Positive. Again.
-
There has been a lot of talk about false positives and false negatives in the testing process. Frankly, I wish we had a bit better test, period.
wrote on 13 Apr 2020, 18:35 last edited by@Jolly said in Positive. Again.:
There has been a lot of talk about false positives and false negatives in the testing process. Frankly, I wish we had a bit better test, period.
Good-Fast-Cheap... Need I go on?
-
@Jolly said in Positive. Again.:
There has been a lot of talk about false positives and false negatives in the testing process. Frankly, I wish we had a bit better test, period.
Good-Fast-Cheap... Need I go on?
wrote on 13 Apr 2020, 18:49 last edited by@LuFins-Dad said in Positive. Again.:
@Jolly said in Positive. Again.:
There has been a lot of talk about false positives and false negatives in the testing process. Frankly, I wish we had a bit better test, period.
Good-Fast-Cheap... Need I go on?
George has taught you well, Padawan...
-
wrote on 13 Apr 2020, 18:54 last edited by
Do we have false positive/false negative rates on the test? Maybe this is merely due to false test results.
-
wrote on 13 Apr 2020, 18:56 last edited by
That would be my guess.
-
Do we have false positive/false negative rates on the test? Maybe this is merely due to false test results.
wrote on 13 Apr 2020, 19:09 last edited by@Klaus said in Positive. Again.:
Do we have false positive/false negative rates on the test? Maybe this is merely due to false test results.
NO clinical test is accurate 100% of the time. I've heard error rates of 10% on the current COVID tests. 5% is around minimum for many tests, even less for others.
-
@Klaus said in Positive. Again.:
Do we have false positive/false negative rates on the test? Maybe this is merely due to false test results.
NO clinical test is accurate 100% of the time. I've heard error rates of 10% on the current COVID tests. 5% is around minimum for many tests, even less for others.
wrote on 13 Apr 2020, 19:21 last edited by@Jolly said in Positive. Again.:
@Klaus said in Positive. Again.:
Do we have false positive/false negative rates on the test? Maybe this is merely due to false test results.
NO clinical test is accurate 100% of the time. I've heard error rates of 10% on the current COVID tests. 5% is around minimum for many tests, even less for others.
Jolly, is it evenly spread? In other words, are false positives as common as false negatives? How do you prove a false negative, if a certain amount of time (week) has passed before the re-test?
-
@Jolly said in Positive. Again.:
@Klaus said in Positive. Again.:
Do we have false positive/false negative rates on the test? Maybe this is merely due to false test results.
NO clinical test is accurate 100% of the time. I've heard error rates of 10% on the current COVID tests. 5% is around minimum for many tests, even less for others.
Jolly, is it evenly spread? In other words, are false positives as common as false negatives? How do you prove a false negative, if a certain amount of time (week) has passed before the re-test?
wrote on 13 Apr 2020, 19:41 last edited by@Rainman said in Positive. Again.:
@Jolly said in Positive. Again.:
@Klaus said in Positive. Again.:
Do we have false positive/false negative rates on the test? Maybe this is merely due to false test results.
NO clinical test is accurate 100% of the time. I've heard error rates of 10% on the current COVID tests. 5% is around minimum for many tests, even less for others.
Jolly, is it evenly spread? In other words, are false positives as common as false negatives? How do you prove a false negative, if a certain amount of time (week) has passed before the re-test?
Depends on the methodology of the test and the person running it. Sometimes, you can have an adequate test, but a lousy tester.
An example: Monoclonal antibody tests for pregnancy are ubiquitous. You can buy them for $1/each at the dollar store. A study was done with known specimens comparing everyday folks with lab testing personnel. Average people got the right answer around 80% of the time. The lab people were right about 95% of the time, with their accuracy equaling the stated accuracy of the test.
It's all about technique, specificity and sensitivity. The more sensitive a test, the less specific. The more specific, the less the sensitivity. Sloppy technique equals sloppy results. And any test with a manual component such as pipetting, is more accurate with somebody who has done the test many, many times.
And don't forget, you have to have a good sample, to get a good test.
Another factor...Sometimes the test is collected correctly, done correctlyand reported erroneously. Doesn't happen much anymore because most things are interfaced, but guys have been known to type in the wrong answer.
So errors occur in three phases - pre-analytical, analytical and post-analytical, with the majority of problems usually occurring in the pre-analytical phase( collection and everything before actual testing).
-
wrote on 13 Apr 2020, 19:50 last edited by
In the tests I've been involved with designing, for use with the droplet digital pcr instruments I work with, a lot of effort goes into the "no answer" category. If the data does not look good enough to the results algorithm, the algorithm refuses to give an answer.
-
@Klaus said in Positive. Again.:
Do we have false positive/false negative rates on the test? Maybe this is merely due to false test results.
NO clinical test is accurate 100% of the time. I've heard error rates of 10% on the current COVID tests. 5% is around minimum for many tests, even less for others.
wrote on 13 Apr 2020, 21:08 last edited by@Jolly said in Positive. Again.:
@Klaus said in Positive. Again.:
Do we have false positive/false negative rates on the test? Maybe this is merely due to false test results.
NO clinical test is accurate 100% of the time. I've heard error rates of 10% on the current COVID tests. 5% is around minimum for many tests, even less for others.
If that test would have a 10% false positive rate, then it would be mostly useless, no? From what I understand, the number of tests with positive results is around or below 10%. At a 10% false positive rate, most of those "positive" cases would in fact be negative.
-
@Jolly said in Positive. Again.:
@Klaus said in Positive. Again.:
Do we have false positive/false negative rates on the test? Maybe this is merely due to false test results.
NO clinical test is accurate 100% of the time. I've heard error rates of 10% on the current COVID tests. 5% is around minimum for many tests, even less for others.
If that test would have a 10% false positive rate, then it would be mostly useless, no? From what I understand, the number of tests with positive results is around or below 10%. At a 10% false positive rate, most of those "positive" cases would in fact be negative.
wrote on 13 Apr 2020, 21:18 last edited by@Klaus said in Positive. Again.:
@Jolly said in Positive. Again.:
@Klaus said in Positive. Again.:
Do we have false positive/false negative rates on the test? Maybe this is merely due to false test results.
NO clinical test is accurate 100% of the time. I've heard error rates of 10% on the current COVID tests. 5% is around minimum for many tests, even less for others.
If that test would have a 10% false positive rate, then it would be mostly useless, no? From what I understand, the number of tests with positive results is around or below 10%. At a 10% false positive rate, most of those "positive" cases would in fact be negative.
I don't know if that is the actual number, I'm working off of gossip, hearsay and the news media. But you would be much better off with false positives than false negatives.