- My Forums
- Tiger Rant
- LSU Recruiting
- SEC Rant
- Saints Talk
- Pelicans Talk
- More Sports Board
- Fantasy Sports
- Golf Board
- Soccer Board
- O-T Lounge
- Tech Board
- Home/Garden Board
- Outdoor Board
- Health/Fitness Board
- Movie/TV Board
- Book Board
- Music Board
- Political Talk
- Money Talk
- Fark Board
- Gaming Board
- Travel Board
- Food/Drink Board
- Ticket Exchange
- TD Help Board
Customize My Forums- View All Forums
- Show Left Links
- Topic Sort Options
- Trending Topics
- Recent Topics
- Active Topics
Started By
Message
re: Coronavirus Disease 2019 (COVID-19) ***W.H.O. DECLARES A GLOBAL PANDEMIC***
Posted on 4/24/20 at 8:09 pm to mattz1122
Posted on 4/24/20 at 8:09 pm to mattz1122
Trump supporters are now claiming that when President Trump was talking about bringing UV light into the body, he was actually talking about this "Healight" device. It's inserted through the trachea and supposedly emits UV that kills the virus in the lungs.
LINK
LINK
1
Posted on 4/24/20 at 8:51 pm to GOP_Tiger
Sycophant Trump supporters are fricking morons. The idiot president just asked if injecting disinfectant is viable way to beat COVID. There’s literally no way for the morons here to defend this claim, but here goes...
Posted on 4/24/20 at 9:24 pm to Lsut81
quote:I posted this in thread on the PT board, but since this is the most inaccurate test I’ve seen, it’s likely all 100% of the positives were false positives:
More antibody testing coming in...
They tested 400 people in one week, and 700 people the next week, yet got the same 6% both times. The sample the second week should be considerably higher than the sample the next week 2+ times higher the R0 was 1 or above.
So that makes me think that 6% is the false positive rate. And it took me some time, but I found the company that provided the tests.
https://www.miamiherald.com/news/coronavirus/article242260406.html
quote:And it cites the accuracy in the article:
The DMG rapid blood test was that was delivered to Miami-Dade in partnership with BioMedomics.
quote:Which is exactly what the companies clinical analysis found:
Because the test relies on the level of antibody production in each individual’s blood, there is an 89 percent to 91 percent accuracy rate.
Biomedics: COVID-19 IgM/IgG Rapid Test
quote:
Of the 397 blood sample from SARS-CoV-2-infected patients, 352 tested positive, resulting in a sensitivity of 88.66%.
quote:So this is by far the most inaccurate test I’ve seen (same as the Chelsea/Boston study).
Twelve of the blood samples from the 128 non-SARS-CoV-2 infection patients tested positive, generating a specificity of 90.63%.
In fact, it’s so inaccurate that we would expect 9.37% positive tests if 0% of the sample was truly negative (100% false positives) with a 05% confidence interval between 14.42% and 4.33%. So 6% is well within that range (within about an 80% confidence interval).
So the fact that they got 6% for both samples when the 2nd week’s sample likely should have been 2 more time larger, and since it’s so inaccurate, we would expected over 9% false positives, but 6% false positives false within the sampling range, all positive tests were likely false positives.
Posted on 4/24/20 at 9:41 pm to buckeye_vol
Do you understand sensitivity and specificity?
Essentially 90% and 90% is a really good test, especially for both at the same time in the same test.
The tests that exist in medicine are not 100% sensitive and specific.
And if it is accurate, you would see 6% positive no matter the sample size.
If you tested 400 and then 4000, you should get 6% both times with a test with high sensitivity and specificity.
Nuclear cardiac stress tests function at that level.
Essentially 90% and 90% is a really good test, especially for both at the same time in the same test.
The tests that exist in medicine are not 100% sensitive and specific.
And if it is accurate, you would see 6% positive no matter the sample size.
quote:That is exactly what would be expected with a test that is both highly sensitive and specific.
