Skip to content

What about this idea of rapid antigen testing?

So, there’s this idea going around that seems to make sense, but then again if it makes so much sense I wonder why they’re not doing it already.

Here’s the background. A blog commenter pointed me to this op-ed from mid-November by Michael Mina, an epidemiologist and immunologist who wrote:

Widespread and frequent rapid antigen testing (public health screening to suppress outbreaks) is the best possible tool we have at our disposal today—and we are not using it.

It would significantly reduce the spread of the virus without having to shut down the country again—and if we act today, could allow us to see our loved ones, go back to school and work, and travel—all before Christmas.

Antigen tests are “contagiousness” tests. They are extremely effective (>98% sensitive compared to the typically used PCR test) in detecting COVID-19 when individuals are most contagious. Paper-strip antigen tests are inexpensive, simple to manufacture, give results within minutes, and can be used within the privacy of our own home . . .

If only 50% of the population tested themselves in this way every 4 days, we can achieve vaccine-like “herd effects” . . . Unlike vaccines, which stop onward transmission through immunity, testing can do this by giving people the tools to know, in real-time, that they are contagious and thus stop themselves from unknowingly spreading to others.

Mina continues:

The U.S. government can produce and pay for a full nation-wide rapid antigen testing program at a minute fraction (0.05% – 0.2%) of the cost that this virus is wreaking on our economy.

The return on investment would be massive, in lives saved, health preserved, and of course, in dollars. The cost is so low ($5 billion) that not trying should not even be an option for a program that could turn the tables on the virus in weeks, as we are now seeing in Slovakia—where massive screening has, in two weeks, completely turned the epidemic around.

The government would ship the tests to participating households and make them available in schools or workplaces. . . . Even if half of the community disregards their results or chooses to not participate altogether, outbreaks would still be turned around in weeks. . . .

The sensitivity and specificity of these tests has been a central debate – but that debate is settled. . . . These tests are incredibly sensitive in catching nearly all who are currently transmitting virus. . . .

But wait—if this is such a great idea, why isn’t it already happening here? Mina writes:

The antigen test technology exists and some companies overseas have already produced exactly what would work for this program. However, in the U.S., the FDA hasn’t figured out a way to authorize the at-home rapid antigen tests . . . We need to create a new authorization pathway within the FDA (or the CDC) that can review and approve the use of at-home antigen testing . . . Unlike vaccines, these tests exist today—the U.S. government simply needs to allocate the funding and manufacture them. We need an upfront investment of $5 billion to build the manufacturing capacity and an additional $10 billion to achieve production of 10-20 million tests per day for a full year. This is a drop in the bucket compared to the money spent already and lives lost due to COVID-19. . . .

I read all this and wasn’t sure what to think. On one hand, it sounds so persuasive. On the other hand, lots of tests are being done around here and I haven’t heard of these rapid paper tests. Mina talks about at-home use, but I haven’t heard about these tests being given at schools either. Also, Mina talks about the low false-positive rate of these tests, but I’d think the big concern would be false negatives. Also, it’s hard to believe that there’s this great solution and it’s only being done by two countries in the world (Britain and Slovakia). You can’t blame the FDA bureaucracy for things not happening in other countries, right?

Anyway, I wasn’t sure what to think so I contacted my epidemiologist colleague Julien Riou, who wrote:

I think the idea does make sense from a purely epi side, even though the author appears extremely confident in something that has basically never been done (but that maybe what you need to do to be published in Time magazine). In principle, rapid antigen testing every 4 days (followed by isolation of all positive cases) would probably reduce transmissibility enough if people are relatively compliant and if the sensitivity is high. The author is quick to dismiss the issue of sensitivity, saying:

People have said these tests aren’t sensitive enough compared to PCR. This simply is not true. It is a misunderstanding. These tests are incredibly sensitive in catching nearly all who are currently transmitting virus. People have said these tests aren’t specific enough and there will be too many false positives. However, in most recent Abbott BinaxNOW rapid test studies, the false positive rate has been ~1/200.

Looking at the paper the author himself links (link), the sensitivity of the Abbott BinaxNOW is “93.3% (14/15), 95% CI: 68.1-99.8%”. I find it a bit dishonest not to present the actual number (he even writes “>98%” somewhere else, without a source so I couldn’t check) and deflect on specificity which is not the issue here (especially if there is a confirmation with RT-PCR). The authors of the linked paper even conclude that “this inherent lower sensitivity may be offset by faster turn-around, the ability to test more frequently, and overall lower cost, relative to traditional RT-PCR methods”. Fair enough, but far from “these tests are incredibly sensitive” in the Time piece.

Two more points on the sensitivity of rapid antigen tests. First, it is measured with the RT-PCR as the reference, and we know that the sensitivity of RT-PCR itself is not excellent. There are a lot of papers on that, I randomly picked this one where the sensitivity is measured at 82.2% (95%CI 79.0-85.1%) for RT-PCR in hospitalised people. This should to be combined with that of rapid antigen testing if you assume both tests are independent. Of course there is a lot more to say about this, sensitivity probably depends on who is tested, when, whether there are symptoms, and both tests are probably not independent. Still I think it’s worth mentionning, and again far from “these tests are incredibly sensitive”. Second, the sensitivity is measured in lab conditions, and while I don’t have a lot of experience with this I doubt that you can expect everyone to use the test perfectly. And on top of that, people might not comply to isolation (especially if they have to work) and logistics problems are likely to occur.

Even with all these caveats, I think that this mass testing strategy might be sufficient to curb down cases if we can pull it off. Combined with contact tracing, social distancing, masks and all the other control measures in place in most of the world, being able to identify and isolate even a small proportion of infectious cases that you wouldn’t see otherwise can be very helpful. We’ll soon be able to observe the impact empirically in Slovakia and Liverpool.

So, again, I’m not sure what to think. I’d think that even a crappy test if applied widely enough would be better than the current setting in which people use more accurate tests but then have to wait many days for the results. Especially if the alternative is some mix of lots of people not going to work and to school and other people, who do have to go to work, being at risk. On the other hand, some of the specifics in that above-linked article seem fishy. But maybe Riou is right that this is just how things go in the mass media.


  1. Andrew, Thanks soooooo much for bringing this subject here on your blog. Wow.

  2. JFA says:

    I’m not sure why people are super concerned with the sensitivity. If it were .05 then sure, that’s a bad test. But what if the sensitivity was .5 and we missed half of those who actually have the virus. Now implement regular testing in schools such that anyone with a positive test has to quarantine for 2 weeks (or less with a negative PCR test). That cuts the number of active cases in the schools by half. But even with a low false positive rate, chances are that most people who come up positive are actually negative (unless the virus is really raging). So a few kids and staff who don’t have the virus have to stay home to reduce the number of infected people walking around the schools by half. That seems like a pretty good outcome while we wait for a vaccine, particularly when the relevant alternative is to keep *everyone* out of school. It seems that for testing we’ve really made perfect then enemy of the good.

