Analysis: It’s hard to make good pandemic policy if the Texas data is fuzzy

_{Editor's note: If you'd like an email notice whenever we publish Ross Ramsey's column, click here.}

_{If you would like to listen to the column, just click on the play button below.}

One of the end-all, be-all coronavirus indicators — one the governor of Texas has said he uses to make decisions about whether and how to respond to the pandemic — has turned out to be pretty rough around the edges.

By the state’s own standards, the Texas calculation of the “positivity rate” tests negative. More than six months into the coronavirus pandemic, the state is changing the way it accounts for the spread of infections.

Gov. Greg Abbott made a big deal of positivity rates way back at the beginning of the pandemic. It was a new and unfamiliar term at the time, a way to measure the spread of the virus by tracking how many COVID-19 tests come back positive.

A positivity rate of 10% would mean 100 positive results for every 1,000 tests. It also just happens to be Abbott’s original danger signal — an arbitrary number he chose as a warning that the coronavirus was spreading too quickly and that a response was needed. By the same logic, a descending positivity rate that drops below 10% might mean it’s time to loosen some restrictions.

That’s great, if the numbers are reliable and if the testing is widespread enough to be meaningful. But according to the Texas Department of State Health Services, the numbers haven’t been as well founded as they seemed. The state isn’t flying blind, exactly, but it’s flying without true readings from its instruments.

From now on, the rate will reflect the date a test was done, instead of the date the test results was reported to health authorities — a date that sometimes falls weeks or months after a test is administered.

The new operation will be more accurate, eventually. Daily rates will be recalculated as new results for a given date are reported.

And the change amounts to an admission that the system up to now has been flawed by the vagaries of how and when results are reported to authorities. That flaw remains, but the new accounting will at least put positive test results on the right dates — no matter how slowly they’re reported.

And therein lies another bug in the soup: Decision-makers will still have to wonder, as they look at daily trends around the coronavirus, whether their current information is accurate. They’ll be able to say — as they’re saying about the system that’s being replaced — that they acted on the best information they had at any given time. But they won’t be able to say that the information was precise at the time it was used to make decisions in response to the pandemic.

The results — and the changes to the calculations — are just stuff for people in lab coats to talk about. The rate produces a precise-looking number, one that’s often reported out to two decimal points, that’s used by elected officials who open and close businesses and schools and all the rest.

Squint at the numbers, recall what was going on at the time, and you see the government’s actions in miniature. The rates rise in the spring, and Abbott closes schools, bars, restaurants and other “non-essential businesses.” The numbers begin to flatten a few weeks later and the reopenings begin. They begin to rise rapidly through May and June, and new closings are imposed, along with pleas to wear masks, to stay socially distanced and to keep scrubbing your hands. They drop again, and public schools and colleges open (at varying speeds) to in-person classes.

Now, the state admits, Abbott and others have been relying on questionable data. Last month, Rajesh Nandy, associate professor of biostatistics and epidemiology at the University of North Texas Health Science Center at Fort Worth, said of the positivity rates from county and state health officials that, “they're not useless, but they are highly qualified and unreliable in terms of studying the trend.”

What’s more, the reliability of the positivity rate depends on how many tests are being conducted. Testing in Texas hasn’t been a steady, statewide effort, and that produces weak spots in the data public health officials are using to track the coronavirus.

The positivity rate is a big deal, but it’s not the only number out there, and never was, in the eyes of epidemiologists and policymakers. Hospitalization and death rates, among other things, are also closely watched, and used, along with all of the other numbers, to figure out how to counter the virus when and where it flares up.

That’s a lot easier if the numbers accurately measure what the coronavirus is doing.