3 ms·

You can get home tests that are similar to PCR. They are more expensive but the performance of the antigen tests since about 2022 is so bad that they’re almost

You can get home tests that are similar to PCR. They are more expensive but the performance of the antigen tests since about 2022 is so bad that they’re almost useless.

jMyles 3y ago
When you say, "so bad", you mean in terms of sensitivity? They seem to be sufficiently selective, but under-sensitive. So, not great for risk aversion.
- cameldrv 3y ago
  Yes sensitivity. False positives are fairly rare.
  - jMyles 3y ago
    Yeah, seems so. So the tests are good as confirmatory devices. But yeah, it was extremely unserious when they were used to gatekeep public events, etc. The reasonable presumption has always been that transmission of respiratory pathogens will occur at public events (which is a strength: humans are quite good at quickly acquiring community immunity - it's one of the reasons our species is so incredibly robust against respiratory pathogens).
    - cameldrv 3y ago
      That is a point of view. In terms of gatekeeping public events, requiring a negative test certainly is going to reduce the number of people who show up with covid, but with only antigen tests, you're reducing the risk maybe 50-60% or so. If you're going to require people to do something annoying, it would be a lot better if the risk reduction were more than that. When you say "humans are good at quickly acquiring community immunity", I don't really think we've done that with Covid. Covid is still over 2% of all deaths in the U.S., and it's causing way more disability and injury. We won't truly know what it's done to us for years.
      - jMyles 3y ago
        > Covid is still over 2% of all deaths in the U.S. The presence of a particular pathogen (or for that matter, any phenomena) coincident with death is not a meaningful metric. Nearly 100% of human deaths involve a person with a belly button. > and it's causing way more disability and injury While it's important and long past-due that post-infection syndromes are finally getting the attention they deserve, there is no evidence that this particular coronavirus is different from other four endemic coronaviruses (and several other pathogens) in this regard.
        cameldrv 3y ago
        > The presence of a particular pathogen (or for that matter, any phenomena) coincident with death is not a meaningful metric. This is where Covid is listed on the death certificate as the primary cause of death. Everyone that dies of cardiac arrest or cancer also has a belly button. > While it's important and long past-due that post-infection syndromes are finally getting the attention they deserve, there is no evidence that this particular coronavirus is different from other four endemic coronaviruses (and several other pathogens) in this regard. Generally there hasn't been all that much research on post-viral syndromes. To some degree that's because they don't obviously affect so many people. It could very well be that other viruses are the cause of more conditions than is currently appreciated though. Certainly the Epstein-Barr/MS results could be the tip of the iceberg. Regardless of what other viruses do, it's looking to me like Covid is likely to be pretty bad long-term, but we won't know for sure for 30+ years I suppose.
        jMyles 3y ago
        > This is where Covid is listed on the death certificate as the primary cause of death. Everyone that dies of cardiac arrest or cancer also has a belly button. Where are you getting death certificate data for 2023? I can't seem to find it yet. If you are looking at WISQARS coded deaths, then I imagine you are already familiar with this critique: https://www.washingtonpost.com/opinions/2023/01/13/covid-pandemic-deaths-hospitalizations-overcounting/ https://www.washingtonpost.com/opinions/2023/01/13/covid-pan... ...so I won't go rehashing it. But in any case, it's a can of worms. If SARS-CoV-2 disappeared magically tomorrow, do we expect overall mortality to suddenly decrease by 2%? Of course not. The deaths will just go back to being coded with their prior ICD-10 code. For example, multiple myeloma deaths caused by respiratory distress will go back to