Wednesday, April 15, 2020

Ok, let's do that math... Responding to whatever the opposite of fear-mongering is during COVID-19

A family friend posted this article from the Washington Times on social media this morning, with the TLDR being that COVID-19 cases have been super low and that it's being blown out of proportion when compared to influenza. Here's a link to said article. Honestly, I didn't know this media outlet existed until this morning.

Anyway, the article opens with a poor presentation of data. Flu is reported in terms of thousands and millions, COVID is reported in tenths and hundredths of percents. It didn't rest well with me, so I did the equivalent of making a two by two table and crunched out the numbers myself. Here's the text from my comment, as I feel I spent too much time working on it to just let it hang out on FB to be forgotten.

I read this article, and the way they presented the numbers didn’t quite sit right with me, but I didn’t just want to “feel” like it was wrong. I wanted to see why - so I put my sleepy med student brain to work.  
1)      The authors of this article compare influenza, which occurred over the course of 6 months, to COVID, which has been on our shores for approximately 4. If we adjust the values given for flu to match the time frame for COVID, we get 25.7-43.6 million confirmed cases and 15.8-41k deaths.  
2)      For COVID, they list the numbers as percentages, which from a data ethics point of view isn’t the right way to do it. It can deceive readers into thinking things are smaller (or larger) than they actually are, which is unfair, in my opinion. Anyway, translating these percentages to real numbers, we get 561k infections in the US, and 23.1k deaths.  
3)      So, let’s say in a 4 month span, flu has 25.7-43.6 million confirmed cases. COVID has 561k new infections. Flu is much more rigorously tested than COVID, mostly because it’s cyclic and we expect and prepare for its arrival each year. Still, okay – 561k is much less than 25.7-43.6 million.  
4)      Now, let’s look at the deaths. 15.8-41k deaths for flu. 23.1k deaths for COVID. 23.1k falls into the range of 15.8-41k, so we can say that the raw number of deaths, based on this article’s data, are not actually that different from each other.    
5)      Finally, let’s calculate the percent of people (or percent mortality) who die from each disease, based on these numbers that I’ve already crunched. Based on these numbers, flu has a percent mortality somewhere between 0.06-0.1% - in other words, it takes about 1000 infections for a person to die of the disease. For COVID, the percent mortality is 4%. This would mean for every 1000 infections, 40 would die – and in fact, COVID has a 40-67 times increased mortality based on the data they presented in this article.  
All this to say, calling the case and death counts “ridiculously low” isn’t true, and is a very callous way to talk about death. Based on the timelines, the same ballpark amount of death has happened due to COVID as would happen with flu in the same amount of time. However, with COVID we are being way more cautious, enacting social distancing, and yes – perhaps sacrificing parts of the economy. The fact that deaths between flu, a disease that has historically not altered social activity very much at all, and COVID have been similar withthese precautions in place underscores the importance of what we’ve been doing.  
Anyway, if you’re still hanging with me, I realize the comment is long… but I did the math as the article instructed, and I’m not impressed.

No comments:

Post a Comment