|
Any and all updates regarding the COVID-19 will need a source provided. Please do your part in helping us to keep this thread maintainable and under control.
It is YOUR responsibility to fully read through the sources that you link, and you MUST provide a brief summary explaining what the source is about. Do not expect other people to do the work for you.
Conspiracy theories and fear mongering will absolutely not be tolerated in this thread. Expect harsh mod actions if you try to incite fear needlessly.
This is not a politics thread! You are allowed to post information regarding politics if it's related to the coronavirus, but do NOT discuss politics in here.
Added a disclaimer on page 662. Many need to post better. |
On August 27 2020 12:15 cLutZ wrote:Show nested quote +On August 27 2020 10:07 Mohdoo wrote:On August 27 2020 09:28 cLutZ wrote:On August 27 2020 09:09 Mohdoo wrote:
CDC saying not to test non-symptomatic is so amazingly dangerous. How utterly stupid. If you are talking about this guidance it seems reasonable. A major problem of the Covid tests is false positives and false. Thus, if you are testing large populations of people you generate a lot of noise. This part seems particularly important. If you have been in close contact (within 6 feet) of a person with a COVID-19 infection for at least 15 minutes but do not have symptoms:
You do not necessarily need a test unless you are a vulnerable individual or your health care provider or State or local public health officials recommend you take one.
A negative test does not mean you will not develop an infection from the close contact or contract an infection at a later time.
You should monitor yourself for symptoms. If you develop symptoms, you should evaluate yourself under the considerations set forth above.
You should strictly adhere to CDC mitigation protocols, especially if you are interacting with a vulnerable individual. You should adhere to CDC guidelines to protect vulnerable individuals with whom you live.
I think we are at a point where I need to ask: what is your technical background? To what extent have you conducted research on viruses? You said testing too many people with a noisy test makes more noise. That is perhaps among the most blatantly incorrect things you have said in this thread. You appear to not have a technical background, but you speak as if you know about this topic. We have seen numerous studies that show whether you are old or young, there are many people with extremely high viral concentrations in their nose and throat. That means those people can spread it, because as air passes by the virus, the virus hops along and is able to travel on particles/droplets. Let me reiterate: you do not need to show symptoms to be infectious. You do not need to cough to pass the virus. Simply speaking is sufficient. Saying a negative tests doesn't always mean someone isn't infected is a simple matter of existence. That is a really poor attempt to discredit testing. That is true for every single test ever made. Since one of the best things about testing is signaling people to quarantine, we absolutely must continue to test non-symptomatic people. It is dangerous not to. My technical background is in biomedical engineering where my specific work was with VADs, so I'm not an expert on this particular test. But in hospital work you generally learn that you need to watch out for both type 1 and type 2 errors, and both come into play with C-19 if you are testing asymptomatic people (not that you universally should not test asymptomatic people, its a judgement call as to when you should as I'll try to explain). In their explanation (which I bolded) they explicitly warn about type 2 errors, that being false negative tests for people who actually have the disease. The risk with this is that if someone is exposed, and you test them, then it comes back negative, they will act like they don't have the virus and not monitor for symptoms or self-quarantine (I am currently doing so because I was at the hospital, even though I had a negative test). They aren't explicitly discussing type 1 errors, false positives, but this is also a big problem for some tests, and C19 testing seems to have a significant level of this. This is why indiscriminate mammograms cause a lot of problems, as another example. Its fine to be indiscriminately testing so long as C19 has high prevalence in the population because you will mostly getting signal, but as it % of prevalence drops false positives will start dominating the statistics (as we saw with the NFL players example). According to this false positives are less than 5%, however this chart makes me suspect false positivity is more like 1%. What this means, is that if your state has a positive test rate of around that mark, most of your positives are actually false. Overall, seeing the 2nd chart, I agree with you that their judgement call is likely wrong at least for the states on the right hand side. False positivity is not a worry for those states, but if your %positive rate is <1% you probably need to re-evaulate procedures. You can use Bayes theorem to estimate what the "target" positive rate should be and test more/less based on this. Lets say you dont want more than 50% of positive results to be false positives. If you assume the false positive rate is 1% (and for simplicity say no false negatives), then your target is for 2% of results to be positive or more. So if %false positive is <1% this call is too early, but it is a call you eventually make. The less accurate you think tests are, the sooner you make it.
It is important to keep in mind that VADs and infectious disease research are very different. Both are biologically relevant, but that's about it.
Are you familiar with the specifics of how you work up samples for PCR, run PCR, and interpret PCR data? If so, you will know that the accuracy of the test is essentially up to the people running it. There are a few different kits that labs can use and each have their own pros and cons. And since different tests are looking for slightly different hings, there is a lot of nuance. But at the end of the day, you essentially are given an S-shaped curve, and based on when this curve takes shape, you can ascertain concentration, which you use to determine infection.
Lots of things can go wrong in PCR, but luckily you can always inspect your data and run repeats. It is easy for a lab to run a 4x repeat on each sample to ensure as close to 100% accuracy as you can get. This is possible for both environmental and human samples. There ABSOLUTELY are labs out there that likely opportunistically tried to do covid testing and do a bad job at it. If you are bad at running a test, the test will suck. That says nothing about the effectiveness of covid tests, as we have seen elsewhere. So it is entirely possible that you or anyone could point to some janky lab that had a 5% false positive rate. That means nothing. A good method being used poorly is nothing new to the world.
Look at reputable organizations that have a lot on the line. They are doing really good testing. The link you provided is an example of why academia needs to stop eating shit when it comes to messaging. A single run of a single test can totally blow ass. But that is why any reputable company is running lots of repeats so that you are able to guarantee accuracy through data analysis and repetition. They are not saying that as little as 2% of all test results conveyed to people are wrong. They are ********NOT******** saying that.
Another thing is that it is possible to detect dead forms of the virus, depending on how you are conducting the test. But you are also able to develop a known profile of what dead virus looks like and screen against that. I need to make it abundantly clear that the core of your argument, that covid results are fundamentally inflated and wrong, is in no way valid.
This is similar to the publications talking about how antibodies last 3 months. In academia, it would be irresponsible to say anything other than "antibodies last 3 months" because they haven't been studied for longer. People think that means antibodies only last 3 months. Great. Or the study that showed antibodies weren't detected in someone who was infected 6 months ago. But that is normal. Human bodies keep memories of antibodies for use later, but in general, antibodies for infections naturally fade away in concentration since they are no longer needed. Publishing a study showing that antibodies are leaving at a normal rate is a normal thing to do in academia. But when Joe Shmoe reads that, he assumes immunity is only temporary. It is a total shit show.
So to that point: academia has a ton of guilt with regards to covid disinformation. Their standard publishing practices needed to be completely abolished as soon as this became a public issue. Everything should be published in a way that common people can understand. Using the terminology typical of virology research is very bad right now.
|
France update :
Case numbers are going up, up, up, around 20% increase every day... 7400 cases yesterday. It's near exponential currently. (and that's not due to the increase in testing, case positivity ratio is up three-fold in a month)
However, ICU and hospitalisations are still on a flat line sooooo.... I'm still torn. Either older people are really cautious and youngsters don't care, or somehow the virus got less aggressive... Probably a combination of both.
Youngsters do NOT care currently from what I see, that's for sure. They don't want to "waste" a year of partying, and it's reckless for their families.
Schools are starting again in a week. My bet is on them being open "normally" at first, with waves of closures due to cases in the first few weeks... followed by a change of plans by the government.
Mask mandate are now in place in most of the big cities, and it should accelerate fast.
At work, we're probably going to fall back to 2 separate teams pretty soon (as we can't work from home, and we need to keep things running at all costs). Probably as soon as the first case hits the camp.
Winter has me worried.
|
On August 29 2020 02:59 Nouar wrote:
However, ICU and hospitalisations are still on a flat line sooooo.... I'm still torn. Either older people are really cautious and youngsters don't care, or somehow the virus got less aggressive... Probably a combination of both.
Could just be that it's not there yet. 7k daily is alarming, but if the test data is very strong - maybe it's just that the people that would be hospitalized are those that got sick two weeks ago, back when the case loads were much lower.
|
Update for the netherlands.
Numbers and hospitalizations have been more or less flat here (the netherlands) as well.