    • Excellent JFA. It’s the frequency or testing that is key to reducing the potential for transmission. Dr. Fauci in an interview with Sanjay Gupta did mention that nearly 40% of the transmission is hypothesized to generate from asymptomatics.

      My observation here in DC is that many individuals may get tested one or twice. That’s it. For the next 6 to 12 months, frequent testing is necessary. This does not mean abandoning masks or other measures to prevent or suppress transmission.

      It seems to me that the FDA favors a point of care rapid test. On December 23rd, I will be paying $90 for the rapid antigent test: which will be swapped in the physician’s parking lot. It’s pricey obviously. It should be way cheaper or free.

      • Just to add. There is so much news about the vaccines, which, in turn, overwhelms interim proposals to stem the tide of contagion. Plus subsets are anchored in the notion that PCR is the diagnostic gold standard, despite delays in test results and because these subsets are not so well informmed about what the tests convey in fact.

    • Meg says:

      The problem with this reasoning is demonstrated by my school district.

      They are going back to a fully remote model for everyone over 3rd grade because they can’t staff enough classrooms. Substitutes are in very short supply. This is usually handled by faculty and staff doing coverages for classrooms where teachers are out and a sub couldn’t be found. But our moderately sized district has too many teachers and staff quarantined and teaching from home, requiring bodies to be in the room with the students. And those bodies don’t exist. So in this case testing that has poor specificity *is* actually a problem.

      The bigger hurdle to the transmission issue is low sensitivity. But as you pointed out, if you have low sensitivity and are testing very frequently in a setting like a school then you will be in a better place then not testing enough. It also provides the benefit of (maybe/hopefully) catching you when you first are infected (and if it were systematic and very widespread people might actually figure out how to gauge risk levels of exposure which is basically unknown right now). And that is really the most important time. My husband (ID specialist) says nothing is very clear about any of this still, but *probably* you are *most* infectious just before you become symptomatic (if you ever become symptomatic). Of course you are still infectious for some unknown time afterward as well.

      However, relying on PCR to confirm a negative antigen test (specifically the FIRST in a series) is problematic and would exacerbate shortages of staff in places like schools and hospitals.

      A PCR will return a positive on patients who WERE recently infected but aren’t infected anymore. My husband is an infectious disease physician in NJ (so fully in the thick of it since the end of February). He said he has patients who have tested positive repeatedly for months after recovering (which is certainly a longer tail than average, but he’s just one guy, who admittedly has seen hundreds of hospitalized Covid patients.) I frequently ask him if they have more info about the sensitivity of the tests and he said he thinks everyone has given up and accepts that its false negative rate is probably below 20%, accounting for variation in biology, test and lab personnel, and qualities of the test itself.

      So, basically no single testing strategy is the answer. The only really useful thing would be an enormous army of contact tracers who can compel cooperation in areas where infection rates are out of control. That would allow you to know your positive result is because of an active infection, not an old one. Or allow you to time a sensitive test (or maybe two if you receive a negative first) to the point when you are most likely to be most infectious but unlikely to have symptoms if you ever get them. And then with properly timed negative tests you could shorten a quarantine and feel pretty confident you aren’t a hazard.

      In addition to properly wearing masks (distance is seeming less useful indoors unless you have an insane amount of ventilation as 6ft is probably not nearly far enough).. and yeah kids sitting in their classrooms with windows open and fans blowing <0 C air around a classroom is not a great situation to be learning (or teaching) in. But that doesn't matter as long as everyone gets to sit in restaurants and drink in bars and go to the gym.

  3. Chris Said says:

    Hi Andrew,
    I’m a longtime occasional reader of your blog, and also the founder of, an all-volunteer organization now directed by Michael Mina.

    A few points on why other countries aren’t doing this as well.
    * In other countries, medical regulators are strongly incentivized to avoid errors of commission rather than errors of omission. That’s reasonable, to a certain extent, but the degree to which regulators are afraid of making a positive mistake is far too high.
    * Related, most medical regulators in other countries are incentivized to follow whatever US regulators do. It’s kind of like how “nobody ever got fired for buying IBM”. No international regulators get fired for just doing whatever the US regulators do.
    * Beyond regulators, it turns out that most medical decision makers do not follow Bayesian Decision Theory! Rather than probabilistically weighting the pros and cons, many focus their attention on “what could go wrong”. Again, that’s good to a certain extent. There’s a lot that can go wrong with rapid testing (bungled early rollouts, false sense of security, etc). But far less is done to weigh them against the benefits, which in my view are far bigger.

    On sensitivity: There’s a lot of confusion here. Michael estimated 98% sensitivity, but others put sensitivity at 80% or in some populations as low as 30%. What’s going on? The answer is that sensitivity as a metric depends entirely on the viral concentrations of the samples. Michael’s claim is that sensitivity is >98% for viral loads >10^5 cp/mL, which can reasonably considered 98% sensitivity in contagious individuals. People can quibble over where exactly contagiousness starts (or where it becomes non-negligible), but there is converging evidence from animal models and viral culture studies that you aren’t contagious if you have a viral concentration < 10^5 cp/mL. Mayyyyybe at worst you could be a little contagious on the upward part of the curve, but that's only for a very brief period. In any case, quibbling over the exact cutoff misses the point. These tests are meant to be perfect. They're just supposed to help make a dent in transmission, similar to mask and social distancing. (

    • Chris Said says:

      Typo in the last line: “These tests are *not* meant to be perfect”.

      Also: I agree with Julien about laboratory conditions. Even though the tests have >98% contagiousness sensitivity in the lab, their real world sensitivity may be lower, especially in the hands of untrained individuals. That’s still ok (nobody says we should ban masks because some people might wear them incorrectly), but it’s still really important to educate people that a negative result is not a free pass to party.

      • > really important to educate people that a negative result is not a free pass to party.
        That is one of the concerns.

        What is the shelve life of a test negative being highly suggestive of not being contagiousness (i.e. a graph of sensitivity versus time after test)?

        Some countries have immunity task forces that might be worthwhile contacting.

    • Meg says:

      With respect to the sensitivity claims in the article Mina cites about the Abbott test, the problem with the figure he reports is that it is not an honest claim of what the paper reported. 14/15 created a very wide margin of uncertainty. Everything you said above is true, but saying the test is 98% sensitive is at best misleading.

      I’m not arguing against using rapid tests in a widespread way (Although, the aforementioned article said they had to retrain the technicians so they correctly read whether a test was positive or not, which is a whole different kettle of fish). I just think the vocal arguments for them should include the limitations and explain why those limitations aren’t as important as their strengths (as you did.)

  4. Tom says:

    The FDA approved the first at-home COVID test this week.

    It’s more expensive than the proposed test but perhaps useful for the regulatory comparison question (“it makes sense, why aren’t we doing this?”).

