11–20 of 88 posts
As a person who had a hypertensive crisis late last year, nothing boils my blood (yes, a pun, I am fine now) more than how people measure blood pressure incorrectly in doctors offices and even hospitals. There are many different things than can increase a BP measurement above the "baseline" including talking, moving around, not having rested but also just waking up, not being in a supported position, only a single va…
The first way has been done in studies for years, maybe even a decade ago.
Earlobes.
I'll search for the papers later and link them.
https://scholar.google.com/scholar?q=earlobe+blood+pressure
https://www.todaysmedicaldevelopments.com/article/wearable-b...
I've always wanted a Garmin linked ear-cuff device that uses the earlobe for heartrate and blood pressure and then doubles as music playback or alerts from the watch. You could even do body temperature from the earlobe reliably.
[dead]
Edit: To clarify, plenty of things have been tried besides the cuff, but most patients who need something more sophisticated than that are already sick enough to be in the ICU, where an a-line can be placed. This is really a solution in search of a problem.
This approach has been tried before. Bill Softky describes a startup he'd worked for using a similar sound-transduction continuous non-invasive blood-pressure monitoring technique. It ... had problems: Our non-invasive device was supposed to measure blood pressure just as accurately [as an arterial line], but without the cutting, using specially-sculpted sonic vibrations and fancy algorithmic analysis, which was my j…
Looking at the diagram I’d suspect they use accelerometer information from both the wrist watch and the upper arm mounted sensor to remove the effects of arm motion. At simplest it could only check when the arm is in a neutral position. But I’d expect they did something more complex/better than that.
I'm obviously distant from the project, but a team of SWEs spending years trying to make nondeterministic data deterministic suggests a fairly deep problem.
Aktiia has a product on the market which supposedly already does this https://aktiia.com/uk/ . It is based on an optical method though.
The experiment has included multiple fasting periods, with a maximum of 7 days as well as changing one variable at a time in categories such as diet and exercise. The results have been very interesting and I intend to continue on this path until at least the end of the year.
As part of the data collection I have been taking my blood pressure a minimum of twice a day, sometimes more. Also blood glucose, ketones and (consumer) EKG.
The first thing that jumped at me was the inaccuracy or variability of these measurements. I even got a Dexcom continuous glucose monitor. Interesting but useless for my purposes. The thing produced 20% error with respect to finger poke measurements. And, then again, when I got a calibration kit to check my finger poke meter, the calibration range is approximately +/- 18%. In other words, unless you hit extremes it feels like these measurements are almost useless. You can kind of tell you are going up or down, yet don't really know where you are.
The same, of course, has been true of blood pressure measurements. I went through three consumer machines. I can't say any of it is accurate because there are too many variables. I have run multiple experiments with regards to where and how to measure BP. All I can determine are relative changes by effectively measuring under as close to the same conditions as possible twice a day, morning and evening (both before meals).
During the last month or so I have been using a protocol I learned from one of Andrew Huberman's presentations (can't remember which one or I would post a link). I believe he was interviewing a researcher who explained the process they use during their studies. In simple terms, they take three measurements and then average. The first is after 15 minutes sitting, feet on the ground, back supported, no movement, no speaking, no activity. The second and third are at 5 minute intervals under the same conditions. In other words, the entire process takes at least 25 minutes.
After adopting this approach I have been seeing wildly different numbers with respect to the single measurement protocol I had been using for two months. In addition to that, the standard deviation of the computed values are much tighter now.
This experience, so far, has made me wonder about just how many people might be misdiagnosed and put on medication every year because of bad data. I can see the value in having more data, of course. Yet, continuous data is only good if it is accurate to within a reasonable margin.