In the setting of COVID-19, almost every country in the world closed its borders, locked down its citizens, and forced businesses to close. Today, most governments still restrict travel, economic activity, and social gatherings.
The justification for these unprecedented measures has been a growing number of COVID-19 cases. This has unleashed an epidemic of COVID testing – with PCR and rapid antigen tests as the means of identifying positive COVID cases. Our very own Dr. John Hunt examines the science behind COVID testing, whether the testing paradigms are effective, and the rationality behind government response to the virus.
What COVID tests mean and don’t mean
RT-PCR tests can be designed to be highly sensitive to the presence of the original viral RNA in a clinical sample. But a highly sensitive test risks poor specificity for actual infectious disease.
Rapid antigen tests are different. They measure viral protein. They do so by reacting a clinical sample with one or two lab-created antibodies that are labeled with a measurable marker. These antigen tests are often poorly specific, meaning they can show as positive in the absence of any actual viral protein or any COVID disease.
For a lab test, what does it mean to be sensitive? What does it mean to be specific?
I’ll use COVID to help explain these terms. In order to do this correctly, we need to avoid using the language of the media and government because those institutions tend to mislead us via language manipulation. For example, they’ve wrongly taught us that a COVID-positive test is synonymous with COVID- disease. It isn’t, as you will soon see.
So for this article, I will use the term “Relevant Infectious COVID Disease” to mean a condition, caused by COVID-19, in which a patient is sickened by the virus or has (in their airways) living replicating virus capable of being transmitted to others. This seems a fair definition of what we should be caring about in this disease. If the patient isn’t sick and isn’t capable of transmitting the disease, then any COVID RNA or protein that may appear in a test is not relevant, nor infectious, and therefore of little to no consequence.
You can think of a test’s sensitivity like this: In a group of 100 people who absolutely have Relevant Infectious COVID Disease, how many people does the test actually report as “positive?” For a test that is 95% sensitive, 95 of these 100 patients with the true disease will be reported by the test as COVID positive and 5 will be missed.
Specificity: In a group of 100 people who absolutely do not have Relevant Infectious COVID Disease, how many will be reported by the test as “negative?” For a test that is 95% specific, 95 of these healthy people will be reported as COVID-negative and 5 will be incorrectly reported as COVID-positive
Sensitivity and Specificity are inherent characteristics of a test, not of a patient, not of a disease, and not of a population. These terms are very different than Positive Predictive Value (PPV) and Negative Predictive Value (NPV). PPV and NPV are affected not only by the test’s sensitivity and specificity but also by the characteristics of the people chosen to be tested and, particularly, the patients’ underlying likelihood of actually having true Relevant Infectious COVID Disease. The Positive Predictive Value—the chance a positive test actually indicates a true disease—is greatly improved if you test people who are likely to have COVID, and, importantly, avoid testing people unlikely to have COVID.
If you do a COVID test with 95% sensitivity and 95% specificity in 1,000 patients who are feverish, have snot pouring out of their noses, are coughing profusely, and are short of breath, then you are using that test as a diagnostic test in people who currently have a reasonable up-front chance of having Relevant Infectious COVID Disease. Let’s say 500 of them do actually have Relevant Infectious COVID Disease, and the others have a common cold. This 95% sensitive test will correctly identify 475 of these people who are truly ill with COVID as being COVID-positive, and it will miss 25 of them. This same test is also 95% specific, which means it will falsely label 25 of the 500 non-COVID patients as COVID-positive. Although the test isn’t perfect it has a Positive Predictive Value of 95% in this group of people, and is a pretty good test overall.
But what if you run this very same COVID test on everyone in the population? Let’s guesstimate that the up-front chance of having Relevant Infectious COVID in the US at this moment is about 0.5% (suggesting that 5 out of 1000 people currently have the actual transmittable disease right now, which is a high estimate). How does this same 95% sensitive/95% specific test work in this screening setting? The good news is that this test will likely identify the 5 people out of every 1000 with Relevant Infectious COVID! Yay! The bad news is that, out of every 1000 people, it will also falsely label 50 people as COVID-positive who don’t have Relevant Infectious COVID. Out of 55 people with positive tests in each group of 1000 people, 5 actually have the disease. 50 of the tests are false positives. With a Positive Predictive Value of only 9%, one could say that’s a pretty lousy test. It’s far lousier if you test only people with no symptoms (such as screening a school, jobsite, or college), in whom the up-front likelihood of having Relevant Infectious COVID Disease is substantially lower.
The very same test that is pretty good when testing people who are actually ill or at risk is lousy when screening people who aren’t.
In the first scenario (with symptoms), the test is being used correctly for diagnosis. In the second scenario (no symptoms), the test is being used wrongly for screening.