    • Chris Said says:

      Hi Tom,
      To my knowledge, everything that the FDA has approved for home use requires a special device, which increase costs and reduces scalability. We could get much greater aggregate transmission reduction if we approved inexpensive antigen tests that do not require a special device, even though their sensitivity is lower. Widespread use and frequency trump sensitivity. See the diagram at the bottom of to see what these look like.

      • Tom says:

        I don’t disagree. That was my point. It is not the same test, but if the issue is a regulatory roadblock then this is something that was approved and those knowledgeable can determine whether it was approved under regular order or something special (Warp Speed) and what hurdles there are to approval of the low cost test (or whether something is already in the regulatory pipeline).

      • I would be very interested in how Michael Mina would respond to Mike Yeadon. Yeadon claims that asymptomatics can’t transmit SARS2.

        • Joshua says:

          FWIW, Yeadon has made a lot of nutty claims related to COVID.

          • The question of the extent of viral load that asymptomatics can transmit has been an ongoing, even among American infectious disease specialists. Yeadon’s view of the utility of the PCR test in some contexts is integral differentiating among several testing options. For example, in noting that a PCR test can yield positive result even after the infectivity period has passed.

            • Joshua says:

              Sameera –

              I’m certainly no expert.

              But I think it’s nuts to think there is no asymptomatic transmission if we’re seeing over 200k positive tests a day.

              I think it’s highly implausible that after initially saying there was no asymptomatic transmission, all over the world people who study this would reverse themselves after seeing tons more real world evidence to say there is asymptomatic transmission, if in fact they were correct in their original viewpoint.

              Yeadon is in cahoots with a crowd that is promoting a lot of sheer nonsense – such as that PCR testing is producing a useless metric and a “casedemic” when the simple fact of spiking hospitalizatons, ICU admissions, and deaths makes it irrefutably obvious that is not the case.

              People in the crowd he’s hanging with have repeatedly made claims that were later clearly falsified – such as his claim that the pandemic was over in the UK almost two months ago and that there’s no need for a vaccine – with the errors always being in the direction of them underestimating the seriousness of the pandemic.

              Of course there have been and will continue to be errors made in the other direction as well.

              Of course people like Mina agree that there are serious limitations to the PCR tests as a surveillance tool, and that the current testing regime captures as positive, many people who are no longer infectious.

              But to feed the dangerous belief that asymptomatic transmission doesn’t occur is dangerous and nutty,imo. I have encountered this belief online quite a bit. I like to think it’s mostly just a relatively small group of online fanatics, but I suspect that the impact is pretty far-ranging – along the lines of “most people are immune because of T-cells,” and “it’s basically just like a bad flu.”

              • Joshua says:

                As to the % of infections that come via symptomatic vs. asymptomatic spread…it certainly seems there’s uncertainty that remains. It’s an important question to resolve.

                But contributing to the propagation of the viewpoint that PCR testing has zero value because there’s no asymptomatic spread…seems highly, highly irresponsible to me. As is promoting the view that vaccines are unnecessary because the pandemic is “effectively over.”

              • I am not ensconced in the UK public health environment. But the debates there are very robust on Twitter. Nor have I followed each and every one of Yeadon’s claims. However Yeadon’s and nearly every infectious disease expert convey that the amount, force, & duration of viral load to which you are exposed has implications for whether or not you contract the virus. Why social distance and masks are encouraged by most. Yeadon seems to suggest that ‘coughing’ would eject more virus.

                Moreover Yeadon is a signatory to the Great Barrington Declaration. It seems to have more gravitas abroad than in the US.

                My own view is that the antigen tests can give us an indication of whether we are infectious or not. I find Michael Mina’s perspective to be one of the most cogent I’ve come across. I plan to advocate for funding of these antigen tests.

              • Addendum:

                Michael Yeadon emphasizes that the PCR yields ‘false positives’. ‘Characterizing the ‘positive’ as a ‘false’ one seems inaccurate if indeed one can test positive for weeks after the infectious period has passed. Now Yeadon attributes some of the inaccuracy to lab contamination as well. So most of us would not be able to evaluate ;ab practices. It’s not an issue that I have seen crop up here in the US. Rather delays in test results have been in the COVID19 news cycle.

                I also saw a thread in Yeadon’s Twitter Feed suggesting that has conveyed that the PCR test is inappropriate. But I have to revisit that thread to be sure.

                It is important to learn from all perspectives if sounding nutty. In fact some of the transformative and accurate claims have been rejected initially and deemed heretical.

              • Joshua says:

                Sameera –

                As you know, I think that the rapid antigen tests would comprise a very useful tool.
                Nonetheless, the bottom line is that the prevalence of hospitalizatons, ICU admissions, and deaths are all correlating with the prevalence of positive PCR test results.

                Of course, many are arguing it’s pure coincidence, and that it’s actually just the flu or something like that, and the diagnoses of COVID are mis-diagnoses, or part of a deliberate plot to use panic about a non-existent threat to “reset” society.

                I’m sure you’ve seen all of that being spread widely on Twitter.

                Yes, it’s important to learn from all perspectives. It’s also important to, at some point, consider plausibility and to consider the ideological biases of people who are potentially spreading nonsense.

    • Steve says:

      This test also requires a prescription. It is ridiculous. We needed a testing system months ago that allowed people to get back to work and trace exactly who was contagious

  5. Anonymous says:

    Step way back and ponder the big picture.
    are turning the world upside down, destroying national economies & the economic/social lives of billions of people

    • Phil says:

      I assume you were going to say “are these things worth it.” But this is only a question worth asking if there is an alternative. Trying to keep the number of deaths in the U.S. to some ‘acceptable’ number like half a million is wrecking the economy, ruining a year or two of important childhood years, ruining people’s social lives (at least for a year or two), and so on. But _not_ trying to keep the number of deaths in the U.S. under control would also have many of those same effects, while also killing a lot of people.

      Tell me what your plan is and how many people would die if we follow it, and I’ll tell you if I think it’s better than what is happening.

  6. FYI: Here is what DC Mayor’s COVID19 Udate has proposed.

    DC Public School Chancellor Ferebee announced the following:

    I consider this announcement a step forward. The Washington Teacher’s Union has expressed justifiable reservations about the reopening of schools. I had anticipated these teacher concerns back in June, which is why I urged DC school administration to conduct at school testing. Even before I knew about these rapid antigen tests.

    Grassroots appeals to Congress would be great. So I am going to play my part in that effort. I have followed many perspectives on testing options. I do support Dr. Mina’s plan. Nor am I being paid for my efforts to discuss this on social media and blogs.

  7. John Williams says:

    The website of lists Rochelle Walensky, who will soon run the CDC, as an “advisor.” I suspect this means less than it seems, or we would have heard more about this idea. The organization does not seem to have a board of directors, which is another red flag.

