Ambiguous covid tests
March 27, 2022 3:50 AM Subscribe
I had Covid symptoms. I got a positive result on a home antigen test. The next day I got a PCR test which came back negative. Looking for data on how to interpret this.
I am going to continue to isolate, so that's not part of the question.
I started taking Paxlovid (anti-viral meds) and as it has risks, it's better not to take it if you don't have Covid.
I know the PCR test is the gold standard but home false positives with symptoms are also rare.
Also relevant: I did a much, much more thorough swabbing than the tester, who did not seem (to me) to be thorough in their swabbing, compared to how PCR tests have felt in the past.
Looking for data that will help me interpret the combination of positive home test, negative pcr test and symptoms.
I am going to continue to isolate, so that's not part of the question.
I started taking Paxlovid (anti-viral meds) and as it has risks, it's better not to take it if you don't have Covid.
I know the PCR test is the gold standard but home false positives with symptoms are also rare.
Also relevant: I did a much, much more thorough swabbing than the tester, who did not seem (to me) to be thorough in their swabbing, compared to how PCR tests have felt in the past.
Looking for data that will help me interpret the combination of positive home test, negative pcr test and symptoms.
It takes a day or two longer for a PCR to work. The result means you acted promptly.
posted by SemiSalt at 4:59 AM on March 27, 2022 [3 favorites]
posted by SemiSalt at 4:59 AM on March 27, 2022 [3 favorites]
The same thing happened to a friend, and it took a few days for the pcr to show up positive.
posted by melamakarona at 5:43 AM on March 27, 2022 [1 favorite]
posted by melamakarona at 5:43 AM on March 27, 2022 [1 favorite]
I did a much, much more thorough swabbing than the tester, who did not seem (to me) to be thorough in their swabbing, compared to how PCR tests have felt in the past.
Like any kind of test, PCR tests have a false negative rate too, and half-assed swabbing technique probably does account for a fair number of those.
The beauty of the home RAT is that it's cheap enough to be repeated, and by the time you've had a RAT result confirmed on three successive days, the chance that it's wrong is low enough to be just about negligible.
posted by flabdablet at 5:57 AM on March 27, 2022 [3 favorites]
Like any kind of test, PCR tests have a false negative rate too, and half-assed swabbing technique probably does account for a fair number of those.
The beauty of the home RAT is that it's cheap enough to be repeated, and by the time you've had a RAT result confirmed on three successive days, the chance that it's wrong is low enough to be just about negligible.
posted by flabdablet at 5:57 AM on March 27, 2022 [3 favorites]
I have seen more than one scandal of some covid testing locations returning ONLY negative test results. You may be able to look into this and see if yours is one (it’s been happening around the US as well, the linked example is not in the US)
https://www.opendemocracy.net/en/ournhs/i-raised-an-early-alarm-on-pcr-test-scandal-but-authorities-ignored-me/
posted by bilabial at 6:01 AM on March 27, 2022 [1 favorite]
https://www.opendemocracy.net/en/ournhs/i-raised-an-early-alarm-on-pcr-test-scandal-but-authorities-ignored-me/
posted by bilabial at 6:01 AM on March 27, 2022 [1 favorite]
Can you take a second rapid antigen test, preferably a different brand or at least a different batch? If that one is also positive, I would assume that you have covid and the PCR was a false negative.
posted by maleficent at 6:58 AM on March 27, 2022
posted by maleficent at 6:58 AM on March 27, 2022
Best answer: This NEJM chart depicts the possible timelines of test accuracy in a way I find super helpful.
posted by chesty_a_arthur at 7:20 AM on March 27, 2022 [6 favorites]
posted by chesty_a_arthur at 7:20 AM on March 27, 2022 [6 favorites]
Please finish your course of Paxlovid unless your doctor tells you to stop. The idea is that it will stop viral replication, which would result in negative test results. That’s the point and an indication that you started taking it early enough in your infection for it to be effective. Stopping it early is a very bad idea in the same way that stopping antibiotics early is a very bad idea. The risks associated with Paxlovid are quite manageable and slight, even compared to antibiotics. If your doctor felt it was the right course of action for you, it is highly unlikely that you need to be personally concerned about possible side effects. Continue taking your medication as prescribed unless your doctor says otherwise. But again, you would EXPECT a negative test result after beginning a course of antiviral. And the chances of a positive antigen test being false are very small. You don’t have ambiguous results. You have symptoms + a positive test.
posted by Bottlecap at 7:47 AM on March 27, 2022 [2 favorites]
posted by Bottlecap at 7:47 AM on March 27, 2022 [2 favorites]
Best answer: The exact answer depends on the rates of false negatives and positives for both tests, given the times you took them, as well as the “prior probability”, the probability you would have given to your having Covid before you took either test — this will depend on your exposure history and whether your area is experiencing a surge.
