I took this antibody test through LabCorp on Friday and tested positive. I have been in strict (see: paranoid) isolation with one other person since March 7 and haven't felt sick or anything close to it since then.
I did have what I thought was a strange flu/cold early February with a sore throat, post-nasal drip, wildly varying body temperature, and lower back pains. In hindsight those are COVID-19 symptoms.
the person I live with took the same Abbott Architect antibody test from LabCorp today so that we can compare results and hopefully eliminate any false positives.
Good anecdote. I also had a "weird cold" in February with symptoms that are unusual for me but typical of COVID-19, and have been waiting to get the antibody test. I kind of expect it to turn out the same way. (Otherwise, I would really like to figure out what the "weird cold" that everyone had in February was. Because it sure was weird, and everyone I know had it.)
"Weird cold" in February checking in. I did not have any temperature variations or fevers, but I got struck with a fatigue like no other. For the first time since I was 3 years old, I took afternoon naps three days in a row because I was just completely wiped out. It's rare for me to take a nap at all (happens maybe once a year)! It took two weeks to get out of this fatigued state. My other symptoms were "chills-like" sensations throughout the day, post-nasal drip and burning feeling in my chest/lungs but very minimal cough. My child did have COVID-19 symptoms back around Feb/March (can't remember exact time-frame) and I even wrote about that here on HN. Maybe I should take this test.
I still think there's a possibility this virus was already in the US before 2020. I don't think they ever found a cause to the "mystery respiratory virus" in Virginia from July 2019:
> Three people have also died, but Dr. Benjamin Schwartz of the Fairfax County Health Department said Wednesday afternoon that those who died were "older" and had complex health problems. Officials don't yet know the extent to which the respiratory illness contributed to their deaths, he said.
I agree with you on being open to thinking that COVID-19 arrived earlier than was initially thought. When I first shared my "weird cold" story on HN, the consensus was that I was a crazy person and there was no way that could be COVID-19. There would be more deaths if it had been here. But now there are some deaths, and a large number of people with "weird cold" stories, and it really makes me think that a lot of people have already had this. Maybe the doctors at the time didn't recognize it as COVID-19 (I didn't when I had my weird cold). Maybe the numbers that we see now are just wrong; i.e. the recorded cases we have are just the tip of the iceberg (and if you tested everyone every week, we would have hundreds of millions of cases).
Or maybe that weird cold was just a weird cold and nothing more. It seems unlikely to me, but I'm open to it. I hate to make public policy suggestions on anecdotes... but someone should really look into that weird cold. Everyone in New York had it. It sure is weird.
people say that because .25% of the population in NYC are dead from covid. we dont see those numbers here -- if it was widespread in feb, what caused the discrepancy?
While it's harsh to say: People die every day. The CDC officially states that, on average, ~880 people per 100,000 die every year in the United States (in 2017). Extrapolating that to NYC, this means that the city can expect to loose about 0.1% of its population every year (possibly even more; I imagine some forms of death are more prevalent in large cities. Possibly less; healthcare is better).
Yes, losing 2x+ the annual number in one quarter should be noticed. But, maybe the virus evolved additional transmissibility over time, or maybe it evolved to be a little more deadly.
It simply seems less likely that this virus was first noticed in November, by China, yet it didn't hit major urban areas in the rest of the world for another four months. We're too interconnected as a species; nearly half a million people came to the US from China alone, after COVID-19 was classified there [1]. It feels more likely that people have been getting it, and dying from it, well before March; these cases were just simply miscategorized as the flu.
I guess we need to know the total number of cases to figure out what the deaths mean, and until everyone has gotten an antibody test, we can't know that total number.
To make the numbers work for the "weird cold" in early February case, I guess we have to work backwards. Assume that when we went into lockdown was the actual peak, 8 million people in New York City had COVID-19. That means that there is one death for every 500 cases. Then we have to pick a reproduction rate, which I have no idea how to pick, so I'll say that it increases by 1.5x every day (so on day one you have x cases, then on day 2 you have x + 1.5x, then on day 3 you have x + 1.5x + (1.5)^2x, etc. Going back a month from March 20 (which is approximately 30 days, and when a lot of people report their "weird cold"), that would mean we'd expect around 8 million / 383500 = 20 cases on Feb. 20. With 1 death per 500 cases, you'd have 0 deaths at that level.
I know I've pulled these numbers out of my nether region and so they are likely very wrong. But with some back-of-the-envelope math, I think we can have some sick people in February without a lot of deaths.
Obviously my 8 million total cases in NYC number is too high, and the 20 cases on 2/20 is too low for me and my friends to be those cases. But that exponential can be tweaked to make something plausible. If we make it 1.2 instead of 1.5, then we should have had about 7000 cases on 2/20, and that means around 14 deaths. That seems quite plausible to me. So I dunno. There was a weird cold. It's weird. It could mean anything.
I signed up to get an antibody test. 1 test is not data, but it will be very interesting to see the results.
> Then we have to pick a reproduction rate, which I have no idea how to pick, so I'll say that it increases by 1.5x every day (so on day one you have x cases, then on day 2 you have x + 1.5x, then on day 3 you have x + 1.5x + (1.5)^2x, etc
> I know I've pulled these numbers out of my nether region and so they are likely very wrong. But with some back-of-the-envelope math, I think we can have some sick people in February without a lot of deaths.
Yeah, in two ways: That's not how the growth rate math works, but if we went by your math instead of the number, that's about 4-5x faster than what we were seeing in March.
A growth rate of 1.5 means if we had X cases on day 1, we'd have X(1.5^1) on day 2, and X(1.5^2) on day 3. This virus's growth rate at the beginning of the pandemic stage was around 1.4 (or to use your math, 0.4? I'm not sure what you meant by "(1.5)^2x", is that a typo of "(1.5^2)x" or did you mean 1.5^(2x)"? The second one is straight wrong).
1 out of 421 NYC residents have died, based on today's numbers, and recent serological tests estimated that 20% of residents were exposed. If that's right, then the death rate would be about 1/84, or about 1.2%. But I've seen other claims that the exposure rate in NYC is higher, which would give a lower fatality rate.
Don't forget to factor this breaking through to long term care facilities. From the last I looked at this, if it had not hit long term care, the number of deaths would have been low enough that it may have gone undetected longer. (In WA, 92% of the deaths are still folks over sixty! I don't know the numbers for how many of them were in long term care.)
Which is to say, you can't just look at the population death rate and really appreciate just how dangerous this is for older populations. The CFR for over sixty is a staggering 15% in WA. That is ridiculously high and completely masked if you look at all cases.
> Don't forget to factor this breaking through to long term care facilities.
If they're not taking any precautions (because there's no knowledge of the virus in the community) then it can spread to those types of facilities very easily. In New Zealand where only 1,500 people have have COVID-19 (likely to be very accurate, 200,000 tests have been conducted) there's already been two outbreaks in nursing homes.
Certainly. I am not implying that we should just let it run its course. Just as I would not let any disease we have a vaccine for loose in a nursing home without the vaccine.
