Ok this might have been discussed but I can't seem to find it posted anywhere.

If only the people who are showing multiple symptoms and are in bad shape are getting tested why are there so many negatives on the tests?

It seems weird that in Texas less than 10% are coming back positive. I would think that number should be much higher if you are already showing enough symptoms to be considered for a test.

I keep hearing, anecdotally of course, that healthcare professionals in DFW are being refused a test despite having all the symptoms because they have had no known exposure to COVID.

Edit to be more specific: by "keep hearing" mean it was two maternity ward nurses my wife knows. They were sent home for 10 days...not 14 mind you. Whoever calls the shots refused to test them.

mccjames said:

---

Ok this might have been discussed but I can't seem to find it posted anywhere.

If only the people who are showing multiple symptoms and are in bad shape are getting tested why are there so many negatives on the tests?

It seems weird that in Texas less than 10% are coming back positive. I would think that number should be much higher if you are already showing enough symptoms to be considered for a test.

---

I can only speak anecdotally but from I've seen there is likely a large false negative rate. Upwards of 30%. I hope this is not true but I've been practicing as this is the case. In addition, hospitals are moving to screening patients sick enough to come into the hospital so they can be cohorted into groups. So not that they necessarily have a high probability of having disease. Both of these alter the prevalence of positive tests.

What we do know is that the prevalence is increasing in the tests provided which is not good news.

So in essence a 30% error rate. I would think for medical tests that is pretty damn high. I wonder if it is the test itself or the admin of the test or maybe the virus itself causing the errors. I guess maybe the lab analyzing the test could also have issues causing the errors.

Yes likely a combination of all those factors. Plus is a nasopharyngeal swab the best location to test? Should we be using sputum samples? Should we be using blood tests? How long after symptoms is the best time to test? A lot of uncertainty at this time. We are flying blind.

mccjames said:

---

So in essence a 30% error rate. I would think for medical tests that is pretty damn high. I wonder if it is the test itself or the admin of the test or maybe the virus itself causing the errors. I guess maybe the lab analyzing the test could also have issues causing the errors.

---

I thought someone in the medical field recently said even just regular flu often has around a 30% error rate. If that is the case and that is something they have had time to perfect, then not sure we will ever see much better with this.

Necrosis said:

---

mccjames said:

---

Ok this might have been discussed but I can't seem to find it posted anywhere.

If only the people who are showing multiple symptoms and are in bad shape are getting tested why are there so many negatives on the tests?

It seems weird that in Texas less than 10% are coming back positive. I would think that number should be much higher if you are already showing enough symptoms to be considered for a test.

---

I can only speak anecdotally but from I've seen there is likely a large false negative rate. Upwards of 30%. I hope this is not true but I've been practicing as this is the case. In addition, hospitals are moving to screening patients sick enough to come into the hospital so they can be cohorted into groups. So not that they necessarily have a high probability of having disease. Both of these alter the prevalence of positive tests.

What we do know is that the prevalence is increasing in the tests provided which is not good news.

---

The sensitivity is around 70%, which as you alluded to, means approximately 30% of people who have the virus may still test negative. This doesn't factor into errors in collection method, storage and running the test. This is one of the main reasons not to just test everybody that walks in the front door. If you've got fever and lower respiratory symptoms but are otherwise well appearing you can bet that I'll be sending you home to self quarantine without a test. If I decide to test you with mild symptoms and the test is a false negative you'll be falsely reassured and no longer be self quarantining. But you would be spreading the virus under this false reassurance.

mccjames said:

---

Ok this might have been discussed but I can't seem to find it posted anywhere.

If only the people who are showing multiple symptoms and are in bad shape are getting tested why are there so many negatives on the tests?

It seems weird that in Texas less than 10% are coming back positive. I would think that number should be much higher if you are already showing enough symptoms to be considered for a test.

---

Not really weird when MSM sets our expectations.

mccjames said:

---

Ok this might have been discussed but I can't seem to find it posted anywhere.

If only the people who are showing multiple symptoms and are in bad shape are getting tested why are there so many negatives on the tests?

It seems weird that in Texas less than 10% are coming back positive. I would think that number should be much higher if you are already showing enough symptoms to be considered for a test.

---

Here's a slightly edited response I wrote to this question elsewhere:

There are so many different reasons a person who actually is infected with covid-19 may test negative in the standard RT-qPCR assay.

The obvious one is they aren't really infected, but rather have some other viral respiratory illness, of which there are many.

The others are numerous:

- quality of the transport media and handling of the sample from swabbing until it gets to the processing lab

- quality of the administration of the swab (nasopharyngeal swab). I've seen/heard a lot of anecdotal evidence that many swabs aren't being correctly performed. The test is only as good as the sample collected.

- The site being sampled may not always contain a high viral titer throughout the duration of infection. This could explain why, in serial testing of a small group of known infected patients done in China, they reported ~70% positive in the first swab, ~90% positive for the second swab, and 100% positive of a third swab. I don't recall the time between swabbings.

- Swabs don't capture much biomass to begin with, and placing them in 2-3mL of viral transport media dilutes the sample. So, people with a low viral load at the sampling site may fall below the test LOD

- The standard RT-qPCR test is complex, and there are small variations in efficiency at every step, particularly RNA extraction. So, samples that start on the brink of the LOD oftentimes test negative. Nasal and nasopharyngeal swabbing for diagnosis of respiratory illnesses is known to be finicky.

