Why LESS Sensitive Tests Might Be Better
Вставка
- Опубліковано 13 січ 2021
- Thanks to Brilliant for sponsoring this video - the first 200 people to click on www.brilliant.org/MinutePhysics will get 20% off a Premium subscription to Brilliant.
This video written & produced in collaboration with Aatish Bhatia, www.aatishb.com
This video is about how cheap, fast, and LESS sensitive rapid antigen tests might be better for screening (& maybe surveillance) than PCR COVID tests due to the nature of contagiousness/infectiveness at various points on the viral load trajectory of symptomatic and asymptomatic COVID sars-COV-2 carriers.
REFERENCES
Thanks to Daniel Larremore for feedback on early versions of this video larremorelab.github.io/
Rapid Antigen Testing:
COVID-19 testing: One size does not fit all. science.sciencemag.org/conten...
Rethinking Covid-19 Test Sensitivity - A Strategy for Containment. www.nejm.org/doi/full/10.1056...
Test sensitivity is secondary to frequency and turnaround time for COVID-19 screening. advances.sciencemag.org/conte...
The effectiveness of population-wide, rapid antigen test based screening in reducing SARS-CoV-2 infection prevalence in Slovakia. (pre-print, not yet peer reviewed) www.medrxiv.org/content/10.11...
Effective Testing and Screening for Covid-19. www.rockefellerfoundation.org...
Brown University & Harvard University modeling of COVID-19 testing shortfall. globalepidemics.org/testing-t...
Fast Coronavirus Tests are coming. media.nature.com/original/mag...
Open letter signed by epidemiologists and infectious disease experts supporting widespread & frequent rapid antigen testing for COVID-19: www.rapidtests.org/expert-letter
More information on various COVID-19 tests: chs.asu.edu/diagnostics-commo...
Field performance and public health response using the BinaxNOW Rapid SARS-CoV-2 antigen detection assay during community-based testing. academic.oup.com/cid/advance-...
Performance of an Antigen-Based Test for Asymptomatic and Symptomatic SARS-CoV-2 Testing at Two University Campuses. www.cdc.gov/mmwr/volumes/69/w...
Asymptomatic Spread:
People without symptoms spread virus in more than half of cases, CDC model finds www.washingtonpost.com/scienc...
(More than half of all) SARS-CoV-2 Transmission From People Without COVID-19 Symptoms. jamanetwork.com/journals/jama...
Three Quarters of People with SARS-CoV-2 Infection are Asymptomatic: Analysis of English Household Survey Data. www.ncbi.nlm.nih.gov/pmc/arti...
Viral Load Curve:
SARS-CoV-2 viral dynamics in acute infections. (pre-print, not yet peer reviewed)
www.medrxiv.org/content/10.11...
Support MinutePhysics on Patreon! / minutephysics
Link to Patreon Supporters: www.minutephysics.com/supporters/
MinutePhysics is on twitter - @minutephysics
And facebook - / minutephysics
Minute Physics provides an energetic and entertaining view of old and new problems in physics -- all in a minute!
Created by Henry Reich - Наука та технологія
I love the fact that you addressed some of the caveats in the end, many people often take everything at face value without questioning the information that didn't make it into the video, adding caveats at the end really helps people realise a 3min video is not all the information
Agreed I honestly took it at face value and immidiatly realized
I love the fact that a smart channel like this attracts the smart af people in the world. And breaks down the facts! BUT MISSES THE 4UCKING POINT!
THIS PLANDEMIC IS ALL FOR THAT MONEY AND CONTROL!
like wtf are these test so expensive when our taxes pay for their development? And the PCR test is not GOOD yet we still need to test everyone! Which everyone seems to have covid. Which is why we need to vaccinate everyone! PCR test was $100+ this taxpayer funded rush for a vaccine but still no discount for us will cost how much?
Oh it’s cost that much cause it needs to be refrigerated..So shipping needs to stored in a specially designed fridge. Plus it expires in a week and if it drops to room temperature it expires too. Oh we have a set tier for vaccination so let’s skip the old people that need it the most and give it to the rich first and medical workers.
Oh less than 10% of medical workers actuality taken the vaccine?
I can go on and on but YOUR 4UCKING PSUEDO INTELLECTUAL MORONS WONT ALLOW THE TRUTH TO SETTLE IN.. that “omfg I was dooped.”
YES YOU BEEN DOOPED I WAS TOO. They don’t GIVE A 4UCK ABOUT YOU!
they don’t they just want that MONEY AND THAT POWER!
Like come on if the whole point was to vaccinate everyone WHY TF ARE WE WASTING MONEY AND TIME TESTING? All this for a supposed plandemic with the chance of death being a 4UCKING MEASLY 0.8%
Really?
Like
WTF
PEOPLE WAKE TF UP!
@@dhananjeyannatarajan5366 ORLY? so please enlighten us on what you took for face value?
@@i992dc I see why you have a problem with intellectuals...
I'll give you this though. That the price of medicine, is a huge problem in general.
But keep in mind, that not every part of the wold has the same political structure, and the same medicine prices as the US. So there's a suggestion for a more productive use, of your anger and energy. But still, every country has a pandemic on their hands. Luckily, in my country, even the nut jobs wear a mask.
I'd suggest you take another look at your own numbers. If the virus had a death rate of 0,8%. You aught to be more concerned with isolation and basic precautions, then you seem to be. But I'll leave that up to you.
@@soul0360 how a as bout you look at the CDC’s yearly death rates for the US. And tell how TF does 2020 the year of the CORONAVIRUS PANDEMIC have less deaths than years prior?
It looks like the corona virus was the cure for the flu death as they seem to have magically gone away.
IF YOURE ABOUT THE FACTS AND THE NUMBERS THAN WHAT DO THE NUMBERS SAY?
It says we be CONvid..
