3 ms·
> If they are not, then in principle nanotainers can not work because the sample size is too small. I am the last person who wants to defend Theranos, but theo
by drcode 11y ago
> If they are not, then in principle nanotainers can not work because the sample size is too small.
I am the last person who wants to defend Theranos, but theoretically if there are only a finite number of ways for tissue contaminants to enter a sample, these contaminants could be corrected for with statistical methods and accurate results could still be possible.
- mbreese 11y agoYou can't correct for inadequate sample sizes with statistics and still expect a clinically relevant result (the confidence interval would be too large). The issue isn't contamination in small samples, it's that the biomarkers they need to measure may not be accurately measurable at such small volumes - in particular with the value ranges required for the test. It may just be that their "nanotainer" tests may only be viable for boolean tests like their herpes test (herpes DNA present - yes / no).
- cbsmith 11y agoWait, what? How do you statistically compensate for a sample size of one, or zero? The issue here isn't contaminants in the sample, it's the markers (i.e., what is being tested for) in the sample. If marker isn't uniformly distributed, then these small samples might not even find it in them.