the C00 series. But getting back to the broader point: even in the worst reading, even falling to 2% of all deaths for a deadly respiratory pathogen shows pretty good immune robustness. Some influenza is so lethal to some birds that it becomes the only cause of death for an entire flock, causing a 90% mortality rate. No other species communicates immunity across subpopulations as fast as humans. As far as the mortality data: the change in practice to code deaths as resulting from a particular coronavirus seems like a good evolution in terms of data richness. But to clarify what it means for population mortality, we'd need to do it for other infectious pathogens, not just SARS-CoV-2. To continue the example, if HCoV-HKU1 is the pathogen resulting in lethal respiratory distress for a multiple myeloma patient, than it makes just as much sense to code that death HKU1 as it does for a patient who does from respiratory distress from COVID-19 as COVID-19. > Everyone that dies of cardiac arrest or cancer also has a belly button. Many of those also have COVID-19. And of those, some are coded as COVID-19 and some are not. If a person is infected with a coronavirus and dies from cardiac arrest, is it properly understood as a coronavirus death? Prior to 2020, the answer was always "no" (in fact, there was no ICD-10 code available for this (and to my knowledge, there won't be any deaths coded for coronaviruses other than COVID-19 in this year's dataset either)). But now, the answer is "yes" for only one of the five endemic coronaviruses. The data needs to catch up to the reality in order to have an actionable picture of population mortality. > Generally there hasn't been all that much research on post-viral syndromes. To some degree that's because they don't obviously affect so many people. It could very well be that other viruses are the cause of more conditions than is currently appreciated though. Certainly the Epstein-Barr/MS results could be the tip of the iceberg. Regardless of what other viruses do, it's looking to me like Covid is likely to be pretty bad long-term, but we won't know for sure for 30+ years I suppose. Hopefully the crazy dearth of research on this topic is something that the COVID-19 pandemic will have changed for the better. It's about time. > it's looking to me like Covid is likely to be pretty bad long-term, but we won't know for sure for 30+ years I suppose. Given the unavailability of data on the matter, I'm not sure how you can draw that conclusion. Are people reporting post-viral syndromes following COVID-19? Yes. Are they more frequent than with the other four endemic coronaviruses? We can't know, because we've never tracked this before. But we do know that what we have suddenly started calling "long covid" has been known to science for decades (and can be evinced by searching, for example, "{hku1|oc43|nl61} cardiovascular" on Google Scholar). And sadly, I don't know that we'll really know much better in 30 years. Hopefully we'll understand more about the prognosis, and have some treatments. But distinguishing between the post-infection syndromes of the five endemic coronaviruses might be impossible.
virusduck 3y ago
There is no "since 2022"--they have performed at relatively the same level to date. They just are not perfect and are unlikely to prevent the kind of non-symptomatic exposures people have, but in most cases, they are good to quickly differentiate a random sore throat from a COVID-related one.
- cameldrv 3y ago
  No the performance is worse with newer variants. One factor seems to be that newer variants tend to replicate a lot in the throat at first and only later move to the nose, but there is probably more than one reason for the change. In 2021, the antigen tests were pretty sensitive if you were symptomatic, but now the false negative rate is 50% or greater if you're on your first day or two of symptoms. They perform better a few days later. The problem is that people will have symptoms consistent with covid and they're going to visit grandma or whatever, so rather than just cancel, they take a test, the test is negative, and they go do their thing. IMO going just based on symptoms is more reliable at this point unless you're using a molecular test.