In cases where the test was taken long enough after the onset of symptoms and the prior probability was low — no known exposure, no surge — the numbers work out that probably the PCR was correct, the person doesn’t have Covid.
But when the test is taken early, that’s increasing the rates of false negatives dramatically, tipping the scale in favor of it really being Covid. And for people who have had prior exposure, either through vaccination or previous infection, symptoms will tend to show up earlier in the course of the infection, meaning that tests at first symptoms are more likely to show false negatives.
So… the math is complicated, and depends on some numbers that are really hard to accurately estimate. Best to play it safe.
posted by wyzewoman at 8:01 AM on March 27, 2022
In cases where the test was taken long enough after the onset of symptoms and the prior probability was low — no known exposure, no surge — the numbers work out that probably the PCR was correct, the person doesn’t have Covid.
But when the test is taken early, that’s increasing the rates of false negatives dramatically, tipping the scale in favor of it really being Covid. And for people who have had prior exposure, either through vaccination or previous infection, symptoms will tend to show up earlier in the course of the infection, meaning that tests at first symptoms are more likely to show false negatives.
So… the math is complicated, and depends on some numbers that are really hard to accurately estimate. Best to play it safe.
posted by wyzewoman at 8:01 AM on March 27, 2022
I think the false-positive rate for LFTs is much, much lower than the false-negative for PCR tests, even without taking testing procedures, timeline and base rates (priors) into account. That is, it's almost impossible to get a false positive on the LFT, while false negatives on PCR is (somewhat) common. I'd also second what Bottlecap said.
I would absolutely believe the LFT and continue the Paxlovid.
posted by Dashy at 8:14 AM on March 27, 2022 [1 favorite]
I would absolutely believe the LFT and continue the Paxlovid.
posted by Dashy at 8:14 AM on March 27, 2022 [1 favorite]
“ the false-positive rate for LFTs is much, much lower than the false-negative for PCR tests.”
This is the opposite of what I understand. I think OP needs to consult with the doctor who prescribed Paxlovid and schedule a second PCR. In the US, there’s a huge spike in non-Covid bugs right now (rhinovirus etc.) My kid has a non-covid snotty cold (confirmed by multiple negative antigen tests and a PCR.)
posted by haptic_avenger at 9:41 AM on March 27, 2022 [1 favorite]
This is the opposite of what I understand. I think OP needs to consult with the doctor who prescribed Paxlovid and schedule a second PCR. In the US, there’s a huge spike in non-Covid bugs right now (rhinovirus etc.) My kid has a non-covid snotty cold (confirmed by multiple negative antigen tests and a PCR.)
posted by haptic_avenger at 9:41 AM on March 27, 2022 [1 favorite]
When testing symptomatic people, false positives are rare with both rapid antigen tests and PCRs.
posted by jesourie at 12:20 PM on March 27, 2022 [1 favorite]
posted by jesourie at 12:20 PM on March 27, 2022 [1 favorite]
I would not discontinue the medication without talking to the prescriber. Yes, false negative PCR tests are a thing, for reasons including faulty technique, storage conditions, and sometimes patient-related factors.
posted by M. at 12:46 PM on March 27, 2022
posted by M. at 12:46 PM on March 27, 2022
“When testing symptomatic people, false positives are rare with both rapid antigen tests and PCRs.”
