4 ms·

Didn’t someone already try to do blood analysis with just a drop of blood and failed miserably?

by Osiris 3y ago

yjftsjthsd-h 3y ago
So? One failed attempt doesn't mean nobody else will ever succeed.
prettyStandard 3y ago
I assume you mean this in jest. I get it. To be fair/serious if you add the "near past" with "near future" you might get enough time for things to change. Was there something fundamentally impossible about blood analysis with a drop of blood?
- dekhn 3y ago
  At the time of Theranos's debunking there were some posts to HN showing that it was mathematically implausible. Something about signal to noise and specificity at extremely small molecule counts. I don't know how to find it, but I assume somebody in the hacker news hivemind knows what I mean and can dig up the link.
  - kevin_thibedeau 3y ago
    Breathalyzers can't serve as a credible proxy for blood alcohol down to 100ppm without any consideration of physiological differences and yet they're accepted as divination tools.
- CamperBob2 3y ago
  One problem is that blood isn't really a homogeneous substance. When dealing with small test samples, you might see different results, perhaps widely-varying ones, from one drop to the next. (Play around with a glucose meter sometime.) Things like where the blood was drawn from, what contaminants were present on the skin at the time, and the age and health of the patient can all have an effect. Also, many tests require a centrifuge, so that's something else you have to account for if your goal is to perform the same tests with only one drop of blood. Unless your plan is to pull some Star Trek-level technology out of your hat. TL,DR: you didn't need to consult a team of PhDs to debunk Theranos's business plan, you could have asked any experienced med tech. The investors didn't bother doing either.
  - tough 3y ago
    They wanted to believe
  - mcbits 3y ago
    > Also, many tests require a centrifuge, so that's something else you have to account for if your goal is to perform the same tests with only one drop of blood. Unless your plan is to pull some Star Trek-level technology out of your hat. That was basically the plan. Or failing that, to convince people that such technology existed or was close to existing. It's not too implausible that the centrifuge could be replaced with diffraction analysis or something since the goal is to test for the same things, not necessarily to "do the same tests." But that doesn't help if you're looking for something present in one in a million blood cells and only have a quarter-million cells to look at.
- rcxdude 3y ago
  There were some fundamental issues with some of what Theranos was claiming to do. Mainly when detecting substances in blood with very low concentrations using very small volumes, especially when the blood is taken from the finger. In those cases even if you could pick apart every atom in the sample perfectly you still would not be able to do what Theranos was claiming to do, because the substances would either not be in the sample or there would be so few molecules that the uncertainty in the real concentration would be too large. real-time monitoring is a dream which in principle does not suffer the same fundamental issues, but does require much larger jumps in technology compared to current means of running tests.

Didn’t someone already try to do blood analysis with just a drop of blood and failed miserably?

by Osiris 3y ago

Didn’t someone already try to do blood analysis with just a drop of blood and failed miserably?

yjftsjthsd-h 3y ago

So? One failed attempt doesn't mean nobody else will ever succeed.

prettyStandard 3y ago

I assume you mean this in jest. I get it. To be fair/serious if you add the "near past" with "near future" you might get enough time for things to change. Was there something fundamentally impossible about blood analysis with a drop of blood?

dekhn 3y ago
At the time of Theranos's debunking there were some posts to HN showing that it was mathematically implausible. Something about signal to noise and specificity at extremely small molecule counts. I don't know how to find it, but I assume somebody in the hacker news hivemind knows what I mean and can dig up the link.
- kevin_thibedeau 3y ago
  Breathalyzers can't serve as a credible proxy for blood alcohol down to 100ppm without any consideration of physiological differences and yet they're accepted as divination tools.
CamperBob2 3y ago
One problem is that blood isn't really a homogeneous substance. When dealing with small test samples, you might see different results, perhaps widely-varying ones, from one drop to the next. (Play around with a glucose meter sometime.) Things like where the blood was drawn from, what contaminants were present on the skin at the time, and the age and health of the patient can all have an effect. Also, many tests require a centrifuge, so that's something else you have to account for if your goal is to perform the same tests with only one drop of blood. Unless your plan is to pull some Star Trek-level technology out of your hat. TL,DR: you didn't need to consult a team of PhDs to debunk Theranos's business plan, you could have asked any experienced med tech. The investors didn't bother doing either.
- tough 3y ago
  They wanted to believe
- mcbits 3y ago
  > Also, many tests require a centrifuge, so that's something else you have to account for if your goal is to perform the same tests with only one drop of blood. Unless your plan is to pull some Star Trek-level technology out of your hat. That was basically the plan. Or failing that, to convince people that such technology existed or was close to existing. It's not too implausible that the centrifuge could be replaced with diffraction analysis or something since the goal is to test for the same things, not necessarily to "do the same tests." But that doesn't help if you're looking for something present in one in a million blood cells and only have a quarter-million cells to look at.
rcxdude 3y ago
There were some fundamental issues with some of what Theranos was claiming to do. Mainly when detecting substances in blood with very low concentrations using very small volumes, especially when the blood is taken from the finger. In those cases even if you could pick apart every atom in the sample perfectly you still would not be able to do what Theranos was claiming to do, because the substances would either not be in the sample or there would be so few molecules that the uncertainty in the real concentration would be too large. real-time monitoring is a dream which in principle does not suffer the same fundamental issues, but does require much larger jumps in technology compared to current means of running tests.