Numbers because we pretty much hit max testing capacity for now while still not much of a selection is beeing made on who to test. Though our sewage data also shows declining numbers.
Many people with no symptoms apply for a test. For example if you come from abroad from a high risk area you are advised (quarantine is still not legally obligated in any situation in the netherlands) to self quarantine and then people get a test because they argue that if the test shows negative there is no need anymore for the quarantine (which as said isnt even legally obligated and enforced to begin with).
Hospitalizations are still low,yesterday on tv there was a discussion about this. One reason given was that the elderly and vulnerable people are pretty much self isolating themselves to be safe. I dont know how much of that is true,some of them obviously are but not all of them. Infections are mostly amongst young people now and we dont have the spike in the nurserys that we had in march/april.
The government casually said during a press conference yesterday that we are on a strategy of maximum control of the epidemic,which is different from the start where the strategy was a controlled spread (and max control and not doing anything beeing mentioned as the other 2 options when they explained which option they chose).
They have changed their strategy (though they never did announce so officially) in that aspect and its not fully clear why. Maybe because it has become clear that herd immunity is mathematically impossible and maybe also the fear for the long term effects of the virus even with mild cases,about which not much is known currently.
.
We had an experiment with masks in amsterdam and rotterdam which has been terminated and masks are no longer requiered except for public transport. Schools have opened/are opening (depends on the area, we slightly spread the holidays for different areas).
Support for the economy (businesses big and small) is relatively generous and will be extended for 9 months,it was announced today. They also said (casually again) that the negative economic effects will remain for a longer time then initially anticipated.
They are now counting on -5% for the economy this year over the whole year which i think is relativly mild,though they also said the worst might yet be to come. (germany will probably end up beeing somewhat similar). The support meassures have held up the economy pretty well untill now.
Its back to where we were 2 years ago which objectivly isnt the end of the world though there is a huge difference between all the different sectors. Some are hurt very bad while some others even got a boost. And as said,the worst economic hit might still be ahead of us.
|
The explanation is quite simple actually. It's summer, which means high humidity and good ventilation of inside spaces (might be worse in US because of air conditioning). Which means people that get infected get a low concentration of the virus. So, it's much easier for their bodies to fight it. If we get in the winter with such numbers however, the hospitals will get crowded quickly.
|
That's true; respiratory infections certainly are historically less severe in summer months, and coronavirus may follow the same pattern.
The number of people dying is very large all the same. So maybe, as bad as things are, there's always opportunity for it to get a whole lot worse?
|
On August 29 2020 00:35 Mohdoo wrote:Show nested quote +On August 27 2020 12:15 cLutZ wrote:On August 27 2020 10:07 Mohdoo wrote:On August 27 2020 09:28 cLutZ wrote:On August 27 2020 09:09 Mohdoo wrote:
CDC saying not to test non-symptomatic is so amazingly dangerous. How utterly stupid. If you are talking about this guidance it seems reasonable. A major problem of the Covid tests is false positives and false. Thus, if you are testing large populations of people you generate a lot of noise. This part seems particularly important. If you have been in close contact (within 6 feet) of a person with a COVID-19 infection for at least 15 minutes but do not have symptoms:
You do not necessarily need a test unless you are a vulnerable individual or your health care provider or State or local public health officials recommend you take one.
A negative test does not mean you will not develop an infection from the close contact or contract an infection at a later time.
You should monitor yourself for symptoms. If you develop symptoms, you should evaluate yourself under the considerations set forth above.
You should strictly adhere to CDC mitigation protocols, especially if you are interacting with a vulnerable individual. You should adhere to CDC guidelines to protect vulnerable individuals with whom you live.
I think we are at a point where I need to ask: what is your technical background? To what extent have you conducted research on viruses? You said testing too many people with a noisy test makes more noise. That is perhaps among the most blatantly incorrect things you have said in this thread. You appear to not have a technical background, but you speak as if you know about this topic. We have seen numerous studies that show whether you are old or young, there are many people with extremely high viral concentrations in their nose and throat. That means those people can spread it, because as air passes by the virus, the virus hops along and is able to travel on particles/droplets. Let me reiterate: you do not need to show symptoms to be infectious. You do not need to cough to pass the virus. Simply speaking is sufficient. Saying a negative tests doesn't always mean someone isn't infected is a simple matter of existence. That is a really poor attempt to discredit testing. That is true for every single test ever made. Since one of the best things about testing is signaling people to quarantine, we absolutely must continue to test non-symptomatic people. It is dangerous not to. My technical background is in biomedical engineering where my specific work was with VADs, so I'm not an expert on this particular test. But in hospital work you generally learn that you need to watch out for both type 1 and type 2 errors, and both come into play with C-19 if you are testing asymptomatic people (not that you universally should not test asymptomatic people, its a judgement call as to when you should as I'll try to explain). In their explanation (which I bolded) they explicitly warn about type 2 errors, that being false negative tests for people who actually have the disease. The risk with this is that if someone is exposed, and you test them, then it comes back negative, they will act like they don't have the virus and not monitor for symptoms or self-quarantine (I am currently doing so because I was at the hospital, even though I had a negative test). They aren't explicitly discussing type 1 errors, false positives, but this is also a big problem for some tests, and C19 testing seems to have a significant level of this. This is why indiscriminate mammograms cause a lot of problems, as another example. Its fine to be indiscriminately testing so long as C19 has high prevalence in the population because you will mostly getting signal, but as it % of prevalence drops false positives will start dominating the statistics (as we saw with the NFL players example). According to this false positives are less than 5%, however this chart makes me suspect false positivity is more like 1%. What this means, is that if your state has a positive test rate of around that mark, most of your positives are actually false. Overall, seeing the 2nd chart, I agree with you that their judgement call is likely wrong at least for the states on the right hand side. False positivity is not a worry for those states, but if your %positive rate is <1% you probably need to re-evaulate procedures. You can use Bayes theorem to estimate what the "target" positive rate should be and test more/less based on this. Lets say you dont want more than 50% of positive results to be false positives. If you assume the false positive rate is 1% (and for simplicity say no false negatives), then your target is for 2% of results to be positive or more. So if %false positive is <1% this call is too early, but it is a call you eventually make. The less accurate you think tests are, the sooner you make it. It is important to keep in mind that VADs and infectious disease research are very different. Both are biologically relevant, but that's about it. Are you familiar with the specifics of how you work up samples for PCR, run PCR, and interpret PCR data? If so, you will know that the accuracy of the test is essentially up to the people running it. There are a few different kits that labs can use and each have their own pros and cons. And since different tests are looking for slightly different hings, there is a lot of nuance. But at the end of the day, you essentially are given an S-shaped curve, and based on when this curve takes shape, you can ascertain concentration, which you use to determine infection. Lots of things can go wrong in PCR, but luckily you can always inspect your data and run repeats. It is easy for a lab to run a 4x repeat on each sample to ensure as close to 100% accuracy as you can get. This is possible for both environmental and human samples. There ABSOLUTELY are labs out there that likely opportunistically tried to do covid testing and do a bad job at it. If you are bad at running a test, the test will suck. That says nothing about the effectiveness of covid tests, as we have seen elsewhere. So it is entirely possible that you or anyone could point to some janky lab that had a 5% false positive rate. That means nothing. A good method being used poorly is nothing new to the world. Look at reputable organizations that have a lot on the line. They are doing really good testing. The link you provided is an example of why academia needs to stop eating shit when it comes to messaging. A single run of a single test can totally blow ass. But that is why any reputable company is running lots of repeats so that you are able to guarantee accuracy through data analysis and repetition. They are not saying that as little as 2% of all test results conveyed to people are wrong. They are ********NOT******** saying that. Another thing is that it is possible to detect dead forms of the virus, depending on how you are conducting the test. But you are also able to develop a known profile of what dead virus looks like and screen against that. I need to make it abundantly clear that the core of your argument, that covid results are fundamentally inflated and wrong, is in no way valid. This is similar to the publications talking about how antibodies last 3 months. In academia, it would be irresponsible to say anything other than "antibodies last 3 months" because they haven't been studied for longer. People think that means antibodies only last 3 months. Great. Or the study that showed antibodies weren't detected in someone who was infected 6 months ago. But that is normal. Human bodies keep memories of antibodies for use later, but in general, antibodies for infections naturally fade away in concentration since they are no longer needed. Publishing a study showing that antibodies are leaving at a normal rate is a normal thing to do in academia. But when Joe Shmoe reads that, he assumes immunity is only temporary. It is a total shit show. So to that point: academia has a ton of guilt with regards to covid disinformation. Their standard publishing practices needed to be completely abolished as soon as this became a public issue. Everything should be published in a way that common people can understand. Using the terminology typical of virology research is very bad right now.
