So I had a sister that got sick, she took a test, not COVID. She was still sick a little while later, another test another negative. She finally took a third one and that returned a COVID positive test.
In the meantime my children have been sick for a little while, and we got one of them tested and it was negative; however right before they got sick we had interaction with someone who had a positive COVID test shortly thereafter.
I am starting to wonder at the efficacy of the tests and is it that they just can't catch Omnicron as well? Are my kids really just sick with some other random thing despite dozens of people in our social circle getting COVID?
I'm just wondering at this point if there is a reason for the mass testing, because it doesn't really let you know anything except whether or not you have COVID and you'll get better or get worse. I can understand in hospitals or something like that, but in the general population I am failing to understand as it seems Omicron is incredibly virulent, and you can spread before you have symptoms, so what is the end game with the mass testing?
If you look at the actual testing data for antigen tests [0], the LoD for them is very high compared with PCR. This means a person needs a high viral load for tests to be effective. For example, the BinaxNOW kits require 140.6 TCID50/mL [1] vs 0.012 TCID50/mL for a PCR test (I picked a random one here) [2]. The FlowFlex tests are even worse, requiring 2500 TCID50/mL [3]. I personally purchased a few of the Lucira tests (not affiliated I promise) because they're more similar to an at-home PCR (they use LAMP) and so the LoD is lower (0.9 TCID50/mL) [4].
Wow those package leaflets are impressive. We (in Germany) get those Chinese test on an emergency admission and all they have is a single table with specifity and sensitivity (no info on viral load at all). They are however cheap and they still work . The responsible agency is actually testing them recently and at least publish the results: https://www.eurosurveillance.org/content/10.2807/1560-7917.E... . Still they allow companies to sell obviously nonworking tests with nonreproducible test results.
Health officials in other countries have found it is much easier to detect omicron in the throat, and not the nose, with these antigen tests. This started with a big wave of people who were PCR positive but nose antigen negative. When swabbing the antigen in the throat, the false negative percentage dropped
Swab throat too during rapid COVID test, Israel's Health Ministry says
Antigen nose swabs are extremely unreliable, unfortunately. Barring public knowledge about this, the campaign could be ineffective at best or actively harm infection rate at worst.
Anecdotal: I was in my 3rd of 4th day of symptoms before I finally came up positive on antigen (at home and at testing sites; the test appeared to be the exact same). PCR showed it from day 0 but the results lag quite a bit especially right now in the US (it took 7-10 days for me to get the PCR back in early Jan).
They're not so much unreliable as ... really not very sensitive, as I understand it. Viral load in covid sufferers follows a curve with a rapid ramp up and a slower ramp down, both PCR and lateral-flow devices require a certain minimum level in order to read positive. PCR is much more sensitive (because it's an amplification technology, so can "see" a lower level of viral load) than lateral-flow devices - which means you can be "safe" on an LFD and "unsafe" on a PCR for a small amount of time on the ramp up and a fairly large amount of time on the ramp down. The level at which you're at risk of passing it to other humans is I _believe_ somewhere between the two tests, so it sounds like you maybe got unlucky with an LFD or two, but the results you report are about what I'd expect :(
LFDs are decent for telling you if you're infectious _right now_ - they're much worse than PCRs for telling you if you have covid at all.
Disclaimer: I'm a software engineer, do not take medical advice from me. My partner is a biochemist and I regularly chat with a couple of council covid-response officers, and the above is the explanation I've had from them when I've asked.
Best I’ve heard it described is that Antigen tests are best for detecting high virulence on days 3-5 (e.g., when the spread is happening from tons of viral production, including in the nasal passages). They are more for public health usage to prevent spread rather than diagnostic like PCR.
In Wake County, NC, rapid high-sensitivity PCR tests are available at about a half-dozen sites with results in 12 hours. Appts are have been available within 24-48 hours depending on demand.
You drive into a parking lot, they hand you a test, you self-swab and give it back to them.
There is other testing available - this is just the county-provided testing. It’s well run.
