Cheap, fast, surprisingly good Covid testing using scratch-n-sniff


It occurred to me, learning that 80% of Covid infected patients lose their sense of smell (Anosmia) that it should be possible to build the cheapest and most effective Covid screener (for use at entrances to schools/airports/buildings/restaurants) with a simple "scratch-n-sniff" card. These cards, which cost pennies, would come with a set of scratch squares, and under each would be boxes with the names of possible scents. A QR code would (encrypted) have the answer. You sniff, check the boxes and then a phone or other device scans your answer. Much more effective than a thermometer, since probably under 20% of Covid patients have a fever. No close contact with anybody needed. Takes longer but faster to process so no waiting in line. Could even be automated but somebody has to be able to tell the people who fail they can't come in.

A bit of research would discover how many scratches you need for a reliable number. An alternate trick would be to have two rows of the same scents in different orders, draw lines between the matches. Something would work. However:

  • This is a limited utility home test, since 20% false negative -- but then some of the rapid tests are not a lot better.
  • Once you get it, anosmia persists long after the infection, so you need a different test to tell you are clean. You would need to get a card to bypass any anosmia checkpoint.
  • A few people already have anosmia, though it's pretty rare. A more expensive rapid test can be used for them.
  • At home, it's easy to test yourself with anything around the house with a scent, so you don't need the card, though many people who have it don't realize they have it -- only about 50% of Covid patients report it.
  • You are infectious for 1-2 days before you lose your smell, just as is the case with fever.

So mainly this is a better alternative to fever checks. While it uses a consumable it's one that costs a few pennies. Fever checks need a person who comes up close to you. This takes a little longer but many people can do it in parallel, the scan of your card can be done fast by a phone or machine. It could be suitable for high-volume locations, put your card in slot, walk through when green.

There are some other cheap and effective tests, including an AI tool that can analyse coughs to detect Covid. A triple test could measure fever, cough and smell to be even more accurate. Of course, people with a cough should not be trying to enter buildings unless they know it's not Covid.

So, the "scratch and sniff" card industry has been thinking about this, but in a much more limited way, but it does not seem to have taken off.


You are infectious for 1-2 days before you lose your smell

Yes, that is the issue with all use of symptoms to diagnose. However, the same problem exists for temperature and cough checks, and in fact, in practice it exists for even PCR tests, because the vast majority are not given until somebody has symptoms.

It is still very useful to be able to filter out people during their symptomatic phase.

Yes, it's basically all COVID theatre. Even most testing is pointless in stopping the spread, once you have significant community spread. Rapid tests with a high false negative rate are close to useless for that. PCR tests that take 3+ days to give you a result are fairly useless for that. Testing can even make things worse, when a symptomatic individual relies on a false positive to think they must have something else.

Testing can sometimes help with contact tracing, though in most places the number of cases has long exceeded the point where contact tracing is very useful. Testing can sometimes help with treatment, though maybe 90% of cases don't need any treatment. But testing is way way overblown in importance. We've wasted billions of dollars on it that could have been spent more usefully.

And yeah, this smell test is even worse than that. I guess on the plus side, it's relatively inexpensive, assuming you have the test administered by someone who is going to be there anyway, and it provides that false sense of safety that people crave. Combine it with the security theater and I guess the only question is whether it's too esoteric to provide people with that false sense of safety.

What size gathering are you going to be willing to hold or attend if there's a smell test at the door but not willing to hold or attend if there isn't? I just don't see that goldilocks zone.

the vast majority [of PCR tests] are not given until somebody has symptoms

Are you sure about that?

I don't have hard data, but usually you need a reason to do a PCR test (and a motivation to wait in line in many places.) There are more reasons than symptoms of course, including contact with an infected person. And we've seen a growth of people testing in order to have a gathering, certainly before thanksgiving. But data would be good.

However you underestimate the value of low efficacy tests. If you're about to go into a mall or get on a plane with a group of people, your risk is proportional to the number of infectious people with you. Anything at all that detects any infectious people and denies them entry reduces that risk. Even a 20% detector like body temperature. It does not make it safe, but it reduces the risk.

Reducing the risk, in turn, reduces the risk of spread. If you can seriously reduce the risk of spread 5-fold with an 80% test, and you did it everywhere people congregated, that would wipe out the disease, actually by taking R0 well below one.

Note that since people with obvious symptoms often will isolate on their own without a test, you don't get as good as 80% since many have pre-removed themselves. And there is the risk that people who don't understand how it works will treat a negative test as a blessing to go out and socialize.

If you're about to go into a mall or get on a plane with a group of people

I'm not. Not until I get a vaccine.

If you can seriously reduce the risk of spread 5-fold with an 80% test

You can't, for the reason I quoted in my first reply to you.

and you did it everywhere people congregated

Like in every home in the USA?

