COVID 19 Update: Is More Testing the Solution?

*I will try to stick to facts as I understand them.  Any opinions I express are not necessarily those of the leadership of my medical group or profession.

Almost from the beginning of the COVID 19 crisis, the subject of testing has been a flash point between the federal government and the states as well as between the government and the news media.  I believe almost everybody agrees test availability was woefully lacking in the early days of the crisis.  Availability improved considerably, but still falls way short of "enough for anyone who wants it" that some seem to be demanding. But is "testing everyone" realistically achievable and is is even desirable?  In my opinion, much of the confusion and misunderstanding stems from a basic lack of understanding of what testing tells us and what it doesn't and more importantly, what we do with test results after we have them.  So, I thought I would start with the basics of laboratory testing to help folks better understand the issues at stake here.

I understand that in law schools they teach students to never ask a question they don't already know the answer to.  In medical schools, we teach students to never order a test if they don't know what action they will take or not take when they know the results.  Will a positive test make you initiate a medication or treatment or additional testing?  Will a negative test lead you to withhold or stop treatment?  If we do not know the answer to these questions, we should not be ordering the test in the first place!  Before we can make decisions about test results, we need to understand several things about the accuracy of the tests we order.  Here are the key terms:

Sensitivity(true positive rate):  The sensitivity of a test is the probability that a patient who has the condition in question will have a positive test.  Let's say you have a group of 100 patients and 10 of them have disease X.  A perfectly sensitive test will be positive in all 10 of the patients who have the disease.  Now, maybe the test is also positive in 10 patients who don't have the disease(false positive), but that does not diminish the sensitivity of the test because you know that all 10 patients with the disease you are looking for are in that group of 20 with positive tests.  You can now tell the other 80 patients they don't have the disease and concentrate on the 20 that are left.  Highly sensitive tests are generally useful as screening tests with patients testing positive undergoing additional testing to identify the true positives from the false positives.

Specificity(true negative rate): The specificity of a test is the probability that a patient who does not have the disease will have a negative test.  If a test is perfectly specific then 100% of patients who don't have the disease will have a negative test. Let's say in our sample population you run a test with 7 positives and 93 negatives.  All 7 of the patients who test positive have the disease because all 90 of the patients who don't have the disease are in the 93 who test negative.  Highly specific tests are generally useful for confirming the presence of the disease you are testing for.

Precision:  Precision is the probability that when you perform the same test on the same person multiple times, you will get the same result or in quantitative tests, how close you get to the same result on repeated testing.

Positive Predictive Value:  Positive Predictive Value(PPV) is the probability that subjects with a positive screening test truly have the disease.  This value is heavily influenced by the composition of the population you are testing.  Let's take our sample population again where you have 10 patients with the disease and 90 who don't.  But, the test you are running has a 30% false positive rate and a zero false negative rate.  All 10 patients with the disease will test true positive but 30% of the other 90 who don't have the disease will also test positive.  Out of the total of 37 patients who test positive there is only a 10/37 or 27% chance that any one of those patients truly has the disease and a 73% chance the test is falsely positive.

Negative Predictive Value:  Negative Predictive Value(NPV) is the probability that a patient with a negative test does not have the disease.

*********************************************************************************

There are two basic types of tests for COVID 19:  1.  Tests that directly test for the current presence of the virus.  2.  Test for antibodies to the virus which can be subdivided into IgM(current or recent infection) and IgG(past infection).

Tests for the current presence of the virus are highly specific but have a 5-30% false negative rate(they are much less sensitive). False negatives can be due to flaws in collection methods, improper handling of the specimen or faulty reagents. If you strongly suspect a patient has COVID based on symptoms and signs but the test is negative, it is wise to run a second test.  There have been reports of patients with classic symptoms and clinical course who did not test positive until the 3'rd test was run!  Assuming the results of the first test do not bias the results of the second and so on, by the time you have 3 negative tests, each with a 30% false negative rate, you can whittle down your false negatives to under 3%.

The major problem that has emerged with antibody tests is they are reported to have 7-30% false positive rates depending on which of many brands of tests are being used.  If only 4% of the population has or had the disease and the test has a false positive rate of 30%, the test has an extremely low PPV and an almost 90% chance that any one positive test is a false positive.