>I am starting to wonder at the efficacy of the tests and is it that they just can't catch Omnicron as well? Are my kids really just sick with some other random thing despite dozens of people in our social circle getting COVID?
Ever since the vaccines came out and were implemented in full force, I am wondering what the purpose of testing even is. At least in the US, even the most restrictive locales have no intention of shutting down commerce to limit transmission. I do not know anyone who is not treating this coronavirus like any other, or the flu, or a rhino, or whatever other sicknesses normally go around.
My kids' daycare (or local government, not sure which is responsible) has a crazy policy of shutting down a class for 2 weeks if a covid case is detected. But not if any of the kids are sick with RSV/flu/rhino/noro/rota, etc. The daycare has not had any reports of covid for months. I am pretty sure all the parents are simply choosing to not get their kids tested because who can afford to drop their obligations for 2 weeks and also force all the other parents in class to do the same. Every single kid walking through those doors has a snotty nose, and surely at least 1 has had covid 19 over the past few months.
if there is a reason for the mass testing, because it doesn't really let you know anything except whether or not you have COVID
If you test positive, you're supposed to stay home and isolate so you don't spread COVID.
Tests aren't perfect, since as you said, there seems to be a significant false-negative rate with Omicron, and you can be contagious before you show symptoms, but the tests help prevent people from spreading confirmed cases, and can be helpful to pre-test before you spend time with someone who is at high risk for COVID.
> Tests aren't perfect, since as you said, there seems to be a significant false-negative rate with Omicron
Its not even Omicron. Rapid tests always had something like a 40% false-negative rate or something stupidly high.
------
That's why you're supposed to take two tests if you're asymptomatic, to reduce the chances of a false-negative.
Fortunately, the false-positive rate is less than 5% IIRC, and fairly reliable.
EDIT: Sensitivity and Selectivity isn't really false-negatives or false-positives. But for the purposes of this discussion, its probably good enough. Besides, I'm just an ignorant programmer who doesn't know the difference between them.
The most recent guidance says isolate for 5 days, then go back to work if you want to, or optionally take another rapid test and if it's still positive, isolate for another 5 days. The CDC also acknowledges that the virus is transmissible for 10 days after symptom onset. These "shoulds" and "supposed tos" are being thrown out the window. Health care workers in CA can go back to work immediately following a positive test if their symptoms aren't severe and in many cases they're being asked to.
Rapid tests give you a result within 15 minutes. PCR tests you gotta ship to a lab, and wait maybe 2 or 3 days before they report back to you.
I think PCR tests (which nominally take 15 minutes) are the best. But we clearly don't have the infrastructure to support it. Every lab doing PCR tests is going at full throttle and yet we still have this 2 or 3 day delay everywhere.
Meanwhile, at-home rapid tests work immediately. They're way less accurate, but its better to have some degree of testing rather than nothing at all.
--------
Lets say you have a workforce consisting of 10 people who are sick with COVID19 right now. PCR tests might catch 9/10 of those people. Rapid tests might only catch 6/10.
But removing those 6-people with Rapid tests in your workforce will slow down the spread. 4-people spreading COVID19 in your workplace is better than 10-people spreading COVID19.
-----
EDIT: More importantly, Rapid tests will catch those 6 people _TODAY_. PCR tests will take ~3 days before they report back, meaning if you're on PCR-tests only, your COVID19 positive people will be around the workforce for 3 days.
If you use both Rapid Tests + PCR tests, you remove those 6 people _TODAY_ from the workforce (meaning 4 people will be spreading COVID19 in your workforce). 3 days later, those 3 other people are removed which is "too late". The Rapid test was in fact more useful.
PCR tests are better for contact tracing, due to their higher reliability. If someone tests positive with PCR tests, you trace their contacts and tell _THOSE_ people to get PCR tests and/or rapid tests, in an attempt to "get ahead" of the disease. (IE: those 3 people who weren't caught by rapid tests were in contact with 15 people total. You now tell those 15 people to rapid test + PCR test).