I just think the total IFR actually undersells how dangerous this is.
I'm sure we would have noticed. I'm not sure we would have noticed as quickly. We certainly would not have tested to find that nearly a quarter of NYC could have it.
Yes, it can do damage and is very dangerous for an at risk group. No, we don't know who that is, yet. Age clearly proxies for a risk factor. But which one?
Unless you are wanting to claim that no children have gotten this. Which, seems highly unlikely. (Or is the risk whether it will provoke an immune response?)
If I'm interpreting this[0] right, it looks like the normal death rate is 3.51 per 421 people per year. 1/421 in NYC (is that just COVID cases or all excess deaths?) is over much less than a year, so it seems fairly serious.
But I also had a truly "weird cold" go through my workplace early this year in SFBA, so I still have to wonder what that was.
That's a great point although obviously its very hard to estimate something that has a horizon of atleast a month. the bay area did shut down quite a bit earlier, it could be that the lack of deaths is simply due to a better response.
> people say that because .25% of the population in NYC are dead from covid
Source? I can only get a rate that high by taking the highest death count I can find (which involves extrapolations to attribute deaths to Covid-19) for New York State and dividing it by the population of New York City (which is clearly invalid).
From https://www1.nyc.gov/site/doh/covid/covid-19-data.page I compute 0.23% of NYC's population has died of Covid-19, counting confirmed + probable deaths and dividing by 8.399 million people. That's 1 out of every 421 people.
hm, care to explain more? that's exactly how I think about it. Total deaths in NYC attributable to covid (~20k) over the total population (8.5M) = ~.25%. In my mind this is a very easy lower bound to estimate, what am I missing here?
In anycase, even a rate of ~.1% would be absurd given ~20% of people are testing positive by antibody tests.
Actually, I was asking <i>you</i> to explain, or at least give your source.
I subsequently found what I believe you are working from here: https://www1.nyc.gov/site/doh/covid/covid-19-data.page
which does specify New York City (as opposed to New York State) so both the numerator and denominator refer to the same thing.
Interestingly, they appear to be using a larger denominator than your 8.5 million (NYC metro vs. NYC proper perhaps?) and get a rate of 0.175% = 175.66/100,000, eg in the "citywide total" line of https://github.com/nychealth/coronavirus-data/blob/master/by...).
But this is at least closer to your figure than I was able to get previously.
Specifically has numbers for NYC. You can sum the boroughs or use the total listed below. Using confirmed by pcr cases you get ~15k/8.6 = ~.175% as you said.
Using probable excess deaths adds an extra 5k deaths, hence ~.25%
We haven't confirmed functional mutations. One of currently known mutations (D614G being a prime suspect) can be functional and we can't rule out the possibility.
We don't yet know why (or even if) more people died in New York than on the west coast, per capita. We certainly cannot yet attribute it to a different strain.
I keep reading people asserting this, and I assume it's because we believe that a virus that becomes less lethal would have an evolutionary advantage.
But, wouldn't lengthening the incubation period also be a successful evolutionary strategy regardless of lethality?
It seems to me that there are many possible strategies that a mutating virus might gain an advantage and we shouldn't just assume that the only one that they would use would be to become more mild.
> Trade-offs between different components of parasite fitness provide the dominant conceptual framework for understanding the adaptive evolution of virulence (Alizon et al. 2009).
...
> By far, the most widely studied trade-off involves transmission and virulence (Anderson and May, 1982; Frank, 1996; Alizon et al. 2009). Transmission and virulence are linked by within-host replication: increasing parasite abundance increases the likelihood of transmission, but also increases the likelihood of host death; mathematically, this assumption can be formalized by making transmission rate β an increasing function of parasite-induced mortality rate ν. Nearly all of the literature we summarize below assumes this trade-off. However, another potential trade-off suggested by an examination of R0 involves virulence and recovery rate (Anderson and May, 1982; Frank, 1996). This trade-off is also mediated by replication rate, with high abundance increasing the likelihood of host death, but also decreasing the likelihood of the host clearing the infection (Antia et al. 1994); mathematically, this assumption makes recovery rate γ a decreasing function of parasite-induced mortality rate ν.
'Tend to' doesn't mean some conscious, directed effort. On large enough scale, this is generally the more probable direction, and that's about it. Its truly random by nature.
It can very well go the other way as with 1918 'spanish' flu. Or it can change in ways that won't affect overall mortality much.
It's interesting you should mention the 1918 flu. That one had two waves, the first one less lethal... Evolution is random, add you say, but that's not quite what happened.
It was in the middle of world war 1. Soldiers were infected, and the ones that became the most ill were sent home to recuperate, or die. Either way they spread the virus better than the ones who weren't as ill, and who stayed in the field.
So the usual evolutionary pressure was upside down here.
You're neglecting the other side of things in that analysis. The more successful deadly viruses tend to eliminate their victims, so the surviving population has a higher concentration of people who are resistant or immune to it, including those who acquire immunity after being exposed to the virus.
A lot of the opinion I've read suggests that so called super-spreaders are seeding the majority of flare-ups. In other words, there may be some individuals with very high R0 compared to the vast majority of others. My theory is that Covid-19 was circulating for some time here until the right people became infected, then it really started spreading rapidly.
> I still think there's a possibility this virus was already in the US before 2020.
Yeah, I don’t understand why people are so against the idea of it being in the U.S. before 2020. There’s no possible way we can trace every American who may have traveled within the China region or interacted with another person who was in that region around the start of COVID spreading. Just because there were no official cases of COVID does not mean it was not here.
I had a roommate visit Japan in November. He came back and was extremely ill for about a week. No hospitalizations but he was out of commission and isolated by himself.
I think the main evidence against it being here siginifgantly earlier is that we see how it spread when it was definitely here, so why wouldn't we have seen hot spots like Atlanta and New York earlier then?
Yeah - if someone could point me to a nursing home where 5% of the population died in a week in December/January I'd believe it - anecdotes about people being having the flu in the winter isn't the most convincing.
> if someone could point me to a nursing home where 5% of the population died in a week
Sure, if we were talking about mortality rates but we’re not. we’re simply talking about prevalence of COVID prior to the major breakout of February and beyond. These are two completely different things to be looking at.
I’ve worked in nursing homes during that time of December where people were dying. Did we think to send bloodwork off to test for some novel coronavirus? Of course not. Secondly, using the example of a nursing home as the sample population is silly as they are not the ones who were traveling. It’d be their family members and caregivers who traveled and then brought the disease into the facility.
I think you are missing the statistical point I and others are making. If the disease was introduced in a meaningful way in December/January it would be a statistical certainty we would see signs of it (such as nursing homes having huge waves of deaths or total mortality in subregions surging).
I'm not sure what your counterclaim is? There are no samples in America tested from that time period that would show that there was Covid nor is there any statistical evidence that would suggest there was Covid.