The law of diminishing returns: 90% sensitivity only costs 10% the price.
solution: amend the laws
@@Brindlebrother the laws?? I don't understand
@@Brindlebrother Lol if we could amend basic statistical laws, entropy wouldn't be a problem.
@@badgoogle9938 the law of diminishing returns
@@HDv2b I was replying to Tarskybull, I understand the OP post, I didn't understand the reply
I like how they said physical distancing over social distancing.
God yes
agree
I perfer Stay out of melee range XD
I just don't get it - what puts the 'social' in social distancing anyway??
@@eccentricOrange, en.wikipedia.org/wiki/Proxemics#Human%20distances
Intimate distance (Closer than 460mm) for embracing, touching or whispering
Personal distance (460mm to 1.22m) for interactions among good friends or family
Social distance (1.2m to 3.7m) for interactions among acquaintances
Public distance (further than 3.7m) used for public speaking
It is worth mentioning that while the rapid antigen test is 80-90% as effective as PCR, it's not like taking it multiple times will much improve its screening capacity. If the virus is present at low viral loads, there is no replacement for PCR. Seems to me that if you are interested in visiting an at risk family member or are getting ready for an operation, go for the PCR.
yep, there's no replacement for a PCR test and self isolation while you wait for the result
@STEMPod Leaders solid channel
just don't drink coca-cola before your test
@@xapemanx No, don't have your mouth full of coca-cola before you take a test. Another really good advice for measuring stuff distances: if you want to know how tall you are, don't cut your measuring tape in two pieces. Actively destroying your instruments can indeed cause wrong measurements.
ONLY Partially - but if you are interested in visiting an at risk family member, or anyone else, the rapid test would likely be much better! Why -
1. Because you get results in 10-30 minutes - not in many hours or more likely days later like with PCR, so you know your CURRENT viral load not what it was days ago. So, test right before coming in contact with anyone - before work, before school, before church, before meeting for dinner. You could do this - because it's also very cheap.
2. Testing multiple times is actually the whole point of rapid testing - and multiple tests do indeed improve it's screening capacity.
See - twitter.com/michaelmina_lab/status/1349712137275527168
3. You have to think about what each test actually is showing you - PCR is saying you have - or have PREVIOUSLY had - COVID, on the day you tested which could have been 1-7 days earlier (due to test turn around time). It can and does detect low viral loads before you are infectious - but it also detects low viral loads AFTER you are no longer infectious, and in fact can show positives with just the remaining virus particles after you have cleared the virus. None of that is a great indication if today I am sick or more importantly - get anyone else sick!
4. Rapid tests become more accurate with higher viral loads - so if it's positive, you are MUCH more likely to be contagious, PCR can not/does not differentiate -
5. Due to the expense, scarcity, and slow turn around - the PCR tests are just not the right tool to fight a pandemic - rapid tests could be.
Please see www.rapidtests.org/
I live in a homeless shelter. They give these tests everyday. I have been tested about 6(negative :) tests so far. Kinda blown away by the what they have spent so far. I’m just one of hundreds here.
what happens to the homeless that test positive?
@@xapemanx They go to the gulag
They are spending way too much resources on this... If they would just build more hospitals to pull people through it, we could build a natural immunity.
@@superghost6 Sure, at the expense of killing 2% of the population.
@@superghost6 lol yeah, like building a hospital and all the labor necessary to run it, and equipment like ventilators is cheaper and easier than *tests?*. South Korea and Taiwan are proof that countries can successfully (and without insurmountable cost) test, trace, and isolate the disease down to eradication, which is now also being supplemented with vaccinations to make even easier and cheaper. when people get sick it costs a *lot* of resources, it's much better and cheaper to prevent it in the first place as much as possible.
The rapid tests vary in sensitivity. There is one where sensitivity is 80% when the person is showing symptoms but only 40% in asymptomatic individuals. Specificity was 98% in both cases
Plus, 2% False Positive Rate is likely too high for these tests to be used en masse. The more tests we give, the lower the positivity rate we’d expect to see. For example, many areas are seeing about a 10% positivity rate at the moment with PCR and only tested exposed people. If we start testing everyone, and the positivity rate dropped to 1%, for example, that would mean that a given individual would only be 29% likely to actually have the virus AFTER a positive rapid test, using Bayes Rule.
@@jackbarbey yes thats why testing people without symptoms is shit.
So it's really "The Case for faster, cheaper tests", *despite* their lower sensitivity.
While the title is pretty much clickbait, he does spend a tiny portion of the video explaining why higher sensitivity can be a problem in and of itself. In a world where both kinds of tests took exactly the same amount of money/time, testing masses of people with the interest of finding out if they are contagious would favor using the less sensitive test. Of course, that isn't the situation we're in, and even if it were, that would be the only use case where lower sensitivity would be a benefit. But even though it doesn't make any practical difference to any normal person, it isn't a completely irrelevant consideration for your average epidemiologist.
@@michaelleue7594 He does mention a slight advantage to lower sensitivity (no positives for post-infectious people), but also a major disadvantage (missing pre-infectious people).
So in a world where both where as expensive and fast, you would want to go with the more sensitive one.
I've heard an argument as well that overly sensitive test may mark as infected, people have come into contact with a small amount of viral load, below the threshold required for the virus to infect their body(they might test as positive even though they are not & will not become infected with covid, barring any subsequent viral exposure of course, because the tests are detecting a very small amount of virus in the nasal passage that is too low for the average person to become infected). There is no proof that the pcr test have been set to be too sensitive, this is merely a hypothesis about how false positives could potentially be higher than expected.
@@piranha031091 He explains why higher specificity would be beneficial. Although it is linked with lower sensitivity it is not the same thing and unfortunately, this video fails to explain the difference.
@@piranha031091 I'd guess pre-infective people could still be indirectly captured under "should quarantine" by the people who are doing the tracking, since they likely have been recently exposed by someone who's infectious. So maybe the main advantage of higher sensitivity would be if the tests were done completely at random, but once there's this tracking of contagion going on, you can kind of assume some "likely positives" for close contacts and achieve a similar containment of contagion. I guess.
