CDC, Georgia to change website reporting after combining antibody and viral test results

ATLANTA — The Atlanta-based Centers for Disease Control and Prevention and Georgia’s state health department will alter their website data to separate testing data combining antibody and diagnostic test totals.

Wednesday evening, Channel 2 Action News learned Georgia included 57,000 antibody tests in their tally of the number of Georgians who had been tested for COVID-19. That accounted for roughly 15% of the Georgians who were reportedly tested for the virus. When removed from the positive case data, it increased the rate of infected Georgians by 2%.

Simply put, the data inflated the number of Georgians who have been tested for COVID-19. That figure stands around 3%.

The inclusion of the data also offered an “apples-to-oranges” look at the testing. Antibody tests give you a window into the past. COVID-19 test results are used to determine current trends and infections.

Wednesday afternoon, the state’s public health department commissioner said she’d just learned of the discrepancy.

“I don’t want you to think I’m hiding anything,” Dr. Kathleen Toomey said. “I’m actually still learning about this, literally over the past few hours.”

In a statement Wednesday afternoon, the Georgia Department of Health said it would remove the antibody testing data from the diagnostic data, although it had been following the current CDC guidance on the reporting data.

Thursday, Toomey addressed the issue during a state news conference.


“I’m actually am grateful that the media are looking at this closely because often you can recognize things that I did not see, or there are some inconsistencies that I didn’t notice,” Toomey said.

“This is a time when we need the public’s trust, and we won’t have the public’s trust unless we can assure them that the data are accurate, timely and do represent what’s going on in the community,” Toomey added.

During an earlier virtual news conference with the Georgia House Democratic Caucus, Rep. Bob Trammell addressed data manipulation concerns, saying he didn’t believe the error was made intentionally, but the information was used as such.

“I don’t know about manipulation, but what I can say is that there certainly hasn’t been a reluctance to try to celebrate data points to fit a pre-made decision,” Trammell said.

“First of all, I want people to know we’re not perfect in the data,” Gov. Brian Kemp said in the state news conference, vowing ownership and fixing of any state errors.

Kemp also stressed that his policy decisions had been guided by a number of data models. He said if the department of public health needed more resources to handle data, he would be prepared to provide them.

“We are committed to full transparency and honesty as we weather this health care crisis,” Kemp said. “Please afford them (DPH) some patience, and please steer clear of personal attacks. We are all in the fight together.”

Thursday, a CDC spokeswoman told Channel 2 investigative reporter Nicole Carr that the agency would also change its online reporting data in the coming weeks.

“Initially, when CDC launched its website and its laboratory test reporting, viral testing (tests for current infection) were far more commonly used nationwide than serology testing (tests for past infection),” the statement read. “Now that serology testing is more widely available, CDC is working to differentiate those tests from the viral tests and will report this information, differentiated by test type, publicly on our COVID Data Tracker website in the coming weeks.”

“It’s biased in a way to make it look like we’re testing more than we are, and that’s not right,” concluded J.C. Bradbury, an economist and professor at Kennesaw State University.

Other states correct data reports, experts weigh in

Georgia is the latest in a string of states that are now removing antibody data from test totals.

“Georgia’s not the only state that has been doing this,” said Dr. Gigi Gronvall, a senior scholar at Johns Hopkins Center for Health Security. “We were first alerted that Virginia had also been doing this.”

Last week, Virginia Gov. Ralph Northam tweeted that he had been unaware of the data mixing.

“When I found out recently that data from all types of tests were being combined, I immediately directed that the diagnostic tests be separated out,” Northam tweeted May 14.

“I am a doctor, and I have said all along that I will act based on science and data-and the data must be reliable and up-to-date so we can make informed decision based on the facts going forward,” he tweeted. “Going forward the (Virginia Department of Public Health) website will break out the number of diagnostic tests.”

A Texas state spokesman also confirmed their changes to Channel 2 Action News on Thursday night.

“Throughout the COVID-19 response, we’ve been requiring labs to report all tests to us, and that’s how we’ve been reporting them out,” said Chris Van Deusen, a spokesman for the Texas Department of State Health Services.

“We’ve had a small number of antibody tests reported, but with that kind of testing becoming more available, we wanted to start separating those numbers out to provide more detailed information about the virus in Texas.”

Vermont has similarly updated its reporting methods.

Gronvall and experts here in Georgia explained the problems that arise from the conclusions drawn when that data is mixed. This comes days after Georgia began removing antibody positives out of case counts.

“When you only count the positives from the viral test, it’s going to look like the number of people who are testing positive as a percentage of all tests is going down, when it, in fact, that may not be the case, or at least it’s exaggerating that effect,” said Bradbury.

“It gets more complicated than that because these tests have different accuracy levels,” Gronvall said, pointing to inconsistencies in current antibody test results. “The percentage of false positives and false negatives is going to be different with each test, so why do you would mix them together? It just clouds what you know about the situation.”

“Mixing these two very fundamentally distinct tests together really was surprising and disappointing,” said Dr. Aaron Levine, a Georgia Tech professor in the School of Public Policy.

Levine studies issues of bioethics and health public policy.

“In the realization yesterday that that total testing number was almost impossible to interpret called, really, those numbers into question and all the policies that have been based on those numbers into question,” Levine said. “And I think rightfully raises concerns in the minds of many Georgians.”

“We really need one coordinated public health response,” he added.