Offficial Death Toll in Colorado Lowered by 24%
So the controversy around the official coronavirus death toll deepens. Skeptics of the official count got a huge boost when the State of Colorado recently lowered their official numbers from 1,150 to 878. A 24% reduction is a pretty significant revision, and even the States’s Governor appeared to throw shade at the official stats.
What Do Total Deaths Look Like in Colorado?
I recently looked at total death statistics in several US States to see if I could detect any evidence that coronavirus deaths were being overcounted. However, Colorado was not one of the States I looked at. Given this big revision in the official record, I was curious to see what the total death statistics in Colorado would show.
In the case of Colorado, it looks like the delta with previous years is not a good estimator for coronavirus deaths. The first coronavirus death in Colorado wasn’t recorded until March 9 which corresponds to week 11. For whatever reason, total deaths in Colorado were trending well above the previous 5 years long before the coronavirus started hitting.
Instead of using previous years, I decided to take the average number of deaths from the first 10 weeks (avg = 829 per week), and use that number as the expected number of deaths for weeks 11-17. I then took the difference between actual and expected values as a rough estimate for the number of coronavirus deaths. It’s easier to show visually than describe:
The shaded area in the graph above leads to an estimated 778 deaths from coronavirus. The official coronavirus death toll thru week 17 (i.e. April 25th) is 881 [source]. Note that the official count is before the revision down. I wasn’t able to locate daily statistics with the newly revised numbers. The total death data appears to support a downward revision. However, there is far too much uncertainty in the total deaths data set to draw any rock solid conclusions.
Conclusions
- The previous official death count in Colorado may have been overestimated.
- Regardless, there have clearly been a significant number of coronavirus deaths in Colorado since March.
- Getting the data right is important.
Caveat on Provisional Data
As mentioned in the previous post, the total death statistics for 2020 are provisional. The CDC has been updating the data set once a week (on Fridays around noon.) One of the criteria I used to select States in the original post was that the data didn’t change too much in the previous couple weeks. In the case of Colorado, the data for the most recent weeks analyzed was still being revised up pretty significantly. If the data continues to be revised up in the continuing weeks, the conclusions reached could be different.
