Cake
• About a week after I came home from Costa Rica via a packed Atlanta airport where we had to stand shoulder-to-shoulder in the security lines and on the train between terminals for 4 hours with 5 young children touching everything, Toni lost her sense of taste and smell for the first time in her life. I had developed a scratchy throat and cough a few days after returning.

Her doc said other viruses can cause loss of taste and smell, true dat, her symptoms weren't very severe and testing isn't available for her level of symptoms.

But today we got word that Stanford is offering free blood testing, first come first serve, one per household. She signed up right away. 10 minutes later they weren't accepting anyone else.

It's funny, we both hope she tests positive so we have immunity, but it's very unlikely. Anyway, it's a simple finger prick and 15 minutes later you know. Apparently it has a sensitivity of 86% and specificity of 90%, whatever that means.

Fantastic image: BioWorld.

• Apparently it has a sensitivity of 86% and specificity of 90%, whatever that means.

Both are a measure of how good the test really is. Thinking of the possible outcomes of the test, there are four of them:

Result negative, your wife didn't have the virus
Result positive, although your wife didn't have the virus
Result negative, although your wife did have the virus

In an ideal world, we'd only ever get the first two results (in bold). We don't live in an ideal world though, so tests can fail and deliver wrong results. This is what sensitivity and specificity refer to:

If your wife really had the virus, there's a 86% probability that the test turns out positive (and 14% that the test result is wrong).

If your wife really didn't have the virus yet, there's a 90% probability that the test will show that (and 10% that it won't).

The worst-case scenario here probably would be getting a positive test result without actually having had the virus, and then prematurely letting down all defenses. So, if the result comes back as positive, celebrate a little between the two of you, but still stay safe! :)

• I was ready to jump in an answer when I first saw your post but @Factotum did a great job of explaining the difference between sensitivity and specificity.

Another way of describing sensitivity is exactly what it sounds like. How sensitive a test is. Low sensitivity means that the test would need high amounts of the target (in this case, COVID-19 antibodies) before being able to provide a positive result. A patient could have antibodies but in a low amount, and low sensitivity tests would return a negative result (false negative). Highly sensitive tests can detect even the smallest amounts of the target.

Specificity on the other hand is how specific a test is at detecting a specific target. Low specificity tests could provide positive results from detecting something other than the intended target (a false positive). This often happens when the testing kit cross-reacts with something other than the intended target, providing a false positive.

• I was ready to jump in an answer when I first saw your post but @Factotum did a great job of explaining the difference between sensitivity and specificity.

Thanks! I just rephrased the two paragraphs starting with "If", because I feel they are more correct this way than the other way around. Please check if that's true. :)

(Before, I had something along the lines of "if the test is positive, there's a 86% probability of really having had the virus".)

• true dat

Exactly as my daughter would have put it !

• I got pretty wacked a few weeks back. I had a fever and my sense of taste and smell got all distorted. I didn't lose totally lose it, but things just tasted weird. I had a fever and I was consistently sick for like 10-14 days. I since have felt totally fine. I don't think I had the virus because I didn't have a sore throat. But I do kinda wonder if I did. Does anyone know: Is a sore throat a necessary symptom of COVID-19?

• So doggie and I drove Toni to a large, open-air park where the testing was happening, away from houses where people live by maybe a mile. The posted signs asked us to keep our windows up. Some very friendly volunteers had us flash our printed docs at them to prove we had a reservation while they stood back 10 feet.

The line of cars waiting was an hour long but that was okay with doggie and I because we had geese to bark at:

• There were 7 tents that served as testing bays so they could process 7 cars simultaneously. The volunteers were so friendly and thanked us for being there. They had very simple protective gear; no N95 masks. A few of them were using painterβs masks they fetched from the painting supplies in their garages.

• Do you know when Toni will get the results?

• Any minute now according to his first post

• They had said the test involved a simple finger prick and lasted 15 minutes. We interpreted that to mean a small amount of blood and fast results.

They clarified: they needed a medium-sized pipette of blood, which takes 15 minutes to withdraw. Results come back in 2-3 days, but only if youβre positive. Then they call you, otherwise you donβt hear. They said false positives are very rare.

They said no matter the result you still have to shelter in place because scientists donβt know yet if you can get reinfected.

The volunteers have to make pretty close contact with you for the 15 minutes it takes to get your π©Έ. You get an Amazon gift certificate at the end. We would have paid to do this.

• So the scientists at Stanford estimated the test accuracy at about 90%. I am sure a lot of things factor into that, but these tweets caught me by surprise. Can any math pros explain?

• 15 minutes to draw - this was finger stick then, NOT by venipuncture. correct?

The statistics you quoted seem pretty remarkable, since 4 known Corona viruses cause common colds, not Covid 19, and this antibody testing needs to avoid reacting to the common cold variety.

Not sure whether I would want a negative or a positive test if it were me. At my age, if I were not ill, I would greatly prefer to be + with antibodies to the virus of Covid 19. I am not alone I think.

And an opinion by Scott Gottlieb MD

• Statistics is a weakness for me as a mathematician: itβs more abstract in that the results you would logically assume to be the answer are often not. But Iβll do my best and hope @Factotum will chime in if Iβve completely botched this explanation.

I believe countries such as Malaysia deal with the limitations of current tests, of whether you are currently infected, by requiring you to test negative three times in a row. If the error rate for a test is 10% then the error rate for all three tests to be incorrect is less than 1%: itβs a probability of 1/10*1/10*1/10.

So multiple tests for having been infected would seem the best way to deal with false positives on an individual level.

Doctor Binneyβs argument seems to be focused on regional or global assessments of herd immunity: the more people who have actually been infected, the greater the reliance that can be placed on the accuracy of the accumulated data.

