The Risks of At-Home Covid Tests

Aug 22, 2021

For the first time, there's widespread availability of over-the-counter Covid tests. The primary being the Abbot BinaxNOW antigen test.

This is also the first time most people have run their own at-home medical tests beyond perhaps a pregnancy test. As a result, there are some potentially dangerous fallacies about these tests and what the test results actually mean.

Key Takeaways for Home Covid Tests:

If the test is Positive you are absolutely infected!
Less than 2 tests per 1000 will return a false-positive1
There is NO reason to "confirm" the test-positive with a PCR test
If the test is Negative there's still a high chance you're actually infected2
Up to 680 tests per 1000 will return a false-negative if asymptomatic3
Up to 250 tests per 1000 will return a false-negative if fully symptomatic.4
These tests generally only work on people who are fully symptomatic
If exposed, Test-negatives should be confirmed with a laboratory-run PCR test5
These tests DO NOT tell you whether you're infectious!
There is NO test for "infectiousness" in existence
These tests are primarily useful for a confirmatory diagnosis on an already fully symptomatic patient

To Summarize:

If you have Covid-like symptoms and the test comes back POSITIVE, then you absolutely have Covid!
If you have Covid-like symptoms and the test comes back NEGATIVE, then you very likely have Covid but must confirm with a full laboratory-run RT-PCR test.
If you have been exposed, but are asymptomatic, DO NOT use this test, quarantine, and go to a clinic for a proper laboratory-run RT-PCR Test.

If you're using these Home Covid Tests in ANY other manner, then you're dangerously misusing the test!

Here are CDC6 guidelines for how these tests are to be used and not used:

The CDC recommends their use primarily in large congregant settings such as nursing homes, large warehouse distribution centers, schools, etc. NOT for keeping Covid out, for the sole purpose of slowing the rate of infection through those populations by isolating and quarantining people earlier during the infection cycle after known exposure through repeated daily testing.

These Tests Cannot be Used to Decide:

Whether the people coming to your party are "infectious" – this is how superspreader events happen!
Whether your child can go to school today – if Jr. has symptoms or has been exposed, then assume it's Covid until an actual doctor says otherwise
Whether or not you are "infectious" – these tests only register a reliable test-positive for patients with a Ct-count BELOW 20! So yes...you ARE infectious!

So if your kid has so much as a sniffle, don't misuse these tests as an excuse to send them to school. Get them to a real doctor for a proper test.

And please throw away the notion of using these tests for next your dinner party or birthday party or dance party!

If someone is sick and coughing they shouldn't be showing up at your party in the first place. If someone is infectious but asymptomatic, these tests will are very unlikely to reveal that.

These Home Covid Tests will provide no useful information that will allow an inherently risky event to be any less risky.

So if you're throwing a party, there's really only one scenario where these tests have utility. If you've already decided what you're going to do and just need some emotional justification for your dangerous risk-related decision, then just having the tests around can certainly make everyone feel much better about themselves!

And in that case, you probably don't even need to waste the tests by actually using them...just assemble them into a tasteful arrangement on the coffee table.

THE TECHNICAL DETAILS

Below, I will go into the reasons why these things are true and the source of these limitations. It does require some minimal level of understanding to be able to safely use these tests...

What Is A “Test”?

Before you can understand how to use a medical test, you need to understand a few things about these tests in general, including the Home Covid Test kits.

The first thing to understand is what a test actually is. A test is just a way to measure something. That "something" is generally NOT anything you're directly interested in, but it should be a reliable predictor of the information you're actually interested in.

You don't take a Covid Test to learn about how many viral fragments were on a nasal swab (what the test actually attempts to measure). What you want to know is whether you're INFECTED.

Unfortunately, what a test measures vs. what you want to know doesn’t always line up!

For example, the speedometer in your car is just a proxy for how fast your car is going. But the speedometer is just a needle that turns around a dial pointing at numbers...it's not "actual" velocity. And if your speedometer is misaligned and measuring 20 mph too slow, your car is STILL going at 80 mph even if the speedometer says 60 mph.

Medical tests are no different.

So, in a perfect world, the tests are flawless and it’ll find whatever is there, always returning a Test-Positive…and if nothing is there, it always returning a Test-Negative.

Unfortunately, that's a cartoon-level understanding of how this actually works! For a variety of often complex reasons, tests can fail to return results that accurately reflect the state of the condition you're interested in. In this case, whether someone is infected with the Covid virus or not.

How do You Test the Tests?

How good is one test versus another? How are they measured and compared?

You see people blithely saying things like "this test is 97% accurate". This is always a good indicator they don't have a clue what they're talking about.

First, when it comes to medical tests, there is no such thing as the word "accuracy". If you see someone using this term, then you can assume they do not understand how tests are evaluated nor how the test results should be properly evaluated.

All molecular tests (including Covid tests) are defined by three key measures:

LOD (Limit of Detection) – how much virus do you need for the test to work
Specificity – how sure are you the test found what it was looking for
Sensitivity – how sure are you the test didn’t miss what it was looking for

These can generally be reduced to a couple of important metrics:

False-Positive Rate – based on “Specificity”, it's an estimate of the fraction of Test-Positive results that are incorrect (meaning you're NOT infected, even though the test says “Positive”)
False-Negative Rate – based on “Sensitivity”, it's an estimate of the fraction of Test-Negative results that are incorrect (meaning you ARE infected, even though the test says “Negative”)

In general, the False-Positive rates for all such tests are EXTREMELY LOW. This is really never something you need to be concerned about.

As an aside...there's been a LOT of nonsense and outright lies ⁠about Covid tests having issues with false positives. This is complete fabrications being passed around by Covid-deniers. Anyone who is actually involved with designing and developing these tests knows this is never an issue with these sorts of molecular tests.

