<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=749646578535459&amp;ev=PageView&amp;noscript=1">

To test or not to test

May 21, 2020


Let’s get to this question of testing. But, if you don’t want to read this and you would rather laugh a bit and get some pretty good science while you are at it, you can watch John Oliver go on a rant about testing. If you know me well, then you know I am clueless about all things pop culture. So, I hadn’t heard of him until he was mentioned on my geek virology podcast TWiV. So yes, I am getting my pop culture intros from a geeky virology podcast, but such is life. Oliver does something I haven’t and will continue to not do on these letters… look backward and assign blame. Not because it doesn’t need to be done but simply because that is not what these letters are about. These letters have been about looking forward and incorporating the best science into our lives. 

Back to testing. The two tests are:

  1. Do you have the virus now (Molecular PCR testing - swab or spit)? 

  2. Did you have the virus (Serologic antibodies test - blood)? 

I’m not going to cover the molecular testing aspect of this as I am getting fewer questions on this. Just know that they are not perfect. They are likely around 70-80% accurate. On a side note, someone at KaiNexus is hospitalized right now with COVID and 2 out of 3 of his tests were negative. So, if you have the right symptoms, I wouldn’t put much weight in a negative test. This is an interesting article about how much testing we need for “opening up” to be considered safe. 

Let’s get into the antibody test. There is one caveat and one piece of science you need to understand (or just accept as fact if you don’t want to understand it). 

The caveat - We don’t know if having the antibodies provides immunity. It likely will, but then we don’t know if it will be for a couple of months or for a lifetime. Worst-case scenario, it could be like dengue fever and the re-infection could be worse. I really think it will confer some immunity for some amount of time but this is conjecture at this point.

The science - If the prevalence of a disease is low and the test is not 100% accurate, then the likelihood that the test is wrong is greater than the likelihood that the test is correct. Please read that sentence again. That is the fact that needs to be accepted or understood. If you don’t want to understand why, no problem but please read it again. If you want to understand why, then I go into detail on this in the 4/29/20 letter about Prevalence, EMRAP talks about it here, TWiV talks about it here in episode #613 and here again in episode #610, and NY Times talks about it here. If you are getting the test, then you should at least understand how inaccurate and essentially meaningless the test is. Please take some time to understand this or just accept it.

CNN writes about both the caveat and science points here. Medscape interviews a smart doctor here. Johns Hopkins writes about it here

What will I do? I will not be getting the test because I wouldn’t know what to do with the information. 

Johns Hopkins has some really exceptional resources. Here is a list of the approved tests. Here is their disclaimer… 

Here we list the manufacturer-reported sensitivity and specificity data, where available. A highly sensitive test should capture all true positive results. A highly specific test should rule out all true negative results. These measures are not independently validated by the Johns Hopkins Center for Health Security.

I bolded the important part. Always be skeptical about self-reported results from a manufacturer.

For purposes of conversation, let’s assume that the test is 100% accurate and we know with 100% certainty that having the illness provides 100% immunity. This poses a more interesting scenario - fictitious, but interesting. Likely, I would still not get the test nor would I want to broadcast the result. I want to model behavior that is part of the solution - limiting physical contact. Despite my immunity in this fictitious scenario, I would still be able to transmit the virus on my clothes, etc. Also, people who I would be interacting with at a store wouldn’t know I have immunity. Again, there are no easy questions left.  

That is not to say I will never get a test. If I don’t get sick and thus never get a COVID + test, I can imagine a time when I would take the test. But, the test would have to have better accuracy, society would have to have a much higher prevalence, and science would need to have a better understanding of what prior infection means for future immunity. Until then, those tests aren’t for me. 

I certainly won’t judge anyone who does take a test, but I hope they understand the science behind what they are doing, they know why they are doing it, and they know if the answer will or won’t change their behavior. 

She (TWiV - virologist) agrees that this is a great opportunity to learn some science. 

Stay emotionally connected and physically distant,


PS: In part 3 of modeling, I referenced an article about each state’s hospitalization rate but didn’t provide a link. Here you go. Sorry about that.