I disagree completely with your last paragraph. There is no way to effectively abolish misinterpretation, particularly misinterpretation of people within a community by those outside it. Academic publishing has no immediate relation to facebook posts.
|
On August 30 2020 08:10 IgnE wrote:Show nested quote +On August 29 2020 00:35 Mohdoo wrote:On August 27 2020 12:15 cLutZ wrote:On August 27 2020 10:07 Mohdoo wrote:On August 27 2020 09:28 cLutZ wrote:On August 27 2020 09:09 Mohdoo wrote:
CDC saying not to test non-symptomatic is so amazingly dangerous. How utterly stupid. If you are talking about this guidance it seems reasonable. A major problem of the Covid tests is false positives and false. Thus, if you are testing large populations of people you generate a lot of noise. This part seems particularly important. If you have been in close contact (within 6 feet) of a person with a COVID-19 infection for at least 15 minutes but do not have symptoms:
You do not necessarily need a test unless you are a vulnerable individual or your health care provider or State or local public health officials recommend you take one.
A negative test does not mean you will not develop an infection from the close contact or contract an infection at a later time.
You should monitor yourself for symptoms. If you develop symptoms, you should evaluate yourself under the considerations set forth above.
You should strictly adhere to CDC mitigation protocols, especially if you are interacting with a vulnerable individual. You should adhere to CDC guidelines to protect vulnerable individuals with whom you live.
I think we are at a point where I need to ask: what is your technical background? To what extent have you conducted research on viruses? You said testing too many people with a noisy test makes more noise. That is perhaps among the most blatantly incorrect things you have said in this thread. You appear to not have a technical background, but you speak as if you know about this topic. We have seen numerous studies that show whether you are old or young, there are many people with extremely high viral concentrations in their nose and throat. That means those people can spread it, because as air passes by the virus, the virus hops along and is able to travel on particles/droplets. Let me reiterate: you do not need to show symptoms to be infectious. You do not need to cough to pass the virus. Simply speaking is sufficient. Saying a negative tests doesn't always mean someone isn't infected is a simple matter of existence. That is a really poor attempt to discredit testing. That is true for every single test ever made. Since one of the best things about testing is signaling people to quarantine, we absolutely must continue to test non-symptomatic people. It is dangerous not to. My technical background is in biomedical engineering where my specific work was with VADs, so I'm not an expert on this particular test. But in hospital work you generally learn that you need to watch out for both type 1 and type 2 errors, and both come into play with C-19 if you are testing asymptomatic people (not that you universally should not test asymptomatic people, its a judgement call as to when you should as I'll try to explain). In their explanation (which I bolded) they explicitly warn about type 2 errors, that being false negative tests for people who actually have the disease. The risk with this is that if someone is exposed, and you test them, then it comes back negative, they will act like they don't have the virus and not monitor for symptoms or self-quarantine (I am currently doing so because I was at the hospital, even though I had a negative test). They aren't explicitly discussing type 1 errors, false positives, but this is also a big problem for some tests, and C19 testing seems to have a significant level of this. This is why indiscriminate mammograms cause a lot of problems, as another example. Its fine to be indiscriminately testing so long as C19 has high prevalence in the population because you will mostly getting signal, but as it % of prevalence drops false positives will start dominating the statistics (as we saw with the NFL players example). According to this false positives are less than 5%, however this chart makes me suspect false positivity is more like 1%. What this means, is that if your state has a positive test rate of around that mark, most of your positives are actually false. Overall, seeing the 2nd chart, I agree with you that their judgement call is likely wrong at least for the states on the right hand side. False positivity is not a worry for those states, but if your %positive rate is <1% you probably need to re-evaulate procedures. You can use Bayes theorem to estimate what the "target" positive rate should be and test more/less based on this. Lets say you dont want more than 50% of positive results to be false positives. If you assume the false positive rate is 1% (and for simplicity say no false negatives), then your target is for 2% of results to be positive or more. So if %false positive is <1% this call is too early, but it is a call you eventually make. The less accurate you think tests are, the sooner you make it. It is important to keep in mind that VADs and infectious disease research are very different. Both are biologically relevant, but that's about it. Are you familiar with the specifics of how you work up samples for PCR, run PCR, and interpret PCR data? If so, you will know that the accuracy of the test is essentially up to the people running it. There are a few different kits that labs can use and each have their own pros and cons. And since different tests are looking for slightly different hings, there is a lot of nuance. But at the end of the day, you essentially are given an S-shaped curve, and based on when this curve takes shape, you can ascertain concentration, which you use to determine infection. Lots of things can go wrong in PCR, but luckily you can always inspect your data and run repeats. It is easy for a lab to run a 4x repeat on each sample to ensure as close to 100% accuracy as you can get. This is possible for both environmental and human samples. There ABSOLUTELY are labs out there that likely opportunistically tried to do covid testing and do a bad job at it. If you are bad at running a test, the test will suck. That says nothing about the effectiveness of covid tests, as we have seen elsewhere. So it is entirely possible that you or anyone could point to some janky lab that had a 5% false positive rate. That means nothing. A good method being used poorly is nothing new to the world. Look at reputable organizations that have a lot on the line. They are doing really good testing. The link you provided is an example of why academia needs to stop eating shit when it comes to messaging. A single run of a single test can totally blow ass. But that is why any reputable company is running lots of repeats so that you are able to guarantee accuracy through data analysis and repetition. They are not saying that as little as 2% of all test results conveyed to people are wrong. They are ********NOT******** saying that. Another thing is that it is possible to detect dead forms of the virus, depending on how you are conducting the test. But you are also able to develop a known profile of what dead virus looks like and screen against that. I need to make it abundantly clear that the core of your argument, that covid results are fundamentally inflated and wrong, is in no way valid. This is similar to the publications talking about how antibodies last 3 months. In academia, it would be irresponsible to say anything other than "antibodies last 3 months" because they haven't been studied for longer. People think that means antibodies only last 3 months. Great. Or the study that showed antibodies weren't detected in someone who was infected 6 months ago. But that is normal. Human bodies keep memories of antibodies for use later, but in general, antibodies for infections naturally fade away in concentration since they are no longer needed. Publishing a study showing that antibodies are leaving at a normal rate is a normal thing to do in academia. But when Joe Shmoe reads that, he assumes immunity is only temporary. It is a total shit show. So to that point: academia has a ton of guilt with regards to covid disinformation. Their standard publishing practices needed to be completely abolished as soon as this became a public issue. Everything should be published in a way that common people can understand. Using the terminology typical of virology research is very bad right now. I disagree completely with your last paragraph. There is no way to effectively abolish misinterpretation, particularly misinterpretation of people within a community by those outside it. Academic publishing has no immediate relation to facebook posts.
No matter how you slice it, the current methods have bad results. The technical terminology and general word choice of technical writing ends up commonly misinterpreted. I wouldn't say academia is guilty, but they have the ability to improve the situation. I don't think its as simple as "not my job", I would argue they have a duty to help prevent disinformation when they can. COVID has been a great example of how our current methods of spreading scientific information could use some work.
|
On August 30 2020 12:33 Mohdoo wrote:Show nested quote +On August 30 2020 08:10 IgnE wrote:On August 29 2020 00:35 Mohdoo wrote:On August 27 2020 12:15 cLutZ wrote:On August 27 2020 10:07 Mohdoo wrote:On August 27 2020 09:28 cLutZ wrote:On August 27 2020 09:09 Mohdoo wrote:
CDC saying not to test non-symptomatic is so amazingly dangerous. How utterly stupid. If you are talking about this guidance it seems reasonable. A major problem of the Covid tests is false positives and false. Thus, if you are testing large populations of people you generate a lot of noise. This part seems particularly important. If you have been in close contact (within 6 feet) of a person with a COVID-19 infection for at least 15 minutes but do not have symptoms:
You do not necessarily need a test unless you are a vulnerable individual or your health care provider or State or local public health officials recommend you take one.
