Understanding COVID-19 Testing: Sensitivity vs. Specificity

You. It's great to be with everybody again. So this is going to be a brief video talking about the testing itself, how reliable, dependable it is, what is sensitivity and specificity? A lot of questions have been fielded to us through social media and a lot of the questions right now are revolving around this. How do I interpret my test results once I get them and what do they actually mean in relationship to if I had disease, can I get infected again, etc. So please keep on putting these questions out there. We may end up doing a live question and answer solely around the testing itself. Reliability, prevalence, data, etc. These are very statistical mathematical concepts and so it's kind of hard for some people to wrap their brains around it. So maybe doing a question and answer where I answer the same questions in different ways might be helpful as well for a lot of people to just grasp better what sensitivity, specificity, true positives, false positives, what these things actually mean. So this is going to be a brief overview, maybe start putting those questions, fielding those questions and we can collect them and then do a live Q and A and just answer some of the questions around the testing itself. I feel like this is going to be the next big topic. Reason being is in the Richmond area in the last week, two of the major lab vendors, LabCorp Quest, just released antibody testing to the general practitioners. Most aren't doing it, some are doing it, but this is going to be a bigger topic as people have the availability to test themselves. So with COVID 19 testing first, what are we actually testing? What are we looking for? Well, there's two major tests. One's the RT reverse transcriptase PCR that's looking for parts of the virus, viral rna. What we know from that test is depending on when you do the test. If you do it on day one to three of the illness, it's pretty accurate, over 90% accurate. But people are being told to shelter in place, so you're probably not going to get a lot of people earlier on in illness, especially since it can take about six days for people to have significant symptoms and six to seven days before they feel like they need to present. About day six, interesting enough, the accuracy drops to less than 80% and by day 14 the accuracy is less than 50%. So with the PCR test you have to have someone who's sick and you have to get them in a certain stage of the illness as well. It depends on where the sample comes from. If someone has disease, the going and getting lung samples is the most accurate. It gets about 95% accuracy. The problem is that's usually a hospitalized patient, it's usually someone who's sick, and they're sticking a tube down your nose to get lung samples. The spit, it's about 72% sensitive as compared to the oral pharyngeal swabs, they're much better, but they, the oral is not as good as a nasal, but again, it depends on when you actually get that test done. The antibody testing, there's different kinds of antibody testing. There's IgG, IgM and IgA, and these are basically looking for. The IgG are looking for different parts of the, of the virus. There are ones to the spike proteins and different parts of the spike proteins, which are those little spiky things that actually attach to cell walls. And there's parts of the receptor that actually binds to the ACE2. So, and then there's the IgM, which tells you, you know, is this acute versus not acute? And IgA looks more at your body's mucosal membranes, GI tract, reaction to the infection. And that's what most stuff is revolving around is the antibody testing. And specifically an igg, just a single igg, which is what most tests are. Currently there are a few places offering more advanced testing, the full spectrum, but I do not know of anybody in central Virginia. I only know one person in the country personally that's doing that, and they're out in the Colorado area. So with these tests, we have to think about when we get the results back. What does sensitivity versus specificity mean? So when I tell you, if you get your test done for the COVID 19 and your sensitivity is 87%, what does that mean? Well, sensitivity is basically the true positive rate. If I say it's 87% sensitive, that means 80% of the time if you have disease, it actually tests positive, which means 13% of the time, if you actually have disease, it tests negative. So that's the thing you have to think about. It's a 13% false positive positive rate with this test. That's sensitivity. Specificity is a true negative rate. If you don't have disease, what are the odds that the test will test negative, True, negative. Now, the thing about it, and the thoughts are it's what percentage of healthy people are correctly identified as not having the disease? That's your specificity, sensitivity. What percentage of actually sick people will be identified as sick? Now, the thing about this testing, a couple things. One is a big range. The range for sensitivity or actual True positives range 67 to 93%. And the actual specificity or true negatives ranges, 93 to 100%. Now why are the. Why do you see these big ranges on an antibody test? And the basic thing revolves around prevalence. How many people actually have this infection? If you go into a population of a million people and there's been no recorded cases there at all, and you do the test and the test is negative, the specificity, you know, odds are you don't have disease because it's a pretty low prevalence disease, no one has it. Basically the test was negative. If your test was positive, you'd probably think, hmm, does this person really have disease? Because it's pretty rare. So you come back to this number and go, the sensitivity is 87%. Well, maybe I'm the 13% that are falsely positive. So to interpret this testing, you actually clinically need to know what is going on in the area, how many people actually have had disease. And that's why it's important for testing regimens pop up testing centers to help get us some more accurate information. That's one thing I think people are struggling with. If you're in New York City and 22% of people, which is what we think now, have already had this, the odds of positive test being truly positive is going to be up. So it's probably going to range close to that 93% sensitivity. And if you're in some rural area where no one's had disease and you're positive, maybe the sensitivity is closer to that 67%. That's a pretty big range of variance between, you know, 2/3 accuracy versus, you know, 7%. So that's where actually knowing what's going on. That's the reason why I said before that this test is probably best for surveillance to figure out how many people have disease in the area, not actual diagnosis because there's just such a big range of variants. The other thing about it is that's where clinical acumen come in. The practitioner you're talking with, you know, which you should be hopefully interpreting these tests with your own personal provider. If you had no symptoms whatsoever, you're around no one at all with disease and this test comes back negative. Do you believe that? Well, probably, yeah, because you weren't around anybody. And that's also the problem. If it came back positive, you wouldn't really know how to interpret that. If you don't know the local data, that's where in Virginia, because we don't know prevalence data as of yet, at least nothing's been reported that I've heard from Our Department of Health, it becomes a little tricky with this data. And so that's the reason why when we see patients, we're saying, hey, were you sick in January? Because there are a lot of people having weird illnesses in January. I personally believe this was here in Virginia, January, December, timeframe. Because we're seeing lots of patients with weird viral infections, bronchitis that weren't getting better for four to six weeks, weird presentations, weird emissions from viral pneumonias. Now that we know this was actually in the United States and the first death was the beginning of February, that means the first infectious cases were probably here in the United States, late December. So knowing the data is important. Having someone help you interpret the data is very important as well. We've been using. I've been seeing patients who've had family members sick who had classic COVID 19 symptoms. They had a little cold, and their antibody testing is coming back positive. So I'm believing that because they're fitting a clinical scenario. I've also had patients who've had classic symptoms of COVID 19 and they've had positive siblings, have a test, their test was negative. So do I believe that? Well, no, because their siblings, their people in the same house had the classic stuff, test was positive. In that case, I think their test is falsely negative. And that's where using this test to determine if we can go back safely into the workplace might not be the best use of it because it's not super accurate. So hopefully that was helpful. Hopefully this also spurred some more questions. Post those below and hopefully we can kind of bring them together and have a more robust Q and A session. For all those who are interested in getting a little deeper in this, because I feel like this is gonna be a big topic as we go forward. If you feel like this is helpful, like us, follow us, share us with friends. That's where we get this kind of information out there, is with you, sharing it. If you feel it's useful and helpful, Continue to follow us on social media. We have lots of educational resources on our website as well. Check us out there and be safe, be healthy, be well, find some good resources of information, stick with them, be calm, and we will get this together. Have a great day and take care of.