My symptomatic, boosted SO received 4 negative nasal-swab PCR tests (1 day apart each), given by a nurse, before finally testing positive on the 5th test. Seems totally unreliable
There are a few news stories out there indicating that because Omicron grows better in the throat that also swabbing the throat helps reduce false negatives. Though the official word would be to not do this since the tests were officially only developed tested with nasal swabbing.
It's worth noting that rapid test results roughly align with contagiousness, which makes them an excellent tool to fight the spread of this thing. [1] (See the 2nd image on this page).
PCR tests will report COVID infections that are way below the threshold of transmissibility, and they'll keep reporting positive results long after the infected person is no longer contagious.
"Am I contagious?" is arguably a more useful question to answer than, "Do I have any detectable amount of COVID in my body, at all, using the best laboratory testing processes available?"
I my country they did a (pre-omicron) comparitive study, testing people with PCR and giving them an antigen test to use at home [1]. There they found a false-negative rate of 16% (sensitivity 84%) for symptomatic people with high viral load, which is a proxy for contagiousness. For people without symptoms false-negative rate was 77% (sensitivity 23%). Not that great.
The answer is probably something along the lines that it's not a magic bullet but 1.) There's a lot of political pressure to do "something" and tests are one of those things and 2.) Whether useful or not testing is required either literally or people want it for the peace of mind (illusory or not) so lack of tests is a headache.
Covid is much more dangerous for the elderly than the young, even vaccinated. The young are so safe that there have been ~250 total covid deaths in the US age 0-4.
Also, there is growing evidence that the lateral flow tests correlate well with being infectious.
Testing reasons:
For lots of staff (medical, school, factories), to avoid unnecessary prophylactic isolations, particularly for people whose work cannot be done remotely. And for the same group of people, testing daily (were we in a competent country with the test manufacturing to support) could significantly reduce community spread. Daily tests are infeasible to do with pcr if for no other reason than the reporting lag.
For visiting high risk people: reasonably reliable, fast indicator of whether you have covid in order to make near realtime decisions about activities to engage in. Like masks, this will reduce community spread.
From what my pharmacist told me a fair bit of the kits can't detect omnicron reliably. The ones that do seem to have about a 7-10% failure rate. He suggested doing two or even 3 kits at once or space them out over a couple of hours to decrease the chance of a false negative.
Tests aren't statistically independent events. Taking multiple rapid antigen swabs doesn't improve their accuracy. Multiple tests spaced out over a few days does.
> I am starting to wonder at the efficacy of the tests and is it that they just can't catch Omnicron as well? Are my kids really just sick with some other random thing despite dozens of people in our social circle getting COVID?
My understanding is:
1) Existing rapid tests are not as sensitive to Omicron as previous variants.
2) Negative results on rapid tests are unreliable, especially single negatives. They're more reliable when taken in series, and the instructions for the ones tell you to take two tests, 3 days apart. However, their positive results are very reliable, so if only 75% of the positive people who take one actually test positive, you've still successfully detected 75% of those infections.
PCR tests are far better at detecting infections early.
With this variant, the tests seem mainly about "how much longer should I quarantine?" It's test two or that finally shows positive, after you're pretty sure it's COVID.
I am really surprised I haven’t heard anyone talk about false positive rates, false negative rates, or bayes’ theorem since covid tests first became available. I mentioned it to family when the stories of conflicting tests showed up on the news and they balked at the idea of using math…we are very different people, my family and I lol
I took a self test after having symptoms and it was positive so I just isolated myself instead. I skipped doing an official test because just as you said it does not provide any relevant information to me other than what I already know.
People like myself would act very differently to a cold than COVID. It changes also needs to isolate/quarantine, and helps us get a handle on asymptomatic cases who otherwise would go around continuing to infect others.
This won't help us get a handle on asymptomatic cases - to do that, we'd have to be able to test everyone on a frequent and consistent basis.