A negative test does not mean you will not develop an infection from the close contact or contract an infection at a later time.
You should monitor yourself for symptoms. If you develop symptoms, you should evaluate yourself under the considerations set forth above.
You should strictly adhere to CDC mitigation protocols, especially if you are interacting with a vulnerable individual. You should adhere to CDC guidelines to protect vulnerable individuals with whom you live.
I think we are at a point where I need to ask: what is your technical background? To what extent have you conducted research on viruses? You said testing too many people with a noisy test makes more noise. That is perhaps among the most blatantly incorrect things you have said in this thread. You appear to not have a technical background, but you speak as if you know about this topic. We have seen numerous studies that show whether you are old or young, there are many people with extremely high viral concentrations in their nose and throat. That means those people can spread it, because as air passes by the virus, the virus hops along and is able to travel on particles/droplets. Let me reiterate: you do not need to show symptoms to be infectious. You do not need to cough to pass the virus. Simply speaking is sufficient. Saying a negative tests doesn't always mean someone isn't infected is a simple matter of existence. That is a really poor attempt to discredit testing. That is true for every single test ever made. Since one of the best things about testing is signaling people to quarantine, we absolutely must continue to test non-symptomatic people. It is dangerous not to. My technical background is in biomedical engineering where my specific work was with VADs, so I'm not an expert on this particular test. But in hospital work you generally learn that you need to watch out for both type 1 and type 2 errors, and both come into play with C-19 if you are testing asymptomatic people (not that you universally should not test asymptomatic people, its a judgement call as to when you should as I'll try to explain). In their explanation (which I bolded) they explicitly warn about type 2 errors, that being false negative tests for people who actually have the disease. The risk with this is that if someone is exposed, and you test them, then it comes back negative, they will act like they don't have the virus and not monitor for symptoms or self-quarantine (I am currently doing so because I was at the hospital, even though I had a negative test). They aren't explicitly discussing type 1 errors, false positives, but this is also a big problem for some tests, and C19 testing seems to have a significant level of this. This is why indiscriminate mammograms cause a lot of problems, as another example. Its fine to be indiscriminately testing so long as C19 has high prevalence in the population because you will mostly getting signal, but as it % of prevalence drops false positives will start dominating the statistics (as we saw with the NFL players example). According to this false positives are less than 5%, however this chart makes me suspect false positivity is more like 1%. What this means, is that if your state has a positive test rate of around that mark, most of your positives are actually false. Overall, seeing the 2nd chart, I agree with you that their judgement call is likely wrong at least for the states on the right hand side. False positivity is not a worry for those states, but if your %positive rate is <1% you probably need to re-evaulate procedures. You can use Bayes theorem to estimate what the "target" positive rate should be and test more/less based on this. Lets say you dont want more than 50% of positive results to be false positives. If you assume the false positive rate is 1% (and for simplicity say no false negatives), then your target is for 2% of results to be positive or more. So if %false positive is <1% this call is too early, but it is a call you eventually make. The less accurate you think tests are, the sooner you make it. It is important to keep in mind that VADs and infectious disease research are very different. Both are biologically relevant, but that's about it. Are you familiar with the specifics of how you work up samples for PCR, run PCR, and interpret PCR data? If so, you will know that the accuracy of the test is essentially up to the people running it. There are a few different kits that labs can use and each have their own pros and cons. And since different tests are looking for slightly different hings, there is a lot of nuance. But at the end of the day, you essentially are given an S-shaped curve, and based on when this curve takes shape, you can ascertain concentration, which you use to determine infection. Lots of things can go wrong in PCR, but luckily you can always inspect your data and run repeats. It is easy for a lab to run a 4x repeat on each sample to ensure as close to 100% accuracy as you can get. This is possible for both environmental and human samples. There ABSOLUTELY are labs out there that likely opportunistically tried to do covid testing and do a bad job at it. If you are bad at running a test, the test will suck. That says nothing about the effectiveness of covid tests, as we have seen elsewhere. So it is entirely possible that you or anyone could point to some janky lab that had a 5% false positive rate. That means nothing. A good method being used poorly is nothing new to the world. Look at reputable organizations that have a lot on the line. They are doing really good testing. The link you provided is an example of why academia needs to stop eating shit when it comes to messaging. A single run of a single test can totally blow ass. But that is why any reputable company is running lots of repeats so that you are able to guarantee accuracy through data analysis and repetition. They are not saying that as little as 2% of all test results conveyed to people are wrong. They are ********NOT******** saying that. Another thing is that it is possible to detect dead forms of the virus, depending on how you are conducting the test. But you are also able to develop a known profile of what dead virus looks like and screen against that. I need to make it abundantly clear that the core of your argument, that covid results are fundamentally inflated and wrong, is in no way valid. This is similar to the publications talking about how antibodies last 3 months. In academia, it would be irresponsible to say anything other than "antibodies last 3 months" because they haven't been studied for longer. People think that means antibodies only last 3 months. Great. Or the study that showed antibodies weren't detected in someone who was infected 6 months ago. But that is normal. Human bodies keep memories of antibodies for use later, but in general, antibodies for infections naturally fade away in concentration since they are no longer needed. Publishing a study showing that antibodies are leaving at a normal rate is a normal thing to do in academia. But when Joe Shmoe reads that, he assumes immunity is only temporary. It is a total shit show. So to that point: academia has a ton of guilt with regards to covid disinformation. Their standard publishing practices needed to be completely abolished as soon as this became a public issue. Everything should be published in a way that common people can understand. Using the terminology typical of virology research is very bad right now. I disagree completely with your last paragraph. There is no way to effectively abolish misinterpretation, particularly misinterpretation of people within a community by those outside it. Academic publishing has no immediate relation to facebook posts. No matter how you slice it, the current methods have bad results. The technical terminology and general word choice of technical writing ends up commonly misinterpreted. I wouldn't say academia is guilty, but they have the ability to improve the situation. I don't think its as simple as "not my job", I would argue they have a duty to help prevent disinformation when they can. COVID has been a great example of how our current methods of spreading scientific information could use some work.
Having recently had articles revised twice only to get more precise definitions simplified language sounds good at first. However the amount of effort spent to really get crystal clear meaning in science (at least if you want to get published in a decent journal) is done because it's very important that you say *exactly* what you mean and nothing else. And usually that can't be done without using the terminology of the field.
I don't think it's possible to simply much without losing quality. And even if you did it will still not be understandable to the vast majority of the people.
It would be better for people who get their news from facebook to understand that even with their masters in X field (and especially if they don't even have one) that still doesn't mean they are competent to understand the science in field Y.
Also never underestimate the ability of journalists to completely misinterpret even things that are way simpler than science. I've never seen a news article tangentially related to my field of work be competently correct and I have examples where major news programs completely misunderstands things (like saying a thing goes backwards when it goes forwards basically) and/or lie about things, and refuse to correct them in order to make a story better. And we are talking things that are way easier than science and borderline public knowledge.
So simplifying scientific articles will likely make science worse and have no impact on desinformation.
What does have good impact in this situation is to have daily national/regional/local press conferences along with government information sites that provide reliable (scientific) sources that explain the situation and measures in an easy way. Something almost all countries (except the US naturally) seems to think is important
|
On August 30 2020 12:33 Mohdoo wrote:Show nested quote +On August 30 2020 08:10 IgnE wrote:On August 29 2020 00:35 Mohdoo wrote:On August 27 2020 12:15 cLutZ wrote:On August 27 2020 10:07 Mohdoo wrote:On August 27 2020 09:28 cLutZ wrote:On August 27 2020 09:09 Mohdoo wrote:
CDC saying not to test non-symptomatic is so amazingly dangerous. How utterly stupid. If you are talking about this guidance it seems reasonable. A major problem of the Covid tests is false positives and false. Thus, if you are testing large populations of people you generate a lot of noise. This part seems particularly important. If you have been in close contact (within 6 feet) of a person with a COVID-19 infection for at least 15 minutes but do not have symptoms:
You do not necessarily need a test unless you are a vulnerable individual or your health care provider or State or local public health officials recommend you take one.
A negative test does not mean you will not develop an infection from the close contact or contract an infection at a later time.
