Read on Twitter
I started this morning annoyed about Kentucky's publicly available COVID-19 data, so I'm going to yell into the void.
First, testing. Beshear said yesterday that "unlike other states" testing is increasing in Kentucky. And that is mostly true. There were more tests in August than any other month of the pandemic.
But over the past two weeks* tests have fallen. In the middle of August, we had around 64,000 tests a week. Last week that dropped off to 52,000 and this week we're only at 30,000 with two days of reports left.
*I measure weeks Sun-Sat based because of how Beshear releases data
*I measure weeks Sun-Sat based because of how Beshear releases data
And this goes to the heart of my problem: it appears the state is going off a different dataset than they release to the public.
You can tell this because of the official positivity rate, which Beshear says is around 4.53 percent.
You can tell this because of the official positivity rate, which Beshear says is around 4.53 percent.
Beshear bases his positivity rate on a "rolling average."
If you were a statistician with an ideal data set, that would mean taking the positivity rate on each given day (number of cases over number of tests) and then taking the average of those positivity rates over 7 days.
If you were a statistician with an ideal data set, that would mean taking the positivity rate on each given day (number of cases over number of tests) and then taking the average of those positivity rates over 7 days.
That's not the formula the DPH uses. They are taking the total number of cases over the past 7 days and dividing it over the total number of tests over the past 7 days.
https://chfs.ky.gov/agencies/dph/covid19/Positivityrateexplanation.pdf
https://chfs.ky.gov/agencies/dph/covid19/Positivityrateexplanation.pdf
I use the same formula. I take the total number of cases over the past 7 days (not the number of newly reported cases, because the state removes duplicates) and divide it by the total number of tests.
My calculated 7 day positivity rate is 9.17 percent.
My calculated 7 day positivity rate is 9.17 percent.
Ok, the reason I have to use that formula instead of taking an average of the daily positivity rate, is because we don't know when the tests were conducted or when the positive case was reported. Which speaks to my other problem: reporting of cases/deaths.
This has been a consisitent issue throughout the pandemic: Beshear has explained it several times. I wrote a story about it in APRIL. https://www.kentucky.com/news/coronavirus/article241624491.html
But six months into the pandemic, there are still discrepancies between the numbers Beshear reports, numbers in the Cabinet for Health and Family Services and numbers by the local health department.
Example 1:
The Cabinet for Health and Family Services reported back in July that 18 people have died in a nursing home in Lincoln County ( https://chfs.ky.gov/agencies/dph/covid19/LTCupdate.pdf)
Beshear's office is still reporting only 10 deaths from Lincoln County. https://chfs.ky.gov/agencies/dph/covid19/COVID19DailyReport.pdf
The Cabinet for Health and Family Services reported back in July that 18 people have died in a nursing home in Lincoln County ( https://chfs.ky.gov/agencies/dph/covid19/LTCupdate.pdf)
Beshear's office is still reporting only 10 deaths from Lincoln County. https://chfs.ky.gov/agencies/dph/covid19/COVID19DailyReport.pdf
The state has not responded to my question about the discrepancy, which I sent August 24.
Example 2:
The Madison County Health Department is reporting 1,451 total cases of COVID-19 since the beginning of the pandemic. https://www.facebook.com/MadisonCountyHD
Beshear only has 1,174 cases from Madison County. https://chfs.ky.gov/agencies/dph/covid19/COVID19DailyReport.pdf
The Madison County Health Department is reporting 1,451 total cases of COVID-19 since the beginning of the pandemic. https://www.facebook.com/MadisonCountyHD
Beshear only has 1,174 cases from Madison County. https://chfs.ky.gov/agencies/dph/covid19/COVID19DailyReport.pdf
I could give a ton of examples between the discrepancies between local health departments and Beshear's administration on death totals and case totals.
But it leads to this question: Why, six months in, are there still massive discrepancies between local and state numbers?
But it leads to this question: Why, six months in, are there still massive discrepancies between local and state numbers?
Most of this is not a huge deal. The numbers are imperfect and they're just meant to give you a snapshot of where things currently stand with the pandemic.
But it is really frustrating when there are conspiracy theorists out there trying to undermine things at every turn and Beshear is dismissive of any data that isn't coming from his office, when his office is using a dataset not released to the public to get the positivity rate.
Anyways, end of my rant.
Here's where I keep my data. I try to show all of my work. Maps and topline numbers are based on local health departments, charts are based on Beshear. https://www.kentucky.com/news/politics-government/article241165146.html
Here's where I keep my data. I try to show all of my work. Maps and topline numbers are based on local health departments, charts are based on Beshear. https://www.kentucky.com/news/politics-government/article241165146.html
Follow up: @JSChisenhall pointed out that the state doesn't include labs that don't report their total tests in the official positivity rate, which would explain why it's lower.
That said, we don't know how many tests they aren't including in the calculation.
That said, we don't know how many tests they aren't including in the calculation.
