Originally published at: A lawsuit alleges Apple Watch's pulse oximeter is racist | Boing Boing
…
While it’s not just melanated skin that has an issue, tattoos also cause the sensor to not perform properly too, and Apple should have properly addressed this issue by now.
I’m not sure any of this can be addressed, frankly. I’m not a medical devices engineer, so take this with a grain of salt. However, what I’ve read is that the skin method of detecting blood oxygen isn’t actually reliable at all and is known to have a lot of issues with skin tone and other factors.
I would argue Apple shouldn’t have included it at all. Most of the health functions of the watch are sketchy at best. They all work by taking a super noisy data set from a questionable sensing method and throwing a trainload of AI pattern matching at it, hoping to create a useful signal. However AI is not a magic bullet to get around the garbage-in-garbage-out problem as so many seem to believe it is.
In Apple’s zeal to create a science-fictiony health monitor you wear on your wrist, they are quickly leaving science behind.
Of course, if you ask Apple, they’ll point out that the health monitor functions aren’t actually for health monitoring – they’re strictly meant for entertainment. Blood oxygen YouTube.
Go after the one with the deepest pockets…
I am an Advanced EMT, and do back country search and rescue, and work with pulse ox on every shift.
Pulse Ox measurement relies on a particular translucent quality. It is a deeply imperfect technique that we use for guidelines of health. Skin color, tattoos, and perfusion (circulation) all matter. If someone’s hands are cold for example, or if they have diabetes that impacts their circulation, the spO2 monitor will be impacted.
Apple can’t fix this without a whole new technology. For now, skin sensing monitors are simply not a one size fits all.
People need to get over the world owing them everything. Get a refund, move on.
“Blood Oxygen app measurements are not intended for medical use, including self-diagnosis or consultation with a doctor,” Apple writes.
My doctors very much relied on my Garmin for telemed visits during the pandemic for the vital signs it tracks. When I got the 'Rona this past May, spO2 was of particular interest.
ETA: As @mototom points out, Garmin, FitBit, et.al. are equally as unreliable as the Apple Watch for vital signs (albeit far less expensive!). So, why did my docs ask for the data?
And yet the Garmin has the same disclaimer about not being intended for medical purposes. This isn’t an Apple problem, it’s a problem with a particular technology. But Apple is where the money is, thus lawsuit.
Haha, indeed.
Apple crossed a line when they moved into medical diagnostics. The standards for correctness and reliability in that space are orders of magnitude higher than anything a smartphone company has ever made. They’re getting away with the “entertainment” dodge for now, but they are on a collision course with regulators. Perhaps as soon as someone dies because of information their Watch gave them.
Totally agree! I’ll edit to say so.
I was with you right up until this.
Technology that works for some people and not others based on the color of their skin is systemic racism incarnate. It is perfectly reasonable to object to this and it is the structural mechanism of the US system for lawsuits to be the de facto mechanism by which corporations are held accountable and encouraged to fix their shit.
It would be better if there was direct accoutability to the populace or to our elected officials. That’s not how the system works. Complaining about people using the shitty system the way it was intended is several layers of shitty in itself.
It looks like one manufacturer has solved this problem in medical environments, because they started testing their SpO2 monitors on a more diverse group of patients in the 1990s and observed the disparity. Then they developed technology to correct for it.
So it’s solvable; someone just has to give a shit.
They have to give enough a shit to spend the next fifteen years in the Eastern district of Texas battling various patent owners.
Or they can license the patent for SET technology from Massimo or wait until this physics postgrad from Brown makes her tech available:
If Patent A is derived from patents B & C and a submarine patent D, the license to use patent A does not necessarily indemnify the licensee from claims from patent holders B, C & D.
Massimo’s patents are all theirs. No derivative patents on their part.
Also, a patent that is derivative from an expired patent stands on its own. The original patent’s holder has no claim on derivative patents.
But does that solution fit into a watch-with-tiny-battery form or the price point to compete with other smart watches? The license cost alone could have priced Apple’s Watch out of the market.
Apple and Masimo are currently involved in litigation
(alleges that Apple poached employees from Cercacor, a spinoff research arm of Masimo)
I think Garmin may have a license for some of these tech solutions-- the effectiveness of such tech in a watch remains to be seen verified in the peer reviewed literature.
This has particular nuance in the time of pandemics when hospitals use O2 measurements using the same faulty technology to measure vitals of infected people to keep them alive. For example, https://www.cnn.com/2021/02/20/health/pulse-oximeters-dark-skin-covid/index.html. It has become a requirement to have to interrogate my doctors with the “what is the racial makeup of the test group” when my doctor says that the efficacy is based on a study.