What specific evidence would you be expecting to see in December or January? The infection fatality rate for this disease is probably below 1%, and we don't know what percentage of the population is infected. We don't truly know how quickly it spreads. Nursing home populations are at greater risk, but even if the IFR is 3-5% for that cohort, the population mortality would be some fraction of that, and most of the deaths would be people with pre-existing conditions, and the deaths would be attributed to those conditions.
If there was a year-on-year increase of 1-2% in nursing home deaths for a month or two, would that register with anybody? Maybe they would have noticed if a lot of people were being put on respirators? But if no one knew about COVID they would probably just chalk that up to it being a bad flu season.
> Nursing home populations are at greater risk, but even if the IFR is 3-5% for that cohort, the population mortality would be some fraction of that,
Looking much higher than that. 80+ is 15-20% and 70-79 is 8% (https://www.worldometers.info/coronavirus/coronavirus-age-se...). It'll be concentrated more in people with pre-existing conditions so I expect nursing homes would see greater figures than that.
> Statistics from Kirkland now appear to tell the national story. Of 129 staff members, visitors and residents who got sick, all but one of the 22 who died were older residents,
So we're not talking about one or two deaths but a large proportion of your residents suddenly getting ill in the same way and of those a large proportion dying over a short period.
> Looking much higher than that. 80+ is 15-20% and 70-79 is 8%
The percentages you're citing are from the Chinese CDC [1] and represent the case fatality rate. They don't represent the infection fatality rate, let alone the overall population mortality. There's a big difference between these numbers which has been repeatedly ignored in the popular media, they keep on taking the scariest one (CFR) and presenting it in an unbalanced context. [2]
(To be fair, CFRs are the numbers we are the most certain about, but when you cherrypick the worst ones like the newspapers do they're also the scariest and least useful.)
The working definition of CFR in the Chinese study is basically people who saw a doctor, were suspected or confirmed of having COVID, and then died. But many other people would have caught the disease and not been seen by a doctor. Most of them would have been milder cases and they wouldn't have died. Key point is this number is not at all indicative of total Covid-related mortality in an exposed population.
The Kirkland story is really tragic, but it's just one data point and doesn't prove that ~15% of all nursing home populations will die. The conditions in other nursing homes could be very different, in fact that nursing home in Kirkland has since been investigated and fined $600,000 for unsafe practices [3].
Why does it have to be introduced in a “meaningful way” to be relevant? If there was at most a single case of it within the U.S. that would show that the disease spread earlier than what is assumed. Again, your argument rests solely on the idea of there must be thousands of death in order for this disease to exist within the states. That is an odd way to think about this.
It’d also be kind of hard to go back and rerun tests for people during those months, which shows why it’s very difficult to pinpoint when this disease started to really spread. I mean, if France and China both had cases during the month of November and December, you’d be hard pressed to say it’s not possible there were similar cases in the U.S albeit undiscovered.
It is not particularly unusual for a nursing home to see very high death rates around flu season. In the UK, where they have an excellent healthcare system, the average lifespan of a nursing home resident is 2.2 years.
somewhat of a dated article (2010) but the average lifespan of those admitted to nursing homes for EoLC is actually only about six months.[1]
however, some may stay longer, but then they have to worry about their condition worsening and eventually not allowing them to do ADLs (bathing, eating, dressing) on their own. the average length of stay before disablement is close to two years [2]
I understand that line of thinking but that would imply the prevalence rate of COVID was in complete parallel with the incidence rate. The former is what we do not know and that's clear when we've had patients labs reran from 2019 and found it was positive for COVID. Government data from China says the first case can be traced back to 17 November 2019 and even then they're unsure if that was patient zero[0]
How many American's were in Asia, China, or surrounding areas between October and December of 2019? How many of them were possibly infected? Nobody knows because there was no plan put in place with regard to travel to mainland China until the end of January 2020; even then since 1 January 2020, there were 430,000 people who traveled from China to the U.S.[1]
I have been wondering why it is considered to have originated in Wuhan rather than Wuhan simply being the first major outbreak where it was identified as a novel coronavirus. Is it always the case that hospitals will be capable of identifying a new type of virus based on some patients who look like they have the flu?
One thing that's known about covid-19 is that it has an unusual spread pattern. In some circumstances it might not spread too much (eg. average R0 is well less than 10) but in other circumstances where air is recirculating a lot or a particular patient is very contagious, it seems to spread a lot more. Couple that with many people being asymptomatic. So, what if it had been spreading in many places but Wuhan was simply the first place where it was detected? It seems to be capable of spreading without being detected for quite a while in many other parts of the world. (eg. Singapore's foreign worker dormitories)
In Russia, virus lab had major incident at Sep 16 2019. Paint job caused blast and fire, so responders come in to fight fire and STOLE SOME EQUIPMENT FROM LEVEL 4 VIRUS LAB, which was caught on video here: https://youtu.be/_w7SAeNcXA8?t=63 (look at man in uniform in background).
Almost correct. The "lab" was actually doing corona virus research, and funded by the US govt.
Corona is so contagious, that combined with Wuhan being an international airport that connects to Shanghai, the #8 international airport in the world, this was the perfect storm for a pandemic.
If by "funded by the US govt" you mean funded primarily by the Chinese government, with additional funding provided by the US and EU, due to the nature of the lab's work...
"Weird cold" in February checking in. Fever, back pain, chills, post nasal drip. Felt bad for ~3 days and then felt better. Getting antibody tested tomorrow, will report back.
Same here. Rushed to the hospital after two nights of not sleeping from the back pain and chills at end of Feb. First time I've ever taken an injection for pain relief.
Crazy, I had the exact same symptoms during last week of Feb / first week of Mar.
I was so wiped out that I was taking naps in a conference room at lunch. A friend of mine caught me and was surprised because normally I work through lunch, let alone sleep.
No cough, no fever but just this terribly uncomfortable feeling of congestion in my throat and upper chest. Best thing I can compare it to is postal nasal drip like you said.
I'm a family physician working in an outpatient office... there's just so many viruses that can explain those types of symptoms that are STILL more common than sars-cov-2.
I saw a 31yo 2/18 with primarily sore throat, aches/feverish (not documented), and again on 2/20 because the sore throat got significantly worse. I started feeling iffy 2/21 (Friday) evening and worse Saturday with a bad sore throat and fever around 102F, lymphadenopathy, this persisted for at least 2 days. I went to work on Tuesday and did a rapid strep (I did not feel I had strep but the other doctor wanted to do it)... I'm not really thinking that was covid19.
This is just about identical to my experience in early March. No (or at least no observed) fever, a distinctive but not awful sore throat, and intense fatigue. Urgent care clinic did not test for the virus, or even for the flu.
took this test this Friday (Quest Direct) in NYC and tested negative.
In late Feb. I had a 'flu like symptoms' with fever and massive headache for few days. What worried me later is that I had the 'can't taste anything' symptom a week later. No coughing though.
If you read online, you see people swear they had it at some point this winter. So, if you had a 'flu like' symptoms in Feb or early March, perhaps it was just that...