This program doesn't even cover one test for everyone in my house, let alone weekly or more frequent testing. And aside from this, I have to spend hours looking for any available rapid tests to buy around here, and if/when I have some, I can't afford to use them on randomized or even weekly testing.
The rapid tests have a very high false negative rate regardless of variant, but a very low false positive rate. They should not be trusted as proof that you do not have COVID, only as proof that you do have it.
I think you can generally reasonably be sure that you dont have it if you test negative on a rapid test, especially if you take multiple. The accuracy of the better ones (which is what we're getting in the US) looks to be about 80% for people who do have the virus. To me, you can trust that to say you dont have covid, especially if youre vaccinated.
Does every submission related to covid have to be an opening for a HNer to recount a personal anecdote about how a thing being done to try and reduce covid spread is questionable?
My post was put in earnest because I am really trying to understand, I live in an area that is pretty much of one mind and believe that the best way to try and work through some of the division we are facing in this country is to understand and listen to the perspective of others. Even if I don't agree, trying to hear the other side may make me aware of considerations and factors I was unaware of before.
Ordering is super easy. I'm glad they're using USPS's existing infrastructure so we don't get a healthcare.gov 2.0. We'll see how actual shipping goes.
Wow… I run uBlock Origin, Privacy Badger, Firefox Tracking Prevention, Little Snitch, and a pi-hole, and I only had to allow one domain on LS for this to work. That's a better experience than like 99% of the internet for me most of the time.
Man, I was so bummed I didn't have to make an account and everything /s
I mean, for what it's worth, this gets high marks from me for ease of use and seems to be consistent with just about every advice you see surface here about getting your users through an ordering funnel. Of course the fact that the cost is 0 here makes it a simpler problem for them, but given that it is a gov't project I could have envisioned a lot, lot worse.
The US post office gets a lot of crap thrown at it when in reality it is quite well run. They're totally bootstrapped (stamp sales pay for everything) and take no money from taxes. They service every household in the country, rain or shine. They've scaled from nothing to hundreds of millions of customers and survived hundreds of years. They move millions and millions of physical goods across the entire country in days. If this were a startup or private company we would be in awe of their track record.
This isn't quite true; the USPS was formed in 1971, out of the United States Post Office Department.[1] It didn't '[scale] from nothing'.
The USPS also enjoys a monopoly on mailbox deliveries, as well as price protections, in the form of a floor imposed on potential competitors. This doesn't seem particularly admirable.
"They have scaled from nothing to hundreds of millions of customers and survived hundreds of years."
Okay, this is not particularly novel, and the implied comparison to scaling digital services is flawed.
"If this were a startup or private company we would be in awe of their track record."
Again, if a startup operated like USPS they would fold real quick. The USPS has continuously received public money to remain in operation. They are not innovative. The only area where they are expanding is in parcel delivery, and they often turn it over to other partners to do the last-mile pickup and delivery.
However, startups can aspire to the ubiquity of USPS. In many towns, the Post Office is the only happening place and fosters community, and occupies an important place in the hearts of people who rely on it in a big way. Many private logistics providers such as UPS and FedEx do not serve these areas because it's uneconomical for them, and the Post Office is the only logistics provider. For this reason, the USPS deserves taxpayer support.
I also think they are important for national security. If you ever needed to roll out a vaccine quickly - the USPS would be the first bit of infrastructure you would leverage.
And it's really nice that they're actually using the physical address for this instead of the given email address. Especially since some imbecile put my email address in when ordering her tests. I was concerned that I would not be able to order my own since my email was already used. But no, I was able to.
USPS gets points for practicality and execution, but is yet another service that performs zero email address verification and ends up with idiots that don't know their own email sending me their info.
Would it matter? The address has to be valid and once input it is the unique key. I didn't see any address validation on my end (which is also an option for other forms).
Literally my only complaint is that they ask for first/last name rather than full name. But I wonder if that’s some kind of legal/regulatory requirement in this case.