    More generally, I think implementing this idea in a major way would be a whole lot harder than its proponents think.

    • Hello John,

      Just to clarify, the FDA & Centers for Medicare/Medicaid has given permission to schools to use the antigen test on asymptomatics.

      See the following explanation on

      ‘Legal status of rapid tests in schools

      Both the Center for Medicare/Medicaid Services (CMS) and the FDA have given tacit permission for schools to use rapid antigen testing in asymptomatic individuals, as long as the facility is authorized under a Clinical Laboratory Improvement Amendments (CLIA) Certificate of Waiver. In particular:

      CMS will not cite facilities for using POC antigen tests on asymptomatic individuals, as long as those facilities have a CLIA waiver (source)

      The FDA states that if PCR is not available or wait times are too long, POC antigen tests may be used for asymptomatics, even if the tests are not specifically authorized for this indication (commonly referred to as ‘off-label’) (source)

      Another alternative, which may not require a CLIA waiver, is to perform non-clinical surveillance-based testing. The National Scientist Volunteer Database has developed a platform for on-site, rapid and affordable COVID-19 surveillance in schools and other community organizations. To learn more about the platform and to discuss potential implementations, please visit or contact testing(at)’

      ——————- is a new volunteer organization. It is clear in its mission: Promote Rapid Tests.

      I think that if the FDA creates another pathway for legal use for rapid antigen test, the public will embrace its use.

  8. Caruso says:

    So we are destroying the world economy and economic/social lives of billions of people — all based upon the very “questionable” accuracy of a couple of COVID-19 “tests”.

    That is irrational behavior.

    At the very least, these COVID-19 tests must be validated beyond any reasonable doubt.

    Specificity and Sensitivity claims are meaningless without a clear evidence of what exactly the tests are “discovering”.
    The PCR test routinely generates a Positive Result on harmless virus RNA fragments … rather than an active COVID-19 infection.
    These false positives greatly inflate the reported COVID-19 cases numbers, causing unjustified panic.
    The morbidity numbers are likewis inflated.

    All the national COVID-19 stats are errobneous.

    • You must be aligned with Yardley Yeadon, a former VP at Pfizer. He claims that the PCR is an entirely inappropriate test for diagnostic purposes because it generates positive results even if transmission is not possible or negligible.

    • Andrew says:


      “Beyond any reasonable doubt” is not a phrase we like to use in statistics. There’s no certainty, we’re all gonna die someday, etc. No doubt people are incoherent about risks, increasing their chances of heart disease and cancer with their diet, driving around and crashing their cars, buying guns and then getting shot, etc etc and then getting worked up over super-rare risks that just happened to be well publicized. That’s something we have to work with when considering social behavior. News reports of dead bodies in China, Italy, etc., along with millions of Americans personally knowing people who’ve died of covid, have caused . . . I don’t know if “panic” is the right word right now, but certainly it’s causing concern. Again, many millions of Americans don’t want to go out into the world and catch the virus, especially people who’d be in jobs where they’d be in contact with dozens of people every day. This is not to say that all this behavior is rational, however that might be defined, just that to move forward it will be necessary to address these concerns.

      P.S. Also, no need to put the word “tests” in scare quotes just cos the tests give errors. Just about all tests have errors. In statistics we model this.

  9. Joshua says:

    I’ve been harping in this for months (including here at this blog).

    I have yet to see what strikes me as a strong counter-argument to rolling this out on a massive scale.

    The closest I’ve seen is thst (1) in being totally unprecedented, the logistical difficulties are being underestimated. To which I say you’ll never know for sure until you try. (2) it relies too much on people doing the right thing if they test themselves and get a positive result. To which I reply, sure, that’s a problem. But I still think in the end it’s better than the status quo.

    I have been hoping to see people smarter than myself come up with strong counter-arguments. Since I this isn’t happening (and doesn’t seem likely to happen) it might give me some form of solace to know there’s actually a good reason it isn’t happening. Given that I think this is a no-brainer, it’s very upsetting to see so little progress on this front. It just feels very arbitrary and devastating in terms of the enormous opportunity cost of the status quo.

  10. One of my colleagues is watching the Liverpool study, I think they said less than 50% have agreed to be in the study and participants are expected to unusual. They are not very hopeful it will pan out, but admit they might be wrong.

  11. Mendel says:

    Ugh, why can’t he get the numbers straight?

    “If only 50% of the population tested themselves in this way every 4 days, ”
    “The cost is so low ($5 billion)”
    “We need an upfront investment of $5 billion to build the manufacturing capacity and an additional $10 billion to achieve production of 10-20 million tests per day for a full year.”

    50% of the population is ~160 million, to test themselves every 4 days they require 40 billion tests per day, so $15 billion for the manufacturing and $30 billion for the tests at $2 apiece at cost, but then there’s logistics and distribution on top of that, so maybe double that if we’re lucky?

    Currently, these tests are priced at ~200€ for 25; some pharmacies offer to perform antigen tests for 20€ apiece (this includes swabbing).

    • David says:

      I’m not sure exactly where all your math is coming from but isn’t that 40 million tests per day, not billion. If we use 40 million tests a day for three months, ~4 billion total tests or ~15 billion tests in a year.

    • Keith E. says:

      If we aren’t lucky, I could easily see this costing a hundred billion and then it fizzles. Budgets can and do explode. It may still be worth the attempt, when compared against the massive ongoing cost of COVID… but, when we (very rightly) are simultaneously developing and preparing multiple vaccines, multiple tests, multiple treatments, we may find that fifty billion here, fifty billion there, starts getting into real money.

    • Michael Mina says:

      Mendel – your math is wrong. 160 million every 4 days. We need only people who are really going out and participating in society the most, so it’s fewer than 160 million. 120 million would go a massive distance. That is, say, 30 million per day. If these are performed in households, we can have 2 people pool their samples (each their own swab) and then run one test together – that’s ~15M tests per day. (The US already makes about 8M per day). The tests can cost about $2 per test if the US government purchases in bulk. OK, so now we are at $30M every day.

      Have you calculated what the virus is costing us? $16 Billion every day. That’s over a half a billion every hour. The orders of magnitude aren’t even close. Just the most recent COVID relief bill sitting with congress is $908Billion. $30M per day – even if it is $50M is nothing We could have 4 months of these tests for $5 Billion dollars. Meanwhile the Return on Investment would be in the hundreds of billions if not more if it means we can go back to school, open up shops and restaurants more safely and generally start to see things progress towards normalcy.

      • Joshua says:

        But there’s another testing pathway as well. Suppose private businesses run these tests for employees or customers and pay for the tests? Yes, there would be logistical obstacles but still better than the status quo.

        Further, I would happily buy tests myself if they were available. And pool them with my pod if I wanted to save money. I think it’s just nuts that I don’t have the ability to do this. Being able to do the rapid tests would enable me relieve tensions around family interactions considerably. Of course there would be risk involved. But as compared to what, exactly?