But there are a LOT of issues with False-Negatives and this is what can lead you astray in interpreting these test results! The rate of false negatives is directly tied to the particular type of test and how rapidly the test returns results (hint: faster = worse).

Types of COVID Tests

There are a variety of tests to determine whether you're infected with the SarsCov2 Covid virus. Each with various tradeoffs. But the key thing to understand is that there is NO FREE LUNCH when it comes to testing.

You cannot get a cheap fast test without sacrificing something critical!

And in general, what you're sacrificing is test Sensitivity. Meaning the faster your time-to-result, the higher the rate of false negatives. The faster the test, the more likely it will say you are NOT infected when you really ARE infected.

This is what led to multiple superspreader events at the White House last year. They were using a so-called 20-minute "Rapid Test" (the Abbott IDNow) and the false negatives were in the range of 30% to 40% percent. As result, they kept letting infected people into events.

Here's a summary of the major types of tests for SarsCoV2 infection and how they compare in these key metrics.

Table 1: Test Performance

Table 2: Examples & Usage

So if you want an answer in 10-minutes (which the Home Covid Tests promise), then you're going to have to give up a LOT. This is the problem with ALL rapid tests, faster results are uncertain results.

And the Home Covid Tests are at the very bottom of that pile!

Antigen Tests

The home Covid tests, like the BinaxNOW, are antigen tests. Meaning they're looking for protein fragments from the viral nucleocapsid. The shell containing the viral RNA.

All other molecular tests are what are called "nucleic acid amplification tests” (NAAT). Meaning they're looking directly for segments of viral RNA. Pieces of the virus's genetic material.

Antigen tests perform poorly because there's no way to amplify (makes millions of copies) of proteins during a test. With an RNA test, there's a chemical process called “polymerase-chain-reaction gene amplification” or PCR for making billions of copies of a specific target sequence so it becomes a great deal easier to reliably detect.

With a protein-based test, all you can do is try to purify away molecules that might get in the way of detecting the target protein. But all you have to work with is some fraction of the amount of protein that was first collected on the swab.

As a result, antigen tests require a MASSIVE viral load before being able to detect anything and reliably trigger a test-positive result. So a very high LOD or Limit-of-Detection.

This is why these sorts of lateral flow-strip antigen tests ONLY work reliably on people who are fully symptomatic within the first 5-days of symptoms. Before then, there's generally not enough virus present. After 5-days, the virus is being actively eliminated from the nose and throat and there's not enough virus.

This is also why you can be quite infectious to everyone around you even if symptomatic, especially with something like the Delta variant, and these tests can still return a false test-negative.

Again...there are NO tests for "infectiousness". The ONLY thing these tests can do is to potentially determine whether or not you're infected.

But the Package Insert Says Something Different!

There can be signficant differences between how these tests actually perform in the real world in the hands of untrained users and how they perform during FDA validation trials.

The validation data submitted to the FDA for getting these tests approved are collected under highly specialized and controlled conditions. Generally, the test subjects are chosen to give the most reliable test-positive (previously diagnosed fully symptomatic patients) or test-negative (fully asymptomatic patients with no know exposures).

And the sample collection, sample prep, and test interpretation are done by Ph.D.'s or at a minimum highly trained medical staff under close supervision by scientists.

The validation trials are meant to determine how well the test, in isolation, will perform under absolutely perfect and ideal conditions. It represents the far upper range of potential performance of the test in isolation from human factors.

Your Mileage May Differ…

And that's a perfect comparison. EPA Offical Average Highway Ratings are absurdly high. I'm pretty sure no one gets 25mpg in their F-150 pickup. Real-world mileage is likely closer to 10-12 mpg.

That's because the mileage tests are done under extremely contrived conditions that represent an ideal upper limit to engine performance, but have little or no bearing on how things will work in everyday usage in the hands of a normal driver.

FDA data is no different. And it's NOT that the FDA processes are somehow flawed...they are not! This sort of data is very important for comparing the relative effectiveness of various tests, for understand the core performance of the test in question and for understanding inherent limitations with different tests.

What's flawed are people’s assumptions about what that data means. It doesn't mean what you think it means!

Conclusion

These at-home Covid Tests have limited usefulness beyond confirming that what you’re in bed sick with is Covid versus a cold or the flu.

Their optimal use is for daily screening of large closed populations at a very low cost.

Screenings NOT to diagnose Covid infections, but to slow the rapid spread of infection through a closed population in places like nursing homes and massive warehouse distribution facilities by forcing early quarantining of exposed individuals.

As a good indication of these limitations, I know of NO biotech companies (who would best understand these tests) that are using these to screen their regular employees. They all rely on normal laboratory-run RT-PCR tests instead.

So definitely use these tests…they’re a valuable additional resource. But everywhere, I see people grossly misusing these tests out of ignorance of how they work and where they are and are not applicable.

Responsible use is knowledgeable use!

References:

https://www.medrxiv.org/content/10.1101/2021.08.04.21261609v1

https://pubmed.ncbi.nlm.nih.gov/33846714/

https://www.medrxiv.org/content/10.1101/2021.08.04.21261609v1

https://www.cdc.gov/coronavirus/2019-ncov/lab/resources/antigen-tests-guidelines.html

About the author:

I am a scientist with 20+ years in the biotech industry. I currently work as a consultant with companies involved in developing molecular diagnostics platforms, including some of the key testing platforms used to detect the SarsCoV2 virus.

So I bring an insider’s perspective that is scientifically oriented but directed to a general audience trying to make sense of the conflicting stories surrounding the Covid pandemic.

To learn more: https://www.linkedin.com/in/dalewharrison/

If you found this article useful, please sign up by hitting the “Subscribe Now” button below or share it with a friend…

Prior Posts...

Covid Updates

Discussion about this post