Exactly how many deaths has Ellis County experienced as a result of the COVID-19 pandemic?
That’s a good question – and one for which there’s no good answer when viewing the public datasets. Depending on which state of Texas dataset you view, there have been anywhere from 22 to 33 deaths due to COVID-19.
If you look just at the nursing home deaths within that data, it’s also impossible to know exactly how many occurred there as the two Texas datasets reflect counts ranging from 19 to 26.
It’s also impossible to reconcile the state’s datasets on nursing homes with a federal dataset – because the Centers for Medicare & Medicaid Services only required nursing homes to report deaths beginning with the week of May 24.
Is it any wonder no one knows what to believe? Indeed, comments posted on the county and the Sun’s Facebook pages alone indicate serious misgivings from the public as to any of the data’s veracity.
Death count not the only problem
The death count is just part of the problems that exist with the data the state is sending to county officials. Other discrepancies exist between the State Department of Health Services’ COVID-19 dashboard and the state’s contracted contact tracing service, Texas Health Trace, which provides information directly back to the county.
This week, the Sun sat down with the county’s information officer, Nathaniel Pecina, and emergency management director, Samantha Pickett, to gain a better understanding as to what is going on with the different numbers being released by the state.
What the Sun ascertained is there’s a daily struggle to reconcile and verify what’s being delivered to the county.
Ongoing paper chase
Pickett, Pecina and another county employee, Ralph Mulvaney, are tasked with taking the receipt of the incoming data and cleaning it up/verifying it for dissemination to the public. Much of the problem with the data exists in that it goes through multiple other hands first before being given to the county. In fact, the data takes quite a journey.
COVID-19 data begins its journey with doctors, medical facilities and testing labs reporting positive case results to DSHS.
DSHS then sends the data to THT, which, when its work involving contact tracing is done, sends it to the county’s designated local health authority and county officials. Two purposes exist for sending the data to the county: 1) for it to be input into emergency management systems for first-responders’ use and 2) for the county to release to the public.
Unfortunately, the county has found numerous issues in the Excel file it receives from THT.
“We’re having to remove duplicated cases, fix addressing errors and other mistakes,” Pecina said as example.
Other errors have involved case ID numbers and cases appearing and disappearing off of the dataset received from THT.
A major issue they’re working on now is trying to determine an accurate death count. They’re working to reconcile that; however, the only information the state gives them is a name, gender and where the person died. Their efforts to reconcile the issue have grown to include reading obituaries as part of their verification process.
Mulvaney has the primary task of reconciling the THT dataset, which requires his verifying each address and chasing down the other mistakes. It’s a process that can take up to a full workday to complete.
The county typically receives a THT dataset late afternoon/early evening. Mulvaney works on it the next day so it can be released that afternoon/evening.
“By the time he finishes with one dataset, another one has come in,” Pecina said.
Adding to the issues is the significant time lag that’s a direct result of the data’s journey from the ground up through the state’s entities before it can head to the county for the final verification.
“It’s not real time data,” Pickett said. “What we’re releasing is probably a week if not two weeks behind when the cases actually occurred.”
As an example, the county hadn’t received any information on a July 15 case until July 29, at which time it was sent to the county with the notation it was already recovered.
“So, our next report will show that as a new case and as a recovery at the same time,” Pickett said.
THT numbers vs. the DSHS dashboard
The above information path involves the DSHS’ transmission of the data it receives from the ground up to THT. What’s unknown is how/where DSHS is procuring the data it is posting to its dashboard – and how it’s coming up with what it’s uploading there. And it doesn’t match the raw report THT sends to the county.
After several calls to the state to ascertain why there are differences, Pickett and Pecina said they were told this week that DSHS would accept the county’s reconciliations to the THT-generated dataset.
“The state has said it will make its dashboard reflect the corrections and verifications the county has made (to the THT data),” Pecina said.
Concurrent reporting requirements
It’s not just data being sent by THT to county officials are working with in their efforts to release accurate information.
There’s also concurrent, separate reporting being done by nursing homes to the state’s Health and Human Services Commission and the federal Centers for Medicare & Medicaid Services. Neither of those are matching up with the THT report to the county.
The county also has to look at the data provided by the North Central Texas Trauma Regional Advisory Council, which is where hospitals directly report.
All are information sets the county is utilizing to keep up with what’s happening locally.
County Judge Todd Little said there are challenges involved.
“We’re getting so much (information), it’s hard to manage,” he said. “It’s not that anyone is giving us wrong information, it’s that we’re receiving information from three different sources, 24 hours a day. We’re navigating through waters we’ve never been in before.”
Additional efforts toward answers
Over the past few weeks, the Sun has repeatedly reached out to state and federal agencies involved with the datasets, with some questions answered and some not. The state’s Health and Human Services Commission also noted the governor’s waiver of some requirements under the state’s Open Records Act.
As of press time, the federal website on nursing home data indicates at least two weeks’ worth of data hasn’t been posted even though the website states it will be updated on a weekly basis. The last update given is that of July 12. A request for an explanation as to the lag in the required federal posting was not responded to by press time.
Editor’s note: Anyone with information relating to what is going on within the county’s nursing homes during this pandemic is asked to reach out to the Sun at email@example.com. Respondents will be confidential unless mutually agreed otherwise.
This story has been updated to correct the spelling of Ellis County public information officer Nathaniel Pecina.