Note that since people with obvious symptoms often will isolate on their own without a test, you don't get as good as 80%

Or anywhere remotely close to it. People with such an obvious symptom of not being able to smell, which is almost certain to be caused by COVID nowadays, already are isolating themselves.

If loss of smell really is a symptom present anywhere near 80% of the time while people are infectious, it'd explain why R0 is already so much lower than you'd expect from a virus that spreads so easily.

And there is the risk that people who don't understand how it works will treat a negative test as a blessing to go out and socialize.

Surely at some point we have to let the people who lose their sense of smell for months (at least, possibly permanently) out if their houses.

Yes, the general rule is that people are no longer contagious 10 days after symptom offset, even if they continue to have symptoms.

There is a difference between testing to help you understand when you have it, and testing as a social benefit. When a virus has an R0 of 2 (not too far from what this one has) if you can get over 50% of those who are infectious to isolate, you will wipe out the disease. If you can get 80% you wipe it out quickly. We could have avoided all the lockdowns if we had been able to do that early.

All the things that stop spread help -- testing to lock down those who have it, social distancing, masks, and various factors of a lockdown. But because people are not robots, and some will refuse to comply with simple principles -- indeed, they defy them for political reasons -- you have to stop far more than half the infections, and it's the fault of those idiots. Without the people who break the rules, who get lockdown fatigue, who deliberately defy the rules -- we could have gotten through this with low economic cost. And with the President encouraging them, making it a badge of honour -- that man quite probably killed more Americans than anybody else short of the biggies, like the Axis, the Confederacy, the Viet Cong and the Triple Alliance.

In the end I suspect we'll find the reason that places like Japan, Korea, Taiwan, China and even Sweden did better has not to do with what their goverments decreed but with whether the population as a whole understands (or obeys without understanding, but understanding is better) the basic principles.

This virus has an R0 of around 2 under what conditions? Where? What season? What are you basing that on?

I'd say the R0 in a heavily populated city, without any mitigation, in the winter, is at least 5 or 6. Given what we know now about what a small portion of infections were being detected, that's probable about what the R0 was in Wuhan before mitigation started taking place. Maybe even double that. It's hard to say what it would be this winter, though, because we really haven't seen what this virus can do in January in the Northern USA. Will it be worse than March (all else equal)? Not as bad? About the same?

What it is safe to say, I think, is that Re will be much much lower than that.

IF you have a virus with a value of 2 -- and this virus has been estimated in that range under certain conditions -- and you can block over 50% of the transmissions, you've got a success. That's the point, not to say what the rate is in every place. Even if you have a value of 5, and you can block 60%, it can still grow exponentially, just a lot slower, and if you can find other mitigations, then you have a better hope. So any cheap test of decent reliability is useful.

We're already blocking way more than 60% of the transmissions we'd have if we behaved as though there was not a pandemic, so I don't understand what your numbers represent.

Well, that just means that since we're in a period of fairly quick exponential growth, that the "life as normal" R0 is much higher.

I don't know if we're really blocking 60% though. Our surge right now doesn't seem that much less than the pre-lockdown surges in various places. Less, but not a 60% smaller exponent. It really seems to depend on our mood and how much we decide as a group to ignore the rules and get fatigue on the rules. This surge is happening in a lot of countries, so it's not government action or change behind it, it's either weather or fatigue or something else. But since it's happening in north and south, it may not be weather.

Well, that just means that since we're in a period of fairly quick exponential growth, that the "life as normal" R0 is much higher.

That's exactly what I suggested a few messages ago, though I would clarify that it's the "life as normal in highly populated areas of the northern USA in the winter" R0 that is much higher than you were suggesting.

This is a virus that, under the proper conditions, spreads through aerosolized particles. The R0 under those conditions is sure to be high. And while we don't know exactly what those conditions are, there are several examples of spread through ventilation systems. You don't get an R0 of only 2 or 4 with an airborne virus that can be spread by asymptomatic individuals in NYC during the peak spreading season without any mitigation.

Yes, that sentence is a mouthful. One of the problems with even talking about R0 is that it's such an oversimplification to think of it as a single number.

I don't know if we're really blocking 60% though.

I wouldn't be surprised if it's 90%. Maybe even 95%. In fact, given the fact that people in the most densely populated areas are doing the most blocking, it's hard to imagine that it's less than 80% averaged by population.

Our surge right now doesn't seem that much less than the pre-lockdown surges in various places.

What do you consider "pre-lockdown"? How much were we already blocking then? How are you measuring the slope of the curve then, considering that, by definition, we weren't testing at all before we started changing our behavior.

Schools started closing temporarily in mid-February. It's really hard to measure the R0 in January.

But since it's happening in north and south, it may not be weather.

What do you mean it's happening in north and south? Hemispheres?