It cost $110 and personally I think it was worth it... (it is a reminder to be more careful as I can still get it)...
Just to put this in some context, but the Quest Diagnostics and other tests currently available are good but not great. For example, from their website there are the following disclaimers:
This test has not been reviewed by the FDA.
Positive results may be due to past or present infection with non-SARS-CoV-2 coronavirus strains, such as coronavirus HKU1, NL63, OC43, or 229E.
Negative results do not rule out SARS-CoV-2 infection, particularly in those who have been in contact with the virus. Follow-up testing with a molecular diagnostic should be considered to rule out infection in these individuals.
Results from antibody testing should not be used as the sole basis to diagnose or exclude SARS-CoV-2 infection or to inform infection status.
Other sources state the following, however its important that not a lot of research has been done to validate the manufacturers specifications, of these tests such as specificity. Only over time and independent review will these numbers be more accurately quantified:
"The company started out using the Euroimmun AG Anti-SARS-CoV-2 ELISA IgG antibody test and then added the Abbott IgG antibody test. It is continuing to use both tests, although the Abbott test has become the primary platform. Euroimmun is a division of Perkin Elmer.
Both tests have received emergency use authorization from the FDA; only 10 serology tests have received this designation from the agency. The Euroimmun IgG test has a reported specificity of 96% and sensitivity of 65%, while the Abbott test claims a sensitivity of 100% and specificity of 99.5%. The FDA has just cracked down on serology tests requiring supporting validation documentation to be submitted within 10 business days of an initial notification."
Similar story here, had COVID-19-like symptoms in mid-March, but the Abbott test (administered at CityMD) came back negative.
It makes me wonder what exactly is going on with this test. Do some people not produce antibodies? Is it overly specific to one strain of the virus and other people in NYC got a different strain that doesn't produce compatible antibodies? Was there some non-COVID-19 respiratory infection that hit NYC at around that same time that went under the radar? Is this test somehow just not as good as we think it is?
> Was there some non-COVID-19 respiratory infection that hit NYC at around that same time that went under the radar?
Not really under the radar, there are always a lot of respiratory illnesses. In my area, after reducing non-covid cold/flu by an order of magnitude due to social distancing, we're testing everyone with cold/flu symptoms, and they're still ~96.5% not covid-19.
Yeah, but I'm in NYC. Since some inflection point in March the vast majority of patients presenting with these symptoms do in fact have COVID-19. Being the global epicenter of the pandemic will do that :/
A lot of those tests were people not presenting with COVID-19 symptoms though. Think of all the asymptomatic people who were getting tested; that group had a much lower test positivity rate than 55%, thus bringing the average down. For people who had all the right symptoms, it was higher than 55%.
That effect doesn't seem like it would skew a larger difference between Alberta and NYC test positivity rates. If anything, that would depress Alberta test positivity rates more than NYC ones, since we've got more spare testing capacity for people without all the right symptoms.
But which other coronaviruses are widely circulating right now is the question. As in, if you were to lose your smell because of a coronavirus, how likely is it that it was not SARS-CoV-2?
I don't know the absolute figures in terms of cases per year, but the four common coronaviruses are... common. From [1], we see that 1 in 200 flu-like infections was the coronavirus HKU1. This [2] paper corroborates and gives figures for the other three common viruses.
We can therefore assume that about 1 in 25 bad-cold or flu-like illnesses are caused by a coronavirus.
> what the "weird cold" that everyone had in February was.
Probably just a different coronavirus or flu. I was sick for weeks in October 2019 with a dry cough and body aches plus lethargy that sound superficially similar, but the fact that whatever it was didn't produce packed ICUs seems to strongly indicate that it was something else.
Same here, had a lingering cough for nearly a month. The other weird part is I had flu-like symptoms for 3-4 days before I had any congestion and a runny nose. I can't recall any other time I've had a cold that didn't result in congestion within 24 hours of the first sign symptoms like a cough or sore throat. Just an anecdote, won't know until I'm able to get tested.
Yeah, I know exactly what you're talking about. I get a lot of colds, and they are all the same. The day before, I have a ton of unexplained energy. Then my body hurts. Then my throat hurts. Then it's like that for a week. It is always the same.
Whatever I had in February was completely different. The cough is all I really remember. It is so rare for me to get a cough that even though I was around people that were coughing, I didn't really expect to get sick. But get sick I did. (Oddly the thought of COVID-19 never crossed my mind even once during the whole thing. I knew about it but I didn't make the connection at all. It seems so strange to me in retrospect.)
I had something weird in March. It was a mix of a mild burning feeling in my chest, loss of appetite, and fatigue. I honestly thought it was just anxiety related to the quarantine manifesting in symptoms, and it very well might have been, but I have no idea.
I got wrecked by a cold in February, too -- late February. Got it from my friend. Neither of us ever get sick, and we both got wanged pretty hard. Same basic symptoms. Could have been the flu, could have been COVID. I'm gonna grab a test as soon as I can.
[edit] Make that tomorrow at 1245p. Turns out "as soon as I can" is pretty soon!
Right there with you. I felt like I got hit by a truck, and wound up going to the hospital for IV + Tamiflu. Took three weeks to feel normal again, and usually I'm over a cold or actual flu within a week. I never skip getting my flu shot, either.
Two things come to mind: The first is that the day I went to the hospital, there were so many of us with similar symptoms that we overflowed down the hallway.
Second, when the triage nurse was talking to me, she off-handedly commented that everyone was saying they'd gotten their flu shots, yet there we were. We both kind of shrugged and guessed it just wasn't a very effective year.
This is really only a step above anecdote in terms of sample size, but I know from nurses and doctors at all the nearby hospitals that they reached their capacity for respiratory / ICU cases (which is exceptionally unusual for them) right when the outbreak was supposed to be flaring up here (and what I mean by that is that there was the outbreak in China, then it was reported breaking out in other countries, then it was reported breaking out in Seattle, NY, and the Bay Area, then it was reported breaking out in our state capital and tourist hotspots, and then it broke out here, all in pretty regular succession).
Now, maybe more people went to the hospital precisely because that sequence of reporting led them to think it was serious when they may not have before, so the timing was a self-fulfilling prophecy. Ultimately, we'll really know so little for certain until we've had ubiquitous testing and a lot more time to study this, and the best we can do until then is extrapolate & theorize from what little data we do have. But the "weird cold" stories I've heard from February and even as far back in November just don't seem consistent with what pneumonia patients hospitals have seen - because that's how this was initially noticed as being unusual in Wuhan.
I swear I had one of these too, but it was in December, in Vegas, at the AWS con. (It started thurs or fri, was kinda pissed cause I booked some extra days in town but ended up staying in)
I don't remember all the symptoms, but I remember thinking how weird it felt and that "with a cold or the flu I usually get X"
Not gonna call it covid because I don't have any proof, but a "weird cold" is definitely going around
I had a weird bout of something in the middle of March. It was a mild fever, general weakness, and a strange pressure in my upper abdomen. It lasted a couple of weeks.
