4 ms·

> so nobody measures it as frequently as they do their pulse. Once we have the technology to accurately and passively measure blood pressure throughout the day

> so nobody measures it as frequently as they do their pulse. Once we have the technology to accurately and passively measure blood pressure throughout the day It's surprisingly tricky. The way a BP cuff works (the auscultatory method[1]) is that the pressure is continuously decreased and you listen for the sound of a pulse. When the cuff pressure is higher than your systolic (peak) pulse, bloodflow is blocked and you don't hear a pulse. When the cuff pressure is lower than your diastolic, the cuff is no longer blocking bloodflow at all, and the pulse suddenly becomes much quieter because it's not restricted. It's only between the two pressures that you hear it extra clearly. So the problem with a sensor is that you can't just measure pressure against the skin. At minimum, it really should actually impede a large artery. You can squeeze the arm as tight as you like but if the brachial artery is not affected then your measurement will be very poor. You can do fancy computer vision etc to look at special veins/arteries, and coupled with some demographic assumptions you can infer somewhat accurate measurements- but they really miss out on outliers, which is really a lot of the point of taking measurements. This also generally will require an upper arm strap or something, and be fairly tight. Not great to wear always At some point, I think we will probably see implantable blood pressure monitoring. We do something similar with implantable glucose monitors- which need regular replacement, unfortunately. BP probably wouldn't. Problem is that direct pressure measurement is usually done by tapping off an artery- something that will never be routine. You'd want to do some combination of acoustic and pressure measurement, and it'd need to be calibrated from a base station regularly. [1]: https://en.wikipedia.org/wiki/Blood_pressure_measurement#/media/File:Blood_pressure_measurement_principle.svg https://en.wikipedia.org/wiki/Blood_pressure_measurement#/me...

jkingsman 3y ago
Worth mentioning that BP is about arterial, not venous pressure. Venous pressure is an order of magnitude lower than arterial pressure; usually 0-10mmHg.
- hwillis 3y ago
  ty, edited. oopsies!
alphanumeric0 3y ago
Yes, the most accurate method is via arterial catheter. It's basically a needle directly into the artery.
boringuser2 3y ago
I'm confused. The technology already exists and is approved by the EU. https://aktiia.com/ https://aktiia.com/
- hwillis 3y ago
  Photoplethysmogram devices like this measure blood volume. That works great for your heart rate. It is pretty terrible for your blood pressure. In theory they measure the speed an artery expands as blood passes. 1. Since they can't measure blood flow, only volume, they can't tell if your BP is low or if they just aren't centered on an artery. They can't measure diastolic, only the difference between systolic and diastolic. 2. Since they don't know the flow, they can't even guess the size of the artery, so they're basically just guessing how much it should be expanding. They also have to know how tall you are to even get close. 3. They measure the derivative of the increase in volume to try to correlate that with blood pressure. That's very variable with the specific artery and the condition of the artery- there's a lot of (primarily age) variance in artery elasticity. Papers indicating accuracy should be subject to high scrutiny. The device basically starts out by guessing that you're healthy and normal for your age and height. If you take a bunch of people that are healthy for their age and height (most people), the device will perform very well. It will also be practically useless. If you have ideopathic large swings in blood pressure, like POTS, or fainting, or migraines, a device like this can give you a bit of warning. It will not give you accurate readings or track long-term changes, and those issues are made much worse because of opaque calibrations etc. https://en.wikipedia.org/wiki/Photoplethysmogram https://en.wikipedia.org/wiki/Photoplethysmogram
  - boringuser2 3y ago
    Do you think the EU regulatory agencies haven't properly accounted for the claims made by the makers of this device? It looks very promising and it is clinically approved. Your theory craft seems to check out, but they could have proprietary technology that Huawei is yet to clone/rip-off, so it's not a well-known methodology. Funnily enough, they actually tested some of their algorithms in China based on their papers, so maybe it is just a matter of time. That being said, if you have industry inside knowledge, that's a little different.
TechBro8615 3y ago
I wonder if there are methods that could take advantage of natural movements throughout the day, where your body naturally emulates what the cuff forces it to do. For example, when you transition suddenly from sitting to standing, is it possible for a sensor to measure some proxy metric that, combined with an accelerometer, can be used to infer blood pressure? If there is some method like this, even if it has large margins of error, maybe cumulative measurements could converge on a fairly accurate reading? (I'm not a medical professional nor do I know much about sensing, or even really the basics of how blood pressure is measured or what it indicates.)
joshvm 3y ago
Maybe a related question, but how good a proxy is cuff pressure for a gold standard method that measures inside an artery? Whenever I've had BP measured for clinical studies (not focused on BP specifically) they used a mix of cuff and continuous fingertip sensors depending on the activity.
- hwillis 3y ago
  It's pretty good, about +/- 5 mmHg. At worst it can bump you up/down a category- drinking a couple glasses of water can do that too, and the medical treatment would not be any different. The electric cuff machines can be over twice as variable- about 1/20 person-measurements will read off by >20 mmHg, which is a hugely important difference.
krzat 3y ago
Continous measurement of body temperature, blood pressure and heart rate via implant could be enormously helpful to health for so many reasons...