According to the real time analytics [0], as of 1PM (central time), there are more visitors to COVIDtests.gov (754,000) than to all other federal government websites combined (1.33 million)
In spite of appearances, I'm not sure if they went with existing USPS web resources, at least on the frontend. Looks like custom build using a very typical modern managed AWS services stack. In any case, interesting new state of the art for government, compared, let's say, to the HealthCare.gov launch.
At first I thought it would just be a new SKU in the USPS.com website. But it looks like they rolled something custom, using CloudFront, S3, Route53 (two new domain names, special.usps.com and special-api.usps.com for POSTing the order form JSON to), API Gateway (and probably Lambda and DynamoDB):
Not sure if USDS (U.S. Digital Services) team worked specifically on the covidtests.gov rollout, but USDS was largely borne out of the dev effort that rescued Healthcare.gov:
> The spectacular failure of Healthcare.gov at launch led to the creation of what came to be known as the Tech Surge, a group of Silicon Valley developers who rescued the website from disorganized contractors and bureaucratic mismanagement. That group gave rise to the U.S. Digital Service and, to a lesser extent, 18F, two government agencies now working to improve the state of federal technology
> Not sure if USDS (U.S. Digital Services) team worked specifically on the covidtests.gov rollout
They did[1]:
> The United States Digital Service, which grew out of the effort to rescue healthcare.gov, is also assisting with the launch. "We have the best tech teams across our administration working hard to make this a success, and we think we're well positioned to do that,"
The US Digital Services was formed to prevent exactly that. The main lesson learned was "follow industry best practices instead of just paying a contractor who sounds like they could do it."
Strange thing is this was supposed to go live tomorrow (19th). Though maybe they were going by Guam where it is already 1/19.
Just speculating, because it's using a suite of AWS managed services throughout (CloudFront, S3, Route53, API Gateway) and it's a common design to use Lambda functions with API Gateway. Knowing that this is a very high profile, high traffic site launch, and knowing something of the attitudes of the USDS team, they are likely to want to minimize the number of moving parts and services they have to manage themselves. Also, observing the interaction of the site, which all boils down to a simple HTTP POST of a small JSON blob to the API Gateway endpoint, I am (again just speculating) assuming they are wanting to do minimal processing and just stuff the JSON into a database as quickly as possible. It could be an RDS database, but that still seems like too much administration, and especially since DynamoDB can do highly scalable writes, it just seems like the least surprising choice. I could be wrong, and it would be great to hear the team share on record what's actually under the hood, but given what we can confirm externally by observation plus what we know about the mission and the team, I feel reasonably confident it is all managed services, at least on the frontend component we can interact with.
It appears that omicron almost always returns false negatives with the rapid antigen tests for mild cases.
Since omicron is the vast majority of cases in the wild, and nearly all cases are mild (and the ones that aren't are probably going to get a PCR test anyway), rapid tests are practically next to useless now.
So in that first 380 million, 1 out of 3 is already good for omicron. Hopefully as they work toward more contracts for the remainder of the 1 billion tests, the government will try to take omicron into account.
I just want to point out this is a false assumption.
> So in that first 380 million, 1 out of 3 is already good for omicron.
The text says those 3 companies have been contracted for a combined 380 million tests. It doesn't say equally, or list a number for them. You could have a worst case scenario (I'm being overly dramatic here) of 1 test from Roche, and 379,999,999 combined tests from Abbott and iHealth.
A comment above says that omicron is hard for rapid tests to detect from nasal fluids and that swabbing one's throat leads to much more reliable tests.
Is it? If there's a significant false negative rate, that will mislead people who may otherwise isolate/quarantine based on symptoms. If I feel kinda sick but test negative and need milk, I'm probably going to the store. Same with taking flight that I can't afford to reschedule, or a Dr. office visit, etc etc.
Readit News
In the meantime my children have been sick for a little while, and we got one of them tested and it was negative; however right before they got sick we had interaction with someone who had a positive COVID test shortly thereafter.