They tested 400 people in one week, and 700 people the next week, yet got the same 6% both times.
If you tested 400 and then 4000, you should get 6% both times with a test with high sensitivity and specificity.
Nuclear cardiac stress tests function at that level.
This post was edited on 4/24/20 at 9:52 pm
Posted on 4/24/20 at 9:44 pm to mattz1122
quote:
Sycophant Trump supporters are fricking morons. The idiot president just asked if injecting disinfectant is viable way to beat COVID. There’s literally no way for the morons here to defend this claim, but here goes...
While conservative, I chose not to vote for the first time since both candidates were terrible.
That said, a Trump supporter does not lose intelligence when Trump says Trump stuff anymore that a progressive loses intelligence every time Pelosi or AOC says stupid shite.
This is not a fricking football game where one team has to constantly shite talk the other. Grow up.
Posted on 4/24/20 at 9:53 pm to buckeye_vol
Can you explain what you are saying like I am a third grader? My take away from your post is that you are saying the Miami testing was so inaccurate it is worthless.
Posted on 4/24/20 at 9:53 pm to Scruffy
quote:Yes. I have a master’s in statistics.
Do you understand sensitivity and specificity?
quote:No it’s not, especially since the true accuracy is so dependent on the base rate of the population they are sampling from.
Essentially 90% and 90% is a really good test, especially for both at the same time in the same test.
With a specificity of 90.63%, any result that yields a positive result below 9.37% means that it’s more likely that all 100% are false positives than not.
And if the infection prevalence was truly 6%, we would expect 14.1% positives, with 8.8% false positives and 5.3% true positives so 62.4% would be false positives.
In other words, DO YOU UNDERSTAND specificity or sensitivity, more importantly do you understand Bayes’ Theorem and its application here? Otherwise, you would know that this test is far too inaccurate to measure a anything with even a marginally low base rate.
Posted on 4/24/20 at 9:56 pm to buckeye_vol
quote:Well, I’ll concede.
Yes. I have a master’s in statistics.
There is a reason I hate statistics.
Posted on 4/24/20 at 9:57 pm to buckeye_vol
quote:
Yes. I have a master’s in statistics.
Whoa, we got a badass in here
Posted on 4/24/20 at 10:00 pm to SloaneRanger
quote:That’s exactly correct.
Can you explain what you are saying like I am a third grader? My take away from your post is that you are saying the Miami testing was so inaccurate it is worthless.
Specificity (true negatives) is used to calculate false positives (1-specificity; 9.38%) and sensitivity (true positives) is used to calculate false negatives (1-sensitivity; 11.34%).
So the lower the true infection rate (base rate) the more problematic a lower specificity becomes and the inverse is true for sensitivity.
In this case, given that 9.38% of negatives would have a positive test, then the very lowest positive sample result we would expect is 9.38% (sample with 100% negatives), so anything less than 9.38% would be expected to be a false positive.
Now the because the specificity was based on a sample of 128, the true specificity could range between about 4 and 14 percent (95% confidence interval). So the fact that they got 6% from both samples a week apart, and we would expect considerably more positives the 2nd week (since infections were growing rapidly), I suspect that the 6% corresponded to the specificity (94%), at least that sample’s population. That means it’s probably both samples were 100% negative and all positives were false positives. At the very least, only a couple could have been true positives.
This post was edited on 4/24/20 at 10:05 pm
Posted on 4/24/20 at 10:03 pm to buckeye_vol
The problem is we don't have the raw data or a report to know what they did Maybe the corrected their data but we don't know that. Basically all of these announcements are suspect until there are more accurate tests available OR the infected numbers get well into the double digits. At least the NYC results were high enough to say that all of the tests couldn't have been false negatives.
That still leaves the issue of cross sensitivity to other common coronavirus ABs.
That still leaves the issue of cross sensitivity to other common coronavirus ABs.
Posted on 4/24/20 at 10:04 pm to buckeye_vol
Not arguing with you at all, as I am well out of my league. But how did these UM researchers miss all this? Are they just blowing smoke up our arses?