Before that I had gone to see my GP for a med update for refills. Apparently my doctor also saw what would be one our area’s earlier coronavirus patients that same day. I’m presuming I got put into the same waiting room as that guy.
The results I got had a big FDA warning that positive results may be a result of other infections (ie, that the tests are not specific to covid).
I feel like the WHO (body may not produce antibodies as part of beating covid, antibodies have not been shown to confer any immunity) and now the FDA may be talking about true but theoretical issues.
What is the REAL rate, in the USA, of folks who have for example MERS antibodies that will trigger a false positive in this type of test. Ie, do these tests really come back positive for MERS antibodies and how likely am I to have had MERS but not covid so that I get that false positive?
Given the FDA warning (which was the biggest thing on my results) I'm guessing 30%+ of positives are a result of these situations? But interested in actual data
These tests are not FDA approved, they're being allowed under emergency use authorization. They're not subject to the same rigger as tests normally allowed on the market and the manufacturer's specifications may be overstated.
Although I think the WHO has been way too cryptic talking about the issues surrounding antibodies and immunity, I think the real questions are really around degrees of immunity.
It might be that 80% of people get antibodies and immunity, but 20% don't. Or some people get antibodies but it doesn't prevent them from being a carrier. Or the immunity is only partial, etc. And then there's the ADE stuff which is theoretical but scary.
> I feel like the WHO (body may not produce antibodies as part of beating covid,
FTR, this assertion has been proven to be false (the other needs further investigation)[1]. Antibodies (IgM and IgG) are produced during the infection and peak afterwards.
Is a 99.9% specitivity good enough though to conclude you had covid?
Assuming you are in the US, your prior on having been infected in early February is well below 0.1% (not sure how much the covid-like symptoms would boost that). So conditionally speaking, it seems quite plausible this is a false positive.
There's a lot of anecdotal evidence for people experiencing strange flu-like illnesses all around the world as early as January. It's possible GP was infected with the weaker strain that was said to have existed (although I'm not sure if that argument's been refuted by now, I remember it was controversial).
I've been waiting for a decent serum test because both my wife and I had "the flu" with symptoms consistent with COVID-19, me in mid-January and her in late February.
$10 for a "virtual assessement" and ~$100 for the test for uninsured, but my insurance covered it. I scheduled it same day and it took a total of 20 minutes in and out with blood draw. I received my results via email in 2 days.
you're right, it's hard to find that definitively online. however, it lists it on my emailed test results as well as a friend's who did it on the same day. that doesn't mean they're not using multiple test types though. maybe worth calling and asking if they use it everywhere?
Interesting. I've long been of the opinion that I was definitely NOT exposed to it, although I had a persistent lingering cold in late feb/early march. (Symptoms took way longer than usual to clear -- but no flu symptoms).
That said, if it was in California before folks thought it was, it's theoretically possible.
My wife works in a hospital and will be tested for COVID-19 and antibodies, FINALLY, sometime this week (we think). If, for some reason, she happens to have antibodies, I'm going right out to get tested myself-- and, actually, I'd like us BOTH to get tested for antibodies (second time for her) because otherwise we'd be left guessing if one of us had antibodies and the other did not.
This is your wife whom you presumably spend most of your time with and sleep next to, right?
With the long incubation period and how contagious it is, I'd assume the chance of a false negative (.1% in the new Abbott test) would be higher than her having caught it and somehow not passing it to you.
How do you know you got that specific test? It looks like it's their second IgG antibody test and their fourth overall[1]. It also looks like the architect moniker just seems to be their testing platform? I assume their previous IgG antibody test was built on it.
I don't know much about this space, would just like to get the test done and would prefer one with high specificity and low false positive rate.
I also had a very sudden onset of a severe body ache around Feb 2020. I started having some pre-flu symptoms. And after one day, I started having a very severe body ace and it came so suddenly. I have never felt sudden onset of body ache from a flu/cold before.
I took a nap for 2 hrs or so and it fortunately went away almost immediately.
I know a family friend who were scheduled to go on a trip in early Feb 2020 but cancelled it because both had very severe flu symptom. This was an important trip (a road trip of few hrs to visit family) that had been scheduled weeks ago.
One of them to told me he had a meeting with a visitor from China who had mask on. And they think they might have had Covid-19 but no way to know until they get the antibody test.
Just took the Quest version today. I traveled a lot in February but only felt mildly under-the-weather once, so it's a long shot. Even if I was exposed, we don't really know the seroconversion rates for mild/asymptomatic cases.
Just back from the Quest draw myself. After a week's worth of drive tests in February with a colleague who had just come back after a 6 week vacation in EU with his family, I caught the nastiest flu I've had. Self isolated till I got better and after 1 day back in the office, the company imposed compulsory WFH.
I had all the COVID-19 symptoms including the weirdly varying body temperature, and I'm in NYC, but my Abbott test (taken 1.5 months after symptoms) came back negative. Who knows what's going on.
Here's my opposite anecdote: I live in NYC and in mid-March I had all of the headline COVID-19 symptoms, including most notably shortness of breath (which I've never before experienced in my life short of high altitude hiking). I got the Abbott test two weeks ago, and the result came back negative. I was so sure I'd had it, but now it's looking like maybe I didn't.
"Because tests to diagnose coronavirus infection were unavailable to most people in New York City in March, the researchers included another 719 people in their study who suspected they had Covid-19 based on symptoms and exposure to the virus, but in whom the illness had not been diagnosed.
In this group, the researchers found a different picture altogether. The majority of these people — 62 percent — did not seem to have antibodies.
Some of them may have been tested too soon after their illness for antibodies to be detectable. But many probably mistook influenza, another viral infection or even allergies for Covid-19, Dr. Wajnberg said."
Yeouch, that could be me in that group. Also:
"The team tested 624 people who had tested positive for the virus and had recovered. At first, just 511 of them had high antibody levels; 42 had low levels; and 71 had none. When 64 of the subjects with weak or no levels were retested more than a week later, however, all but three had at least some antibodies."
It's been well past the 3 weeks for me that they're recommending. I'm gonna get another test and if that also comes back negative, it seems likely that I didn't actually have it.
Yeah, I'm pretty sure I had it and didn't have bad symptons back in Feb. too. Just a light sore throat and generally feeling gross. It'll be interesting to see how many people really caught it once this is all over.
The test can also be conveniently requested through LabCorp.com using an independent physician service, PWNHealth, to determine if the test is right for you.
It is refreshing to read a headline about an antibody test that conveys useful information about specificity and sensitivity rates. This really helps to inform about how useful a test could be for serological surveys.
It's interesting -- in the paper it says that the samples they used to assess the sensitivity (it was a different population for specificity) were from people who had (mostly) been elderly and hospitalized, and that they do not know how it would perform in subclinical or asymptomatic cases. I admit being super not-objective about my reading, since I had this test done last week and came back negative despite losing my sense of smell completely in late March - early April (I live in Boston and was exposed to people who tested positive). It will be interesting to see more data on the Abbott and Roche tests on the subclinical / asymptomatic population.
