The state health department is including some antibody test results in its case totals, potentially clouding information on the current spread of the virus.
Texas health officials are combining some antibody tests along with more common viral tests in statewide COVID-19 tracking data. Experts say this muddies the data and potentially helps pad testing numbers while giving the public a distorted view of the current spread of the virus.
As of Friday, Texas reported administering nearly 646,000 COVID-19 tests since March, and almost 45,200 cases of the virus. The data, published by the Texas Department of State Health Services (DSHS), appears to show the number of viral tests—that is, a swab of a person’s nose or throat to detect an active infection—and the number of positive cases connected to those tests. But mingled with the viral test numbers are antibody tests—which use a blood sample to determine if a person has developed antibodies to fight coronavirus—and their results. These tests are generally used to determine whether someone may have had the virus in the past, not necessarily whether they are currently infected.
It’s unclear what portion of the state’s case counts so far are based on antibody tests, but the co-mingling of the two types of tests raises questions about the validity of the state’s data, which has influenced efforts, largely by Republican government officials, to quickly reopen the state’s economy.
“If antibody tests are included in here, we really need to know what proportion of those positive test results are antibodies, because it changes how we understand the timeline of infections,” said Rebecca Fischer, an infectious disease epidemiologist at the Texas A&M School of Public Health. Fischer, who’s well-versed in the state health department’s data, was surprised to hear that antibody tests were being included without differentiation. “Certainly that data should be made public on their website, given how people are using it,” she said. “It would change our whole understanding of when infection events were happening in the community, and that would be really important to know and be reported in the data.”
A spokesperson for DSHS confirmed to the Observer that the agency includes “some antibody results” in its official statistics. “Now that antibody tests have become more available, we are working to provide data by type of test and whether cases are confirmed cases of active infection identified by [viral] test or probable cases identified by antibody tests,” Lara Antone, the spokesperson, wrote in an email Thursday evening. She did not answer specific questions about how many of the tests or positive cases included in Texas’ count so far are based on antibody tests, why they are being included, or when these test results will be delineated.
There’s no accurate way to meld the two kinds of tests. Adding just the antibody test numbers and not the results inflates the state’s testing numbers and artificially lowers the infection rate, while grouping “presumed positive” cases from antibody tests with active positive cases from viral test totals conflates the two.
This opacity is part of a broader patchwork of county and state data that fails to accurately track COVID-19. DSHS data excludes some positive cases in prisons, significantly undercounting total infections in at least two East Texas counties, the Observer reported last week. State officials have also declined to release information about infections in other major outbreak centers, like nursing homes and meat processing plants, and have not released detailed racial and ethnic breakdowns of cases and deaths around the state.
The number of daily tests and the rate of positive cases are among the main metrics public health experts point to as critical in determining whether it is safe to start reopening public spaces. Texas has lagged far behind, ranking near the bottom of all states in tests per capita, and trailing Governor Greg Abbott’s goal of 30,000 tests per day. Yet the governor accelerated the state’s reopening this month, despite still-limited testing and warnings from health experts and local officials that true case counts around the state were likely higher than reported.
This week, coronavirus testing in the state appeared to finally jump: 49,000 tests were reported on Wednesday, a marked increase from the previous day’s total of 12,000. On Friday, the daily number of reported tests dropped back to about 23,000, while the number of daily cases and deaths hit record highs on Thursday. The testing increase came amid rapid growth in antibody testing statewide, though it’s unclear if the two are correlated. Antibody testing sites have sprung up in Dallas and Austin. The tests are even being offered as a drive-through service in Dallas and Collin counties.
Antone said the state health department does not attribute the increase in results this week to antibody tests, but would not say how many were antibody tests, or when the state began combining these with the viral tests.
Catherine Troisi, an epidemiologist at the University of Texas Health Science Center at Houston, says the majority of antibody tests in use today are unreliable and cannot detect an active infection. “I will tell you right now that antibody tests that are available right now are not very good,” Troisi says. A more reliable variant of the antibody test, one that can actually detect an active infection, is not in wide use in Texas, she says. Antone did not say which variety of antibody test the state is including in its numbers. “They need to be more specific,” Troisi says.
In Virginia, state officials engaged in a similar practice, according to an Atlantic story published Wednesday, which called the approach a “new low in data standards.” The Virginia governor’s office said that it felt pressured to include antibody tests in overall totals because the state was lagging in test numbers. State officials reversed course yesterday after the article ran, saying they would discontinue co-mingling the two types of tests and results.
In Texas, the inclusion of antibody tests in statewide totals appears to vary by county. In Lubbock County, Public Health Director Katherine Wells told the Observer that she does not include the handful of presumed positives she’s received from antibody tests as positive cases in her reports to the state. But until recently, they were included in counts from Taylor County, three hours west of Dallas. Officials there announced last week that they were discontinuing inclusion of antibody test results in the county’s positive case total, striking 82 cases from the tally. The county is still including antibody tests in its total test count, though, with the caveat that the number may be inaccurate for that reason.
Fischer, the infectious disease epidemiologist at Texas A&M, says antibody tests are not reported to the state health department from Brazos County, where she’s working with the local health department. The most urgent concern right now is “looking for infections so we can create roadblocks, so the virus can’t make it to the next person,” which can be more effectively identified through viral tests, she says. “Very few clinicians here are even offering [antibody tests],” she says. “It might tell you when somebody is done being sick, but how useful is that when an epidemic is evolving?”