April 28, 2020
I thought the next topic in my series “for dummies” was going to be about mortality but I think that talking through the concept of prevalence makes more sense.
For the COVID conversation, it refers to the number of people that have had the infection. We certainly have a pretty good understanding of the number of people that have had moderate to severe cases of COVID as they are the ones that come to the ER and potentially get admitted. What makes it so hard to determine the real prevalence of COVID (or any emerging infection) is that we don’t know how many asymptomatic cases there are. There is a huge difference in how quickly we can go through the Dance if the number of people that have asymptomatic-infection is 90% vs 10%.
The reason that this is important is that it is presumed that having the illness confers some immunity for some period of time. Hopefully, it is 100% immunity for life but, it is likely something less than that. More on what we don’t know here regarding immunity.
A few weeks back, I wrote that until we have widespread seroprevalence testing for COVID we would have to continue to use the blunt instrument of severe social distancing with a slow incremental easing of restrictions to ensure that our hospitals don’t get overwhelmed. We have suspected since the start that the real number of cases was way higher than confirmed positives due to the lack of testing.
Well, some of the first studies are starting to come out that may help answer this question. One that has received a lot of attention was the one from Santa Clara County from the Stanford group and another one from the USC group. These indicate that we are undercounting the current number of cases perhaps by an order of 50 times. They indicate that around 2-4% of that population has been infected.
We need to understand that there are potentially a lot of issues with the tests that look for prior infection especially when there is a low prevalence of a disease. In the prior studies mentioned, they are apparently using very, very accurate tests - 99.5% accurate. But, even with a test that has a 99.5% accuracy (the one they used) in a population with a 2% prevalence of the disease, then 23% of the time you would be telling a person that they have had the disease when they have not had the illness. And, if you have a test that has the characteristic of being 80% specific or accurate, then 92% of the time you tell a person that they have had the disease when they have not had the illness. So, unless you know exactly the characteristic of the test, it may be worthless. If you want to understand this concept further, I highly recommend you watch the EMRAP from 4/21/20 and start here. I found it fascinating and highly educational.
I think the take-home points for me are two-fold:
We are nowhere close to herd immunity (if you didn’t read the email entitled Herd Immunity for dummies and don’t know what herd immunity is please, go back and read it)
If you are getting an antibody test, you really need to understand the quality of the test before you make any decisions with the results.
For the next “for dummies” topic (likely in a few days), I’ll discuss what we know and what we don’t know about the mortality of COVID. You can probably already guess that without really knowing the number of asymptomatic cases, we won’t really know the true mortality. I will do my best to have meaningful information so you can have meaningful conversations about it with your friends and family to help you sort out how you are going to handle the Dance.
Stay emotionally connected and physically distant,
PS: In that EMRAP, I referenced, they go through a therapeutics update. If you haven’t heard, hydroxychloroquine is out.