Good analysis.
So if I have this right, there is an error rate of approx 30%, 10% are testing positive that do not have symptoms because we are not testing those who have obvious symptoms, leaving us with 10%-40% with error factor, that have it?

Does that mean 40% asymptomatic?

mccjames said:

---

Good analysis.
So if I have this right, there is an error rate of approx 30%, 10% are testing positive that do not have symptoms because we are not testing those who have obvious symptoms, leaving us with 10%-40% with error factor, that have it?

Does that mean 40% asymptomatic?

---

I think something baked into that 10% positive rate are screening samples. In my lab, (small local lab in the texas medical center), the majority of samples we've processed to date are screening healthcare workers who are being exposed daily, but not necessarily sick. We are concerned about asymptomatic carriage, and need to pull any workers who test positive out so that they don't expose others. So, not everyone who's being tested is a symptomatic, hospitalized patient.

The "30% false negative" number is anecdotal. Basically it's a clinical diagnosis of covid-19 (diagnosis based on symptoms only) with a negative RT-qPCR test result. You can't know the true false positive rate without confirmatory testing of some kind. Part of that 30% is a "miss" by the test (likely the majority), and part of that will be misdiagnosis.

I don't think asymptomatic carriage can be derived from these numbers. However, in a few isolated populations around the world, testing has shown significant asymptomatic infection. I think it was something like 17% on that cruise ship, and it's been higher in some other small groups. Widespread testing of the population is needed to generate that number, because people who feel perfectly fine aren't being tested unless it's a high-risk patient facing group, and even many of them aren't being tested (yet)

eidetic78 said:

---

mccjames said:

---

Good analysis.
So if I have this right, there is an error rate of approx 30%, 10% are testing positive that do not have symptoms because we are not testing those who have obvious symptoms, leaving us with 10%-40% with error factor, that have it?

Does that mean 40% asymptomatic?

---

I think something baked into that 10% positive rate are screening samples. In my lab, (small local lab in the texas medical center), the majority of samples we've processed to date are screening healthcare workers who are being exposed daily, but not necessarily sick. We are concerned about asymptomatic carriage, and need to pull any workers who test positive out so that they don't expose others. So, not everyone who's being tested is a symptomatic, hospitalized patient.

The "30% false negative" number is anecdotal. Basically it's a clinical diagnosis of covid-19 (diagnosis based on symptoms only) with a negative RT-qPCR test result. You can't know the true false positive rate without confirmatory testing of some kind. Part of that 30% is a "miss" by the test (likely the majority), and part of that will be misdiagnosis.

I don't think asymptomatic carriage can be derived from these numbers. However, in a few isolated populations around the world, testing has shown significant asymptomatic infection. I think it was something like 17% on that cruise ship, and it's been higher in some other small groups. Widespread testing of the population is needed to generate that number, because people who feel perfectly fine aren't being tested unless it's a high-risk patient facing group, and even many of them aren't being tested (yet)

---

I would imagine that asymptomatic carriage has a lot to do with the cultural health of a population. For instance, the overweight/unhealthy population in the US would have a lower asymptomatic rate than the Netherlands. Healthier people's immune system may be able to quickly fight the virus and produce the necessary antibodies?

eidetic78 said:

---

mccjames said:

---

Good analysis.
So if I have this right, there is an error rate of approx 30%, 10% are testing positive that do not have symptoms because we are not testing those who have obvious symptoms, leaving us with 10%-40% with error factor, that have it?

Does that mean 40% asymptomatic?

---

I think something baked into that 10% positive rate are screening samples. In my lab, (small local lab in the texas medical center), the majority of samples we've processed to date are screening healthcare workers who are being exposed daily, but not necessarily sick. We are concerned about asymptomatic carriage, and need to pull any workers who test positive out so that they don't expose others. So, not everyone who's being tested is a symptomatic, hospitalized patient.

The "30% false negative" number is anecdotal. Basically it's a clinical diagnosis of covid-19 (diagnosis based on symptoms only) with a negative RT-qPCR test result. You can't know the true false positive rate without confirmatory testing of some kind. Part of that 30% is a "miss" by the test (likely the majority), and part of that will be misdiagnosis.

I don't think asymptomatic carriage can be derived from these numbers. However, in a few isolated populations around the world, testing has shown significant asymptomatic infection. I think it was something like 17% on that cruise ship, and it's been higher in some other small groups. Widespread testing of the population is needed to generate that number, because people who feel perfectly fine aren't being tested unless it's a high-risk patient facing group, and even many of them aren't being tested (yet)

---

I'm sure someone with a stronger statistics background can (and will) correct me, but:

If the percentage of asymptomatic carriers is 17%; and the percentage of positive testing by symptomatic or known exposed people is 15% or less, it would seem to calculate that the odds of testing positive are higher if you're asymptomatic.

It's as though you would get a higher positive rate by randomly testing people than we are getting by testing only exposed and/or symptomatic people.

This being the case; it looks to me as if testing is a waste of time and resources. It is only necessary for (false) statistical purposes to help governmental organizations justify receiving more resources to fight the pandemic.

I would think we would be better off tracking reported symptoms rather than the results of what seem to be greatly flawed tests.

Where are all these tests we were supposed to have by now?