I think it's crystal clear that weather is a very large factor. Maybe more of an indirect factor (weather causes people to change their behavior), but it's clearly a factor.

If you just are saying that cases are going up in some states in the southern USA, first of all, it is getting colder there, second of all, restrictions have eased there a lot, and third of all, the spikes are much much smaller.

But maybe you're aware of some data that I'm not.

I'm talking about the growth in Wuhan in January, and in Bergamo in Feb-March. Bergamo was the last place in Italy to lockdown, and they held a big sporting event before they realized they needed to lock down.

And yes, I mean southern states vs. northern. Yes, it's getting colder in the south too but to a much lesser degree. Also, the Northeast is not surging the way some warmer places are -- yes, in part because they know better and were burned before. But it's pretty complex.

Little is known about the spread in Wuhan in January. (There was likely also exponential spread in the USA in January, but wedont know a whole lot about that either.)

What was the R0 in Bergamo in February? You agree that it wasn't "life as normal" in Bergamo in February, right?

Northeast is not surging the way some warmer places are

It's worse there, right?

I'm not sure what you're talking about. Growth in cases is worse in New York than Texas or Florida. Daily deaths in Texas doubled between October 18 and November 24. Daily deaths in Florida are down over the same period. Daily deaths in New York have more than tripled over that period. And this despite New York doing much more to slow the spread. Weather is clearly playing a very large role, probably a larger role than government restrictions.

Which warmer places are you thinking of? Maybe Louisiana in an exception. They were an oddity in April too.

Which is sadly a common lament.

Right now though I see heavy case levels in west Texas and New Mexico, light in Maine and Washington, which would run counter to temperature theories. The Euro map shows Spain and Italy worse off than Finland and Norway, but Sweden in real trouble though it is most clearly anomalous as to policy. Scotland is better off than England. Northern Norway better than southern. Germany better than Spain. There are also comparisons which do point to climate as a cause, but the correlation is not so solid, that's all.

Of course there's no single factor that explains everything. Lots of things matter.

Clearly weather is a large factor. Earlier, you said it might not be.

(Incidentally, Texas, Florida, and New York are three of the four most heavily populated states. I excluded California because they spread across such a large number of latitudes. New Mexico, Washington, and Maine are small potatoes in comparison. I wonder if you're focusing on them because the news media is. By the way, cases are surging in Washington, and it's likely going to get much worse over the next month or two.)

Just wondering how much of the surge is that and what is other stuff?

I guess I read too much into "But since it's happening in north and south, it may not be weather."

"about 77% of COVID-19 patients who were directly measured had smell loss, but only 44% said they did, researchers found"

Okay. I take back what I've said. Make a cheap test, get FDA emergency approval (unfortunately that's probably required), and give one to every person in the USA to take every morning.

I'd prefer the self-assessment and voluntary quarantine method. But if you want to also have mandatory tests at high traffic places, that's fine with me, if it's in addition to the self-test-at-home.

300 million people times 200 days times $2/test equals $120 billion. I think it's worth it.


If you seriously think you can create the test for $2 a piece in huge quantities, and you start a Kickstarter (or whatever), I'll personally donate $100. Give me the link. (This offer is only for one Kickstarter that Brad personally endorses.)

Do you think you can go into mass production, with FDA emergency approval, by the end of January?

I'm totally serious. Let's do it. I don't think I'd be of any assistance beyond my $100 donation, but if you think I might, and you really want to do this, let me know and I'll email you.

We (you) just need a proof of concept, and I think a Kickstarter would be able to fund that. I'm sure we'll (you'll) get grants from there to figure out how to produce in the tens to hundreds of millions (plus legal advice on how to get FDA emergency approval, contracts for distribution, etc.). With the ultimate funding coming from Congress once we (you) get FDA emergency approval.

Have you heard about the Corowell Pass rapid smell test? Soon it will be available in the USA. We are already in mass production in many countries. You can learn more by visiting our site

Corowell Tickets are low-cost plus you get a reliable electronic pass via the Corowell Pass app when you complete the smell test. The Corowell Pass can be scanned and validated by the Corowell Guard app, it only takes a second.

Some more info below:

Corowell offers an easy, rapid & low-cost way to objectively identify suspected COVID-19 infections by testing the sense of smell.

Rapid Mass Testing
Corowell’s accurate screening method is non-invasive, suitable for mass-markets and can be performed in seconds on site at the airport for example without additional infrastructure.

Detect Early Symptoms Accurately
The Corowell Symptom test can identify with high accuracy (~85%) real time infected people by diagnosing the COVID-19 caused symptom of Anosmia. Anosmia is considered as one the earliest COVID-19 symptoms and usually occurring anytime between 1-3 days after infection.

Corowell tests are low-cost compared to PCR, antibody and antigen tests. Furthermore the test can be self-administered and taken anytime and anywhere, at home for example.

Add new comment