      • Joseph Candelora says:

        Dr. Mina,
        Your article itself said, “If only 50% of the population tested themselves in this way every 4 days, we can achieve vaccine-like “herd effects”. Are you now revising that to only 50% of those “really going out and participating in society” to achieve that effect?

        You certainly can do so, but it’s inappropriate to state Mendel’s math was wrong when you’re revising your claim.

        Further, how did you come up with 50% as the bogey for your herd-immunity-equivalent? I believe the general estimates have been in the range of 60% – 75% immunity for normal herd immunity. Why is this testing approach more effective than actual immunity?

        It’s an intriguing idea for sure.

        • RE: Your article itself said, “If only 50% of the population tested themselves in this way every 4 days, we can achieve vaccine-like “herd effects”. Are you now revising that to only 50% of those “really going out and participating in society” to achieve that effect?

          I don’t see any contradiction between the two hypotheses. And there is no sin in refining one’s hypothesis either. We have to measure the success of testing and that takes experimentation. I would guess that those who are likely to get tested are workers. Those who have to report to workplace.

          I could be wrong.

          • Joshua says:

            I agree with Joseph – I think that applying the 120 million number, if it wasn’t made explicit earlier on, shouldn’t be framed with “your math is wrong.”

            • Hi Joshua,

              I see your point. However Mendel wrote:

              ‘Ugh, why can’t he get the numbers straight?’
              A tad provocative initially. Anyway, I gather that some of these rapid antigen test startups will be submitting their investment figures to the government and private sector. Michael Mina has emphasized the cost savings in pooling in a few interviews.

              I’m sure there are degrees of frustration on all sides. We should push through it if inclined.

        • Joshua says:

          Joseph –

          > Further, how did you come up with 50% as the bogey for your herd-immunity-equivalent?

          There is a lot of ambiguity and variability and a lack of preciseness in much of the discourse around “herd immunity” “herd immunity threshold,” etc. From the article:

          > If only 50% of the population tested themselves in this way every 4 days, we can achieve vaccine-like “herd effects” . . . Unlike vaccines, which stop onward transmission through immunity, testing can do this by giving people the tools to know, in real-time, that they are contagious and thus stop themselves from unknowingly spreading to others.

          There could certainly be a difference between “herd effects” and “herd immunity,” with the former being a significant difference in the rate of infections and the latter being a particular point at which the rate of spread begins to diminish. Indeed, as I understand it, “herd effects” could theoretically start as soon as any significant number of people are infected – so the term as applied seems rather arbitrary. Further, it’s unclear as to whether the term “herd effects” might have been used as a substitute for “herd immunity” or “herd immunity threshold.”

    • Mendel says:

      Aside: the AstraZeneca vaccine costs the EU about $2 per dose. That’s a much better investment than rapid tests would be.

  12. jonathan says:

    I like to think about my response to these kinds of arguments. So I pretended I’ve never heard of Michael Mina.

    First, my distrust flag goes up when I read something that purports to be a categorical solution without pointing to any flaws. That says to me the writer has constructed a tight narrative in an advocate’s manner, meaning you pretend that objections dont exist to what you’re saying. That is, in my experience, the essential lawyerly skill isnt defeating your opponent’s arguments head-on, but rather making them disappear. That is, anyone can argue facts, but to creatively shape the journey through the space of the case requires linking a series of complex relationships together. You actually build regression spaces in chains. What makes it fun and where the distrust really attaches is that it has a known goal, and the knowability of that goal induces backwards so you are fitting chains from both directions, and that builds your best story.

    That is how to be persuasive. Problem is it is also sleight of hand because it is by definition a chain that has a goal which has value from their perspective. There is an interesting issue in there: the inherent value of perspective. You can approach from below, like with Shapley, or from above. Approaching from above means you’re influencing the complex space the others see, which enables you to unravel how they are being persuasive. That is why going second can be a huge advantage. It is debate class but it’s different because there are specific legal hooks along the way, meaning there is a modular structure imposed.

    But before that, you can see how you construct chains if you consider the line as a function you are approximating so the smoother a section, the less variability, the more cleanly connected the endpoints, the more you erase the other paths. The chain inducing backwards imposes a perspective value cost, which is an advantage as well, which tilts the complex space so you see the sides of the points they want you to see and not the rest. Since it’s regression lines, you can use sticks with 1 and 0.

    Continuing, there is a modularity to the argument which tells me each piece has been specifically aligned so it fits exactly to the next without wavering. That could either be someone who is correct and knows it or someone who has some reason for wanting to appear correct, including self-delusion that he is correct. The former, I think, would know that something phrased to appear inevitably obvious would be received skeptically by those who didnt think of it themselves, so that makes me question the audience this is intended to reach. That matches the surface, which is technical enough for non-experts, but which offers a clue to the actually technical that there is more to this.

    This is an interesting spot because you posted about work done for Trump of laughably bad quality. I pointed out the guy was paid to produce a defensible argument, not to produce an objective statement about votes. That is another example of the endpoints being pinned so you can line up the complex segments to make an argument. One of my goals is to expose how arguments are made so they can be decomposed into basic repeating methods that map to actual math, not to constructed labels. In other words, they have to be phrased in such a manner that they can be counted. That’s basically what I work on.

    If I were playing Holmes, which is another example, I’d sniff out that this piece was designed to appeal for a champion in the public space. The reason that would infer is the words disclose at multiple levels, meaning there is communication to both lay and experts included, which means someone with influence could have this read by someone who knows more (who knows someone who knows more), and get back a positive. That is basic strategic thinking, one chess move past the visible obfuscates, meaning your strategic choice begins to disappear one step away, which means influence on your choices one step away reduce to your visible choices, as the potential within the shared space encloses you.

    As you can see, I’ve been thinking about your really interesting article. I’m trying to describe how imposing endpoints allows you to treat the functions an results we can measure as having not only been generated by either forces within or external forces acting on them, but as those are affected by the changes and shape of complexity that occurs over any complex field when that is treated as other perspective. The most basic conceptual statement seems so far to be the halving of uncountable infinity. I can derive everything from that one statement, but then there’s a whole lot in uncountable infinity. One goal is to be able to view models, particularly noisy ones, by the level of obfuscation they reach. This I think is a nice label for the problem: like with ESP (which as a soccer fan I read as Spain), the counting model of that research describes it as reaching an obfuscation level of an event reduced to background noise, which is definable as you can turn off every part of the experiment as long as the cards are actually random, so this would be the obfuscation of 2 randoms within a complex field that is not random at all. The results thus show you cant entangle random events by containing them. The model says that would be true until they become co-generated. An example of that might be twins who can follow the other’s slightest hints, who may even match patterns, especially on any exposure to the game because this game is basic survival when you’re a bird or a fish or you’re the baby or the momma.