You should monitor yourself for symptoms. If you develop symptoms, you should evaluate yourself under the considerations set forth above.
You should strictly adhere to CDC mitigation protocols, especially if you are interacting with a vulnerable individual. You should adhere to CDC guidelines to protect vulnerable individuals with whom you live.
I think we are at a point where I need to ask: what is your technical background? To what extent have you conducted research on viruses? You said testing too many people with a noisy test makes more noise. That is perhaps among the most blatantly incorrect things you have said in this thread. You appear to not have a technical background, but you speak as if you know about this topic. We have seen numerous studies that show whether you are old or young, there are many people with extremely high viral concentrations in their nose and throat. That means those people can spread it, because as air passes by the virus, the virus hops along and is able to travel on particles/droplets. Let me reiterate: you do not need to show symptoms to be infectious. You do not need to cough to pass the virus. Simply speaking is sufficient. Saying a negative tests doesn't always mean someone isn't infected is a simple matter of existence. That is a really poor attempt to discredit testing. That is true for every single test ever made. Since one of the best things about testing is signaling people to quarantine, we absolutely must continue to test non-symptomatic people. It is dangerous not to. My technical background is in biomedical engineering where my specific work was with VADs, so I'm not an expert on this particular test. But in hospital work you generally learn that you need to watch out for both type 1 and type 2 errors, and both come into play with C-19 if you are testing asymptomatic people (not that you universally should not test asymptomatic people, its a judgement call as to when you should as I'll try to explain). In their explanation (which I bolded) they explicitly warn about type 2 errors, that being false negative tests for people who actually have the disease. The risk with this is that if someone is exposed, and you test them, then it comes back negative, they will act like they don't have the virus and not monitor for symptoms or self-quarantine (I am currently doing so because I was at the hospital, even though I had a negative test). They aren't explicitly discussing type 1 errors, false positives, but this is also a big problem for some tests, and C19 testing seems to have a significant level of this. This is why indiscriminate mammograms cause a lot of problems, as another example. Its fine to be indiscriminately testing so long as C19 has high prevalence in the population because you will mostly getting signal, but as it % of prevalence drops false positives will start dominating the statistics (as we saw with the NFL players example). According to this false positives are less than 5%, however this chart makes me suspect false positivity is more like 1%. What this means, is that if your state has a positive test rate of around that mark, most of your positives are actually false. Overall, seeing the 2nd chart, I agree with you that their judgement call is likely wrong at least for the states on the right hand side. False positivity is not a worry for those states, but if your %positive rate is <1% you probably need to re-evaulate procedures. You can use Bayes theorem to estimate what the "target" positive rate should be and test more/less based on this. Lets say you dont want more than 50% of positive results to be false positives. If you assume the false positive rate is 1% (and for simplicity say no false negatives), then your target is for 2% of results to be positive or more. So if %false positive is <1% this call is too early, but it is a call you eventually make. The less accurate you think tests are, the sooner you make it. It is important to keep in mind that VADs and infectious disease research are very different. Both are biologically relevant, but that's about it. Are you familiar with the specifics of how you work up samples for PCR, run PCR, and interpret PCR data? If so, you will know that the accuracy of the test is essentially up to the people running it. There are a few different kits that labs can use and each have their own pros and cons. And since different tests are looking for slightly different hings, there is a lot of nuance. But at the end of the day, you essentially are given an S-shaped curve, and based on when this curve takes shape, you can ascertain concentration, which you use to determine infection. Lots of things can go wrong in PCR, but luckily you can always inspect your data and run repeats. It is easy for a lab to run a 4x repeat on each sample to ensure as close to 100% accuracy as you can get. This is possible for both environmental and human samples. There ABSOLUTELY are labs out there that likely opportunistically tried to do covid testing and do a bad job at it. If you are bad at running a test, the test will suck. That says nothing about the effectiveness of covid tests, as we have seen elsewhere. So it is entirely possible that you or anyone could point to some janky lab that had a 5% false positive rate. That means nothing. A good method being used poorly is nothing new to the world. Look at reputable organizations that have a lot on the line. They are doing really good testing. The link you provided is an example of why academia needs to stop eating shit when it comes to messaging. A single run of a single test can totally blow ass. But that is why any reputable company is running lots of repeats so that you are able to guarantee accuracy through data analysis and repetition. They are not saying that as little as 2% of all test results conveyed to people are wrong. They are ********NOT******** saying that. Another thing is that it is possible to detect dead forms of the virus, depending on how you are conducting the test. But you are also able to develop a known profile of what dead virus looks like and screen against that. I need to make it abundantly clear that the core of your argument, that covid results are fundamentally inflated and wrong, is in no way valid. This is similar to the publications talking about how antibodies last 3 months. In academia, it would be irresponsible to say anything other than "antibodies last 3 months" because they haven't been studied for longer. People think that means antibodies only last 3 months. Great. Or the study that showed antibodies weren't detected in someone who was infected 6 months ago. But that is normal. Human bodies keep memories of antibodies for use later, but in general, antibodies for infections naturally fade away in concentration since they are no longer needed. Publishing a study showing that antibodies are leaving at a normal rate is a normal thing to do in academia. But when Joe Shmoe reads that, he assumes immunity is only temporary. It is a total shit show. So to that point: academia has a ton of guilt with regards to covid disinformation. Their standard publishing practices needed to be completely abolished as soon as this became a public issue. Everything should be published in a way that common people can understand. Using the terminology typical of virology research is very bad right now. I disagree completely with your last paragraph. There is no way to effectively abolish misinterpretation, particularly misinterpretation of people within a community by those outside it. Academic publishing has no immediate relation to facebook posts. No matter how you slice it, the current methods have bad results. The technical terminology and general word choice of technical writing ends up commonly misinterpreted. I wouldn't say academia is guilty, but they have the ability to improve the situation. I don't think its as simple as "not my job", I would argue they have a duty to help prevent disinformation when they can. COVID has been a great example of how our current methods of spreading scientific information could use some work.
Why does academia have to change its methods to cover media entities that sensationalize rather than interpret? It's the job of scientific journalists, not scientists, to interpret technical writing in a way that laypeople can understand. Requiring that scientists do this only serves to constrain the amount of information these publications can convey.
Publications like Nature and JAMA don't exist to be readable to laypeople, they exist as a means for researchers to share their work with other researchers and professionals. It makes no sense to compromise that purpose because others are misusing them.
|
On August 29 2020 00:07 TheTenthDoc wrote:
As an epidemiologist doing healthcare research in the United States there's two major reasons you're not going to see anything beyond small-scale anecdotal evidence about long-term effects of COVID-19 for quite some time.
1) The obvious one: even an omniscient person couldn't know them yet. Even if you believe the virus was circulating in December in the US (which, personally, I don't), that's still a hard stop of 9 months at most of surviving after the virus. Anything about complications past that point must by necessity drawn from relatives in the viral family.
2) The data resources being leveraged for COVID-19 research right now are generally coming straight from hospital systems with a sprinkling of general medical records, with a huge focus on hospitalizations themselves. This data has a lot of advantages-for one, it tends to be extremely timely almost to a fault, and it cannot be beat in terms of understanding exposures-but it's not great for tracking long-term consequences of health conditions when dealing with people that rarely engage directly with the healthcare system. The exception is conditions that more or less force heavy direct engagement (e.g. HIV/AIDs).
Usually, long-term complications are better tracked in something like insurance claims data or prospective cohorts. The former typically has a pretty significant lag time and its own issues brewing on the horizon, and the latter isn't exactly something you can set up on a dime (and I'm not sure how open people are to the necessary interviews right now).
If someone claims to know at a population level what COVID-19 means for your body after three + months (whether they say it breaks your lungs forever or everyone is totally fine unless they were unhealthy at the time of diagnosis), they are either blowing smoke, making big jumps from a small data set, or have access to some really awesome data that is almost certainly not from the United States.
Was more or less going to say this.
The other issue with tracking long-term consequences of COVID is that a lot of the things we anecdotally believe to be affected are not necessarily related to respiratory symptoms, but are more nonspecific systemic issues. If someone who previously had healthy kidneys gets his lab work done for a primary care visit 6 months after a COVID infection with a creatinine of 1.5, is it COVID-related? Possibly. But it'll be years before we have enough data to actually say anything definitive.
