Questioning the numbers

Chad Frey
The Harvey County website's COVID-19 dashboard is where USD 373 adminstration turns for data to inform the board of eduation as it makes decisions on if students will be in school, or not. This week board members questioned the validity of the data on the website, as it routinely is different than data from KDHE.

As the board of education met to discuss what mode of education to offer next week, members questioned the data being used to make the decision.

Tuesday board member Mallory Morton pointed out discrepancies between the county health department and the Kansas Department of Health and Environment.

“I have been looking at our local dashboard every day, kind of stalking that throughout the day, quite frankly,” Morton said. “I have also been using the KDHE website where they have the school gating section. ... I have been noticing that the numbers have not been adding up between what Harvey County is reporting and the state is reporting.”

Morton said the numbers have not agreed for several months. She called KDHE on Tuesday morning to voice her concern before meeting with the board.

According to Morton, KDHE confirmed the discrepancy.

She pointed specifically to the “percent positive” numbers.

“Sometimes, often times, KDHE is reporting hundreds of tests more than Harvey County is reporting,” Morton said.

That leads to different numbers for percent positive — one recent day the county reported a 10.2% positive test rate, while the state calculated that rate at 8.5% positive. The Harvey County dashboard number stood at 9.1% on Monday night.

The numbers suggested from the Harvey County dashboard suggested Monday night that the district should continue school under a “Yellow” designation — meaning that pre-K to grade six will be in school buildings all day, every day while seventh through 12th grade are placed in hybrid mode (two days of in school and three days of online learning for each student).

Morton said there is a difference in how the county and state report numbers — but does not know what those differences are.

“I am getting concerned with Harvey County and how they are doing their system,” said Matt Treaster, president of the board of education. “This is very troubling. They have to get this figured out. This should match up, day for day.”

What board members are seeing could be coming from what Lynette Redington calls “data cleaning.”

For example, according to the Harvey County Health Department, 4,672 individuals have been tested in Harvey County, equivalent to 6,902 total tests.

“As we look at the information we are receiving from the Kansas Department of Health and Environment epidemiology, those total tests need some cleaning,” Redington to said. “Some of them are coming from other counties. For some reason they got dedicated to Harvey County. We totally understand the ones from our Bethel College and Hesston College. But we do have cleaning of that data to do to get it as accurate as possible.”

The data will likely never totally agree — in part because of the system used to collect it all.

Health care providers are required to report test results within four hours of learning results. They can report those to the local health department, and/or KDHE.

That, along with private labs reporting to KDHE, can create a lag for data unification.

“We have very strong healthcare provider partners in Harvey County with COVID-19 testing ability, and they are consistently sent positive to us, and [KDHE]” Redington said. “This allows us, locally, to get started on investigating. We can get individuals and other people quarantined .... it also means that KDHE does not always have it in their database system before we start.”

Results coming from private labs, sometimes out of state, are reported directly KDHE. Those reports can happen at anytime, and any hour. Harvey County checks those once each day.

“We do not have the staff to do that over and over,” Redington said.

Cases are vetted as well.

“We also vet every case to make sure it is a Harvey County resident. Sometimes it is not,” Redington said. “We notify KDHE they need to move that to another county. ... You will see a number one day, and it will change.”

Redington said the county has also found a number of negative test results that should not be attributed to Harvey County in the database of KDHE.

Staff is reviewing all reports going back to March 13 to update data.

“Presently we have more than 100 negative test results that are noted to Harvey County, but their full demographic information is not in there. We are trying to figure out why they were noted to Harvey County. Some of those 100 maybe are, but we already know some of them are not.”

The data, once prepared, is placed on the Harvey County online dashboard “as quickly as possible” by the county public information officer and GIS staff. As the review of data is performed, the two-week positive percentage will be updated on the county website.

“Schools are looking at that today as they decide what to do for the next two weeks,” Redington said.

Even then, the two week positive percentage will differ between the county and the state — as the two organizations calculate the percentage differently.

The county uses the number of individuals tested, while KDHE is using the total number of tests as their baseline number.

By example, Tuesday the health department announced that 6,902 total tests have been administered in Harvey County to 4,674 individuals.

In the end, Newton USD 373 chose to remain in a “yellow” mode for another week as the board evaluates data sources.

“If there is any discrepancy, to me, it would be great to continue to work with Harvey County to bring the numbers closer to together,” said Luke Edwards, member of the board of education. “For one more week of working in the mode that we are already in, I do not know that we should already get away from using Harvey County’s numbers.”

Newton USD 373 adminstration and the board of eduction met via zoom on Tuesday to decide the educational delivery system for Sept. 14 through 18. After questioning the validity of the county health department data, the board chose to continue with pre-k through sixth grade students in buildings, and seventh through 12th grade students in a hybrid model.