Worth mentioning that in medicine sensitivity refers to the "true positive" rate of a test, ie of how many people with the disease will the test come back positive.
This means that RT-PCR tests are actually not very sensitive at all, rather their advantage is the specificity, which is the "true negative" rate.
I know this is pedantic and jargon, but medical jargon is nevertheless important.
Which country do you come from? I haven't heard this use in medicine in Germany before and haven't seen it in academics.
I don't think it's pedantic. It is hugely important that tests are able to confirm that people are not infected. Using tests with a lower sensitivity will lead to infected people believing they are safe because of a false negative, likely leading to higher risk taking and increased spread. If tests have too low of a sensitivity they could increase spread rather than reduce it due to the behavioural effects of false negatives
Also: In the medical sense (and in the COVID case), false positives are way less bad than false negatives. Testing someone positive means they have to go to quarantine. If they are not actually infectious, this might lead to some smaller monetary issues, but that's that. Testing someone negative means they can continue to walk around publicly. If they are actually sick but tested negative, that's a real issue. Because they will be super likely to infect others then.
The next point is: Some tests have a super high false positive rate. That's due to the fact that (while COVID has infected a lot of people by now), there are still way more healthy people than sick people. So if you test a thousand people, have two positive results and only one of them is legit, then you still have a false positive factor of 2. One person is sick, two persons have been tested positive -> 100% false positive rate. This means that a high false positive rate doesn't mean that the test is bad. It still had the correct results for 999 people. This value "luckily" improves with the amount of people who are sick at the same time. I say "luckily", because it's obviously not good when a lot of people are infected. It's just "lucky" for science, as it means that tests still work as intended.
Thanks! I was confused as well! Don't think its pedantic, it's about making sure we are all taking about the same thing.
@@Sandsack2311 en.wikipedia.org/wiki/Sensitivity_and_specificity
PCR test takes about 8 hours lab time to be processed. It's not days slow, unless the lab is overloaded and samples needs days to be transported.
PCR sensitivity is definitely not a problem. You can exploit it to test groups, instead individuals. You mix the samples of 5-10 people into a single test sample. If the result is negative, then nobody in the group is infectious. If the result is positive, then you have to test the individual samples.
Which is why I wish they did double swabs in the united states to allow bulk testing. This would make the effective cost of PCR per person much closer to antigen but with better accuracy.
@@trepidati0n533 doesn`t mixing samples reduce the effective sensitivy? I would assume that when I mix 5 samples, that it has at least some effect on the sensitivy (not dividing it by 5, but at least a measurable effect), as there is more random genomic material in general that is using reactants and more genetic material creating more random effects ("noise").
@@sarowie No, because it is scarily sensitive. A single virus fragment in the sample is enough. And for this initial scanning, you'd adjust the test a little to make it more sensitive, so you are 100% sure to catch a positive at the risk of a false positive, which would mean only having to test the second samples unnecessarily.
But even if one is positive, the other 8-9 people can become positive even if they were earlier negative and safe
@@housesuits What do you mean?
What you have written could happen only if you have time inverted swabs, like the inverted objects in SciFi movie "Tenet".
One small note from me as a citizen of Slovakia, which study was mentioned here. Implementation in Slovakia was done in such way, that there were really hard punishments for people who decided not to join nation wide screening tests, therefore lot of people got tested and were willing to even wait hours in queue. Problem is that these queues didn't followed social distancing since there was literally not sufficient space for queue following social distancing. Therefore after this nation wide screening rate infected people rise dramatically. I personally got infected between 2 rounds (was negative after first screening but positive after second one, yet following all precautions)
i love how you masterfully condense important information into easily digestible chunks. great job and thank you
In Australia we get pcr test results back within a day
Here in the states, we failed to roll out test in any meaningful way. This is likely why we are also failing at getting the vaccines out in any organized manor.
In the UK it's also meant to be a day, almost always within 2. It was a scandal when most results were taking longer than a day to come back
Same in France where I live
@@reddragon3132 I can be within a few hours (I went in at 7am and knew by midnight), the delay isn't in the processing it's the wait to start processing.
Not during the 2nd wave in Melbourne.
The quick turnaround time is a function of "Go early Go hard strategy". That is, when, the capacity availability for PCR test can cope with the demand. In US, it is too late, the capacity availability is now overloaded and they even talk about "pooled PCR test" - sad case really.
Literally amazed that you provided caveats in the end and how you provide the information by graph.
I adore that music
Dr. Schroeder Silent City
one of the best tracks to ever exist
Thank you for sharing!
Don’t mind me. I’m just gonna thumbs up this comment.
Thanks for the kind words :)
Where can I find it?
@@Coryn02 Soundcloud is where I go to listen to it. Just search
silent city (minute physics)
into the search bar for it, and you should find it
Thanks for the sauce
for countries where covid is out of control this makes some sense, in countries like Australia, China, NZ, Singapore etc. the more sensitive test is essential in the screening to prevent foreign case influx in lieu of local transmission.
I think it’s still a huge advantage to identify infected people in the very late stages since you could then test or quarantine their close contacts and catch up the chain that was kicked off by that person
Exacly, there might not be a need to quarantine that person (if they are no longer infectious) but their contacts can still be tested and isolated (and if positive, also their contacts).
So in other words, the binary value isn't all we need - the quantity of viral RNA (and trend thereof) also matters.
agreed. And if your are more than 10 past the onset of symptoms, most jurisdictions won't quarantine you because you are past the infectious period
But a large-scale testing spree would quite likely catch those close contacts anyway. In particular when you'd PCR test these contacts at least five days later, so perhaps ten days after the first person maximal infectiousness
To a point, but it probably doesn't seem very advantageous to all the people in cohort 2, who were infected because cohort 1 couldn't remember who all was in the grocery store after they caught it from patient 0, who is over a week past their infectious stage and just tested positive.
it only makes sense to me to implement antigen testing when the infection rate is very high. However, once it is not as prevalent in the community then it makes more sense to use pcr.
yes- but no. you forgot about the asymptomatic infections. if you don't catch those it'll spread unnoticed
Which source is related to the "Increase of COVID testing needed" per country?