It would be (relatively speaking) easy if it was just pulmonary complications. But anecdotal post-COVID complications are arising with renal dysfunction, blood clotting, joint pain, etc. and it will be years before we can sort out what is and isn't directly caused by COVID.
|
On August 30 2020 15:57 TheYango wrote:Show nested quote +On August 30 2020 12:33 Mohdoo wrote:On August 30 2020 08:10 IgnE wrote:On August 29 2020 00:35 Mohdoo wrote:On August 27 2020 12:15 cLutZ wrote:On August 27 2020 10:07 Mohdoo wrote:On August 27 2020 09:28 cLutZ wrote:On August 27 2020 09:09 Mohdoo wrote:
CDC saying not to test non-symptomatic is so amazingly dangerous. How utterly stupid. If you are talking about this guidance it seems reasonable. A major problem of the Covid tests is false positives and false. Thus, if you are testing large populations of people you generate a lot of noise. This part seems particularly important. If you have been in close contact (within 6 feet) of a person with a COVID-19 infection for at least 15 minutes but do not have symptoms:
You do not necessarily need a test unless you are a vulnerable individual or your health care provider or State or local public health officials recommend you take one.
A negative test does not mean you will not develop an infection from the close contact or contract an infection at a later time.
You should monitor yourself for symptoms. If you develop symptoms, you should evaluate yourself under the considerations set forth above.
You should strictly adhere to CDC mitigation protocols, especially if you are interacting with a vulnerable individual. You should adhere to CDC guidelines to protect vulnerable individuals with whom you live.
I think we are at a point where I need to ask: what is your technical background? To what extent have you conducted research on viruses? You said testing too many people with a noisy test makes more noise. That is perhaps among the most blatantly incorrect things you have said in this thread. You appear to not have a technical background, but you speak as if you know about this topic. We have seen numerous studies that show whether you are old or young, there are many people with extremely high viral concentrations in their nose and throat. That means those people can spread it, because as air passes by the virus, the virus hops along and is able to travel on particles/droplets. Let me reiterate: you do not need to show symptoms to be infectious. You do not need to cough to pass the virus. Simply speaking is sufficient. Saying a negative tests doesn't always mean someone isn't infected is a simple matter of existence. That is a really poor attempt to discredit testing. That is true for every single test ever made. Since one of the best things about testing is signaling people to quarantine, we absolutely must continue to test non-symptomatic people. It is dangerous not to. My technical background is in biomedical engineering where my specific work was with VADs, so I'm not an expert on this particular test. But in hospital work you generally learn that you need to watch out for both type 1 and type 2 errors, and both come into play with C-19 if you are testing asymptomatic people (not that you universally should not test asymptomatic people, its a judgement call as to when you should as I'll try to explain). In their explanation (which I bolded) they explicitly warn about type 2 errors, that being false negative tests for people who actually have the disease. The risk with this is that if someone is exposed, and you test them, then it comes back negative, they will act like they don't have the virus and not monitor for symptoms or self-quarantine (I am currently doing so because I was at the hospital, even though I had a negative test). They aren't explicitly discussing type 1 errors, false positives, but this is also a big problem for some tests, and C19 testing seems to have a significant level of this. This is why indiscriminate mammograms cause a lot of problems, as another example. Its fine to be indiscriminately testing so long as C19 has high prevalence in the population because you will mostly getting signal, but as it % of prevalence drops false positives will start dominating the statistics (as we saw with the NFL players example). According to this false positives are less than 5%, however this chart makes me suspect false positivity is more like 1%. What this means, is that if your state has a positive test rate of around that mark, most of your positives are actually false. Overall, seeing the 2nd chart, I agree with you that their judgement call is likely wrong at least for the states on the right hand side. False positivity is not a worry for those states, but if your %positive rate is <1% you probably need to re-evaulate procedures. You can use Bayes theorem to estimate what the "target" positive rate should be and test more/less based on this. Lets say you dont want more than 50% of positive results to be false positives. If you assume the false positive rate is 1% (and for simplicity say no false negatives), then your target is for 2% of results to be positive or more. So if %false positive is <1% this call is too early, but it is a call you eventually make. The less accurate you think tests are, the sooner you make it. It is important to keep in mind that VADs and infectious disease research are very different. Both are biologically relevant, but that's about it. Are you familiar with the specifics of how you work up samples for PCR, run PCR, and interpret PCR data? If so, you will know that the accuracy of the test is essentially up to the people running it. There are a few different kits that labs can use and each have their own pros and cons. And since different tests are looking for slightly different hings, there is a lot of nuance. But at the end of the day, you essentially are given an S-shaped curve, and based on when this curve takes shape, you can ascertain concentration, which you use to determine infection. Lots of things can go wrong in PCR, but luckily you can always inspect your data and run repeats. It is easy for a lab to run a 4x repeat on each sample to ensure as close to 100% accuracy as you can get. This is possible for both environmental and human samples. There ABSOLUTELY are labs out there that likely opportunistically tried to do covid testing and do a bad job at it. If you are bad at running a test, the test will suck. That says nothing about the effectiveness of covid tests, as we have seen elsewhere. So it is entirely possible that you or anyone could point to some janky lab that had a 5% false positive rate. That means nothing. A good method being used poorly is nothing new to the world. Look at reputable organizations that have a lot on the line. They are doing really good testing. The link you provided is an example of why academia needs to stop eating shit when it comes to messaging. A single run of a single test can totally blow ass. But that is why any reputable company is running lots of repeats so that you are able to guarantee accuracy through data analysis and repetition. They are not saying that as little as 2% of all test results conveyed to people are wrong. They are ********NOT******** saying that. Another thing is that it is possible to detect dead forms of the virus, depending on how you are conducting the test. But you are also able to develop a known profile of what dead virus looks like and screen against that. I need to make it abundantly clear that the core of your argument, that covid results are fundamentally inflated and wrong, is in no way valid. This is similar to the publications talking about how antibodies last 3 months. In academia, it would be irresponsible to say anything other than "antibodies last 3 months" because they haven't been studied for longer. People think that means antibodies only last 3 months. Great. Or the study that showed antibodies weren't detected in someone who was infected 6 months ago. But that is normal. Human bodies keep memories of antibodies for use later, but in general, antibodies for infections naturally fade away in concentration since they are no longer needed. Publishing a study showing that antibodies are leaving at a normal rate is a normal thing to do in academia. But when Joe Shmoe reads that, he assumes immunity is only temporary. It is a total shit show. So to that point: academia has a ton of guilt with regards to covid disinformation. Their standard publishing practices needed to be completely abolished as soon as this became a public issue. Everything should be published in a way that common people can understand. Using the terminology typical of virology research is very bad right now. I disagree completely with your last paragraph. There is no way to effectively abolish misinterpretation, particularly misinterpretation of people within a community by those outside it. Academic publishing has no immediate relation to facebook posts. No matter how you slice it, the current methods have bad results. The technical terminology and general word choice of technical writing ends up commonly misinterpreted. I wouldn't say academia is guilty, but they have the ability to improve the situation. I don't think its as simple as "not my job", I would argue they have a duty to help prevent disinformation when they can. COVID has been a great example of how our current methods of spreading scientific information could use some work. Why does academia have to change its methods to cover media entities that sensationalize rather than interpret? It's the job of scientific journalists, not scientists, to interpret technical writing in a way that laypeople can understand. Requiring that scientists do this only serves to constrain the amount of information these publications can convey. Publications like Nature and JAMA don't exist to be readable to laypeople, they exist as a means for researchers to share their work with other researchers and professionals. It makes no sense to compromise that purpose because others are misusing them.
I agree with your second paragraph but not your first. It is a scientist's job to publish precise and specific technical articles in scientific journals. It is not there that scientific language should be compromised for the sake of simplicity.