I took this test this Friday (Quest Direct) in NYC and tested negative.
In late Feb. I had a 'flu like symptoms' with fever and massive headache for few days. What worried me later is that I had the 'can't taste anything' symptom a week later. No coughing though.
If you read online, you see people swear they had it at some point this winter. So, if you had a 'flu like' symptoms in Feb or early March, perhaps it was just that...
It cost $110 and personally I think it was worth it... (it is a reminder myself to be more careful as I can still get it)...
Not to put too fine a point on it, even if it is about politics and business it's practically crazy not to fully fund these kinds of activities. The ROI is staggering. It's pretty hard to imagine a price tag where it doesn't work out well for you.
When NYC said it would give free meals to anyone with no id check or possibility of being turned down to be refreshing and a solution without politics or hidden agendas.
I don't know that fully subsidizing testing would necessarily be enough to satisfy demand for testing, but at $100 per test x 300M Americans x call it 1 test per week, you're looking at $30B per week to test everyone every week, if testing really can scale up that far. Which is (extremely) expensive but still less than the current stimulus packages and economic damage.
I actually kinda wonder whether the ROI is high enough that it makes sense for some of the bigger investment funds to subsidize these tests. There are $4T under passive management in the US, so it would cost less than 1% of that number to test every American once at current market prices. If doing so would boost the market by at least 1%, it seems like it would have positive ROI to do so (and I for one would be happy if the index funds I'm invested in did so).
> So, if you had a 'flu like' symptoms in Feb or early March, perhaps it was just that
Please do remember that this was apparently a bad flu season before everybody locked down and the flu vaccine was less effective than average.
It was odd, because, anecdotally, in Southern California flu season seemed to be much better than average, but apparently other parts of the country were getting pounded.
Okay - I'm going to reveal my statistics ignorance here - but if they did get a false positive, how does 0.0% fall within the range of possibilities (or is that allowing for it might be 0.0499..% or lower?)
Not ignorant at all... I believe the issue is that the statement ("95% confidence interval of 0.0 to 0.5") has just slightly different/counterintuitive meaning in this case. In general, you can always make the interval larger, but you want it to be smaller. With the understanding that the 95% is a lower bound, [0.0, 25%] is also a valid confidence interval.
But, yes, you are entirely correct as to the practical meaning. 0% itself is impossible, but 0.0...1% could still happen.
Correct, I rounded to the appropriate number of significant figures. I hate extra significant figures which suggest that the data is more precise than it really is
All of the sibling answers are correct, but ignoring another possibility: that the timeline is simply wrong, and SARS2 was present in the US early enough that the "false positive" is in fact a true positive.
Which can't lead to a "true 0%", since there's no such animal. But it could mean that all of the tests gave correct answers, in the trial.
We know that 0% isn't a real value for a test's error rate, but the statistical procedure being used doesn't have any "look, it's never zero" criteria built into its math. All it says is that 0 is one of the numbers consistent with the data given these criteria.
TLDR: whether you're a frequentist or Bayesian, the Jeffrey's interval is pretty good. Use Clopper-Pearson if you're a frequentist who is scared of undercovering.
I get 0.011% for the 2.5% quantile and 0.46% for the 97.5th quantile.
> The sensitivity of the assay from the estimated day of symptom onset for the 125 patients included in our chart-review study was 53.1% (95%CI 39.4%-66.3%) at 7 days, 82.4% (51.0-76.4%) at 10 days, 96.9% (89.5-99.5%) at 14 days, and 100% (95.1%-100%) at day 17 using the manufacturer’s recommended cutoff of 1.4.
> The sensitivity from the date of PCR positivity was: 88.7% (78.5-94.4%) at 7 days, 97.2% (90.4-99.5%) at 10 days, 100.0% at 14 days (95.4-100.0%), and 100.0% (95.5-100.0%) at 17 days using the manufacturer’s recommended cutoff of 1.4.
There seems to be no analysis of whether these are neutralizing antibodies. The idea of using serology for immunity certificates or "golden tickets" is never going to go well. Even with 99.9% specificity, if the population prevalence is 1%, 10% of positives will be false positives. If in real world testing, specificity is 99% and population prevalence is 1%, then 50% of positives are false positives.
But the population prevalence is much more than 1%: 80k deaths at a 1% infection fatality rate (and I believe this is high, but I'm being conservative) implies 8,000,000 infections so far. This is more like 2.5%. So far. 95% CI for specificity is 99.5%, so you can be reasonably confident that you're doing better than 85%.
It may not be a perfect intervention, but you could really reduce risk. If there's an 85% chance that someone is immune, they do not share a household with a vulnerable person, and they are not in a high risk group themselves-- you've reduced the risk of death to basically nothing.
I disagree with it for other reasons (it incents people to go get sick to be free/be able to work/etc).
> I disagree with it for other reasons (it incents people to go get sick to be free/be able to work/etc).
On an individual basis yes, but doign large scale scientifically/statistically relevant antibody tests to see just how many people have actually had covid-19, would be very beneficial.
Would you or someone else mind expanding on this thought a little? Why does 99.9% specificity mean that 10% will be false positives?
{added: great answers below} well worth understanding this point. In short specificity measures the % of the population tested which had false positives, but doesn't give you the ratio of false positives to positives or the probability that a positive test means you actually have the anti-bodies.
Let me give it a try. Suppose we have 100,000 people in a statistically representative town.
If 1% of people have had COVID-19, then that's 1000 people who have had it, and 99,000 people who haven't.
The test has a sensitivity of 100%, which means all 1000 people who've had it will test positive.
The test has a specificity of 99.9%, which means 98,901 of the 99,000 people who haven't had it will test negative; but that leaves 99 people who haven't had it, but test positive anyway.
That gives us 1099 people who look like they have immunity; but only 91% of those people are actually immune: 9% of the people are false positives.
If instead we have a specificity of 99%, then only 98,010 of the 99,000 people who haven't had it will test negative, leaving 990 people who haven't had it but test positive anyway.
That gives us 1990 people who look like they have immunity; but only 50% of them actually do -- the other 50% are false positives.
So if I'm understanding this correctly, with this test.
If you test negative, you are clear, guaranteed, no false negatives.
If you test positive, there is a 10% chance it's a false positive.
I guess my follow up question, does a retest of the positive population make that false positive rate drop to 0.1%, or is the reason for false positive significant to an individual and not random chance?
> If you test positive, there is a 10% chance it's a false positive.
Well, don't misunderstand -- it's got nothing to do with the test per se, but with the probability that you had the disease in the first place.
The test itself has two probabilities:
1. If you've had COVID-19, the probability that it will report positive (sensitivity)
2. If you haven't had COVID-19, the probability that it will report negative (selectivity)
But those probabilities give you a mapping from reality -> test_result. What you want is the reverse of that -- and find the probability from a test_result -> reality. When you do that, you have to factor in the probability that you have the disease in the first place.