I am starting to wonder at the efficacy of the tests and is it that they just can't catch Omnicron as well? Are my kids really just sick with some other random thing despite dozens of people in our social circle getting COVID?
I'm just wondering at this point if there is a reason for the mass testing, because it doesn't really let you know anything except whether or not you have COVID and you'll get better or get worse. I can understand in hospitals or something like that, but in the general population I am failing to understand as it seems Omicron is incredibly virulent, and you can spread before you have symptoms, so what is the end game with the mass testing?
[0] https://www.fda.gov/medical-devices/coronavirus-disease-2019...
[1] (pdf) https://www.fda.gov/media/141570/download
[2] (pdf) https://www.fda.gov/media/142193/download
[3] (pdf) https://www.fda.gov/media/152698/download
[4] (pdf) https://www.fda.gov/media/147494/download
Swab throat too during rapid COVID test, Israel's Health Ministry says
https://www.reuters.com/business/healthcare-pharmaceuticals/...
So what's better, getting your throat spit up your nose, or getting your nose mucus in your throat?
Anecdotal: I was in my 3rd of 4th day of symptoms before I finally came up positive on antigen (at home and at testing sites; the test appeared to be the exact same). PCR showed it from day 0 but the results lag quite a bit especially right now in the US (it took 7-10 days for me to get the PCR back in early Jan).
LFDs are decent for telling you if you're infectious _right now_ - they're much worse than PCRs for telling you if you have covid at all.
Disclaimer: I'm a software engineer, do not take medical advice from me. My partner is a biochemist and I regularly chat with a couple of council covid-response officers, and the above is the explanation I've had from them when I've asked.
You drive into a parking lot, they hand you a test, you self-swab and give it back to them.
There is other testing available - this is just the county-provided testing. It’s well run.
The place I use in Austin is $85 for a same day PCR test.
Ever since the vaccines came out and were implemented in full force, I am wondering what the purpose of testing even is. At least in the US, even the most restrictive locales have no intention of shutting down commerce to limit transmission. I do not know anyone who is not treating this coronavirus like any other, or the flu, or a rhino, or whatever other sicknesses normally go around.
My kids' daycare (or local government, not sure which is responsible) has a crazy policy of shutting down a class for 2 weeks if a covid case is detected. But not if any of the kids are sick with RSV/flu/rhino/noro/rota, etc. The daycare has not had any reports of covid for months. I am pretty sure all the parents are simply choosing to not get their kids tested because who can afford to drop their obligations for 2 weeks and also force all the other parents in class to do the same. Every single kid walking through those doors has a snotty nose, and surely at least 1 has had covid 19 over the past few months.
If you test positive, you're supposed to stay home and isolate so you don't spread COVID.
Tests aren't perfect, since as you said, there seems to be a significant false-negative rate with Omicron, and you can be contagious before you show symptoms, but the tests help prevent people from spreading confirmed cases, and can be helpful to pre-test before you spend time with someone who is at high risk for COVID.
Its not even Omicron. Rapid tests always had something like a 40% false-negative rate or something stupidly high.
------
That's why you're supposed to take two tests if you're asymptomatic, to reduce the chances of a false-negative.
Fortunately, the false-positive rate is less than 5% IIRC, and fairly reliable.
EDIT: Sensitivity and Selectivity isn't really false-negatives or false-positives. But for the purposes of this discussion, its probably good enough. Besides, I'm just an ignorant programmer who doesn't know the difference between them.
Rapid tests give you a result within 15 minutes. PCR tests you gotta ship to a lab, and wait maybe 2 or 3 days before they report back to you.
I think PCR tests (which nominally take 15 minutes) are the best. But we clearly don't have the infrastructure to support it. Every lab doing PCR tests is going at full throttle and yet we still have this 2 or 3 day delay everywhere.
Meanwhile, at-home rapid tests work immediately. They're way less accurate, but its better to have some degree of testing rather than nothing at all.
--------
Lets say you have a workforce consisting of 10 people who are sick with COVID19 right now. PCR tests might catch 9/10 of those people. Rapid tests might only catch 6/10.