However, it is *also* a scientist's responsibility to ensure their science is correctly simplified for press releases and other outreach to the general public. Especially if they are researching something as relevant to society as control of the current pandemic (it's less worrying if you are misinterpreted if your research topic is on the light-slowing properties of Bose-Einstein condensates, which while interesting have no current immediate application in society). At least here in Europe, a part of any research proposal is outreach, and that includes not just how you propose to divulge your research to the scientific community, but also the general public. And that is becoming increasingly important, even though scientist's curricula are still exclusively judged on the publishing in the scientific community part, so it is unsurprising that they (we) aren't necessarily very good at communicating findings to the general public.
|
On August 30 2020 16:30 TheYango wrote:Show nested quote +On August 29 2020 00:07 TheTenthDoc wrote:
As an epidemiologist doing healthcare research in the United States there's two major reasons you're not going to see anything beyond small-scale anecdotal evidence about long-term effects of COVID-19 for quite some time.
1) The obvious one: even an omniscient person couldn't know them yet. Even if you believe the virus was circulating in December in the US (which, personally, I don't), that's still a hard stop of 9 months at most of surviving after the virus. Anything about complications past that point must by necessity drawn from relatives in the viral family.
2) The data resources being leveraged for COVID-19 research right now are generally coming straight from hospital systems with a sprinkling of general medical records, with a huge focus on hospitalizations themselves. This data has a lot of advantages-for one, it tends to be extremely timely almost to a fault, and it cannot be beat in terms of understanding exposures-but it's not great for tracking long-term consequences of health conditions when dealing with people that rarely engage directly with the healthcare system. The exception is conditions that more or less force heavy direct engagement (e.g. HIV/AIDs).
Usually, long-term complications are better tracked in something like insurance claims data or prospective cohorts. The former typically has a pretty significant lag time and its own issues brewing on the horizon, and the latter isn't exactly something you can set up on a dime (and I'm not sure how open people are to the necessary interviews right now).
If someone claims to know at a population level what COVID-19 means for your body after three + months (whether they say it breaks your lungs forever or everyone is totally fine unless they were unhealthy at the time of diagnosis), they are either blowing smoke, making big jumps from a small data set, or have access to some really awesome data that is almost certainly not from the United States. Was more or less going to say this. The other issue with tracking long-term consequences of COVID is that a lot of the things we anecdotally believe to be affected are not necessarily related to respiratory symptoms, but are more nonspecific systemic issues. If someone who previously had healthy kidneys gets his lab work done for a primary care visit 6 months after a COVID infection with a creatinine of 1.5, is it COVID-related? Possibly. But it'll be years before we have enough data to actually say anything definitive. It would be (relatively speaking) easy if it was just pulmonary complications. But anecdotal post-COVID complications are arising with renal dysfunction, blood clotting, joint pain, etc. and it will be years before we can sort out what is and isn't directly caused by COVID.
Off course it will take years before there is hard scientific evidence for long term effects of the virus. And 100% certainty that some effects are caused by the virus will probably never be there. But that doesnt mean policy makers should wait years before taking those potential effects into consideration when it comes to making policy. It would simply be to late to do anything if they would wait for hard scientific evidence,millions of people will already have been effected and the long term effects are irreversable for those people.
As soon as there is at least some indication of long term effects it should be taken into consideration when it comes to making policy even if hard proof is lacking,thats how you stay ahead of the curve. When you wait for the hard evidence you will always be behind.
Sometimes misjudgement will occur wich is inevitable,but i think that is less damaging overall then always waiting till there is 100% scientific proof.
Policy makers shouldnt jump to conclusions obviously. Not act on ever indication no matter how small.
Policy has a huge effect economically and socially so everything has to be weighted very carefully. But the other extreme,waiting years for 100% scientific evidence for long term effects before taking them into consideration when it comes to making policy isnt an option either imo.
You have to go by the data that is available even when its incomplete and not 100% proven. Its the best indication there is right now and ignoring everything that isnt 100% scientifically proven when it comes to making policy is not an option. (and it isnt happening either). You always have uncertaintys in many different areas when you make policy.
|
On August 30 2020 15:57 TheYango wrote:Show nested quote +On August 30 2020 12:33 Mohdoo wrote:On August 30 2020 08:10 IgnE wrote:On August 29 2020 00:35 Mohdoo wrote:On August 27 2020 12:15 cLutZ wrote:On August 27 2020 10:07 Mohdoo wrote:On August 27 2020 09:28 cLutZ wrote:On August 27 2020 09:09 Mohdoo wrote:
CDC saying not to test non-symptomatic is so amazingly dangerous. How utterly stupid. If you are talking about this guidance it seems reasonable. A major problem of the Covid tests is false positives and false. Thus, if you are testing large populations of people you generate a lot of noise. This part seems particularly important. If you have been in close contact (within 6 feet) of a person with a COVID-19 infection for at least 15 minutes but do not have symptoms:
You do not necessarily need a test unless you are a vulnerable individual or your health care provider or State or local public health officials recommend you take one.
A negative test does not mean you will not develop an infection from the close contact or contract an infection at a later time.
You should monitor yourself for symptoms. If you develop symptoms, you should evaluate yourself under the considerations set forth above.
You should strictly adhere to CDC mitigation protocols, especially if you are interacting with a vulnerable individual. You should adhere to CDC guidelines to protect vulnerable individuals with whom you live.
I think we are at a point where I need to ask: what is your technical background? To what extent have you conducted research on viruses? You said testing too many people with a noisy test makes more noise. That is perhaps among the most blatantly incorrect things you have said in this thread. You appear to not have a technical background, but you speak as if you know about this topic. We have seen numerous studies that show whether you are old or young, there are many people with extremely high viral concentrations in their nose and throat. That means those people can spread it, because as air passes by the virus, the virus hops along and is able to travel on particles/droplets. Let me reiterate: you do not need to show symptoms to be infectious. You do not need to cough to pass the virus. Simply speaking is sufficient. Saying a negative tests doesn't always mean someone isn't infected is a simple matter of existence. That is a really poor attempt to discredit testing. That is true for every single test ever made. Since one of the best things about testing is signaling people to quarantine, we absolutely must continue to test non-symptomatic people. It is dangerous not to. My technical background is in biomedical engineering where my specific work was with VADs, so I'm not an expert on this particular test. But in hospital work you generally learn that you need to watch out for both type 1 and type 2 errors, and both come into play with C-19 if you are testing asymptomatic people (not that you universally should not test asymptomatic people, its a judgement call as to when you should as I'll try to explain). In their explanation (which I bolded) they explicitly warn about type 2 errors, that being false negative tests for people who actually have the disease. The risk with this is that if someone is exposed, and you test them, then it comes back negative, they will act like they don't have the virus and not monitor for symptoms or self-quarantine (I am currently doing so because I was at the hospital, even though I had a negative test). They aren't explicitly discussing type 1 errors, false positives, but this is also a big problem for some tests, and C19 testing seems to have a significant level of this. This is why indiscriminate mammograms cause a lot of problems, as another example. Its fine to be indiscriminately testing so long as C19 has high prevalence in the population because you will mostly getting signal, but as it % of prevalence drops false positives will start dominating the statistics (as we saw with the NFL players example). According to this false positives are less than 5%, however this chart makes me suspect false positivity is more like 1%. What this means, is that if your state has a positive test rate of around that mark, most of your positives are actually false. Overall, seeing the 2nd chart, I agree with you that their judgement call is likely wrong at least for the states on the right hand side. False positivity is not a worry for those states, but if your %positive rate is <1% you probably need to re-evaulate procedures. You can use Bayes theorem to estimate what the "target" positive rate should be and test more/less based on this. Lets say you dont want more than 50% of positive results to be false positives. If you assume the false positive rate is 1% (and for simplicity say no false negatives), then your target is for 2% of results to be positive or more. So if %false positive is <1% this call is too early, but it is a call you eventually make. The less accurate you think tests are, the sooner you make it. It is important to keep in mind that VADs and infectious disease research are very different. Both are biologically relevant, but that's about it. Are you familiar with the specifics of how you work up samples for PCR, run PCR, and interpret PCR data? If so, you will know that the accuracy of the test is essentially up to the people running it. There are a few different kits that labs can use and each have their own pros and cons. And since different tests are looking for slightly different hings, there is a lot of nuance. But at the end of the day, you essentially are given an S-shaped curve, and based on when this curve takes shape, you can ascertain concentration, which you use to determine infection. Lots of things can go wrong in PCR, but luckily you can always inspect your data and run repeats. It is easy for a lab to run a 4x