FSCJ scientist says Florida COVID-19 data can be misleading

JACKSONVILLE, Fla. – Chris Perle, a biology professor at Florida State College at Jacksonville, has been watching the COVID-19 data coming from the state since the early days of the pandemic.

Perle told News4Jax on Wednesday that after watching months of data being released by Florida, he doesn’t think the numbers tell the whole story and have the potential to mislead the public.

He said he sent his concerns about the data the state is presenting to Mayor Lenny Curry in early April.

“I wanted to let everybody on the mayor’s Task Force know that if you’re using this portal, too, that you need to know certain things about how they’re presenting the data that the last several days are laggy they’re incomplete,” said Perle, who teaches population biology and biostatistics and holds a Ph.D. from Stanford University.

That’s why he started interpreting the available COVID-19 data on his Twitter feed.

“If I didn’t think I had the background to do this, I would have never started,” Perle said.

Anyone who is curious about the number of COVID-19 cases around the state can access the public dashboard run by the Florida Department of Health.

The dashboard is updated every day and shows how many people have died from the illness, and how many confirmed cases are in the state.

But Perle said in order to know what is really happening, it takes a little bit more work. Every day he says he checks the New York Times database, as well as Worldometer and the Florida community dashboard designed by Rebekah Jones.

Jones, a former health department employee who helped design the dashboard, has been a vocal critic of the state’s dashboard, which led to her eventual firing. She points out the total number of deaths on the state’s dashboard does not include non-residents who died in Florida.

“These are people who actually died in a Florida hospital and I think that people care more about where the virus is in their community than they do whether or not somebody has a permanent legal address in Florida,” she said on CNN on Wednesday.

Jones does acknowledge the state publishes this information, but it’s not outright on the dashboard and you have to know where to look to find it.

“The other problem is the numbers are calculated based on the day that a test result was reported, not on the day that the test was actually taken. So we’re actually commingling tests that were taken on different days. At that point, who knows what’s going on?” Perle said.

Something else you’ll notice missing from this dashboard is the percentage of people testing positive out of each day’s results. Infectious disease experts believe this is a good way to track how the virus is spreading right now and while the state does release those numbers, they’re not as easy to find.

In recent weeks, as new cases have surged, News4Jax is pulling those numbers from county-by-county reports and adding that perspective to our daily stories.

Through tweets, Perle has also drawn attention to potential issues with the percent positive calculation.

“If we’re tracking percent positive every single day, you need to know what the underlying population that was tested is, well they never provide any of that information,” he said. “So if you’re testing a population that is 50% asymptomatic people that percent positive will be different from another day where you test the population and it’s only 20% asymptomatic people. So the fluctuations in percent positive, don’t necessarily reflect the spread of infection. It just reflects the makeup of the people that were tested on a given day.”

Perle said he likes to check on the ZIP code break down on the city’s COVID-19 dashboard, but even that doesn’t tell the full story.

“We can click on your ZIP code and find out that my ZIP code has 75 cases. But what it doesn’t tell you is that that’s 50 new cases this week. That’s what matters, what’s happening now. It’s been a lot of focus on sort of just updating the cumulative number,” he said.

State health officials have said the recent spike in Florida cases have come as testing has increased. On Thursday, the stated reported over 5,500 new COVID-19 cases, a record for one day. But an analysis by the Miami Heard found that when more than 60 counties moved to reopen portions of the economy in early June, there was a 42% increase in new cases the week before, and only an 8% increase in testing.

Perle said in order to get a real sense of how widespread the infection is, there needs to be random testing, not just volunteer testing.

“A lot of people think random means mandatory. It doesn’t mean mandatory wanting to test, everybody,” he said. “We just need a representative sample of the people out there to do across all sectors and ages and poverty and ethnic background. That is representative of, you know what that is, you know, all these demographics.”

“I feel like there’s an important information gap here,” he continued. “And that’s what I’m trying to do is I’m trying to fill in this information. I’m not trying to presuppose anything in terms of what should be done regarding reopening or not getting into political decision to include an objective review the data will understand that you might be able to use it.”