• I wonder though how much of the accuracy has to do with variable issues like how long has it been since the infection began. In HIV testing, there is an idea that the test accuracy is higher if you get longer after the infection, or so I read on Healthline. I guess it can take the body time to build up enough antibodies to create a positive test.

• The intuition is as follows:

If you test a random individual of, say, the US population, and the test comes back as positive - what does it mean? As I outlined earlier, it can mean one of two things:

a) the individual had the disease, and the test worked correctly
b) the individual did not have the disease, but the test failed and led to a positive result anyway.

Let's split the difference between sensitivity and specificity, and let's just say that the test is "correct" 95% of the time. However, let's also say that only a minority of people really had the disease already, perhaps 1% of the population. Now, if we test the whole population, how many positive results do we expect, and how many of them really had the disease?

Case (a): 95% of 1% --> 0.95% of the population
Case (b) 5% of 99% --> 4.95% of the population

Overall, we will get a positive result for 6% of the population - but only 1/6th of them had the disease, while 5/6th did not.

If we repeat this after, say, 50% of the population had the disease, these values will have changed to:

Case (a) 95% of 50% --> 47.5% of the population
Case (b): 5% of 50% --> 2.5% of the population

The test now delivers a much more accurate result. Of the 50% positive results, an overwhelming number of people really did have the disease, and only some of them are "false positives".

The question is if the initial assumption of testing a random person really holds true in each case that's being tested. If there have been symptoms, for example, or a known contact with a positive person, then the person we're testing likely is not completely random.

• Well it was supposed to detect antibodies. A new Chinese study seems to indicate that young people who had mild symptoms did not develop sufficient antibodies to detect. That would seem to predispose them to getting infected again and may make a vaccine less effective. And that would make herd immunity harder to achieve.

• I think we always have low levels of circulating antibodies anyway, even if we aren't "sick". It's how our bodies "remember" which infections we've had, and is able to mount a quicker immune response in the event of reinfection.

The part about having a severe reaction in the case of reinfection is something I brought up a while ago, I think in the main COVID-19 conversation, using dengue as an example. So far I don't think any cases of reinfection have been reported, and this paper wasn't peer-reviewed either, so while it's wise to remain cautious, I don't think we need to worry about reinfections of COVID-19, yet.

• If young people get infected, clear their infection without detectable antibody responses, then does that mean they are not immune? I don't know. Nor does anyone I suspect, quite yet.

If they get reinfected at a later date , 6 months or more, will they have the same benign illness they had initially, or will it return significantly more severe the second time like can happen with dengue? Again, I don't know, nor does anyone else yet.

There is still much to learn about the Covid 19 virus. One factor that is not understood yet is how many safe harbors it will find - There is some evidence that it can infect tigers, lions, and house cats. If this is true, that might be a real concern - we're not really likely to interact closely with lions or tiger ( the Netflixs Tiger King notwithstanding ) but domestic cats live in lots of folks houses and some roam neighborhoods at night without surveillance. Can humans catch Covid 19 from domestic cats - who really knows yet? But maybe???....

Typically after a viral infection one gets a spike in immune globulins within 2-4 weeks which gradually falls unless one is re-exposed to the virus. So this pattern of minmal Ab titer seems unusual to me but I am not a virologist or immunologist. We need more authoritive answers I think. They will be forthcoming but it may take a while. The folks at Stanford may have some answers they haven't published yet.

One interesting point in the article I linked is the possibility of ivermectin inhibiting viral replication of Corana virus.

Thinking along these lines, reminded me of a virus exposure of mine back in 2001. After September 11 and the destruction of the towers in New York City, the excitement didn't stop then - someone began to mail around weapons grade anthrax spores, and there apparently was some serious concern that small pox might be added to the viral exposures - Small pox was offically declared extinct in human populations by WHO on May 7 , 1980.

Despite the fact that no small pox was in human populations, there were two small samples retained - one by the USA and one by the USSR. Physicians were quietly polled, nationally, to see who would be willing to examine potential smallpox victims, and specifically they were looking for ophthalogists who could biomicroscopically exam human corneas for signs of small pox. I volunteered as I KNEW I was successfully vaccinated for small pox as a young child ~ 55 years earlier. One of the requirements for being accepted as a volunteer, was that I had to be willing to be re-vaccinated for smallpox again.....

I said sure, I had great faith in my childhood vaccination.

So I reported to my local county health nurse who pulled out a tiny glass vial from the 1950s containing viable vaccina virus and proceeded with scarification - multiple tiny punctures in my skin to allow the virus access to my tissues. She put a bandaid on it and off I went home.

I didn't think too much about it until about 36 hours later when the viremia poured through me and my temperature rose to 101 or higher and I felt like I was getting a very very bad case of the flu. This viremia lasted less than one day and my immunity popped up and I fully recovered. But that viremic phase was more severe than any flu I have had in any of the following 18 years since.

I still remember it, and the brief, but temporary, seconds thoughts I had about the wisdom of volunteering....

Reading about small pox is still sobering, and ought to be required of all who deny the great value of vaccination. Almost nothing medicine has done in the last 200 years, has saved as many lives as vaccination, starting with Jenner, and vaccination with cow pox virus.

• Well darn it, you leave a post and an hour later you read something new again -

Researchers in S Korea are reporting patients who recovered from Covid 19 ( and recovered in S Korea means two negative tests for virus on two separate days ) who experience a reactivation of the virus, and a return of symptoms of Covid 19.....

That sequence kind of make me wonder about the accuracy, reliablity, sensitivity of the tests being done. But that just a WAG of mine. Obviously, this needs further clarification/verification, and the S Koreans are working on that too.

One topic I have not seen discussed is that of the size of the innoculum - does the degree of initial exposure to the virus offer any insight into the severity of the patients illness, and or chances of recovery.