Really thanks for this video! Currently Im working with a research group in Imperial College London to develop fast graphene field effect transistor lateral tests with sensitivity greater than 90% and results come within few minutes (reusable*). This video will help us to build stronger arguments for GFET lateral tests! Many Thanks.
So great!!! I wish we had this video 6 months ago!
PCR test here in UAE costs about 25 $ only and the test results come before 24 hours
So they are the best indicator isolation and treatment
While I agree with the premise of this video, and love the explanation, here in NSW, Australia, we have routinely been getting our PCR results back within 12 hours. This has made in invaluable in knowing whether someone is infected or not in a timely manner. Also, mad props to our government for making it free. Socialised healthcare is the best healthcare.
Very good video. The caveat part was especially important, I initially wanted to write a question about the rate of false positives.
I like that you addressed this topic because I have seen lots of (mostly anecdotal) evidence that the PCR is unnecessarily sensitive and so, like you said, people are being told to quarantine when they don't really need to. Lots of people, myself included, have been told by employers that we can't return to work after testing positive until we test negative but the high sensitivity of the test means that is completely unfeasible.
In the UK if you've tested positive on a pcr then you shouldn't have another pcr for 90days (most likely for sensitivity issues mentioned), but the lfd (rapid test) will show if you are infectious
Extra Caveat, there’s also the issue of public trust. A person who finds out they got a false positive or false negative will be less likely to trust testing in the future. Even just being aware of the lower accuracy might mean they shirk testing because of the uncertainty of the test. Broadly speaking getting the public to trust epidemiological experts is important both for this pandemic and future pandemics.
Well, we've seen how that turned out.
I tested positive a month after I was symptomatic and I had a negative quick test a week before
So the virus’ infection was already on its way out. Hope you’re doing well now after everything!
A negative quick test (aka a negative rapid antigen test) doesn’t mean you are not infected. It means you don’t know. The quick test false negative rate averages about 30% but can be as high as 50% depending on the specific test kit, timing and methodology used. In other words, one in two infected people might incorrectly be told they don’t have the virus. Multiple quick tests are called for to avoid false negatives.
www.fda.gov/medical-devices/letters-health-care-providers/potential-false-positive-results-antigen-tests-rapid-detection-sars-cov-2-letter-clinical-laboratory
____________________________________-
“One of the kits was evaluated … the test detected infections between 51% and 84% of the time”.
www.nature.com/articles/d41586-020-02661-2
So you didn't have a cough or still loss of smell that would mean you still had symptoms.. and it takes about a month for you to really get symptoms free.. because you can get a reinsurangs of it happens about 25 days after first symptoms show up.. documented on many cases
@@robrocco5420 Or - it goes away quickly. Also documented in many many cases. However - There are no confirmed asymptomatic cases. There are lots of people that test positive with no symptoms, but we find that those are are a result of false positives. Out of 10 million tests - 300 "asymptomatic" cases; easily accounted for with margin of error in tests.
Keep thinking they're telling you the truth ;)
@@jamesbond1231 there was an asymptomatic case if you remember in the early day's.. a woman from China going back home from visiting wuhan.. she developed symptoms and gave it to her husband.. they insisted having their child tested and but was running around and playing the doctors said he didn't have covid19 but parents still wanted to have their child tested so they so the doctors took child to have CT scan.. and yep scan showed covid19 in the lung's... this was the first asymptomatic patient in the USA.. so yes asymptomatic people do exist with covid19
I'm glad you mentioned Caveat #3 as it is a common assumption that viral load is the same as infectiousness. This is most frustrating when seen in research papers and seminars
0:25 "In the fall of 2020", For a moment there I thought minute physics was talking about something else.
i don't get it
@@adithyan9263 "Fall" means both falling down and the autumnal season
@@pursuitsoflife.6119 thanks
Interesting points Henry! To play devil’s advocate, how would we ensure trust in the tests if they were less accurate? If the general public can distrust the result, how would we better deal with non-compliance?
Also, I love how you talk through caveats which I feel counterintuitively increases trust in your video rather than the opposite. More science channels should do this! It makes us understand that there are always downsides & that the world is not black & white but instead nuanced.
How can one be in compliance with incompetence?
How do we deal with non-complicance to current regulation and current test rates? In my expirience, law enforcment has a very hard time to actually enforce current regulations, as the reality is fuzzy and two officers have a hard time to "overwelm" a not complient group of 10 people.
A less accurate test testing multiple times would yield a more accurate an answer.
Like with pregnancy tests. Taking 2 or 3 should be good
Here in California, you can get the slow tests for free, but the rapid tests can cost around $120...
Gov should pay 100 percent of both.
The big advantage to rapid testing is the effect on the R value. All measures taken have an effect. If rapid testing is widely adopted it would enable us to relax other, more disruptive, measures.
I work in a molecular biology lab specifically built for testing COVID samples, and one thing it’s taught me is the importance in healthcare and global health of having many different ways of accomplishing the same goal. We have strict SOPs but-within those-our approaches have to be highly flexible to accommodate the many types of clients we serve and fulfill each of their individual needs. PCR and antigen tests both have gaps that the other helps fill, and both are vital for different outcomes. Just because vaccines are developed does not mean research is over-developing a vaccine stable at room temperature will allow for more distribution in lower-income communities or developing nations, whereas a vaccine that’s easier to develop and roll out but needs refrigeration has its benefits for us right here right now. This video really helps solidify that notion to me.
