Hacker News new | past | comments | ask | show | jobs | submit login
Clinical validation of a blood-pressure sensor for continuous health monitoring (wiley.com)
87 points by bookofjoe on May 22, 2023 | hide | past | favorite | 18 comments



I'm not surprised this works - accurate cuffless blood pressure has been validated from other physiological sources like PPG / pulse oximetry. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7141397/

and is even now FDA-cleared for clinical use: e.g. https://www.bio-beat.com/post/the-biobeat-cuffless-wrist-blo...


FDA cleared is more/less meaningless vs. FDA approved.

Pulse ox and HRM such as Apple Watch and the cheapy blue and white finger readers are ubiquitous and not absurdly inaccurate (usually).


Wait isn’t PPG for heart rate?


Not specifically. Photoplethysmography just means, roughly, light based volume measurement. It can measure pulse when done crudely. If you have a baseline for blood pressure and a sensitive enough PPG sensor you can detect different levels of volume and infer pressure. It is analogous to measuring water pressure based on the diameter of a filled rubber hose.

To the GP's point, the largest blood pressure monitor company is Omron. Portable monitors have been around. https://www.omron.com/global/en/technology/omrontechnics/vol... is a good place to learn more about the advantages and disadvantages of different approaches. Piezoelectric and PPG would have the advantage of lower cost, lower mechanical complexity, and continuous monitoring instead of point measurements.


> Portable monitors have been around.

Not reliable/valid cuffless monitors, though. That's relatively new, if I understand correctly.


This is a very early stage study; at 175 data points over 35 patients, that's just 5 data points per person. Barely enough to make a convincing linear regression.


Off to one side is over-diagnosis. It's good to have ubiquitous sensing, and it will be really good for the detection of issues, but alongside that will be an increase in associated medical "costs" (I don't mean US $ alone: providing health services in an NHS model incurs costs) dealing with false positives.

It's a balancing act. I've now had one, and my partner has had one instance of going to see a specialist who says "you shouldn't have been sent here, this is the effect of improved imaging and diagnostics without updating the criteria for triage" -The imaging and related diagnostic tools now "find" artifacts which are sub-clinical for risk, but the triage requirements on the technician require them to report on, since "it might not be"


Blood pressure medicines have mild side effects and cost pennies a day, and lots of people do in fact have hypertension.


Sounds like we should be updating criteria then


I'm yet to see a study that convinces me of the potential of this concept. All of them when you get down to the details have at least one of:

  - An unrepresentative test set , i.e. a population with similar blood pressure at recording in which merely predicting a constant 'typical' reading would perform well in terms of accuracy.
  - A major methodology flaw, such as 'demonstrating' an ability to detect fluctuations from a baseline calibration cuff reading in situations where you would expect no significant fluctuations to occur.

I don't have access to this particular paper, but suspect it has the same issue owing to the relatively small sample size.


I find the topic interesting because it also seems to be one of these areas where a metric is being invented with no existing "gold standard", which is a different problem than just validating a new sensor to replace an existing one in the same methodology.

What does it mean to do continuous blood pressure monitoring? Do we have any idea how BP is supposed to behave continuously, when clinical standards are around a random sampling in particular resting postures? And do you have to account for all the fluid movements and pressure changes that happen during activity, with muscles contracting and body parts accelerating?


I love wearables. It is a pain in the neck to take a measurement. Anything that can be worn and sense automatically is awesome. I studious update an SQLite database of my Fitbit data. I just wish I had more measurements to track!


You could probably (depending on how much of the information you want is retrievable easily via fitbit's api) write a basic app that restructures that data and syncs daily to your database without having to studiously (=manually?) do it.


You can generally start with a full data export[1] then use the API[2] to keep up with new data.

1. https://www.fitbit.com/settings/data/export

2. https://dev.fitbit.com/build/reference/web-api/


It'd be interesting to read the study limitations, but, at 35 participants, a larger and diverse cohort would be warranted to see how this translates in the real world.

Aktiia has done some studies for its product and works by taking a BP cuff measurement to calibrate its PPG sensor-based algorithm. If you can get rid of the cuff and can sell an affordable wearable device, that would be a promising future.


The Aktiia device only required periodic recalibration with the cuff. We'll see if it makes it through the FDA.



Need to break the firewall to read this. Yes, not a large sample size but the accuracy is impressive for both systolic snd diastolic estimates.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: