctor recommended I see an orthopedic guy and get some scans for my wrist pain, but also recommended I stop using a mouse with that hand and take NSAIDs until I had more information. And… that actually cured the problem, which the specialist then had trouble diagnosing, because it was mostly better by the time I saw him.
Moderator Note
Let’s refrain from political jabs in this forum.
Colibri
QZ Moderator
Opinions, sure. Diagnoses, not so much.