But removing those 6-people with Rapid tests in your workforce will slow down the spread. 4-people spreading COVID19 in your workplace is better than 10-people spreading COVID19.
-----
EDIT: More importantly, Rapid tests will catch those 6 people _TODAY_. PCR tests will take ~3 days before they report back, meaning if you're on PCR-tests only, your COVID19 positive people will be around the workforce for 3 days.
If you use both Rapid Tests + PCR tests, you remove those 6 people _TODAY_ from the workforce (meaning 4 people will be spreading COVID19 in your workforce). 3 days later, those 3 other people are removed which is "too late". The Rapid test was in fact more useful.
PCR tests are better for contact tracing, due to their higher reliability. If someone tests positive with PCR tests, you trace their contacts and tell _THOSE_ people to get PCR tests and/or rapid tests, in an attempt to "get ahead" of the disease. (IE: those 3 people who weren't caught by rapid tests were in contact with 15 people total. You now tell those 15 people to rapid test + PCR test).
When my son tested positive, they refused to let us re-test to see if it may have been a false positive. He's 2 and was completely asymptomatic.
Deleted Comment
There are other possibilities like that she had a cold and the 5th test was a false positive or she had a cold and got covid several days later.
PCR tests will report COVID infections that are way below the threshold of transmissibility, and they'll keep reporting positive results long after the infected person is no longer contagious.
"Am I contagious?" is arguably a more useful question to answer than, "Do I have any detectable amount of COVID in my body, at all, using the best laboratory testing processes available?"
[1] https://yourlocalepidemiologist.substack.com/p/antigen-tests...
[1] https://www-umcutrecht-nl.translate.goog/nieuws/corona-zelft...
Covid is much more dangerous for the elderly than the young, even vaccinated. The young are so safe that there have been ~250 total covid deaths in the US age 0-4.
Also, there is growing evidence that the lateral flow tests correlate well with being infectious.
Testing reasons:
For lots of staff (medical, school, factories), to avoid unnecessary prophylactic isolations, particularly for people whose work cannot be done remotely. And for the same group of people, testing daily (were we in a competent country with the test manufacturing to support) could significantly reduce community spread. Daily tests are infeasible to do with pcr if for no other reason than the reporting lag.
For visiting high risk people: reasonably reliable, fast indicator of whether you have covid in order to make near realtime decisions about activities to engage in. Like masks, this will reduce community spread.
My understanding is:
1) Existing rapid tests are not as sensitive to Omicron as previous variants.
2) Negative results on rapid tests are unreliable, especially single negatives. They're more reliable when taken in series, and the instructions for the ones tell you to take two tests, 3 days apart. However, their positive results are very reliable, so if only 75% of the positive people who take one actually test positive, you've still successfully detected 75% of those infections.
PCR tests are far better at detecting infections early.
If you are in the US, flu is a possibility. Unlike 2020-21, the 2021-22 flu season is coming on strong: https://www.cdc.gov/flu/weekly/usmap.htm
This won't help us get a handle on asymptomatic cases - to do that, we'd have to be able to test everyone on a frequent and consistent basis.
This program doesn't even cover one test for everyone in my house, let alone weekly or more frequent testing. And aside from this, I have to spend hours looking for any available rapid tests to buy around here, and if/when I have some, I can't afford to use them on randomized or even weekly testing.
> QuickVue At-Home OTC COVID-19 Test correctly identified 83.5% of positive specimens and 99.2% of negative specimens.
My bet is that it's probably using the same infrastructure and is built by the US Digital Service (https://www.usds.gov/) and/or 18F (https://18f.gsa.gov/)
I mean, for what it's worth, this gets high marks from me for ease of use and seems to be consistent with just about every advice you see surface here about getting your users through an ordering funnel. Of course the fact that the cost is 0 here makes it a simpler problem for them, but given that it is a gov't project I could have envisioned a lot, lot worse.