A diagnostic test is used to diagnose a patient the doctor thinks has a reasonable chance of having the disease (having symptoms like fever, cough, a snotty nose, and shortness of breath during a viral season).
A screening test is used to check for the presence of a disease in a person without symptoms and no heightened risk of having the disease.
A screening test may be appropriate to use when it has very high specificity (99% or more), when the prevalence of the disease in the population is pretty high, and when there is something we can do about the disease if we identify it. However, if the prevalence of a disease is low (as is the case for Relevant Infectious COVID) and the test isn’t adequately specific (as is the case with PCR and rapid antigen tests for the COVID virus), then using such a test as a screening measure in healthy people is forcing the test to be lousy. The more it is used wrongly, the more misinformation ensues.
Our health authorities are recommending more testing of asymptomatic people. In other words, they are encouraging the wrong and lousy application of these tests. Our health officials are doing what a first-year medical student should know better than to do. It’s enough of a concerning error that it leaves two likely conclusions:
1) that our leading government health officials are truly incompetent and/or
2) that we, as a nation, are being intentionally gaslighted/manipulated. Or it could be both.
(Another conclusion you should consider is that my analysis of these tests is incorrect. I’m open to a challenge.)
So what if you, as an individual, get a positive PCR test result (one that has 95% specificity) without having symptoms of COVID-19 or recent exposure to a true Relevant Infectious COVID Disease patient? What do you do? Well, with that positive test, your risk of having COVID has just increased from less than 5 in 1,000 (the general population risk) to about somewhere perhaps 5 in 55 (the risk of actual Relevant Infectious COVID Disease in asymptomatic people with a COVID-19-positive test). That’s an 18-fold increase in risk, amounting to a 9% risk of you having Relevant Infectious COVID Disease (or a 91% chance of you being totally healthy). That may be a relevant increase in risk in your mind, enough that you choose to avoid exposing your friends and family to your higher risk compared to the general population. But if the government spends resources to contact-trace you, then they are contact-tracing 91% of people uselessly. And they are deciding whether to lock us down based on the wrong notion that COVID-positive tests in healthy people are epidemiologically accurate when indeed they are mostly wrong.
For the 50 asymptomatic low-risk people falsely popping positive out of each group of 1,000, what makes them pop positive? For a rapid antigen test, it is because the test is never meant for use as a screening test in healthy asymptomatic people because it’s not specific enough. For a PCR test, positivity confidently means that there was COVID RNA in that sample, sure, but your nose or mouth very likely just filtered some dead bits of viral debris from the dust particles in the air as you walked through CVS to get the test before you learned you were supposed to use the drive-through. PCR can be way too sensitive.
A few strands of RNA are irrelevant. Even a few hundred fully intact viral particles are not likely to infect or cause disease. Humans aren’t that wimpy. But keep in mind that there is a very small chance that the test popped positive because you are about to get sick with COVID-19, and the test caught you, by pure luck, just before you are to become sick.
On top of this wrong use of diagnostic tests as screening tests, the government has been subsidizing hospitals for taking care of COVID-19-positive patients. Let’s say a hospital performs a COVID test 4 times during a hospital stay as a screening test in a patient who has no symptoms of COVID. If that test pops positive once and negative three times, the hospital will report that patient as having COVID-19, even though the one positive result is highly likely to have been a false positive. Why do hospitals do this testing so much? In part, because they’ll get $14,000 more from the government for each patient they declare has COVID-19.
When we see statistics of COVID-19 deaths, we should recognize that some substantial percentage of them should be called “Deaths with a COVID-19-positive test.” When we see reports of case numbers rising, we should know that they are defining “case” as anyone with a COVID-19-positive test, which, as you might now realize, is really a garbage number.
- We have an epidemic of COVID-positive tests that is substantially larger than the epidemic of identified Relevant Infectious COVID Disease. In contrast, people with actual, mild cases of COVID-disease aren’t all getting tested. So the data, on which lockdowns are supposedly justified, are lousy.
- The data on COVID hospitalizations and deaths in the US are exaggerated by a government subsidization scheme that incentivizes the improper use of tests in people without particular risk of the disease.
- Avoid getting tested for COVID unless you are symptomatic yourself, have had exposure to someone who was both symptomatic and tested positive for COVID, or have some other personal reason that makes sense.
- Know that getting tested before traveling abroad puts you at a modest risk of getting a false-positive test result, which will assuredly screw up your trip. It’s a new political risk of travel.
- There is a lot more to this viral testing game, and there are a lot of weird incentives. There are gray areas and room for debate.
- Yes, the COVID disease can kill people. But a positive test won’t kill anybody. Sadly, every COVID-positive test empowers those politicians and bureaucrats who have a natural bent to control people—the sociopaths and their ilk.