This is probably true when in a Covid surge and other viruses aren’t going around. But if you check CDC surveillance, all the spring respiratory bugs are surging, and Covid rates are low most places.
posted by haptic_avenger at 1:17 PM on March 27, 2022
This is probably true when in a Covid surge and other viruses aren’t going around. But if you check CDC surveillance, all the spring respiratory bugs are surging, and Covid rates are low most places.
posted by haptic_avenger at 1:17 PM on March 27, 2022
PCR false negative rates are very high - starting at 20% and can get as high as 66%. They are still better than rapid tests, but nowhere near perfect.
posted by todolos at 1:31 PM on March 27, 2022
posted by todolos at 1:31 PM on March 27, 2022
The false negative rate for PCR in that article is related to time since infection (more accurate as viral load increases). Antigen tests are less sensitive than PCR and will give false negatives when PCRs give positives early in the illness. So the sequence of positive antigen - negative PCR would make me consider a false positive. But again OP should redo the PCR and ask their doctor.
posted by haptic_avenger at 1:45 PM on March 27, 2022
posted by haptic_avenger at 1:45 PM on March 27, 2022
Despite what some people are saying above, it's very unlikely that one gets a false positive on an LFT. Their lack of sensitivity (compared to PCR tests) means they are more likely to give false negatives- but that doesn't work both ways. LFT's are highly specific, meaning that a positive test is extremely likely to be correct; and in fact you are more that 96% likely to to have COVID if you test positive on an LFT.
Can you then test negative on a PCR? As chesty_a_arthur pointed out, timing can make a difference, but PCRs also have a false negativity rate of around 5%*. That means one of every 20 people who truly have COVID could test negative on a PCR test- a much larger number that people getting a false positive LFT.
* compare this to the chances of getting a false negative on a LFT which by some estimates is around 30%. This is why they are said to be less sensitive than a PCR test: they do not always detect the virus when it is actually present. Imagine two people, one with a poor and one with an excellent sense of smell. Maybe the person with a poor sense of smell needs more smoke to be present in order to detect a fire with their nose than the super-smeller, but they are extremely unlikely to be wrong if they do detect smoke.
posted by oneirodynia at 11:34 PM on March 27, 2022 [2 favorites]
Can you then test negative on a PCR? As chesty_a_arthur pointed out, timing can make a difference, but PCRs also have a false negativity rate of around 5%*. That means one of every 20 people who truly have COVID could test negative on a PCR test- a much larger number that people getting a false positive LFT.
* compare this to the chances of getting a false negative on a LFT which by some estimates is around 30%. This is why they are said to be less sensitive than a PCR test: they do not always detect the virus when it is actually present. Imagine two people, one with a poor and one with an excellent sense of smell. Maybe the person with a poor sense of smell needs more smoke to be present in order to detect a fire with their nose than the super-smeller, but they are extremely unlikely to be wrong if they do detect smoke.
posted by oneirodynia at 11:34 PM on March 27, 2022 [2 favorites]
Despite what some people are saying above, it's very unlikely that one gets a false positive on an LFT.
Except in real-world analysis, the false-positive rate of LFTs/antigen tests has been much higher than the approval data submitted by manufacturers.
For PCR false negativity rate, at least one study has attributed it to low viral loads. So it would be unlikely that someone who got a false negative on a PCR (likely due to low viral load) would have tested positive earlier on an LFT/antigen (which requires a higher viral load than PCR.)
posted by haptic_avenger at 6:55 AM on March 28, 2022
Except in real-world analysis, the false-positive rate of LFTs/antigen tests has been much higher than the approval data submitted by manufacturers.
For PCR false negativity rate, at least one study has attributed it to low viral loads. So it would be unlikely that someone who got a false negative on a PCR (likely due to low viral load) would have tested positive earlier on an LFT/antigen (which requires a higher viral load than PCR.)
posted by haptic_avenger at 6:55 AM on March 28, 2022
I did a much, much more thorough swabbing than the tester, who did not seem (to me) to be thorough in their swabbing, compared to how PCR tests have felt in the past
I don't know if you'll find this reassuring, but: at the sites I'm familiar with, there's a great variety in the types of swabs that are available any given day. Two tests can be done equally well but feel very different to the patient.
posted by The corpse in the library at 4:43 PM on March 28, 2022
I don't know if you'll find this reassuring, but: at the sites I'm familiar with, there's a great variety in the types of swabs that are available any given day. Two tests can be done equally well but feel very different to the patient.
posted by The corpse in the library at 4:43 PM on March 28, 2022
« Older where to buy new battery for Lenovo Thinkpad E450 | 10”+ Android tablet that is fast with good drawing... Newer »
This thread is closed to new comments.
posted by lesser whistling duck at 3:52 AM on March 27, 2022