    That brings up the issues of what is co-generation. It’s easier to say what isnt: it’s not co-generation when the objects contained in a room are humans making choices. Humans are complex machines. So any choice of card involves complex rotations that generate a choice that is then said or otherwise expressed. And let’s say there is a visualizer, not merely a drawer of random cards: how exactly does one visualize a card without thinking? So complex rotations on that side. So the generation that is hoped to entangle requires some value above background noise, which is a backwards way of saying there would need to be signaling of some sort or such familiarity that both may somewhat follow similar patterns. You dont need card drawing for that: just run twins through numbers, letters, pictures, etc. and see which ones they tend to pick. It isnt then that they have the exact same preferences but that the preferences relate because the modularity of their minds relate. This should cause more coincidence, not only at the primary level of this is exactly the same but at the pattern change levels where you can see related modularities fluctuate.

    I talk about this with one of my daughters: a tiny inflection in her voice draws an endpoint which allows me to infer the existence of an interpretative space that I fill backwards to meet the answers to any questions I ask. People who really think the same like we do, know we think the same. I cannot do that with my other daughter because her mind draws spaces I cant see, and that is because she’s way too complicated, and our communicative method requires defining complex spaces around each other.

    • Hey Jonathon,

      Michawl Mina has amply discussed the pros and cons of the rapid antigen tests. Some of us have followed his perspectives assiduously since July or so. I will review each YouTube presentation and article at least 3 or 4 times. So I have a sense of the evolution of Michael Mina’s perspective. Moreover, he is an immunology and infectious disease specialist. At no time has he presented it as a categorical solution. Rather he has had to debate skeptics of the utility of the at home rapid antigen tests.

      The test kit itself proposed contains one or two confirmatory tests.

      OK so, Michael Mina happens to be a great marketeer too. I speculate he has been given pointers simply through Harvard’s own vast connections. That doesn’t mean that what he is proposing is without merit.

      In any case Jonathan, if you can use examples of your reservations from Michael Mina’s hypotheses, I would be more comfortable.

      • Joshua says:

        I think in particular his argument about the different applications of the different tests – and the surveillance/public health counter individual health/treatment functionality of the different types of tests respectively is particularly interesting. It’s not a framing that I’ve encountered before. It’s not like he thinks that the antigen test should replace the PCR tests – but that they are effectively fit for different purposes. Imo, thst makes a whole lotta sense.

        A commenter here argued that the antigen tests wouldn’t match Mina’s rhetoric because repeated tests would be subject to the same flaws and their effect of reducing the probabilities of false negatives wouldn’t be additive because they wouldn’t be effectively independent tests.

        I get that logic and would like to see that question addressed – but I my sense is that the argument that they aren’t independent was overstated, and then also that issue could be addressed in that there are a variety of methods available for the antigen tests which could help address that issue if it were really established that is a problem.

        • Navigator says:


          “A commenter here argued that the antigen tests wouldn’t match Mina’s rhetoric because repeated tests would be subject to the same flaws and their effect of reducing the probabilities of false negatives wouldn’t be additive because they wouldn’t be effectively independent tests. ”

          That would be a fair argument if we knew nothing about the viral load over time. Mina’s point is that only the very sharp peak (not on a log scale) is when anybody is highly infectious for others, followed by a sudden drop off. That’s the case for everyone, plus/minus.

          Current PCR tests show positivity on an upward and downward slope without the ability to determine where in the cycle anyone is. Scale it up and you are talking about many PCR ‘positives’ that are in reality towards the tail end of an infection cycle and not a menace to society. Even worse, negative PCR may only mean the peak will occur in a couple of days and by the time you get your ‘negative’ PCR results back, you may be highly infectious to others.

          Imperfect antigen tests administered longitudinally give us more info than an expensive cross-sectional PCR without immediate results.

          It’s been an excrement show since day one with all of this, for sure.

          • Joshua says:

            Navigator –


            > Imperfect antigen tests administered longitudinally give us more info than an expensive cross-sectional PCR without immediate results.

            Yup. Especially when you add in the futility of PCR testing if it takes something like 5 days or more to get the results.

          • RE: ‘Mina’s point is that only the very sharp peak (not on a log scale) is when anybody is highly infectious for others, followed by a sudden drop off. That’s the case for everyone, plus/minus.’

            This is one of the points that has been a stumbling block for those who haven’t followed Michael Mina’s explanations closely. I’m glad you flagged it.

        • Kenneth Chiu says:

          False negatives can be tolerated. In fact, some percentage of the population that simply refuse to test can also be tolerated.

    • OK, so what if you had followed Michael Mina’s perspective in real time. Any specific examples that align with your analysis?

      • jonathan says:

        If I had followed Mina’s specific perspective, my assumption is that it would follow the general trend of well meaning attempts to provide guidance in a complicated situation. The issue is not Mina, not in the slightest. In fact, my conclusion was that this is an attempt to do good (and that even aligns with conceptions of self-interest). The issue which I see the piece attempting is reaching through the confusions that develop in even simple situations but which have run haywire now. It’s hard to imagine how complexity could not develop and could not confuse given the time period. Our confusion now is technically different from the confusions of the past but they’re all confusions about what’s going on and what to do. Forms of confusion have become more concrete, but it’s the same general space called ‘confusion’, and what that means is specific pathways through to a safe haven or result appear to generate but dont actually reach across – which gets to the complexity suppressed in any argument – or there are simply other pathways which work about as well or maybe even better. If you can admit to ‘better may exist’, then you admit that in some perspectives your path is worse.

        As to examples, I note that the post raises specifics about test specificity and the conditions under which the test doesnt perform as well as advertised. That means the question was answered before it was asked.

        I wanted to add an example of obfuscation. I was in a conversation with someone who insisted that Israel flooded Gaza. My response was that they’ve repeatedly said they dont and that there arent any dams to open. He provided pictures of a couple of large irrigation ponds. This to him constituted proof. To me, it constituted a pin that could connect directly to his conclusion but if it did, that should require excluding other possibilities. First, you might think that after years of this allegation, that someone with a phone might have a photo of the reservoirs emptying or emptied. They arent isolated or inside secure facilities: people post photos of the view on the internet. Second, does flooding occur in the area? Yes, they have ravines, gullies, etc. and those around the world come with flash flood warnings. Third, is the flooding connected to actual known flooding? Yes, it occurs when there’s flooding in Israel. That means this: one step from photo to cause, and this reinforces through the complex loop so the existence of flooding on its own, just because there’s flooding, is displaced by the cause they want to see.

        That’s a problem with inference generally: it can infer strongly in one direction, in one step, and the strength of that inference can be maintained by weak inference in the other direction, and that hides the likely answer because, and this is ironic, the math generates the inverse of the answer and you read that inverse because you are so influenced up close that you flip it over at a distance. That occurs at specific steps of complexity in a chain.