jkingsman 3y ago

Worth mentioning that BP is about arterial, not venous pressure. Venous pressure is an order of magnitude lower than arterial pressure; usually 0-10mmHg.

hwillis 3y ago
ty, edited. oopsies!

alphanumeric0 3y ago

Yes, the most accurate method is via arterial catheter. It's basically a needle directly into the artery.

boringuser2 3y ago

I'm confused. The technology already exists and is approved by the EU. https://aktiia.com/ https://aktiia.com/

hwillis 3y ago
Photoplethysmogram devices like this measure blood volume. That works great for your heart rate. It is pretty terrible for your blood pressure. In theory they measure the speed an artery expands as blood passes. 1. Since they can't measure blood flow, only volume, they can't tell if your BP is low or if they just aren't centered on an artery. They can't measure diastolic, only the difference between systolic and diastolic. 2. Since they don't know the flow, they can't even guess the size of the artery, so they're basically just guessing how much it should be expanding. They also have to know how tall you are to even get close. 3. They measure the derivative of the increase in volume to try to correlate that with blood pressure. That's very variable with the specific artery and the condition of the artery- there's a lot of (primarily age) variance in artery elasticity. Papers indicating accuracy should be subject to high scrutiny. The device basically starts out by guessing that you're healthy and normal for your age and height. If you take a bunch of people that are healthy for their age and height (most people), the device will perform very well. It will also be practically useless. If you have ideopathic large swings in blood pressure, like POTS, or fainting, or migraines, a device like this can give you a bit of warning. It will not give you accurate readings or track long-term changes, and those issues are made much worse because of opaque calibrations etc. https://en.wikipedia.org/wiki/Photoplethysmogram https://en.wikipedia.org/wiki/Photoplethysmogram
- boringuser2 3y ago
  Do you think the EU regulatory agencies haven't properly accounted for the claims made by the makers of this device? It looks very promising and it is clinically approved. Your theory craft seems to check out, but they could have proprietary technology that Huawei is yet to clone/rip-off, so it's not a well-known methodology. Funnily enough, they actually tested some of their algorithms in China based on their papers, so maybe it is just a matter of time. That being said, if you have industry inside knowledge, that's a little different.

TechBro8615 3y ago

I wonder if there are methods that could take advantage of natural movements throughout the day, where your body naturally emulates what the cuff forces it to do. For example, when you transition suddenly from sitting to standing, is it possible for a sensor to measure some proxy metric that, combined with an accelerometer, can be used to infer blood pressure? If there is some method like this, even if it has large margins of error, maybe cumulative measurements could converge on a fairly accurate reading? (I'm not a medical professional nor do I know much about sensing, or even really the basics of how blood pressure is measured or what it indicates.)

joshvm 3y ago

Maybe a related question, but how good a proxy is cuff pressure for a gold standard method that measures inside an artery? Whenever I've had BP measured for clinical studies (not focused on BP specifically) they used a mix of cuff and continuous fingertip sensors depending on the activity.

hwillis 3y ago
It's pretty good, about +/- 5 mmHg. At worst it can bump you up/down a category- drinking a couple glasses of water can do that too, and the medical treatment would not be any different. The electric cuff machines can be over twice as variable- about 1/20 person-measurements will read off by >20 mmHg, which is a hugely important difference.

krzat 3y ago

Continous measurement of body temperature, blood pressure and heart rate via implant could be enormously helpful to health for so many reasons...