Journalists did ask. The answers.

Quote:

---

"There's a reason why this is happening, I believe, and that is some people feel if they're just with family members and that turns out not to be the case."

---

Quote:

---

"We are seeing what is being seen in many parts of the country right now and that's a significant decrease in the testing that's being done. This is not because tests are not available, it's because less individuals are signing up for testing. "

---

Poor data management.

They probably thought they would be able to manage the "pending assignment" category close to real-time but then July happened. By the time they got back around to it, it screws up the current rates.

They should have just left them out from the beginning and only added them in when confirmed to be a valid test.

Ag$08 said:

---

The positivity rate doesn't make sense as a benchmark. As the virus slows, the demand for tests will slow along with it. If the factors that drive folks to request a test change proportionally, wouldn't the positivity rate stay the same?

---

Hospitalizations are really the only legitimate metric, IMO, and even those could probably use some standardization.

deadbq03 said:

---

Ag$08 said:

---

The positivity rate doesn't make sense as a benchmark. As the virus slows, the demand for tests will slow along with it. If the factors that drive folks to request a test change proportionally, wouldn't the positivity rate stay the same?

---

Hospitalizations are really the only legitimate metric, IMO, and even those could probably use some standardization.

---

Generally agree re: hospitalizations being much more informative than "cases".

There is also quite a bit of value in syndromic survey data from ERs if you want a "current situation" snapshot.


(via Texas2036.org)

This is a years old system that is used to monitor the flu season, but the data is nearly real-time, so there's no multi-day lag.

Note that Texas CLI/ILI peaked in the first few days of July, and unsurprisingly, fatalities seem to be peaking two weeks later.

I agree Abbot has suspect credibility but Governors doubling down on numbers that make them look bad isn't the thing that usually drives more reporting. I also agree the numbers are messed up, but it's a Texas screw up first.

Keegan99 said:

---

If you follow the Texas metrics, you've noticed that the number of "new tests" has fallen off a cliff of late.

But that's NOT what seems to be happening. Texas is in the process of removing over a million old test results from their database as part of an ongoing audit. Some of those test results are assigned to counties. Some are removed. And when one is removed, the test result that is removed is debited from today's total.

As an example, today may have 50k new results, but the audit process removes 30k old results, so the "new tests for today" number becomes the net change, which is only 20k.


TL;DR: The state is taking a very simple calculation and butchering it, producing garbage metrics.


A pretty solid explanation:

So, do NOT believe what the press parroting "Record high rates of positive tests in Texas!" are telling you. They're credulously copying and pasting state numbers and not thinking critically. Nor are they asking any of the state officials "Why?", which would involve doing actual journalism.

---

I've read many people on here stating that waiting lines for various testing facilities have gotten far less in recent weeks. If less people are feeling sick, or around sick people, or not panicking - they will not seek out tests. You can't force tests on people, people have to want to be tested. I believe the reports that people aren't seeking out tests, I don't understand why what you have posted has come out given the reports that far fewer people were seeking tests.

Who is ALex Rodriguez and why are you posting "what he has to infer"?

Keegan99 said:

---

The state is not being transparent about the nonsensical totals.

See the Bloomberg report above as well.

---

Totals, or percent positives?

WHo cares about percent positive?

If only people who feel very sick try to get a test, percent positive might go up. Who cares.

What decisions are being made off percent positive? What battle are we fighting here?

Are you doing ok?

But is the original post an attempt to say Texas is manipulating the data to make positive rate appear high? It seems conspiracy related - if they were deleting old tests from June and subtracting them from August 12th totals, not June totals, wouldn't they also subtract a positive result, if the June test was positive? I just tend to believe what people have reported first hand - far less Texans are lined up at testing sites, for good reason, so tests go down and positivity rate goes up if you assume it is more likely that only those with obvious symptoms are now getting tested. I've heard countless accounts of "when I drove by a few weeks ago there would be 50 cars not just 2 or 3", so I believe tests have gone down.

No school or local decisions around me are being made off positive test rate, and I'm immune to televised media. If they are using anything other than hospitalizations and deaths to make decisions, or create fear, then I'd hope everyone could see through that.

Are testing methods standardized lab to lab state to state? Good post thank you.

GAC06 said:

---

Take a break, you're embarrassing yourself

---

So you agree with ALex Rodriguez that testing is just as high as four to six weeks ago, when there was a bit of a Texas outbreak, and that Texas is intentionally or unintentionally manipulating the current testing numbers?

You also agree that positive test rate is driving decisions in your area?

Keegan99 said:

---

Quote:

---

if they were deleting old tests from June and subtracting them from August 12th totals, not June totals, wouldn't they also subtract a positive result, if the June test was positive?

---

You are thinking there are historically binned totals for each day.

There are not. At least not published by the state.

There is a single running total for all dates.

The daily total is simply the difference between each day's running total and that of the day prior.

There is no way to delete a test from the "June totals". If a test is deleted it can only affect the running total for all dates, as that is the only figure kept.

---

I was going with that exact premise, it appears Alex ROdriquez was suggesting there were most tests run say, August 10th, than reported by Texas because Texas took the new actual tests run on August 10th and subtracted tests they were deleting from June, or some date in the past. I think that is what he was saying, and you re-stated it above. Are they only deleting tests that were negative in June (or some past date)? If they are deleting some June tests that were positive, would they also delete that positive result from August 10th? While I don't think any of this is impossible, I still believe tests are truly down based on first hand reports from testing sites - and I think that is good news.

In my county, new tests are actually up and testing positivity rate is at it's lowest percentage since early June - good news for our county. The Texas percent positive rate may be because testing is down, and only testing those more likely to be sick, and I also see that the percentage of total tests that are antibody tests is up quite a bit so we are counting positives today that may have been sick a month or more ago.

In most cases, test positivity rate is a good indicator for the ratio of cases that you are catching. Empirically it is observable and additionally it makes sense, given that the higher the outbreak, the more tests are rationed to the sickest people, etc. Ideally, you'd be able to test people who not only have symptoms but contact with confirmed positives, and that would lead to a very low positivity rate. covid-19projections uses this relationship:


Therefore, if recorded cases were flat, but test postivitity was dropping, that means that you could infer that the actual number of cases was dropping. This is a common scenario when the number of daily tests being reported is increasing steadily. I think that is why Texas and other states tried to use test positivity as a key metric in evaluating the virus, because they did not want to penalize themselves for increased reported cases in an increased testing environment.

The problem in addition to reporting as Keegan has outlined is also that less people are going for tests. The Houston Health twitter has been begging people to go get tested at their free drive through sites. Whereas a month ago you had to have an appointment, appointments were extremely hard to get at the free drive through testing sites, and there were reports of people waiting hours for their tests, now you can drive up with no appointment and have a very short wait (<30 minutes).

I think a big problem is just the lack of confidence in the testing program in general. You hear about people struggling so much to get a free test and how long it took to get the results, you decide it's not worth the trouble. From a public health stand point, you'd like more people getting tested and tracing programs to branch off of that to mitigate further spread, but it's just not working that way in Texas at the moment, it seems.

Ideally public policy makers would be looking at all the indicators (cases, positive test rate, hospitalizations, and deaths) and inferring the situation based on that. I think a good metric to use for estimating the severity of the virus would be the estimated % of people who have the active virus. According to the covid-19projections model, we were below 0.5% currently infected in Harris County, which would be 1 in 200 people. We peaked at 3% and now we are at about 2.5%. That would suggest 1 in 40 people in Harris County have coronavirus right now.