        • To me this isn’t all that complicated. If one has the wherewithal to ascertain more rapidly whether one is infectious, one can quarantine or isolate responsibly. Some experts have hypothesized that asymptomatics constitute some significant minority of transmission. Typically individuals have sought tests when they experience symptoms or suspect that they have been exposed to a COVID19. Test results are delayed: an infected person may test positive for a month or two after contracting COVID19.

          Michael Mina has characterized the rapid antigen as a test of contagiousness. Practically easy to understand.

  13. Rui Viana says:

    I took a rapid antigen test earlier this week, and it was a nose swab. I don’t know if all of the rapid antigen tests are nose swabs, but if you’ve taken a test with a nose swab before, it’s easy to understand why they’re not approved for a home use. Your first thought is: “you want to put all of that up my nose?”. Now imagine doing it to yourself. The most likely outcome is that you won’t put it in far enough and fail to collect a good sample.

  14. A says:

    I’ve been following Mina on twitter for a few months now and I think the overstating he does in the article you linked is consistent with his other communications. I do wish he was more measured but I understand why he isn’t. Paul Romer was another person who was strongly advocating for such tests early on.

    Regarding why other countries haven’t implemented this: I think there’s some correlation with FDA shenanigans and other countries’ equivalents. The FDA seems to be pretty mainstream in that other equivalent agencies behave similarly and with similar principles, particularly in the West. And maybe some countries that *would* implement widespread testing like this just haven’t needed to because transmission rates are so low. If there aren’t many cases in a country then this type of testing is too risky – you’re already close to zero so benefit is capped but it could potentially cause adverse behavior. The calculation is different in a country like the U.S. where the virus is rampant. And finally, for the U.S. the cost is a drop in the bucket but for countries with smaller economies the calculation is perhaps more difficult.

    I do share your concern though Andrew that it does seem weird that this idea isn’t more popular. Besides a few like Mina I haven’t heard epidemiologists banging on the door for this (or arguing against it either). That’s pretty strange. But perhaps they were banging on the door earlier in the pandemic and have just given up since it doesn’t seem possible with the current administration. Or perhaps this is like masks and aerosol transmission which also took a while to get mainstream acceptance but eventually did.

    • A

      What has Michael Mina overstated? I have followed the evolution his perspective. Consistently, Mina has conveyed the advantages and disadvantages of types of tests. I grant that he is enthusiastic about AT HOME tests that doesn’t require necessarily reporting to health sites.

    • Joshua says:

      I would also like to know what you think he has overstated.

      I do wonder about the unseen logistical obstacles of such a massive effort, and I suspect that a lot of people won’t respond optimally to running a test and getting a positive result. But much of life is sub-optimal those aren’t really reasons, imo, not to do this considering the relative ineffectiveness of the status quo.

      • A says:

        I agree that in the U.S. widespread rapid antigen testing should be done. In saying that Mina sometimes overstates his case I think I was overstating my disagreement with him – I generally agree with Mina! And I’m definitely glad that he is vocal about this strategy. I guess my qualm is that he seems to be a bit too optimistic about mass manufacturing, compliance, and general effectiveness.

        Take his pinned tweet for instance:

        Or this recent retweet that says “@michaelmina_lab says rapid antigen tests could control the spread of the virus by Christmas.” (

        Again, I support Mina and this issue. Maybe this unabashed enthusiasm is justified – I’m not an epidemiologist and haven’t gone through the subject super closely. But I’m always skeptical about these statements that look sensationalist. So maybe “overstate” was not the right word to use.

        Regarding your other point, I’m not worried so much about what people do with a positive result. Not everyone will respond optimally to a positive result but enough will that it will be beneficial. The concern is more with a false negative – it’s the “false sense of security” argument. If people get a negative then they may behave in more risky situations. I don’t completely buy this argument (“false sense of security” seems to be a bit of a zombie idea in this pandemic) but I think it’s a good enough reason for countries with low transmission to not implement this testing (i.e. not the U.S.).

    • jrkrideau says:

      Regarding why other countries haven’t implemented this:

      We seem to have been a bit slom getting moving but it appears that Canada is implementing the rapid tests.

      Rapid covid-19 testing in Canada

      Note that in Canada, health is a provincial matter as I believe it is in the USA. It looks like most testing is/will be covered by provincial health plans but there is at least to one private testing place at Montreal’s Trudeau International Airport.

  15. Jonathan (another one) says:

    Alex Tabarrok at Marginal Revolution has been talking about this since April. His most recent post from two weeks ago discusses how much of the rest of the world is finally coming around. And it links to the phrase that has been his mantra since July: “Frequent, fast and cheap is better than sensitive.”

    His simple answer to why this isn’t done is that the FDA sucks. (That’s my phrasing, not his.)

  16. jrkrideau says:

    slow not slom!

  17. “…but then again if it makes so much sense I wonder why they’re not doing it already.”

    One of the saddest things about the past eight months has been the realization that this perspective doesn’t apply: there are many things that make sense, and “they” are limited by glacial inertia, extreme risk-avoidance, innumeracy, and a host of other flaws. Whether its something as simple as making a lot of masks, or more complex like testing as if we meant it, we seem very good at not doing it.

  18. Sanford says:

    Mina makes a lot of very good points but overplays his hand. He’s correct that many people get too caught up in questions of analytic sensitivity when what we really want is clinical sensitivity – antigen tests can probably accurately detect the most infectious people, which is really the goal. But expecting people to reliably test themselves at home and honestly report their results and follow quarantine rules – well, he hasn’t been reading the papers lately. That’s a fatal flaw.

  19. Rick G says:

    Mina and his approach are obviously correct here, and the only thing holding it back is status quo bias. The response of FDA and other authorities is the same as Andrew: “Wow this sounds great but why isn’t anyone else doing it?”. And so (almost) no one does it. This is really frustrating.

  20. Simon says:

    This report says that the proposed testing regime is unlikely to be the silver bullet many think it is:
    The headline is: “Covid-19: Innova lateral flow test is not fit for “test and release” strategy, say experts”

  21. Michael Mina says:

    Andrew – and others.

    Just coming across this blog. You ask “Why aren’t other epidemiologists talking about this”. I’d suggest you pose to your readers a different question: Why is this particular epidemiologist talking about it and not others.

    There’s a simple answer to both – and it is because I am one of the very few infectious disease epidemiologists who studies infectious disease dynamics across populations (based on within-host virus kinetics) who ALSO is a clinical virology lab testing director that understands testing, what is possible, what is doable, what the metrics are.

    Other infectious disease dynamics epidemiologists don’t generally get into the weeds of testing. Most are not also MDs. It’s a niche field. They study vaccines, sure – but not testing. I am one of a very small number. I have my research lab in epidemiology, I run a wet lab in development of high throughput antibody techniques to understand vaccines, and I am a physician that runs a virology lab at a hospital. Those three pieces place me in a unique position to think through strategies that utilize tools, like testing, that other infectious disease epidemiologists are simply not comfortable with. These tests are not something most anyone has ever considered as a public health tool in the past – testing is almost always limited to medicine.