It might have been helpful to people's understanding of the discussions around this to distinguish between the statistical notion of sensitivity, which is what is usually used by medical professionals, and the notion of sensitivity that you are referring to. For sure the two are highly related in practical terms but they are also quite different concepts.
Madrid showed how effective rapid antigen tests are in comparison with PCR tests during the second wave. While the rest of Spain was doing community screenings with PCR tests, Madrid did it with rapid antigen tests. Madrid reduced drastically the number of new cases in a couple weeks, while the numbers for the rest of Spain were still growing.
Love the content as always, however I need to point out something regarding the voiceover quality. Your microphone/post processing chain really emphasises sibilance, which makes the video a bit hard to listen to sometimes. Most video editing software have built in De-essers, which would help the problem. Minor thing, but I think it's important. Listening on DT-770 headphones.
Hey I work in a lab, and the techs say the various testing machines take 8-12 hours to run about 100 samples. They 2-5 day discrepancy comes from the laboratory's load. They say it takes 2-5 days for all the administrative stuff to take place, as well as getting up to that place in the queue of samples to be run.
One thing I'd worry about with less sensitive tests is that people who get a false negative might take less precautions than someone who didn't get tested at all. Even if they know the test isn't always 100% accurate, they might still get a false sense of security.
The sad thing is that Michael Mina already explained this on a podcast with Virologist on July 16th. This week in Virology: episode 640 www.microbe.tv/twiv/twiv-640/
Nothing has come of it. Which is painful because if you read his actual paper, it is obvious that even if it works just half as well as promised - there's always some real-world issues, such as compliance, mistakes - it would be tremendously effective. It could have stopped the second waves all over Europe and North America.
The tests were there already in June, but too many medical professionals didn't understand what their use-case was, and bad-mouthed these majorly useful tests, and they still have to overcome the stigma of being a "bad" test, while it is just the right tool for the job, and ridiculously cheap compared to the economic damage and the costs of stimulus packages world-wide.
It's nice that you post this video, I think it will come in handy for the next pandemic...
Great video, I do want to add though that there is a benefit to catching people on the late end of their infectiousness. It allows for contact tracing to determine who else was exposed while they were more infectious.
Interesting and still timely video one year later. I actually assumed this was new until I saw the date you released it! Might have been nice to use the language of sensitivity and specificity, but not really an issue. Thanks for the great content!
1:45 in Germany when you got tested you are quaranteened until the test comes back. So there is no chance to "miss the peak".
@Kartoffelbrei
Yea germany seems to forget that stupid selfish idiots exist.
@Kartoffelbrei
I dont know. But I have a wierd feeling you are going to tell me the answer.
@@ccox7198
"Threaten people with forcible imprisonment if they submit. Yeah, good way to catch flies."
Not everyone sees this as a threat. You just got the wrong basic thinking how to act as a society.
@@ccox7198
I can see that you are angry because you cant follow the rules with your logic. Im sorry, but I think you made up your mind.
@@ccox7198
"What's the point of taking a test and being forced to quarantine while waiting for results?"
So you dont have the chance to miss a peak.
"The second you leave after getting a negative you'll need another test."
Thats not true. We usually only test people with symptoms or first grade contacts of those people within a certain timeframe. We dont test second grade contacts - So if you met someone who met a person who was positive at the time, you wont be tested.
I guess "The case for faster, less expensive tests" wasn't click baity enough...
Yes, I probably would not have clicked
This is common with many channels nowadays, I believe the idea is a less click bait, more informative title when the video is first published to bring in your subscriber base and then a click bait, but in this case still true, title to bring in the masses to the content
To be fair, there was an actual point for having lower sensitivity in the video, regardless of price: Not testing positive those people who are no longer infectious, and thus should not be quarantined. Yes, that's at the cost of not finding pre-infectious people, for which the lower price is important, but the lower price is not important for that first point.
Maybe we have different definitions of clickbait (or at least clickbait that needs to be called out), but for me a clickbait title has nothing to do with the actual content of the video (like those "number 7 will blow your mind" when there's nothing special about number 7, to give an example). And this video did in fact answer the question it posed in its title, making it an interesting title, but not a clickbaity one, in my opinion. Other opinions are available 😉
The sensitivity of the test is certainly the main topic here, you can’t tell me that it doesn’t at least seem unintuitive at first that a less sensitive test could have pros.
that graph/graphic at 3.20 is fantastic
Great video all round! I'll be sharing it with friends if the topic comes up again.
Unfortunately though, if it even takes MinutePhysics....several minutes to explain this nuanced topic, it'll be difficult to get it across to the general public :P.
The growth of your channel is amazing and validated by every upload!
I'm extremely grateful for all your videos because it improves the quality of information available to everyone on the internet & is fun to watch!
Thank you so much for another amazing video!
This is one of the oldest educational channels... i wouldn't say the growth is not as fast as i would call "thats fast growth"
@@megaeliminator3260 Yeah, this growth being fast is.... really questionable. I think, he(henry) lost some portion of his fan base after _that_ footnotes video where he says something political(?).... like "men are overpaid" or something... But, overall this channel is quite good!
That's the difference of epidemiology and conventional medicine.
Australia makes PCR tests free to the patient, and they are often returned within about 5 hours. Some infections still slip through, because not everyone with symptoms goes for testing, but this would not be improved by switching tests. Reducing the barriers is crucial, but we also need the PCR sensitivity to catch and trace before clusters get out of hand.
The point of the video is that the sensitivity of the tests doesn't actually matter that much if it's offset by testing more people more often. In Australia the tests may be free to the patient and returned quickly, but how many people are being tested and how often? It's not practical for the labs or the country's budget to be testing large groups of people continuously with PCR, so only suspected cases (symptomatic or close contact to known infected) are being tested. That greatly limits the ability to catch new cases coming in. Regardless of how sensitive a test is, failing to administer a test and assuming you are not sick is a 100% false negative rate. If tests can be administered much more frequently - including to people without specific risks - then the likelihood of failing to test a sick person is reduced significantly. If they are sick, then the sensitivity of the test matters, but the differences in sensitivity aren't that large. A 90% sensitive test gives false negatives to 10% of the people who are sick, which means (assuming everyone is tested regularly), you miss 10% of the sick people. If only those with symptoms are being tested (~50%), then even with a 100% sensitive test, you are missing 50% of the sick people. That's why gaining much more testing is worth it for a small reduction in sensitivity - the main contributor to sick people not being identified is that they aren't tested in the first place.