repeat on each sample to ensure as close to 100% accuracy as you can get. This is possible for both environmental and human samples. There ABSOLUTELY are labs out there that likely opportunistically tried to do covid testing and do a bad job at it. If you are bad at running a test, the test will suck. That says nothing about the effectiveness of covid tests, as we have seen elsewhere. So it is entirely possible that you or anyone could point to some janky lab that had a 5% false positive rate. That means nothing. A good method being used poorly is nothing new to the world. Look at reputable organizations that have a lot on the line. They are doing really good testing. The link you provided is an example of why academia needs to stop eating shit when it comes to messaging. A single run of a single test can totally blow ass. But that is why any reputable company is running lots of repeats so that you are able to guarantee accuracy through data analysis and repetition. They are not saying that as little as 2% of all test results conveyed to people are wrong. They are ********NOT******** saying that. Another thing is that it is possible to detect dead forms of the virus, depending on how you are conducting the test. But you are also able to develop a known profile of what dead virus looks like and screen against that. I need to make it abundantly clear that the core of your argument, that covid results are fundamentally inflated and wrong, is in no way valid. This is similar to the publications talking about how antibodies last 3 months. In academia, it would be irresponsible to say anything other than "antibodies last 3 months" because they haven't been studied for longer. People think that means antibodies only last 3 months. Great. Or the study that showed antibodies weren't detected in someone who was infected 6 months ago. But that is normal. Human bodies keep memories of antibodies for use later, but in general, antibodies for infections naturally fade away in concentration since they are no longer needed. Publishing a study showing that antibodies are leaving at a normal rate is a normal thing to do in academia. But when Joe Shmoe reads that, he assumes immunity is only temporary. It is a total shit show. So to that point: academia has a ton of guilt with regards to covid disinformation. Their standard publishing practices needed to be completely abolished as soon as this became a public issue. Everything should be published in a way that common people can understand. Using the terminology typical of virology research is very bad right now. I disagree completely with your last paragraph. There is no way to effectively abolish misinterpretation, particularly misinterpretation of people within a community by those outside it. Academic publishing has no immediate relation to facebook posts. No matter how you slice it, the current methods have bad results. The technical terminology and general word choice of technical writing ends up commonly misinterpreted. I wouldn't say academia is guilty, but they have the ability to improve the situation. I don't think its as simple as "not my job", I would argue they have a duty to help prevent disinformation when they can. COVID has been a great example of how our current methods of spreading scientific information could use some work. Why does academia have to change its methods to cover media entities that sensationalize rather than interpret? It's the job of scientific journalists, not scientists, to interpret technical writing in a way that laypeople can understand. Requiring that scientists do this only serves to constrain the amount of information these publications can convey. Publications like Nature and JAMA don't exist to be readable to laypeople, they exist as a means for researchers to share their work with other researchers and professionals. It makes no sense to compromise that purpose because others are misusing them.
I agree with everything you are saying here. I am a published scientist and I know the purpose of journals is not for casual reading for joe shmoe. However, I think we still need to say "there is a huge problem with legitimate research being weaponized as disinformation". It is a problem that is bad enough that it needs to be fixed, perhaps even in ways that don't make sense at first glance. Too many people are dying from disinformation. If the people actually writing the research changed what they wrote, the problem could go away. Believe me when I say I know this problem isn't caused at its core by the researchers. But if they can defend their work against being twisted, they ought to.
|
Perhaps a solution would be to create two abstracts: one aimed at pseudo journalist - explaining reasearch in simple words and the other working the way it works now. Not perfect - takes space in publication and time for writing/reviwing, would offer some protection.
|
On August 30 2020 23:42 Silvanel wrote:
Perhaps a solution would be to create two abstracts: one aimed at pseudo journalist - explaining reasearch in simple words and the other working the way it works now. Not perfect - takes space in publication and time for writing/reviwing, would offer some protection.
Let me give an example:
"Study shows antibodies last 3 months"
Rather than that, during a pandemic, maybe us academic folks can get a few of the sticks out of our ass and make an exception during a pandemic when the public is watching, and instead title a paper:
"Great news, antibodies continue to be present in study monitoring long term immunity"
^that title breaks a lot of rules, but are they really rules that matter? Maybe it is ok, during a pandemic, to publish what we mean rather than our general methods.
Another example, this time simply messaging rather than publishing:
"We do not have data indicating masks are effective against covid"
Instead, they could say "Every single piece of information we have over the last 200 years indicate a mask of any sort will help reduce infection in some non-zero way. Go ahead and everyone wear masks for now while we figure out the specifics."
I would like to reiterate: Not a single chemist or microbiologist wondered if masks would be helpful. I was telling my friends to wear masks in ***February*** It was obvious it would help, just a matter of how much. There was never a legitimate argument for the idea that masks would not be effective. But stupid dogshit academic guidelines punched ourselves in the face.
|
Isn't that what the abstract and conclusions is for? Casual reading. The journalist doesn't care either way, they will always find a way to pick and choose and present information that fits their interest. A journalist who wants to present information in an impartial and informative way will find and interview a panel of relevant academics or equivalent to provide the information rather than pick and choose whatever sounds most sensationalist.
And guess what, just from TL.net, which we think have higher standards, we can see that most people don't even bother reading the very articles and pdfs they proclaim information from. They just type some random words into google, and post links of the articles that sounds like it might support their preconceived notion and call that an argument. And it works, because the avergae person can't be bothered to go through 100 pages of extremely dry reading when they aren't mentally equipped with the tools to understand even simple things like how to read a graph? So it is with TL keyboard warriors, so it is with how most journalists present science. The problem is journalistic integrity, or rather the lack of it and the lack of the mentality from consumers to choose media with journalistic integrity.
|
On August 31 2020 00:21 Dangermousecatdog wrote:
Isn't that what the abstract and conclusions is for? Casual reading. The journalist doesn't care either way, they will always find a way to pick and choose and present information that fits their interest. A journalist who wants to present information in an impartial and informative way will find and interview a panel of relevant academics or equivalent to provide the information rather than pick and choose whatever sounds most sensationalist.
And guess what, just from TL.net, which we think have higher standards, we can see that most people don't even bother reading the very articles and pdfs they proclaim information from. They just type some random words into google, and post links of the articles that sounds like it might support their preconceived notion and call that an argument. And it works, because the avergae person can't be bothered to go through 100 pages of extremely dry reading when they aren't mentally equipped with the tools to understand even simple things like how to read a graph? So it is with TL keyboard warriors, so it is with how most journalists present science. The problem is journalistic integrity, or rather the lack of it and the lack of the mentality from consumers to choose media with journalistic integrity.
Yes, that's how it is supposed to work. What I am saying is that the current model is clearly failing. If we know disinformation is currently killing people, we should be examining all ways to prevent disinformation. Scientists may not be at fault, but they can help. So they should.
|
On August 30 2020 23:51 Mohdoo wrote:
Instead, they could say "Every single piece of information we have over the last 200 years indicate a mask of any sort will help reduce infection in some non-zero way. Go ahead and everyone wear masks for now while we figure out the specifics."
I would like to reiterate: Not a single chemist or microbiologist wondered if masks would be helpful. I was telling my friends to wear masks in ***February*** It was obvious it would help, just a matter of how much. There was never a legitimate argument for the idea that masks would not be effective. But stupid dogshit academic guidelines punched ourselves in the face.
Regarding this, the argument used by certain countries is not that masks doesn't work. They are very effective at what they're supposed to do, which is to reduce droplets and particles in the air. No one is saying otherwise. The question however is if they're effective at reducing the spread of a pandemic desease when used by the public and this is where is gets a bit murkier. IF it was such a clear cut case, why don't we see drastic difference in curves and cases in those countries and areas where they don't use them on a wider scale? I don't know, but the countries in Europe with the largest increase in cases are quite mask heavy while other countries with more targeted mask usage don't seem to do any worse. It's not as obvious as some people make it out to be.
|
Scientists can try and help, but it doesn't matter since journalists and other media can always pick and choose whatever narrative they want to feed to their audiences. The audience isn't going to shift through a thousand published papers on researchgate or whatever, they get their info from where they get their info, and the providers of that info who may happen to be journalists will continue to choose how to angle that same info. Frankly I don't know how many in USA got the impression that masks don't help, but I suspect that how research is published would change nothing.
|
|
|
|
|
|
EPT