You can get home tests that are similar to PCR. They are more expensive but the performance of the antigen tests since about 2022 is so bad that they’re almost useless.

jMyles 3y ago

When you say, "so bad", you mean in terms of sensitivity? They seem to be sufficiently selective, but under-sensitive. So, not great for risk aversion.

cameldrv 3y ago
Yes sensitivity. False positives are fairly rare.
- jMyles 3y ago
  Yeah, seems so. So the tests are good as confirmatory devices. But yeah, it was extremely unserious when they were used to gatekeep public events, etc. The reasonable presumption has always been that transmission of respiratory pathogens will occur at public events (which is a strength: humans are quite good at quickly acquiring community immunity - it's one of the reasons our species is so incredibly robust against respiratory pathogens).
  - cameldrv 3y ago
    That is a point of view. In terms of gatekeeping public events, requiring a negative test certainly is going to reduce the number of people who show up with covid, but with only antigen tests, you're reducing the risk maybe 50-60% or so. If you're going to require people to do something annoying, it would be a lot better if the risk reduction were more than that. When you say "humans are good at quickly acquiring community immunity", I don't really think we've done that with Covid. Covid is still over 2% of all deaths in the U.S., and it's causing way more disability and injury. We won't truly know what it's done to us for years.
    - jMyles 3y ago
      > Covid is still over 2% of all deaths in the U.S. The presence of a particular pathogen (or for that matter, any phenomena) coincident with death is not a meaningful metric. Nearly 100% of human deaths involve a person with a belly button. > and it's causing way more disability and injury While it's important and long past-due that post-infection syndromes are finally getting the attention they deserve, there is no evidence that this particular coronavirus is different from other four endemic coronaviruses (and several other pathogens) in this regard.
      - cameldrv 3y ago
        > The presence of a particular pathogen (or for that matter, any phenomena) coincident with death is not a meaningful metric. This is where Covid is listed on the death certificate as the primary cause of death. Everyone that dies of cardiac arrest or cancer also has a belly button. > While it's important and long past-due that post-infection syndromes are finally getting the attention they deserve, there is no evidence that this particular coronavirus is different from other four endemic coronaviruses (and several other pathogens) in this regard. Generally there hasn't been all that much research on post-viral syndromes. To some degree that's because they don't obviously affect so many people. It could very well be that other viruses are the cause of more conditions than is currently appreciated though. Certainly the Epstein-Barr/MS results could be the tip of the iceberg. Regardless of what other viruses do, it's looking to me like Covid is likely to be pretty bad long-term, but we won't know for sure for 30+ years I suppose.
        jMyles 3y ago
        > This is where Covid is listed on the death certificate as the primary cause of death. Everyone that dies of cardiac arrest or cancer also has a belly button. Where are you getting death certificate data for 2023? I can't seem to find it yet. If you are looking at WISQARS coded deaths, then I imagine you are already familiar with this critique: https://www.washingtonpost.com/opinions/2023/01/13/covid-pandemic-deaths-hospitalizations-overcounting/ https://www.washingtonpost.com/opinions/2023/01/13/covid-pan... ...so I won't go rehashing it. But in any case, it's a can of worms. If SARS-CoV-2 disappeared magically tomorrow, do we expect overall mortality to suddenly decrease by 2%? Of course not. The deaths will just go back to being coded with their prior ICD-10 code. For example, multiple myeloma deaths caused by respiratory distress will go back to the C00 series. But getting back to the broader point: even in the worst reading, even falling to 2% of all deaths for a deadly respiratory pathogen shows pretty good immune robustness. Some influenza is so lethal to some birds that it becomes the only cause of death for an entire flock, causing a 90% mortality rate. No other species communicates immunity across subpopulations as fast as humans. As far as the mortality data: the change in practice to code deaths as resulting from a particular coronavirus seems like a good evolution in terms of data richness. But to clarify what it means for population mortality, we'd need to do it for other infectious pathogens, not just SARS-CoV-2. To continue the example, if HCoV-HKU1 is the pathogen resulting in lethal respiratory distress for a multiple myeloma patient, than it makes just as much sense to code that death HKU1 as it does for a patient who does from respiratory distress from COVID-19 as COVID-19. > Everyone that dies of cardiac arrest or cancer also has a belly button. Many of those also have COVID-19. And of those, some are coded as COVID-19 and some are not. If a person is infected with a coronavirus and dies from cardiac arrest, is it properly understood as a coronavirus death? Prior to 2020, the answer was always "no" (in fact, there was no ICD-10 code available for this (and to my knowledge, there won't be any deaths coded for coronaviruses other than COVID-19 in this year's dataset either)). But now, the answer is "yes" for only one of the five endemic coronaviruses. The data needs to catch up to the reality in order to have an actionable picture of population mortality. > Generally there hasn't been all that much research on post-viral syndromes. To some degree that's because they don't obviously affect so many people. It could very well be that other viruses are the cause of more conditions than is currently appreciated though. Certainly the Epstein-Barr/MS results could be the tip of the iceberg. Regardless of what other viruses do, it's looking to me like Covid is likely to be pretty bad long-term, but we won't know for sure for 30+ years I suppose. Hopefully the crazy dearth of research on this topic is something that the COVID-19 pandemic will have changed for the better. It's about time. > it's looking to me like Covid is likely to be pretty bad long-term, but we won't know for sure for 30+ years I suppose. Given the unavailability of data on the matter, I'm not sure how you can draw that conclusion. Are people reporting post-viral syndromes following COVID-19? Yes. Are they more frequent than with the other four endemic coronaviruses? We can't know, because we've never tracked this before. But we do know that what we have suddenly started calling "long covid" has been known to science for decades (and can be evinced by searching, for example, "{hku1|oc43|nl61} cardiovascular" on Google Scholar). And sadly, I don't know that we'll really know much better in 30 years. Hopefully we'll understand more about the prognosis, and have some treatments. But distinguishing between the post-infection syndromes of the five endemic coronaviruses might be impossible.

virusduck 3y ago

There is no "since 2022"--they have performed at relatively the same level to date. They just are not perfect and are unlikely to prevent the kind of non-symptomatic exposures people have, but in most cases, they are good to quickly differentiate a random sore throat from a COVID-related one.

cameldrv 3y ago
No the performance is worse with newer variants. One factor seems to be that newer variants tend to replicate a lot in the throat at first and only later move to the nose, but there is probably more than one reason for the change. In 2021, the antigen tests were pretty sensitive if you were symptomatic, but now the false negative rate is 50% or greater if you're on your first day or two of symptoms. They perform better a few days later. The problem is that people will have symptoms consistent with covid and they're going to visit grandma or whatever, so rather than just cancel, they take a test, the test is negative, and they go do their thing. IMO going just based on symptoms is more reliable at this point unless you're using a molecular test.