@@reverse_engineered I understand all of that. I did not miss the point. What I'm saying is that in Australia the primary limit on the number of people getting tested is neither financial nor temporal. The limit is because some people still consider getting (any) test to be somewhat inconvenient, and know that there is an extremely low chance of a positive. Therefore switching to cheaper faster tests would not increase the testing quantities or frequencies much. In answer to your specific questions, we are currently doing 2000 tests per million per day, and for a 0.02% positivity rate, that's pretty solid compared to other countries, with only Mongolia and Singapore looking better at the moment. ourworldindata.org/grapher/covid-19-daily-tests-vs-daily-new-confirmed-cases-per-million?tab=chart&stackMode=absolute&time=latest&country=®ion=World
I like how this is based on the current situation, but simultaneously explains how it would be if the variables were different
So this is where my ASAP science went.
I think that MinutePhysics can really spread awareness about Covid with his channel, which he is doing.
Except that the people who need to hear it won't watch it or believe it.
@yoza L Were you one of the ones breaking lockdown and whining about the vaccine and tests and government conspiracies?
@yoza L One of many what? Trolls?
@@greenisnotacreativecolour oh my gosh, I'm so glad I can't beelie these people won't even wear a mask, don't they want peopel to be safe. I think the government should take them all away so we can finallyget over this virus!
Well, I'm SICK of Covid. The People make a Huge Deal out of Covid as if it's the Bubonic Plague and it's making my Life a Living Hell and the Consequences of this Fear created the conditions for Chaos that will result in many Deaths and Suffering in the long Run.
So sick of this insanity. I even see people shooting at each other and running through High-Ways in peak hour while wearing covid masks -_- What the fuk? Covid is more scary than Death apparently.
I was confused why the information and tactics you were talking about were so outdated until I saw that the video is literally 1 year old.
The new way of testing is really good! First Antigen-Tests and if those are positive a PCR!
That's why the company I work for also produces special screening tests for qPCR that test multiple people simultaneously. It can't tell you the specific person who is sick, but if you need that you can retest just those people individually.
I have more trust in minutephysics to correctly convey information from verified sources than any actual mainstream news media :')
Exept the Onion news website
PCR test result come back within 24 h in the Netherlands, so then the second point is not really valid.
we can't all be blessed enough to live in real countries unfortunately
It was a problem for a while though.
Yeah, in Spain and Germany they say that you will have your results in the next 24h, but I always got the PCR results back (could see them online) in the afternoon of the same day.
In Brazil, people I know tell me that the get the results only a few days after the test.
Interesting! Here in Australia, the word is that the reagents for the rapid antigen test are in very short supply, and so they’re only used in very certain circumstances. Given that we have public healthcare(and the fact that governments have committed to making tests free even at private clinics), the cost isn’t so much of a concern and as we’re doing quite well, there’s not too much of a load on the labs. Most people will get the results from PCR either the same day or next day. While waiting for the result, you must self-isolate at home apart from family, and if you’ve been at a location of specific concern, continue isolating for 14 days regardless of result.
The lab I was tested at in Arkansas had a rapid PCR. Only place to do such a test in Arkansas I believe got the benifits of a pcr and got the results in 15 minutes.
So the issue is not that it's too sensitive, it's that it's too slow. Gotcha
and expensive, which is, unfortunately, a valid reason given an adequate perspective
Don't do a game where whenever you hear "test" you take a shot
Why would I need that many vaccines?
@@gammarayneutrino8413 a shot of vodka or some drink, not vaccine
Thanks for including that Slovakia shoutout!
With the omicron wave everyone is suddenly talking about Rapid tests again (and they are sold out everywhere) and I happened to find your video! Great information thanks for the details, and helps me understand how we should best try to mitigate this.
1:12 Kind of true, but anyone who has done qPCR knows this is price gouging
Me,
Sees minutephysics video
Also me: now that's a name I haven't heard in a long time
True😭😭😭😭
We got minutephysics then Bill Wurtz, what a time to be alive
I live in Slovakia where we had two nationwide screenings done with antigen tests. The main problem was that people who tested negative had exemptions from quarantine which would make sense for pcr tests but not for antigen as there is still decent chance of getting false negative. This resulted in Slovakia being one of the worst countries just couple weeks after the testing.
We use PCR here in NZ, results turn around time is 48hrs.
Have drive in stations where one can walkin or drive in to get tested.
It works if the people of the country actually work together and follow lockdown rules etc.
We did our 3 month stint and now we've been enjoying life pretty much back to normal.
A PCR only takes 6h, but it takes time to write those reports.
As well as transporting the samples, and those samples "waiting in line" at the lab to be tested. Then that report needs to get back to the doctor's office, and _they_ need to open the reports and call everyone while still scrambling to see all the patients (at least that's how it works in my backwater country of Germany... At least the test is free 😉)
I don’t think the reports and transport are really that big a deal. If they were, it’d apply to rapid tests too. I’ve had 4(?) rapid tests done so far (required for college) and I got results for all of them in less than 36 hours
@@ehsper392
The tests themselves only takes 6h for PCR (from heat treatment, taq enzyme, centrifuge and gel running) and 15min for rapid antibody tests. I`ve done a few PCR test myself so I know the time.
The problem is you have massive amount of testing to do and once the results get out, the info has to be relayed accurately to whoever it concerns. That's why it is slow.
