How Good are Florida's Coronavirus Numbers?
Continuing the theme of yesterday's posting, where the impeachment of Florida's Governor was suggested, is this op-ed piece from Sunday's Palm Beach Post (July 12) by Rebekah Jones, the State employee who had been charged with maintaining Florida's web site containing Coronavirus data. She had been fired for not agreeing to post numbers making the virus' impact on Florida less severe that it was (and is). Obviously, in any impeachment hearing, she would be a witness.
CORONAVIRUS PANDEMIC: FLORIDA DEPARTMENT OF HEALTH DATA
COVID-19 data reporting by state of Florida is unreliable
By Rebekah Jones
Florida has a data problem. While the rest of the world drowned in third-party, speculative and non-scientific coronavirus data at the onset of the pandemic, the Florida Department of Health (DOH) stood proudly and steady in its position as the singlepoint- source of information in the state.
That monopoly over COVID data in the state is partly my fault. I worked very hard for several months to ensure DOH was the only authority over COVID-19 data, and since I was the sole creator and publisher of that data, I trusted its authenticity and accuracy above all else.
I learned a hard lesson about data integrity when I was fired in May for refusing to manually manipulate that data at leadership’s request. Data is only as trusted as its keeper, and DOH’s credibility evaporated faster than dew on the grass in the Florida morning sun after news of my firing spread across the globe.
A few weeks later, on the day Gov. Ron DeSantis announced the full reopening of K-12 schools in August, Scott Pritchard, the lead epidemiologist for Florida’s COVID-19 response since January, abruptly quit after 15 years of service.
The Miami Herald cited a former DOH employee, speaking on the condition of anonymity, who said Pritchard was afraid DeSantis would use him as a scapegoat once cases “started exploding.”
Trust in DOH tanked — also partly my fault — and consequently, secondary and nonauthoritative data resources have popped up left and right.
Public distrust in the officials who are supposed to lead us opens up too many avenues for amateurs and pseudoscientists to take advantage of the data hole by filling it with unreliable, unvetted and unsourced information. The public information sphere became flooded with, based on my experience at DOH, gross misuse of data.
Still, there were other reliable state agencies pushing back against disruptive disinformation.
The state Agency for Healthcare Administration (AHCA) emerged as a more prominent figure in publishing COVID data during the last several weeks, its data services established in April but growing in prominence only since June. AHCA tracks the number of available beds in all of its reporting hospitals (307 as of July 3). AHCA data tells us exactly how many ICU beds, isolation beds, pediatric beds, etc. are available in every hospital in the state, updated every 30 seconds, 24 hours a day.
As the only state data available regarding hospital capacity, AHCA’s data should have been added to the state’s COVID-19 dashboard months ago, but never was. I was told the public data wasn’t “ours” to use. Two weeks ago, that data became an issue of intense controversy.
That notoriety is probably partly my fault, too. Once I published AHCA data front and center on my new Florida COVID Action dashboard, the public began asking questions about it, demanding answers from the state about why political messaging wasn’t matching the state’s own data.
As a government agency, its authority over data relating to hospitalizations should be unquestionable.
After the governor’s unexpected intervention in hospital reporting, however, even that data elicits more questions than it answers.
And during a review of the data others are publishing, including several colleges and universities publishing their own data supposedly based on AHCA data, myself and several other scientists found numerous inconsistencies with DOH case data, and far-fetched proclamations of how all-encompassing AHCA and DOH case data is. Much of the case data doesn’t match the AHCA data, and the AHCA data these third parties are publishing doesn’t match the data AHCA publishes themselves.
Adding to the confusion, the DOH communications team revealed recently that some sort of actively-hospitalized case data will be published “alongside other publicly available case data”.
But DOH has never maintained data on the number of cases actively hospitalized. They did not monitor admission and discharge dates, and would often only learn of a case hospitalization weeks or even months afterward. We had syndromic data from hospitals through the Merlin reporting system, but it was limited to key words listed in ER visit notes.
When asked directly about the communications team’s announcement about adding active hospitalization data, the DOH epidemiologists confirmed that they’ve never had active hospitalization data for cases and would not be adding it to their daily reports. They added they did not know about and couldn’t speak to what data other agencies were publishing.
That lack of communication and data sharing between DOH and AHCA causes a headache for those who know the data well and
Rebekah Jones, who helped create the state Department of Health’s virus statistical dashboard, said she was fired because she would not change data. [PHOTO PROVIDED]
who see the inappropriate use of DOH data as a proxy for active hospitalizations.
The misuse of DOH hospitalization data stretches from bloggers all the way to the New York Times.
Taking the number of people hospitalized today versus the number reported yesterday doesn’t tell you how many more people were hospitalized in the last day — it tells you how many reports of hospitalizations DOH received, which could include hospitalizations from cases back in March. The same is true of death data, as well.
On top of the issues about data completeness and accuracy of conducting AHCA and DOH data, there remains a void of other data sets that are either not reported at all or reported incompletely.
The state currently publishes zero data about testing and cases in state prisons. They also don’t publish data about testing, cases and deaths in jails.
DOH stopped publishing demographic data about who’s being tested after firing me in May, so we have no idea if recent surges in cases by age, gender, race or ethnicity are truly spikes or if they’re proportional to the number of people being tested within those demographic groups.
DOH has never published testing or death data by zip code, either. And they don’t publish dates of deaths, making it difficult to measure and monitor trends.
Another black hole in data publishing centers on contact tracing, which DOH claims to be doing, but won’t publish any data on at all.
If Florida is to come out of this pandemic stronger, they need the help of scientists like myself and those across the state. To do that, we need data.
It’s time for Florida officials and step up fix the data problem they created.
Rebekah Jones, a data scientist who helped create the state Department of Health’s virus statistical dashboard, said she was fired because she would not change data.
* * *
'Positivity' is What Counts
Over the past few postings I have stressed the importance of data showing what percentage of all those tested for the Covid19 virus tested "positively." An increase in that percentage shows the virus is still spreading. A decrease would show that the spreading has been reduced.
Increasingly, that number has appeared in news reports (have they seen my blog?) indicating that a reasonable goal would be for the 'Positivity' percentage to be 10% or lower. Right now, in Palm Beach County, Florida, it is 20%. That is bad news.
It is the result of social and business activities being opened prematurely by the Governor. The only way to reduce it is by adhering to the following guidelines, some of which are required by law:
· Wear a mask when out of your home.
· Wash you hands frequently.
· Use disinfecting gels.
· If necessary to go outside of your home, observe social distancing.
· Ignore efforts to "reopen" businesses, bars, restaurants, etc. despite any governmental okay to do so.
· Pay no attention to anything Governor DeSantis or President Trump say on this subject. They know nothing about this.
· The best source of information is Dr. Fauci.
* * * *
JL
No comments:
Post a Comment