If 50% of the population have had COVID-19, then a positive test means a 99.9% probability of having had the virus. If 1% of the population, a positive test means 91% likely you have it. If only 1 in a million people had COVID-19, then the number of false positives would completely overwhelm the number of true positives.
This is sometimes called the "Base rate fallacy": forgetting to factor in the base rate when determining something like this.
It's important for things like, say, systems which automatically detect terrorists at airports. How many travelers at an airport are actually terrorists planning to attack a plane? It's got to be one in hundreds of millions, if not billions. With that low of a base rate, even if you had a system that was 99.999% accurate, the vast majority of people it flagged up would be innocent.
I had the same question about retesting. Here’s a quote from Scott Gottlieb (former FDA commissioner):
“While all of these tests can still generate false positives—a finding that you have the antibodies when you don’t—that risk can be sharply reduced by repeating the test if it comes back positive. The predictive value of two consecutive positive tests is high enough that you can be confident antibodies are present.”
Let's say you have 1000 people, 1% or 10 have had the virus and are seropositive (have antibodies).
Of the 900 people who do not have ABs, 99.9% or 899.1 are correctly identified as not having them, 0.9 is identified incorrectly as having them when they actually do not.
Of the 10 who actually have antibodies, 100% are correctly identified.
So 10.9 are identified as having antibodies, in 0.9 person's case incorrectly which is about 10%.
If the number of true positives is 10/1000, and the test gives you 11/1000 positive results, then 1/11 of your tested positive results are false positives. (Actually closer to 9% than 10%).
via Bayes Rule: “Assuming an underlying infection rate P(I), what’s the probability that a person is actually immune (=was infected), given that they test positive, i.e. P(I|+)?”:
Bayes theorem assuming 1% actual positive rate. If you test 1000 people, you will get roughly 10 positives and 1 false positive for a false positive rate of 10%.
I think it's easier to understand when you take it to the extremes: Assume nobody has the thing you're testing for. You test 100k people at 99.9% specificity, which means you get 1k positives just because of the rate. Since nobody hass the thing you're testing for, they're all false.
When the thing you're testing for is very rare, it's just as rare that the people who tested positive will actually have it.
We have good data that the IFR is in the 0.1-1% range, putting cases in say MA in the 7% range a couple of weeks ago (time from infection to death), which based on confirmed cases would put it well above 10% now
That means you’d have 1k false positive and 10k true positive from a test.
If we want to ensure that the active population is always above herd immunity threshold, 10% or 20% false positives would be acceptable, since the population in contact would still be 80 or 90% immune.
I think this is more about finding hotspots. Where could a seeding event lead to a large outbreak. Clearly 1 person entering into NYC with COVID today would not have the same effects as 4 months ago because supposedly 20% are immune. At least for the time being at least the assumption is. So there is that effect, but also how big could the next wave be based on that. Of course that all has to be proven that detectable immunity or levels above a certain threshold will prevent you from getting infected again or at least reduce the severity over some period of time. So how many antibodies, which type, how long they last, and how much do they have an effect on reducing reinfection or severity of the next infection - all yet to be determined.
Unless everyone's immunity stops at about the same time, a year or two of immunity would be sufficient to prevent large scale pandemics from recurring from this virus.
If you want neutralizing antibodies you can do serum virus neutralization assay. It's more complicated and takes longer.
Testing doesn't need to be 100% to be effective, but it does need to be better than random chance. A mixture of contact tracing, PCR testing, antibody testing and effective quarantines could be used to make the virus go away, but would require a coordinated strategy that the US has not attempted to implement, much to my dismay.
The people from the Duke-NUS Medical School have designed a "surrogate virus neutralization test"[1] which might be used to perform neutralization tests without the security requirements of using live virus.
Because with a false positive antibody test then you can still contract and be a carrier. This is a problem for the immunity passport proposals (which I think are a terrible idea anyway, but that’s beside the point)
It doesn't have to be perfect. If 90% of the at risk population where it's allowed to spread has immunity, then that's high enough that the virus would logarithmically decay. That's often referred to as "herd immunity". So long as on average cases result in less than one additional case you're fine. 1000 cases with an uninhibited net reproductive rate of 4 with a population that's 90% immune would result in the next "generation" having 400, then 160, then 64, then 26.... You get the point.
This is especially bad if there's no government-sanctioned method for infection - then you end up with a mess of asymptomatic people trying to get infected while they're already infectious.
The person you are responding to is suggesting a scenario like the movie 'contagion' where people marked clean are given a wristband or pass of some kind.
In reality testing can be used as an effective tool regardless of whether or not people can be 'certified'
It seems like a voluntary test for someone without symptoms is of limited value to the individual (beyond their individual curiosity) since we don't know enough either way about reinfection risk at this point, and the result is not terribly actionable - you can't just walk into stores without a mask because you have a positive antibody test - nobody is going to trust your lab result printout as a reason to avoid public health measures.
But that said, there should be a public health benefit to broad antibody testing to understand the true infection rate, and for that reason alone, seems like tests like this
should be covered by insurance or the public purse for everyone - at least in areas with outbreaks or at high risk for outbreaks.
People hav held off seeing their parents. If I knew i had antibodies, I’d plan a trip later in the summer to live with them and do all the shopping to decrease the risk of them getting infected.
People with antibodies are the best people to provide services with the at risk population.
I think it's pretty a pretty safe guess that given what we know about viruses work, someone who's already been infected travelling to live with parents and do their shopping for them will lead to a net decrease in infection risk.
Just because you have antibodies does not mean you cannot spread the disease. Antibodies relate to your own body's ability to resist an infection from what I understand.
The FDA suggestion of doing two tests to reduce the percentage of false positives (since you can't rule out cross-reactivity with another coronavirus) looks like a good idea for broad testing.
That can't be discounted, but if you have two tests that target different epitopes of the virus (hence, independent), the probability of them being wrong at the same time is
p_false1 * p_false2
(probability of independent events), hence much lower.
Cool, it's nice to see progress on antibody tests - the early Chinese ones in use here in Czech Republic are reportedly (accordingly to local news articles) rather noisy. Still preaty nice when they work, as you only need a drop of blood & get results in 10 minutes vs sample swabbing & complicated PCR sample processing taking up to a day.
A bigger issue than the specificity of the antibody tests is that they're generally not positive until you're nearly recovered. This confines their use to specific purposes (assessing whether someone may be immune, estimating true infection fatality rates, etc).
The authors don't list an affiliation with Abbott. Usually such a paper will have a statement about conflict of interest or competing interest. This paper has none that I can find.
> It is unclear what the prevalence of antibody is in individuals with subclinical or asymptomatic infections and how this assay performs in an asymptomatic population.
> Our serological validation was chiefly limited by use of excess serum specimens from a mostly hospitalized population known to be very recently infected with SARS-CoV-2.
Basically, this test may not be as sensitive (or patients may not seroconvert as often) with mild or asymptomatic cases. To avoid this problem we should run a two-part study, the first part a PCR test of a large random sample, and the second part a follow-up to measure antibodies. Has anyone heard about a study of this kind that may be in progress?