Please list your source for the claim that ~50% of infections are from asymptomatic, not mentioned but also presymtomatic, carriers?
Thank you. This claim is definitely worthy of a source.
www.cnbc.com/2020/06/10/dr-anthony-fauci-says-whos-remark-on-asymptomatic-coronavirus-spread-was-not-correct.html
To be clear this is June, but asymptomatic spread has been repeatedly described as not being the main driver of infections. Even if it’s part of the spread, 50 percent seems high.
These studies claim that asymptomatic transmission is very weak to nil!
www.nature.com/articles/s41467-020-19802-w - Post-lockdown SARS-CoV-2 nucleic acid screening in nearly ten million residents of Wuhan, China
pubmed.ncbi.nlm.nih.gov/32513410/
The cnbc article seems more of opinon and doesn't list sources. I've read the nature article which suggests asymptomatic spread is not a primary vector. I guess you can chaulk this up to no one is sure. So seems irresponsible to make the ~50% claim in this video.
At my university (University of Illinois) we have a saliva PCR test from which we usually get same-day results. The cost per test is somewhere around $15. As such, the entire student body tests multiple times a week and that seems to be doing a good job at quelling outbreaks.
A video after such a long time! Joy!
He forgot to add "cheaper and faster" in the title.
It's a sad fact that non-intuitive titles get more clicks in this strange world.
Cheaper and faster would make people say “Well duh, they’re cheaper and faster, those are already good things” and not watch the video. They’d miss out on the important message of why the less sensitive tests are actually important
The pcr testing was never intended to be used the way it is. That comes from the creators of the test themselves.
Incorrect. Kary Mullis is the inventor of PCR and said PCR is not a test, which is correct. He is not the inventor of qPCR, which was nvented by Russell Higuchi and intended and great as a test and a distinct separate invention. Stop spreading misinformation.
You do realize PCR is literally just amplification in a thermal cycler. They send it out for sequencing after and see if it matches target viral sequence. The question to answer is only ever does this sample contain SARS-CoV-2 yes or no.
I appreciate these deeper level explanations. Thank you
Very USA centric! In Australia our PCR tests are much cheaper (less profit for companies) and results are known between 24 - 48 hours. So it's about the same price (we have high admin and low process costs), about the same speed (factoring dissemination of results), and still much more sensitive and specific.
The fact that Australia tests symptomatic and exposed people (as ideally everyone would), also reduces the relative effectiveness of low specific tests for screening.
Amazing video as always ! :) I love how you mention the caveats like precision vs accuracy and have a full analysis of the situation. PCR tests are made for one thing, rapid tests for a slightly different thing. Let's use both wisely and help ourselves until we get enough vaccinations out.
Stay safe all
I freaking love the soundtrack!
It actually gives me the same vibes as one of the songs from "Rising of the Shield Hero" but I digress!
I loved it, too. Very calming while speaking on a series subject.
I love the vibe it gives.
I dont think its a choice between the two. It makes sense to use PCR tests where people have symptoms or have had known exposures (Here in Australia the results are generally known within 24 hours, during which time all tested people must isolate while they wait). The Rapid test should be used for wider screening of asymptomatic people, such as workplaces undertaking daily screening of staff, or passengers before they get on planes.
Finally, a plausible explanation for my case where my early PCR came out negative due to an early contact case whereas my daily antigen caught my infection (which ended up coming 2 days later from another source). I can attest to the fact that frequent antigen testing somehow ends up being more efficient than sparse sensitive PCR testing.
In India PCR tests costs around 30$ with results in 24hrs....
Okay, but I bet in India the other tests are also cheaper than in the US. I know Vietnam developed a test that costs about 1$.
@@mirror452 Yes. A few firms have made Rapid tests and PCR tests cheaper than what government labs charge....
@@keyurunadkat in Switzerland you can test for free with a pcr and get a result in 15 minutes
Joke of the day:
Why did the doofus get fired from the M&M factory?
*Because he threw away the W's!*
So there are two key takeaways:
- highly sensitive PCR tests are underspecific wrt actual contagion hazard
- high frequency can offset sensitivity issues
Also, I wonder if we could have the test center return the Bayes factor instead of a binary positive/negative, and let the community substitute in its own threshold depending on purpose. Having a number would also enable a whole bunch of purposes for which PCR tests are woefully unsuitable
Great video. @minutephysics Just a small note that should be mentioned, PCR tests can also measure viral load based on the number of cycles needed to detect the virus. This means that the information gathered can also be used to reduce the sensitivity of the test. Nevertheless as you say, if we consider the price and time it takes to get the results there is still a good case for going with more rapid tests.
Damn, this is the first time I have seen my country (India) do better in a statistic than the US.
Lol.
Very interesting. But I have to play at 0.75% speed to hear normal speech.
Too slow
If you haven't done one already, could you do a video on how time travel would overlap with alternate universes? Like, if you do something different do you switch universes? Is another universe created or duplicated? I don't know how I started thinking about this, but it's had me curious.
Really good content and a well-reasoned case.
This would have been a great video for governments everywhere in March 2020
When there was no cheap quick test?
The background music is so good!
Yea would really like to know the artist, sounds like great tunes to work to.
@@ayuminor The artist is Nathaniel Schroeder. The song in this video is called Silent City
@@adisonlay8287 Thank you very much! :D
Mass testing here in Austria is done exactly like that: antibody tests that need only around 15 minutes, and if you're positive you do a PCR test afterwards to confirm the result. Around half of antibody positives are also positive on the PCR, as it is less sensitive and you can also be positive post-covid, but you get your preliminary result directly there.
So I'm subscribed and I look at my subscriptions many times a day, yet this video hasn't appeared until three days after
"its expensive and slow"
As an Australian I have to disagree 8h and done in batches so cheap.
And this is PCR tests?
Well Mr NotAsian Australian, I'm a New Zealand veterinarian. And I can tell you that a PCR test is exponentially more time consuming, labour intensive, resource consuming than a simple patient side snap-test (the less sensitive test).
