Read on Twitter
Wow. As others have already reported on today, MAJOR improvement in COVID-19 reporting by DHHS with a new "Community Profile Report"
I wanted to compare it to what I compile daily for Florida.
A
https://beta.healthdata.gov/National/COVID-19-Community-Profile-Report/gqxm-d9w9
1/
I wanted to compare it to what I compile daily for Florida.
A
https://beta.healthdata.gov/National/COVID-19-Community-Profile-Report/gqxm-d9w9
1/
I'll first focus on the "States" tab.
First, I love the simple state-level comparison of:
1. overall population as a % of nat'l pop
VS
2. cases as a % of nat'l total in the last 7d
Ex: CA is 11.9% of pop, but 17.3% of recent cases. FL is 6.5% of pop & 4.8% of recent cases.
2/
First, I love the simple state-level comparison of:
1. overall population as a % of nat'l pop
VS
2. cases as a % of nat'l total in the last 7d
Ex: CA is 11.9% of pop, but 17.3% of recent cases. FL is 6.5% of pop & 4.8% of recent cases.
2/
FL reports 71,974 cases in the 7-day period from 12/10 - 12/16.
This is based on date cases are reported (similar to @COVID19Tracking), not what I report: based on date confirmed as a case (74,203).
The diff in these 2 approaches is inconsequential for tracking purposes.
3/
This is based on date cases are reported (similar to @COVID19Tracking), not what I report: based on date confirmed as a case (74,203).
The diff in these 2 approaches is inconsequential for tracking purposes.
3/
Deaths in the prev 7d are also based on date reported, also similar to @COVID19Tracking.
Although not the most accurate (impacted by reporting lags), it's only way to actually report deaths recently.
Deaths by dt of death most accurate, but very incomplete for recent 4 wks.
4/
Although not the most accurate (impacted by reporting lags), it's only way to actually report deaths recently.
Deaths by dt of death most accurate, but very incomplete for recent 4 wks.
4/
Cumulative cases and deaths
I agree that on 12/16 (data verified through 12/15), there were 1,155,335 cumulative cases to date.
However, HHS reports 20,204, whereas I have 20,490.
My guess is that, for some reason, HHS is reporting only FL-resident deaths.
5/
I agree that on 12/16 (data verified through 12/15), there were 1,155,335 cumulative cases to date.
However, HHS reports 20,204, whereas I have 20,490.
My guess is that, for some reason, HHS is reporting only FL-resident deaths.
5/
PCR lab testing.
Now I'm jealous. The data made available publicly by FL DoH only allows person-level cases (+ve tests) to be differentiated by testing type.
This excludes antigen testing and reports on positivity for RT-PCR.
I cannot confirm/compare numbers, BUT...
6/
Now I'm jealous. The data made available publicly by FL DoH only allows person-level cases (+ve tests) to be differentiated by testing type.
This excludes antigen testing and reports on positivity for RT-PCR.
I cannot confirm/compare numbers, BUT...
6/
...they report 624,246 RT-PCR tests from 12/8 - 12/14.
I count 951,917 total people tested during the same time frame. That would make about 34% of viral tests of the antigen "flavor" and about 31% of cases recently have been confirmed w/ antigen vs. PCR (so makes sense)
7/
I count 951,917 total people tested during the same time frame. That would make about 34% of viral tests of the antigen "flavor" and about 31% of cases recently have been confirmed w/ antigen vs. PCR (so makes sense)
7/
Hospitalizations!
From @AHCA_FL, we only get current hosp patients with a primary dx of C19, by county.
Using a "Unified Hospital Dataset", we are now getting confirmed & suspected C19 ADMISSIONS, though they acknowledge it may not be a "full representation of the true #"
8/
From @AHCA_FL, we only get current hosp patients with a primary dx of C19, by county.
Using a "Unified Hospital Dataset", we are now getting confirmed & suspected C19 ADMISSIONS, though they acknowledge it may not be a "full representation of the true #"
8/
Though I can't compare numbers to those reported by FL (since it's new admissions vs. currently hospitalized) looks like a lot of good data for FL, and metrics that can be compared across states.
I stacked for the screenshot - this is all one row for FL in the spreadsheet.
9/
I stacked for the screenshot - this is all one row for FL in the spreadsheet.
9/
Then we get COUNTY-LEVEL info in the "Counties" tab!
Similar comments as above since similar data elements are reported for counties as for states.
DHHS does have a 7-level classification system for each county, which is a nice addition IMO.
10/
Similar comments as above since similar data elements are reported for counties as for states.
DHHS does have a 7-level classification system for each county, which is a nice addition IMO.
10/
In FL, among its 67 counties
- 56 "sustained hotspot"
- 8 "moderate burden" (all smaller counties)
- 1 "emerging hotspot" (Union)
- 1 "hotspot" (Hardee)
- 1 "moderate burden resolving" (Gilchrist)
Small counties seem susceptible to more frequent category change.
11/
- 56 "sustained hotspot"
- 8 "moderate burden" (all smaller counties)
- 1 "emerging hotspot" (Union)
- 1 "hotspot" (Hardee)
- 1 "moderate burden resolving" (Gilchrist)
Small counties seem susceptible to more frequent category change.
11/
Positivity rates appear to be consistently higher in this DHHS report than reported by FL DoH in their weekly report, but that's the "% positivity for new cases" measure, whereas this one seems to be among all people tested.
12/
12/
I also love that the spreadsheet includes some demographic details for each county that would be vital in assessing county-level comparisons:
- % uninsured
- % in poverty
- % over 65 years
- % NH-black
- % Hispanic
- community vulnerability index
- social vulnerability index
13/
- % uninsured
- % in poverty
- % over 65 years
- % NH-black
- % Hispanic
- community vulnerability index
- social vulnerability index
13/
There's so much more to this PDF report & multi-tab spreadsheet, but on surface view, the TL;DR
1. w/ rare exception, data look accurate on measures I can compare for FL
2. new data elements appear in this report (hosp admissions)
3. Lots of useful classifications made
14/
1. w/ rare exception, data look accurate on measures I can compare for FL
2. new data elements appear in this report (hosp admissions)
3. Lots of useful classifications made
14/
Where has this data been? Love that it's being compiled now.
As @alexismadrigal stated, "This is another huge transparency win."
More in the future as I continue to digest these data.
15/
As @alexismadrigal stated, "This is another huge transparency win."
More in the future as I continue to digest these data.
15/
