Florida changes COVID-19 data on the percent of tests coming back positive

Before the pandemic, so-called “test positivity” was a little-known infectious disease statistic. But over the past few months, scores of concerned people have had their eyes on the data point, which measures the rate at which Florida’s COVID tests come back positive and can help spot upticks in virus transmission.

There are numerous ways to calculate the metric, and Florida has used a relatively rare method of filtering out residents who have already tested positive before. That, in effect, resulted in a lower rate and created a discrepancy in total testing numbers compared with other popular pandemic data trackers, namely the COVID Tracking Project, which also feeds the Johns Hopkins University dashboard. Both trackers have consistently posted a higher rate of positive tests in Florida than the state dashboard and daily reports, leading many to question Florida’s data.

On Tuesday, the Florida Department of Health published a backdated data set of “test encounters,” which will provide the total number of people tested per day among those who have never tested positive. The COVID Tracking Project, a nonprofit and volunteer-powered effort, will use that field as opposed to its old category of testing which only counted people tested for the first time.

The end result? The tracker is adding 3.7 million tests to its Florida data, and the positivity rate on the Johns Hopkins website will go down, matching the state’s. That has prompted Florida COVID data skeptics on social media to accuse the state of covering up unfavorable virus metrics, but Jason Salemi, an biostatitician at the University of South Florida who closely tracks the data, doesn’t see it that way.

“The take-home message is that we aren’t losing anything,” Salemi said. “The measure popularly reported by Johns Hopkins only assesses positivity in people who are being tested for the first time, and more recently in Florida that doesn’t capture positivity in about two-thirds of the people tested on any given day.”

Salemi said, however, that the data change should have come with an explanation from the health department to tamp down skepticism.

“Problems emerge when you make changes and you don’t tell people why you make them,” he said.

The health department has yet to elaborate much on why it made the changes or further explain its aim in working with the COVID Tracking Project, but released a statement on Tuesday saying the previously reported number “did not reflect the current status of the pandemic in Florida.”

A spokesman for the department did not elaborate on the partnership in response to questions from the Miami Herald.

Alexis Madrigal, a journalist who helped launch The COVID Tracking Project, said the group worked with the Florida health department to help make the changes. He said the change in how Florida reports the data and how the tracking project relays the information will help make Florida’s numbers more comparable with other states and more accurately capture testing volume.

Madrigal said the positivity rate was higher before the change, when it measured unique people instead of test encounters, because that only counted people who tested negative once, even if they took numerous tests on different days.

“It’s pretty logical if you think about it: I’ve myself taken three COVID tests now,” he said. “In a unique people denominator, I’d be counted as one. In an encounters denominator, I’d be counted as three.”

Listen to today's top stories from the Miami Herald:

Subscribe: Apple Podcasts | Spotify | Amazon Alexa | Google Assistant | More options

The COVID Tracking Project used to track “unique people” as a unit for tests, which public health advisors favored in the early days of the pandemic because people match up with actual cases of COVID-19. But as repeat testing took off in the later stages of the pandemic, the number began to fall out of alignment with the number of tests performed, Madrigal said.

“These additions to our data set are important because they allow for a more comprehensive view of Florida’s testing utilization and strategy,” Madrigal said. “They will also make comparisons between Florida and other states more useful.”