A “secret” simple way to detect the coronavirus without a test for everyone

5 min readApr 13, 2020

There is a “secret” simple way to detect the coronavirus without a test for everyone. The secret is math. It uses math and is based on balance puzzles or the counterfeit coin puzzle; puzzles that most programmers or math teachers are familiar with.

This is related to two other posts: here and here.

Secret #1: Pooled-testing (AKA batch testing)

Used by Stanford Medicine for early detection:

Stanford researchers did it in Jan-Feb when they had limited tests available. They wanted to see if COVID-19 cases were in the bay area at that time so they took the ~3000 patient samples and combined them in sets of 9 or 10. Of the 292 groups, 2 were positive so they further tested the ones in those groups.
SOURCE: http://med.stanford.edu/news/all-news/2020/04/testing-pooled-samples-to-track-early-spread-of-virus.html

Testing pooled samples for COVID-19 helps Stanford researchers track early viral spread in Bay Area

The prevalence of the virus that causes COVID-19 was tracked during the early days of the pandemic by pooling samples…

med.stanford.edu

Used by Germany for detection/monitoring of nurses. They call it BLOCK TESTING.

Medical staff, at particular risk of contracting and spreading the virus, are regularly tested. To streamline the procedure, some hospitals have started doing block tests, using the swabs of 10 employees, and following up with individual tests only if there is a positive result. 
SOURCE: https://www.nytimes.com/2020/04/04/world/europe/germany-coronavirus-death-rate.html

Here’s a description.

Gather a group of 10 people unlikely to have infection (<1% likely, maybe includes doctors/nurses/HCPs)
Do swabs of all 10 people.
Do the RT-PCR test on all 10 people together.
This won’t tell you which person has the virus, but if the result is negative, it means that all 10 people are negative at that time*. Very useful for detection.

If everyone is negative, this means only 1 test is used. If only one person is positive, you can run pooled batches of size 2. That also saves tests.

You can do the math better and devise better strategies. Ask any mathematician. (Update: I asked a mathematician at U. of Wisconsin and he said the problem isn’t trivial.)
You don’t save tests if you do pooled testing if the 10 people are likely to be positive. (For example: If the 10 people are each 50% likely to be positive, 10 individual tests is the way to test. Doing pooling doesn’t save tests; it actually uses more tests.)

UPDATE 4/13: According to a redditor: `Yes this is being done in Michigan at the Henry Ford Hospital system as well as others I’m sure.` But, I haven’t been able to track down a published source.

Tip1: There is a way to use this to reopen schools.

Did you know:

Currently, the rule of thumb according to ER nurses is that there is a 30% false negative rate for the RT-PCR test. This means that if 100 who have Covid19 are tested, then 30 of them are expected to come back with a negative result!

Microbiotically, it is all about viral load in the nasopharyngeal swab area, which varies over time. (Academic source)

Now that I’ve got your attention, there is an even more clever way to detect coronavirus using no test at all.

Secret #2: Detection Pods

Use batch-incubating AKA “detection pods”. Please see the hyperbolic example below.

Gather a group of 10 people unlikely to have infection (<1% likely, don’t use doctors/nurses/HCPs)
Have them interact in a way that they would infect each other. (Example: group living, shared meals, partying). For illustrative purposes, assume that the infection rate is perfectly high, 100%.
If anybody has the virus at the start, now everyone in the “pod” has it. Isolate the group for the 14 (or 10) day incubation period to let symptoms develop.
You actually don’t have to wait all 14days. In fact, at the first sign of symptoms, you should test 1 person. And, with that 1 test, if the infection rate is 100%, then you know everyone has it. In this case: 10 detections with one test.
Even if each person has a 25% chance of being asymptomatic, we can use math to check how likely all 10 people are asymptomatic.

Assuming statistical independence (“iid”), there is only a 1 in 1Million chance that they are all asymptomatic. This is acceptable risk. (Technically, it is 0.953 out of 1Million or 0.25¹⁰ of a chance.)

Accidental detection pods: church choirs, cruise ships, group homes, kindergarten schools if the kids physically interact.

Now, we have already “accidentally” run this detection pod experiment in the wild. Schools incubate and promote the spread of the virus. Prisons, cruise ships, and nursing homes too. The infection rate isn’t 100%, so not everyone gets it. But the “detection” idea is the same. If you amass a group of people, you will more easily detect the symptoms even if you don’t have a test.

Cruise Ships are accidental detection pods
This is, in fact, what happened on the cruise ship. Before they got tests on the ships, the high incidence of the symptoms in the accidental detection pod was enough to make it likely that someone there had the virus. Other cruise ships didn’t have the virus, and you could tell because that accidental detection pod showed no symptoms.

Counterintuitively, if you limit the spread (which is a good thing for health), you decrease the ability to detect it in a pod.

Here’s the example: If you have 500 people on a cruise ship who are perfectly quarantined, no symptoms doesn’t mean nobody sick. Because of asymptomatic carriers, there could be a good chance that ship IS a carrier.

However, have the same 500 people and let them party and physically connect with their microspit, then no symptoms likely means nobody is sick. Otherwise, you’d have shipwide infection.

PRO-TIP: This idea partly came from reading the US military idea of “safety bubbles” reported on here: https://www.defenseone.com/threats/2020/04/exclusive-us-army-wants-train-hundreds-soldiers-coronavirus-safety-bubbles/164326/

I’m re-purposing the idea of “safety bubbles” to become “detection pods”

Details for nerds (sorry, the writing below is less clear to general readers)

Importantly, these people would need to not interact/expose themselves to other people for the 14 day incubation period.

If someone in the pod is sick, then the outside people will be infected.

and

If an outsider interacts with the pod, then we wouldn’t be able to know if the infection came from within the pod or from the outside.

You would NOT do this with people who were suspected of being sick. For them you quarantine OR test or batch-test (see Secret#1).

This actually forms the basis of a plan: “28 hard days to reopen America” that I’ll be writing up soon. By batch-testing AND incubating/detectionPod’ing people, we can be very certain at the end of 14 days who does/doesn’t have the virus. Double that to 28 days, and you’re even more certain.

ENGINEERS: Engineers will recognize the image I selected as a “MUXing” approach.We are multiplexing the testing and also doing signal amplication of sorts.