Saturday, June 27, 2020

Published 8:55 PM by with 0 comment

Why is the COVID-19 Death Rate So Low in Texas?

Texas has massive COVID-19 infection numbers but relatively low death rates. Why is that?
At this moment, here are the stats for the states with similar positive test results (source: https://www.worldometers.info/coronavirus/country/us/ on June 27, 2020):



Why is the death rate for Texas so low? It could be that Texas is testing better than the other states so the other states have missed more positive people and thus have more deaths per positive test. It could also be that the demographics of positive patients are different in Texas. It could also be that they're undercounting the death totals, and that's what I'll look at here.

I am basing this on excess deaths. I'm using two sites for this:
You can also use this CDC tool but I found it less user-friendly (https://data.cdc.gov/d/muzy-jte6/visualization).

Excess death reporting lags, so I'll use numbers through May 31. At that point, Texas had between 1700 and 4800 excess deaths on the year. ~1700 is 'number of deaths - upper bound of 95% confidence interval' and ~4800 is 'number of deaths - number of deaths in an average year'. The real number of excess deaths is likely somewhere between those two. At that time, they were reporting ~1670 COVID-19 deaths (source: https://dshs.texas.gov/coronavirus/additionaldata/). Why is it conveniently at the low end of the estimate? Check out this plot for unclassified deaths for Texas:


It's interesting. It's like there's an outbreak of some other disease at exactly the same time as COVID-19 that spreads at the same rate but is conveniently not COVID-19. 

In that original list, California seems oddly low also. Doing similar analysis, California had between 3100 and 6800 excess deaths by end of May. They were reporting ~4200 COVID-19 deaths. If we look at the plot for unclassified ones, we see this:



Similar trend. You can check other states on those sites. This general technique (looking at excess deaths) seems like a decent way to track the pandemic in areas given that we don't have great testing, and I wish it were more heavily reported.



      edit

0 comments:

Post a Comment