The USPS also enjoys a monopoly on mailbox deliveries, as well as price protections, in the form of a floor imposed on potential competitors. This doesn't seem particularly admirable.
[1] https://en.wikipedia.org/wiki/Postal_Reorganization_Act
Okay, this is not particularly novel, and the implied comparison to scaling digital services is flawed.
"If this were a startup or private company we would be in awe of their track record."
Again, if a startup operated like USPS they would fold real quick. The USPS has continuously received public money to remain in operation. They are not innovative. The only area where they are expanding is in parcel delivery, and they often turn it over to other partners to do the last-mile pickup and delivery.
However, startups can aspire to the ubiquity of USPS. In many towns, the Post Office is the only happening place and fosters community, and occupies an important place in the hearts of people who rely on it in a big way. Many private logistics providers such as UPS and FedEx do not serve these areas because it's uneconomical for them, and the Post Office is the only logistics provider. For this reason, the USPS deserves taxpayer support.
Deleted Comment
USPS gets points for practicality and execution, but is yet another service that performs zero email address verification and ends up with idiots that don't know their own email sending me their info.
The 18F group is the US Government's "elite" IT-team, including web programming.
-------
If it wasn't 18F, who did it?
Deleted Comment
[0] https://analytics.usa.gov
At first I thought it would just be a new SKU in the USPS.com website. But it looks like they rolled something custom, using CloudFront, S3, Route53 (two new domain names, special.usps.com and special-api.usps.com for POSTing the order form JSON to), API Gateway (and probably Lambda and DynamoDB):
https://docs.google.com/document/d/e/2PACX-1vQP9G1cWiICGOoBA...
https://www.theatlantic.com/technology/archive/2015/07/the-s...
> The spectacular failure of Healthcare.gov at launch led to the creation of what came to be known as the Tech Surge, a group of Silicon Valley developers who rescued the website from disorganized contractors and bureaucratic mismanagement. That group gave rise to the U.S. Digital Service and, to a lesser extent, 18F, two government agencies now working to improve the state of federal technology
They did[1]:
> The United States Digital Service, which grew out of the effort to rescue healthcare.gov, is also assisting with the launch. "We have the best tech teams across our administration working hard to make this a success, and we think we're well positioned to do that,"
[1]: https://www.protocol.com/bulletins/white-house-covid-test-we...
That being said I think they should really really should maybe give some thought to renaming their group.
Strange thing is this was supposed to go live tomorrow (19th). Though maybe they were going by Guam where it is already 1/19.
Curious why would we assume it's running on Lambda & DynamoDB vs any other compute & storage? Previous known gov projects?
Since omicron is the vast majority of cases in the wild, and nearly all cases are mild (and the ones that aren't are probably going to get a PCR test anyway), rapid tests are practically next to useless now.
Three key points from the article:
(1) Rapid tests probably do have a harder time detecting omicron.
(2) Two new rapid tests (from Roche, Siemens), granted EUA in December, "happen" to work "really well" against omicron.
(3) Some of the tests the government will distribute are Roche tests.
Another recent article (https://www.medtechdive.com/news/biden-administration-buy-50...) says the government has contracted with Abbott, iHealth, and Roche for its first 380 million test kits.
So in that first 380 million, 1 out of 3 is already good for omicron. Hopefully as they work toward more contracts for the remainder of the 1 billion tests, the government will try to take omicron into account.
> So in that first 380 million, 1 out of 3 is already good for omicron.
The text says those 3 companies have been contracted for a combined 380 million tests. It doesn't say equally, or list a number for them. You could have a worst case scenario (I'm being overly dramatic here) of 1 test from Roche, and 379,999,999 combined tests from Abbott and iHealth.
An at home test is better than nothing.
Is it? If there's a significant false negative rate, that will mislead people who may otherwise isolate/quarantine based on symptoms. If I feel kinda sick but test negative and need milk, I'm probably going to the store. Same with taking flight that I can't afford to reschedule, or a Dr. office visit, etc etc.