    To your larger skepticism. It’s OK to be skeptical. I’m good with that. Scientists need to be skeptical. But I’d say that pointing out that ” no one else doing it or talking about it” as evidence of anything to suggest that it might not be a very good and powerful idea is not particularly scientific. There’s a reason it was published in the New England Journal and in one of Science’s main journals – and that’s because it’s a new idea. We can’t expect everyone to be pushing something that’s new. But that does erode the utility of the idea.

    The pushback that this approach gets is astounding and generally coming from people who either don’t understand epidemiology and/or virus kinetics, or people who don’t understand testing and how to most appropriately evaluate a test (i.e. it is plain ignorant to evaluate an antigen test against a PCR test – PCR remains positive 5 times as long as there is viable virus / antigen left in the body – yet FDA mandates and most clinical microbiologists expect a virus detection tool like an antigen test to detect remnant RNA – which is remarkably short sighted). If you understand both then you begin to see just how clear the picture becomes. If you pull just 30% of new infections out of circulation daily, you’ll pull back cases rapidly. instead of 100 cases becomeing 130 cases –> 160 –> 200… you have 100 cases become 85 cases –> 70 –> 55. It doesn’t need to be perfect to get outbreaks to decay. It just needs to get R < 1. It's simple math. It doesn't need everyone to perfectly isolate. It just needs a fraction to do it for a few days after a positive antigen test result (not 10 days – just a few days – but to know that you'd have to understand the true transmission windows and their relative weights). If everyone was able to test themselves in a home every few days that would be enough to vastly decrease onward transmission and drop R below 1 (it's only around 1.3 right now).

    So, glad that the blog is here – and I'm glad to see your skepticism – but frankly, we have many voices serving to dilute out actual expertise. It's easy to criticize and be skeptical. Fun even. But we're in a pandemic where we have thousands and thousands of people dying.

    Until people propose a better plan than "lets wait till a vaccine" or until congress appropriates the funds to actually enable the nation to lockdown like some of the other epidemiologists are suggesting we do (which Congress will not do because it would be literally trillions of dollars and cause immense social problems) – perhaps lets go with the avenue that actually could work, that doesn't require lockdowns, that could allow places to open and that would not cost trillions of dollars.

    Or maybe there are better options on the table that I'm not aware of? I don't see any – and frankly I don't see any epidemiologists – or people on this thread – offering better options. But that doesn't mean this idea is wrong. It means too many people are unwilling / unable / or simply haven't thought in this particular direction for a new new avenue.

    People should really read the scientific papers and the op eds I've written – they answer essentially every concern or criticism on the post or the comments. It doesn't take a ton of reduced transmission to reverse upward grown and bring it under control.

    • Thanks for joining this blog which is unique among blogs. I’ll post some of your videos here.

    • bjs12 says:


      I’m not asking why this, that, or another epidemiologist is or isn’t talking about antigen testing. I’m asking why are we obsessing over sensitivity when it is the positive predictive value that matters?

      Is the sensitivity 0.93 or 0.98? Who cares? Let it be 1.0! What is the specificity and what is the prevalence?

      I understand (from the director of testing of a large hospital system) that the Abbott antigen test has a specificity of 0.97 or lower. That’s the nasal swab and clinical lab version. The spit and paper at home version will be lower still. I understand the nationwide prevalence of covid to be 0.05 or lower. That’s for the PCR test. For the subset that is highly contagious, it must be much lower.

      So, best case: sensitivity = 1.0, specificity = 0.97, prevalence = 0.01, then PPV = 0.25. Less good case: sens=1.0, spec=0.9, prev=0.01, then PPV=0.09. The latter scenario is much more likely than the former.

      The (rough numbers) implication of this scenario is that you test 100 people, label 10 of them as positive and only 1 of them is. This is absolutely horrible test performance! I agree that following this procedure and quarantining the 10 positives would likely help slow the spread of the virus. However, it is also true that you could quarantine a random 10% of the population on a rolling basis and it would be almost exactly as effective (and cost $5B less).

    • Ben says:

      > Other infectious disease dynamics epidemiologists don’t generally get into the weeds of testing

      Part of the critique above is in your summary of the tests. In particular:

      > Antigen tests are “contagiousness” tests. They are extremely effective (>98% sensitive compared to the typically used PCR test)

      The critique was this 98% number doesn’t have a citation and doesn’t mix with a citation you give later (

      And now in your response you say:

      > i.e. it is plain ignorant to evaluate an antigen test against a PCR test

      So what is it? You may be making the assumption that the details and particulars of the tests and their rollout just aren’t that important. In that case, just say that’s an assumption. That’s what this sounds like:

      > It doesn’t need to be perfect to get outbreaks to decay. It just needs to get R suggest that it might not be a very good and powerful idea is not particularly scientific

      It’s as scientific as the three opening paragraphs you spend summarizing your resume:×10%E2%88%92264-1-in-a-quadrillion/

      > which Congress will not do

      They did it once before?

      • Ben says:

        Wow I butchered this:

        > > It doesn’t need to be perfect to get outbreaks to decay. It just needs to get R suggest that it might not be a very good and powerful idea is not particularly scientific

        That’s broken into two things:

        > It doesn’t need to be perfect to get outbreaks to decay. It just needs to get R [less than] 1

        That is attached to what is above.

        And then there was something about how this is simple, and I meant to say that sounds great. And mass testing also sounds great. But it wouldn’t bother me if the covid solution was simple, etc. etc.

        And the second half of the line:

        > suggest that it might not be a very good and powerful idea is not particularly scientific

        Is connected to “It’s as scientific as the three opening paragraphs”

        Anyway confusing addendum. I assume some HTML or something ate my comments, but misquoting something isn’t gonna help my argument lol.

  22. sam says:

    Just as the early mask advocates were probably a little overzealous, the rapid test advocacy might be painting too rosy a picture. But I’m honestly baffled why no state health official has dared to run a trial. Antigen tests are established science and at-home tests have worked super well for pregnancy testing. It seems like a no-brainer to at least *try* this. It’s hard for me to comprehend how giving people information about their infectivity could make things *worse* then waiting until people experience symptoms (at which point, they’ve already been contagious for a couple days).

    But, as an experimental scientist, I am aware that the world is a lot messier than our models. So I fully support a large trial—maybe a governor of some state could send antigen tests like the Abbott PanBio ( to millions of its residents and see how the transmission rates change in the real world.

  23. Sam says:

    Yeah, that is a huge barrier. But my understanding (from what Mina and others have said) is that governors have some leeway to do their own thing. Overall, it just seems highly incurious of the FDA and states not to try something.

Leave a Reply