re: Antibodies ...why are they not getting any love?

quote:

---

^More phony double-talk from our Big Pharma shill

---

Do you think the test the Stanford Drs. gave was developed by Stanford?

Nope. It was developed by Hangzhou Biotest Biotech.

quote:

---

It was developed by Hangzhou Biotest Biotech.

---

Indeed. and although Stanford evaluated the tests for accuracy, Denmark arrived at dramatically different (poorer) results for the same test. Takeaway: There may have been significantly overstated AB positives in Stanford's study.

quote:

---

1,045 tests. 35 true positives and 10 false positives. The false positives are equivalent to 28% of the true positives. This is too much of an error to base policy decisions off of.

---

Your example uses an awfully small sample of 1035 people.

Ramp that up to 1 million. 2.5% sick would give 25,000 sick. From that if 99% accurate, you would get 24,750 true positives and 250 false positives. This would seem reasonable to base decisions from.

This is what I was trying to tell texrider, but he failed to grasp the concept.

quote:

---

a 99% accuracy

---

Doubtful I'm posting anything you do not know, in fact quite the opposite. But the key element in a low incidence environment is specificity. Sensitivity error is less important. Normally, "accuracy" is a combo of specificity and sensitivity.

The Stanford study likely overstated specificity of its Chinese test capacity. Minor errors there could lead to significant errors in outcome.

quote:

---

There will always be naysayers but I hope they can lock down their methodologies and power up the statistics. I am most anxious to see the resulting CFR drawn down into a more realistic number.

---

Well this is why the New York data is by far the best antibody data we have thus far, largely because (a) there is much more randomization (those randomly selected at grocery stores) that is absent in the Stanford study (targeted facebook ads; people were sharing it; and self-selection to volunteer) which should minimize the potential bias (although those grocery shopping may be more likely to be infected).

In addition, besides the overall accuracy of the test, the biggest statistical problem with these types of studies is the base rate of infections. As that gets larger, even a less accurate test is far less problematic. Because New York (especially NYC) clearly has a significant portion infected, the results are much more reliable and the uncertainty around the test accuracy has only marginal impact relatively speaking.

Finally, after reading the updated Stanford study, besides the potential bias AND the questions surrounding their representation of the test accuracy, is the impact (and their decisions) regarding demographic adjustments. It’s possible the adjustments for one individual’s (Hispanic, male, 18-64) will be equivalent to 4.36 individuals in the sample or the equivalent 2,524 individuals extrapolated to the county data while another individual’s result (white, female, 65+) will be equal to 0.39 individuals in the sample of 227 individuals extrapolated to the county data.

In addition, the groups adjusted upward by a significant amount (Hispanics; 24.9% from 8%) have a significantly higher crude rate (4.9%) while the groups adjusted downward by a significant amount (whites; 35.4% from 64.1%) have a significantly lower crude (1%). And if their adjusting for test accuracy at the group level, this will only amplify the differences, but this could be just amplifying group biases or sampling/measurement error.

In other words, their entire study, especially without the raw data, has so many layers of uncertainty and bias potential, built on top of one another, that it’s impossible to actually know what the approximate infection rate is of the sample itself, let alone how that extrapolates to the county population. But in all likelihood, it’s far lower than their results indicate since it’s so far incompatible right the far better NY data (or another decent study from Germany), that it’s obvious the authors aren’t trying to falsify their hypothesis and are merely aiming to confirm their own biases they had been presenting for weeks before they even started it.

The same test was us d. In the Miami test.

(no message)

quote:

---

Using terms like sensitivity and specificity will really get texridder upset.

---

You're a fool. Tell me, why did you apologize to Wednesday and I when you didn't know (or even consider) the significance that false positives would have on her decision to rely on an antibody test to clear the way for her to visit her elderly father.

Now you're embarrassed having been such a dumbass that you post a dig about me almost every post to deflect from your ignorance.

quote:

---

biomedomics is the.U.S. distributor for Hangzhou Biotest Biotech, which actually makes the test.

---

I don’t believe this is true. Their own website says they manufacture it and another company distributes it.

quote:

---

The new test, developed and manufactured by BioMedomics, will be available through BD and distributed exclusively by Henry Schein, Inc. to health care providers throughout the United States.

---

And the two companies that were distributing the unapproved Chinese tests (Premier’s was used in Stanford and USC study), were distributing it BEFORE BioMedomics test was first released anyways.

quote:

---

Two U.S. companies — Premier Biotech of Minneapolis and Aytu Bioscience of Colorado — have been distributing the tests from unapproved Chinese manufacturers, according to health officials, FDA filings and a spokesman for one of the Chinese manufacturers.

---