Specificity and sensitivity feel really poorly chosen and difficult to remember. You could swap their definitions and the names of the terms would make just as much sense. I much prefer true positive rate and true negative rate.
Seems to me like pretty intuitive terms. If the test is not sensitive enough, it may miss some cases and have false negatives. If the test is not specific enough, it may be triggered by something different and have false positives.
Those named could be confusing to. True positive rate = .95 could be interpreted as, if you get a positive result, that's a 5% chance of a false positive, when the percent chance of the test being a false positive given a positive result is actually a function of the true positive percentage in the population and not innate to the test.
These stats are generated just a couple weeks after symptoms or PCR test positivity. I'm still wondering how specificity might drop off for much earlier exposures. Anecdotally: people on my Fb feed are getting positive tests they attribute to illnesses they had back in January and February.
(This concern has nothing to do with the benefit these tests have in "real time" testing).
Don’t you mean sensitivity dropping off? The way it’s phrased sounds like someone is more likely to get a false positive as time passes after... no infection happened?
Do you know that all the Quest tests are mfg. by Abbott? I didn't see that anywhere, they didn't seem to show which test manufacturer you got when you receive results either.
The Abbott test may not be used everywhere. I've called them to confirm their labs local to me use the Abbott test and that's probably what you should do if you're looking to get this test. While it does not guarantee that they use the Abbott test in every region, you can see that they reference Abbott in the footnotes for question #3 in their FAQ on COVID-19 testing here: https://education.questdiagnostics.com/faq/FAQ219
Nice! Last year, in July 2019, in Virginia, there were a bunch of strange flu like illnesses, leading to pneumonia, and deaths [1]. When will these people get tested for the antibodies?
How can one measure such high sensitivity and specificity rates when no other test has such high rates?
Surely the only way to measure a test is to compare against ground truth data, which can only come from other test processes, and therefore can never reach 100%?
The Quest website warns that having the antibodies does not necessarily mean you are immune. Is it really likely that reinfection is possible less than a year after initial infection, or is that just something they have to write as a disclaimer?
We just don't know right now. Some antibodies could work better than others, so for some people it could mean yes and for others no. Could be different for symptomatic vs asymptomatic patients. Some antibodies could make the disease worse through ADE. Immunity is probably conferred for at least a couple months since we dont see a huge number of reinfection, but we dont know whether it trails off after a year or two or goes below an important concentration threshold since there physically hasn't been enough time since the start to measure immunity out to a couple years. In general, you're probably fine, but maybe not, and the test could also be giving you a false positive.
As a lay person, I think it is somewhere in the middle. In all likelihood antibody presence does mean some level of immunity or at least resistance. But testing is still only measuring presence of antibodies, not their efficacy in resisting future infections. Furthermore somebody who is reinfected but has antibody driven resistance may end up behaving as an asymptomatic case.
People who have diabetes often do blood tests daily, at home, on their own, drawing a drop of blood from their finger with a lancet. If they can get it down to using a drop of blood, I think people could be convinced to test at least one member of the household daily. That should at least give us enough information to measure spread.
It would be new for most people but it's not unprecedented and I think people would figure it out quickly enough. Conspiracy theorists would ruin it for everyone of course.
I think the conclusion I would draw from the Theranos debacle is "doing hundreds of complex tests from a drop of blood is impossible", not "doing any complex test from a drop of blood is impossible."
That's like asking, "could Enron have been an honest energy company..." sure, with a different board, different executives, different strategy, different technology, different IP, different market sector, and if they sold ice cream cones for fair prices at the local pool.
I know everyone is saying that getting reinfected with covid isn’t technically ruled out, but is it safe to assume so? It seems very difficult to find a case of someone truly being reinfected.
>Alternative index value thresholds for positivity resulted in 100% sensitivity and 100% specificity in this cohort.
This would be true of a coin flip, wouldn't it? "Head means positive, tails means not negative" would be 100% sensitivity and "Tails means negative, heads means you're not positive" would be 100% specificity. So those would be the "alternative index value thresholds."
You misunderstand. A test that labels everyone positive would be 100% sensitivity but 0% specificity, and a test that labels everyone negative would be 100% specificity but 0% sensitivity. Deciding every case at once by a single coin flip would thus be useless.
As the article says, "A perfect predictor would be described as 100% sensitive, meaning all sick individuals are correctly identified as sick, and 100% specific, meaning no healthy individuals are incorrectly identified as sick." Deciding every case at once with a single coin flip does not meet both critera of a perfect predictor.
This is encouraging. Two negative viral tests and a positive antibody test, and people should be able to get an ID that says they're exempt from movement and mask restrictions. (To discourage faking this, getting that ID should also turn eligibility for the Emergency Care Benefit, since they can work safely.)
There's still argument over how long immunity lasts, but it has to be at least two or three months, or there would already be widespread (large numbers) reports of people getting the disease more than once. So, for now, IDs could be good for three months. They can be extended later as more info comes in.
First step in the US should be to test first responders in NYC. NYC already has maybe 21% antibody-positive people, and those on the front lines are probably higher. Knowing who's immune will be a great relief to them.
>>> ..they're exempt from movement and mask restrictions ...
There's a negative repercussion of people around such folks getting a false-impression that masks are not a serious need (if this is in question, that's a different discussion).
It also creates a possibility of visible 'haves' and 'have-nots', which can cause trouble.
I think having uniform messaging for movement/masks irrespective of antibody presence will remove variability in response from people.
Maybe. If only 1% of your employees could work safely, it's probably not worth opening the doors and paying them to try to get something done in the absence of the other 99%.
> A special document to exempt people from “movement and mask restrictions”...
I just find this idea viscerally revolting. All I can think of is armbands with the Star of David.
I don’t believe there can ever be any justification for mass restriction of movement in society, because such is a roadmap for tyranny.
However I need only make the case there clearly isn’t justifiable ends in our current pandemic. What we have found is a minuscule and highly stratified extreme-risk population (mostly, nursing homes) and a vast majority for which COVID is the least of their concerns. Much more pressing a concern for the masses is getting back to their jobs, their so-called “elective” procedures, and resuming healthy socialization and recreation patterns.
What we need, and what the Constitution requires are strictly targeted measures to keep COVID out of nursing homes, and for everything else to re-open.
Mass restriction of movement and mandatory testing is neither legal, nor is it conscionable.
I did have what I thought was a strange flu/cold early February with a sore throat, post-nasal drip, wildly varying body temperature, and lower back pains. In hindsight those are COVID-19 symptoms.
the person I live with took the same Abbott Architect antibody test from LabCorp today so that we can compare results and hopefully eliminate any false positives.
more info on the test: https://www.corelaboratory.abbott/us/en/offerings/segments/i...
FDA article from 05/07/2020 on serology test performance including the Abbott Architect from OP: https://www.fda.gov/medical-devices/emergency-situations-med...