PCR - ain't nobody got time for that!
MinutePhysics might need to educate some politicians worldwide.
8 hours is can't really compete with 20 minutes for unscheduled testing.
@@fredecarlslund5407 Yes
@@benedict6962 Only if time is the only measurement factor.
Do a video on the lockdowns effectiveness that would be really interesting
They sadly won't let you question that stuff with actual science as a normal person.
If anything, what he said here today made more sense in making me understand testing than the government did in 10 months...
China and Australia have shown the obvious - if people stay inside, disease doesn't spread. The kind of research that would need to be done to evaluate the effect of lockdowns in the real world are too complex for a youtube channel - things like the effect of the informal economy, public messaging, political disunity, etc.
@@BlackGateofMordor the more we say information is above people to discuss, the more you must believe that it is never even meant to be understood by the common person.
The ultranuclear option seems to be chasing a dream. Because you'd have to immediately shutoff peoples ability to go anywhere if you wanted draconian to work. If everyone is going to the supermarket because that's the only place they can legally go, then they will be packed and spread the virus even with PPE, because of the sheer number of people in that one location.
The fact we are even debating who should be doing research today is how screwed we are as the human race.
@@AshnSilvercorp Nobody ever said you can't discuss this
@@AshnSilvercorp I didn't say they shouldn't, I'm saying they can't. What youtube channel has the enormous resources to pull off all the research needed for proper evaluation of lockdowns? Papers on the individual parts I mentioned, and many more, will launch the doctorate careers of many. A small group of papers suitable for even a highly educated youtuber to use as sources for the topic won't exist for many years. Providing their uninformed opinion isn't suitable.
where I live they are using both. You can go to the drop in clinic and get the basic rapid test whenever you want. If you have had contact you get 1 basic rapid test asap and then a more accurate test a few days later (whatever the first result). I live in italy.
Btw I would think that from the beginning a testing strategy with many layers would be best. It would be good to have a very fast test that doesn't have high specificity (so it doesn't distinguish much between different viral infections) and might deliver false positives as the first layer. These would include testing for particular symptoms, although that's not enough. Those who are positive here would move onto another test, which can be the "rapid" antigen test, which would be the second layer. And they could be tested more than once as well. Those who go on to develop more serious symptoms or come into contact with sensitive groups can be monitored with the PCR test, which would be the third layer... and such a strategy can be modified by adding more testing strategies as well.
But it doesn't make sense to have a single one-size-fits-all test that's itself scarce and pricey. Quick and cheap and more widespread tests are better for managing a pandemic than slow and expansive tests.
Let's keep in mind that PCR testing will not give a good epidemiological picture as well, since the tested sample is not random and also it shouldn't be random. People get tested if they have symptoms, or think they have been exposed, or have to travel, or can pay more money for a pcr test, and local government policies and the availability of tests significantly changes testing... which is for example why the positive test rate cannot be used to compare different countries with different testing strategies.
Quick, tell the people in charge! Oh, wait they don't care.
But at least I had fun watching this video.
@test test I don't think so
lol yeah i'm not from the states so there's that. if your expecting change with the rotation of an administration ... please don't hold your breath
@test test
Lol I wish
Fauci is happy again. No Trump muzzle. Biden could make it worse?...If he does nothing at all. He has done better than Trump. Malaria pills, Ingesting disinfectant, Trump was a menace.
I think, that a much less confusing way to say the same thing would be "we need cheaper, more specific tests" rather than "we need less sensitive tests".
After all, you can make a "less Sensitive" test by just taking a PCR test and if you test positive rolling a die and disregarding the positive PCR result of you roll a 6.
This is really amazing! I have been wondering why some people that have already recovered (more than 14 days have passed since their pcr tests came out) from covid-19 still got positive results, and in some cases, even after the 21st day.
What makes you choose between bass improv vs something else as music for your videos? Bass for the win!
This should totally not be a controversial video, right?
- __ -
It actually isn't, considering the like ratio is high. Even in my own life, pretty much everyone I spoke to agreed that the Covid number are inflated and misleading.
@@Miranox2 I know 3 people near me that have Covid...out of...thousands of people. Numbers look VERY inflated to me. And I know some people that confirmed that Labs here in Florida are faking numbers or cheating. There are even Web Sites to get a fake covid certificate to skip work. I have 2 friends using that to live off the government for the past months.
I also know of 3 people that died from Covid. Not close tho.
Either way, I think we all gonna get this or we already got it. Nobody is doing the tests to see if we already got it so... numbers can be inflated endlessly.
@@joshanonline There might be some inflation happening in the US but comparing the ratio of (positive/tested) for the new cases we see that it's pretty similar to countries in western Europe, so either all are inflating or the inflating is somewhat insignificant
nvm just realized that you can inflate both positives and tested, but don't you guys have an open database for these things to check though?
@@jinjunliu2401 It's not just the US. I remember some testing labs in France that got busted for inflating the numbers by doing improper testing.
I love the Physics in the video.
JK Henry, I love whatever you make.
Long over, but what also helps to surpress an outbreak (not relevent once a epidemic is in full swing) is PCR testing with pooling.
You can mix many different samples and test only once. Because PCR is so sensitive, you get a positive result, even when only 1/100 is infected. And only when you get a positive result, you go, and test every single of the 100 samples.
Like this, you can catch (and trace) a very few infections in a very large group of people.
But as shown, this logic breaks down, as soon as you get many infections.
Not every method is as relevent in any situation. It changes depending on the state of an epidemic.
The biggest problem with this approach is behavioral. If people get a test and it comes back negative, some portion (possibly a majority) will take that to mean that they are not infected and will then go and drop all precautions, *even if they were initially in contact with someone who had Covid* . This is at least what I've seen in the areas in the US where rapid antigen testing is used. This is amplified in cases where employers require a positive test to give the exposed employee time off work.