On Monday, Dallas County recorded its fewest new cases of the coronavirus since April. In a single day, just 98 tests had come back positive. Then, on Tuesday, we were back over 200. If that rapid up and down left you feeling confused, you’re not alone. This thing bends and contorts and teases. Following the daily data that is being released by county, state, and federal health officials can often raise more questions than it answers.
The pandemic reached us nearly half a year ago. Schools and colleges are beginning to reopen. Over the last five months, we have learned a lot about what seems to keep the virus at bay: masks, physical distancing, and avoiding indoor settings as much as possible.
But we’ve also been introduced to a bevy of new data sets that purport to offer evidence of how effective these measures are in slowing the spread of the virus, but can also be difficult to decipher. For one, the up and down daily totals of positive cases shows why it’s important to focus on seven-day rolling averages. But then there was the massive backlog of cases from the state, which dumped hundreds of thousands more from the summer months. That meant that Gov. Greg Abbott’s treasured metric—positivity rate—that he’d been using to inform his reopening decisions had been inaccurate this entire time.
Now there may be new confusion as the federal government pushes to increase a new kind of testing, rapid result antigen tests. Think of them like a strep test—but they’re counted as “probable” cases because they’re not thought to be as reliable as the diagnostic tests that get lab confirmation. Some states group these together. Texas does not.
There is much to parse and it’s easy to get lost in the weeds. How can we see how successful our efforts are in slowing the spread of the disease, and what data do we pay attention to at this point in the disease? We caught up with two of the most important public health leaders in Dallas: Dr. Philip Huang, the head of Dallas County Health and Human Services; and Dr. Kelvin Baggett, the city’s COVID-19 czar. We hope it answers some of your lingering questions.
What about that state backlog?
The Texas Department of State Health Services revealed a massive backlog in positive cases linked to coding errors and server problems via three private labs: Quest Diagnostics, Walgreens, and CHRISTUS Health. In the two backlogs, the state added over 500,000 new tests that dated back to the beginning of the pandemic, in March. Most were attributed to June and July, but the state also didn’t separate the old results from the daily case counts. That falsely inflated the positivity rate, which may have alarmed you if you looked at Dallas County through this New York Times map.
The coronavirus was much more prevalent in our community than previously believed. While the governor pushed to reopen restaurants and bars and other businesses, the positivity rate would’ve been much higher than he was hanging his bets on. He has said anything 10 percent or higher would be cause for action; that almost certainly would have been, although the positivity rate has been above that even without the backlog and it hasn’t spurred much action.
Locally, Huang says the backlog didn’t affect policy decisions or change the story of prevalence in the community. He’s looking at a whole slew of metrics, and they’ve always pointed to there being an unsafe level of community spread. (The county is also correcting for the date the sample was collected.)
But test results have been a problem since this began.
Let’s talk about those test results.
Private labs are not reporting their total tests directly to Texas’ counties. They’re only confirming positive cases. But they are sending those numbers to the state. As the backlog flowed thousands of new cases in, the state was misidentifying the date of the lagged results. So it looked like our positivity rate had skyrocketed. In reality, it would’ve been higher than previously thought in previous months.
The data are noisy. Meanwhile, testing ebbs and flows. During the first week of August, the state shows about 44,000 tests happened in Dallas County. The next week, there were about 100,000. The third week, there were just 29,000. That leads to variability when leaning on the positivity rate. It’s important, but it’s part of a constellation that our public health officials are monitoring.
“People aren’t having the symptoms, indicating that that universal masking and things to that effect have been effective,” Huang says. “There are not as many people with symptoms, not as many people going in for testing universally across all the testing sites.”
Too, Baggett says the weekslong delays in returning test results has shrunken to about 72 hours.
Inconsistent Data, Consistent Messaging
The numbers are shifty and have been since the pandemic began. The county wants to see a sustained two week drop in new cases, hospitalizations, deaths, and ER visits before shifting its advisory of “stay home, stay safe” to something lighter.
In their public messaging, Huang and Baggett choose to simplify the situation. The numbers are just too high across the board. Until we get two weeks straight of sub-200 new daily cases, Baggett’s message to the public is vigilance. Wear a mask, maintain six feet of distance, wash your hands, and avoid unnecessary trips. This virus is being spread through the community.