Posted on 4/24/20 at 10:10 pm to SloaneRanger
quote:Well I found it interesting that the only press release was from the health department and not UM. And universities are quick to publish press releases of questionable, and even poor, studies, especially when they’ll garner a lot of attention (and this topic garners a ton of it).
But how did these UM researchers miss all this?
So my guess is that the health department is more likely to lack the same scientific acumen to understand the problems with their results, and they have more political motivations on top of that.
The UM researchers probably realizes the problems, which is why they haven’t been so quick to make a quick to publish a press release, especially since this is an ongoing study and they may get more valid results in later weeks to be able to control for these problems a bit.
Posted on 4/24/20 at 10:10 pm to SloaneRanger
quote:
Can you explain what you are saying like I am a third grader?
If the test has a false positive rate of 10%, 10 out of 100 healthy people would show as positive even though they aren't. If California tested 3000 people and around 300 tested positive, you can't extrapolate that 10% of all Californians had once had the virus. In this example, a result of around 10% would be meaningless. That is, of course, unless you're on the OT and then every number is proof that this is just a cold.
Posted on 4/24/20 at 10:11 pm to Scruffy
quote:
Do you understand sensitivity and specificity?
Essentially 90% and 90% is a really good test, especially for both at the same time in the same test.
The tests that exist in medicine are not 100% sensitive and specific.
That's terrible for an antibody test. The one made by Abbott Labs is far, far better.
"Abbott claims that its new test, when performed at least two weeks after a patient has first exhibited symptoms, has sensitivity of 100% and specificity of 99.5%."
Posted on 4/24/20 at 10:15 pm to borotiger
Got it. Thanks. If the article is to be believed, the study was headed up by a UM professor of medicine and public health. That's pretty embarrassing if he missed something that basic.
Posted on 4/24/20 at 10:19 pm to slacker00
quote:Yeah. And given they provided more detailed results from around the state, and provided the range of specificity (93 to 100 percent) we can at least infer that the 32% of the state that had 3.5% positive rate, roughly means the specificity is around 96 to 97 percent (and 96% is the specificity of the first FDA approved test).
At least the NYC results were high enough to say that all of the tests couldn't have been false negatives.
So using those results, we can infer that NYC’s prevalence is probably around 19-20 percent (down from 21.2%) and the entire state is around 11-12 percent (down from 13.9%).
The Stanford study is the only terrible study thus far for an infer of reasons (recruitment, adjustments, etc.), inexcusably terrible. While the other studies have been better from a design standpoint (LA, NY, San Miquel County, Miami) only the NY (large enough base rates; still concerns about potential sample bias) and San Miguel (testing entire county; differentiates borderline results more likely to be false positives) are the only ones where we can get at least infer the results are close enough to accurate to generalize and 0% is not in the range.
Posted on 4/24/20 at 10:37 pm to borotiger
quote:
a Trump supporter does not lose intelligence when Trump says Trump stuff anymore
Posted on 4/25/20 at 12:38 am to SloaneRanger
quote:
Can you explain what you are saying like I am a third grader? My take away from your post is that you are saying the Miami testing was so inaccurate it is worthless.
Then scruffy says it’s plenty accurate and my dumb farming arse if left here clueless
ETA: as I read forward I see that Scruffy conceded and Buckeye is one of those stats assholes
This post was edited on 4/25/20 at 12:46 am
Posted on 4/25/20 at 12:52 am to borotiger
quote:
This is not a fricking football game where one team has to constantly shite talk the other. Grow up.
Trump is in power and in charge so he is the subject. It's not necessarily about being on anyone's team. Trump is a total moron who thinks Californians won't have bad wildfires if they rake forest floors. Trump thinks Americans must show ID to buy a loaf of bread. Trump thinks the British had airfields during the revolutionary war.
This post was edited on 4/25/20 at 12:54 am
Page 1079 of 1190
Popular
Back to top
Follow TigerDroppings for LSU Football News