It’s working. Here is UT Southwestern’s most recent forecasting model: “In both Dallas and Tarrant counties, hospitalizations are projected to continue to decline over the next two weeks, given the compliance with physical distancing and masking. The percent of COVID-19 tests coming back positive continues to drop but is still high enough to suggest that more infection is present in the community than can be seen through testing data alone.”
The second part of that sentence is why Baggett and Huang prioritize personal behavior. As hospitals have begun testing more non-COVID patients prior to other procedures, their positivity rate has plunged to 12.5 percent, down from a high of about 30 percent in early July. Hospitalizations for COVID-19 patients have dropped 24 percent over the last two weeks. Personal behavior affects everything. The “R naught” reproduction rate, the rate at which an ill person spreads the virus to others, gets knocked down when you stay home and wear a mask. So does the positivity rate.
“What I’m trying to get across is that we have to be vigilant, we still have a long way to go,” Baggett says. “We still have yet to reopen to the extent that all of us would like. We have yet to get our kids back in school. We have yet to enter the fall. We have yet to get through Labor Day weekend.”
And be sure to get your flu shot.
What about reopening schools?
The understanding of how COVID-19 affects children is evolving. Initially, it was believed that this novel coronavirus posed little risk for children since they were largely asymptomatic. When it came to school, health officials were mostly concerned with the role children may play in facilitating the spread of the virus, possibly leading to more cases that affect vulnerable populations, including teachers, school employees, and older members of students’ families. That’s still a concern.
Dr. Huang pointed out that since schools shut down last spring just as the virus was first making its way through the community, we don’t know the extent to which reopening schools will contribute to a new surge of cases. “You’re seeing these examples in other communities, other places where they are opening up schools or colleges, seeing how it spread, and shutting them down again,” he said. “So that is definitely a variable here that we anticipate.”
But there is also new evidence that COVID-19 could pose greater health risks to children than previously thought. One study has shown that hospitalizations for children are greater for Black and Hispanic children, who make up more than 90 percent of the Dallas Independent School District population.
Like adults, Baggett said that underlying conditions likely drive COVID-19 health risks for children. “Just saying, ‘Well, it’s safe for children, they’re just transmitters’ is no longer the educated response,” Baggett said. “We believe they’re more than just asymptomatic transmitters of the virus, and we believe that there may be some associated health implications depending on the status and age of the child.”
These are some of the considerations that have led the Dallas ISD to push back the option of in-person instruction until October, but private schools have already begun to reopen. To shape its guidelines around reopening procedures, Huang says the county has put together an advisory group that includes experts in child health, childhood infectious disease, child development, and mental health.
Antigen and PCR Tests
Get ready to hear a lot more about antigen tests. These are the rapid tests, the “point of care” tests that can be deployed in nursing homes, schools, and other such settings. You get results in 15 minutes. They are not as accurate as the PCR tests, which are the viral tests where the samples are sent to a lab for confirmation. But they will make things speedier and identify people who are likely infected and need further action.
Dallas County counts these tests as “probable.” The lion’s share of positive cases have come from the PCR tests, but you’re starting to see these antigen tests tick up. Adding to the confusion is the fact that different states are treating antigen test results differently, with some including positive results in their totals and others leaving antigen results out altogether. This, naturally, has an impact on how states report changes to their positivity rates. Texas does not group PCR and antigen test results together, which raises concerns about whether the most accurate information is being used to guide policy decisions.
Adm. Brett Giroir, the federal director of diagnostic testing for coronavirus, says 40 million antigen tests will be available throughout the country next month. It’s best to think of these as an element of care: false positives are possible, and sometimes multiple tests will be needed to get an accurate result. But it does give additional insight into who may need to stay away from others.
Imagine You Had a Magic Wand.
The locals want more local control—in testing, closures, all of it. If we can do it here, our response can be quicker.
“How do we build local capacity so that we can conduct these tests here?” says Baggett. “We’d need the reagent supplies (the chemicals for the tests), the machines. If I had my magic wand, I’d have all of it here. We do all of it here. And one we’d have a greater consistency in the tests that are being performed. And we’d also have, I believe, a quicker on turnaround time because we’re not having to work through the labs or at least eliminate the transport across state lines.”
