Slashdot Mirror
Apple Watch Can Detect An Abnormal Heart Rhythm With 97 Percent Accuracy, UCSF Study Says (techcrunch.com)
According to a study conducted through heartbeat measurement app Cardiogram and the University of California, San Francisco, the Apple Watch is 97 percent accurate in detecting the most common abnormal heart rhythm when paired with an AI-based algorithm. TechCrunch reports: The study involved 6,158 participants recruited through the Cardiogram app on Apple Watch. Most of the participants in the UCSF Health eHeart study had normal EKG readings. However, 200 of them had been diagnosed with paroxysmal atrial fibrillation (an abnormal heartbeat). Engineers then trained a deep neural network to identify these abnormal heart rhythms from Apple Watch heart rate data. Cardiogram began the study with UCSF in 2016 to discover whether the Apple Watch could detect an oncoming stroke. About a quarter of strokes are caused by an abnormal heart rhythm, according to Cardiogram co-founder and data scientist for UCSF's eHeart study Brandon Ballinger. Cardiogram tested the deep neural network it had built against 51 in-hospital cardioversions (a procedure that restores the heart's normal rhythm) and says it achieved a 97 percent accuracy in the neural network's ability to find irregular heart activity. Additional information available via a Cardiogram blog post.
> the Apple Watch is 97 percent accurate in detecting the most common abnormal heart rhythm when paired with an AI-based algorithm. TechCrunch reports:
What the AI say when asked how to make a cup of tea?
If it gave a senible answer, it might be an AI.
Otherwise it was yet another overmarketed expert system.
Flag as Inappropriate
You're wearing it wrong
Is there a scientific basis for your opinion or do you make up your opinions based on what your can fish out of your ass crack?
Read any tutorial on Bayes theorem. Chances are most of the positive results will be false positives, but neither patients/consumers nor their doctors understand that, they hear "97 percent accuracy" and "You tested positive".
Well, if the alternatives are wearing nothing at all or a $50k medical device your insurance won't cover...
Also, the Apple Watch was useful as a development platform; HealthKit makes it easier to run informal trials and collect medical data from a large number of users. The software they've developed with that data could likely now be ported to other devices.
How can I believe you when you tell me what I don't want to hear?
Most patients can do it too. You don't really need an electronic device to tell you that you don't feel right. On the bright side most abnormal heart rhythms are harmless and quite common as you get older. The ones you have to watch for are the ones associated with effort/exercise, the ones that last more than a few minutes, the ones associated with pain, dizzyness or shortness of breath, or the ones that keep recurring.
Seven puppies were harmed during the making of this post.
Which $50k medical device would that be? I have a bicameral pacemaker/defibrillator and it only cost me $25k - INSTALLED. And I'm pretty sure your Apple watch isn't equipped to correct your abnormal rhythm if it was necessary. At best it's a device that will tell you what you already know because you felt it, kind of like the light that comes on to tell you you're out of gas.
Seven puppies were harmed during the making of this post.
Show the FDA 510K filings and clinical trails proving this for professional health care use, then this claim as merit.
Most blood pressure cuffs cost $50 and most of us trust those just fine.
The fact that Apple Watches aren't held to the same standard as medical equipment is a really good place to start. You can stick that in your ass crack and smoke it.
As usual, some very important data is missing. For example, how many false positives? How many missed negatives that a $50k machine would find? What's the cost of the finds and misses? Time frames? etc.
A heartbeat is basically a one dimensional list of pairs of numbers (strength of beat, time since last beat). Creating an algorithm to figure out when something like that starts getting fucky doesn't sound like a problem that needs the full power of deeplearningAIneuralnetothermarketinggibberishAPPSMOTHERFUCKER brought to bear on it.
Because they've been thoroughly tested millions of times and use a proven design that's been around for decades.
BeauHD. Worst editor since kdawson.
Going to bite:
You are right that the Apple Watch should not be held to the same standard, but if having it increases the survival rate by at least 10%, then it may be a nice bonus feature. The alternative is either a more expense device few people buy or having no indicator at all.
It would be good to have this study done on other watches.
At no point should such a device be mistaken for a specialised medical device.
Jumpstart the tartan drive.
I can make an HIV test that is over 99% accurate by classifying all tests as negative. Accuracy is a stupid metric.
It is NOT the accuracy!
It is the FAILURE!
AI algos do not represent any kind of validated benchmark to anything!
AI algos are always WRONG.
Jajajajajajajajajajajajajaja
Bravo, sir. Much more informative than this retarded article. May I say wee-todd-did awtikel.
Apple watch did not detect anything except electrical impulses.
Electrical impulses were recorded, so the iFruit was used as a data logger. Nothing more. Researchers used the iFruit with an off-label application. Misleading article accepted by an IDIOT MILLENIAL "editor."
5958 / 6158 = 0.9675
It probably classified everyone as the negative case. I couldn't find the paper or the confusion matrix, but this seems like a lot of noise. Accuracy is a useless metric for class imbalanced data.
Cardiogram is also available on Android devices. Is TFA paid for by Cardiogram or Apple?
my original Apple Watch matches what an EKG shows in real time, I think that plus scientific study is good enough
Absolutely. You are one.
Now every numbnuts apple worshipper will think their itoy watch is medical grade equipment.
Well, yes. We can't let anything get in the way of medical monopoly profits, can we?
Better get used to it: wearables are going to get more and more sensors, and will soon be a heck of a lot more effective than a doctor prodding you and trying to figure out what symptoms to look up on Google.
So let's say a false negative rate of 3% - no mention of false positives.
So for every 100 people who *do* have a abnormal heart rhythym, 3 won't know it.
So if you sell it to 100,000 people, 3,000 people won't get the proper answer.
At a certain point, it feels like those false negatives are going to affect a lot of people, who might look at Apple as liable for giving them incorrect information...and we haven't even looked at consequences from false positives.
> About a quarter of strokes are caused by an abnormal heart rhythm
But what about the opposite? How frequently does an abnormal heart rhythm result in a stroke? TFA doesn't mention it.
If this is a low proportion, then there will be many false positives, making detection of abnormal heart rhythm useless in terms of stroke prediction. It will only serve in increase anxiety of users.
My friends heart rate monitor that cost him £10 can detect AF as well. No fancy AI needed...
(He has AF on and off)
You want to put your life in the hands of a itoy then have at it.
The way I see it, having a watch tell you that you have an abnormal heart rate might in some cases potentially help prevent problems but more often than not it's simply going to scare the shit out of people for no good reason. While it is true that a number of problems can be related to an abnormal heart rate, it's also a fact that the majority of people with an abnormal heart rate have no problems as a result of it. Besides, my regular low cost, no-frills blood pressure monitor detects my abnormal heartbeat 100% of the time, without the absurd price tag or fancy AI.
This is going to be the problem with this, and Personal Injury Sharks will take note. People will "rely" on these devices for monitoring critical health issues when strictly speaking they should not. It really doesn't matter how big and bold Apple, FitBit, or whoever makes the disclaimer that it's not certified by the FDA for this sort of thing, the layers will still sue.
If you want news from today, you have to come back tomorrow.
I'm pretty sure your Apple watch isn't equipped to correct your abnormal rhythm if it was necessary
Indeed, if it was I wouldn't buy it. It's one thing to spend a few hundred bucks on a toy, it's another if that toy jolts you in the wrist to correct your heart. I am pretty sure that will tickle a bit
kind of like the light that comes on to tell you you're out of gas.
Not true at all, quite a few people I know, including both of my siblings had absolutely no idea they had heart arrythmias until it was discovered by accident. In my brother's case, he was in the hospital for other reasons, alseep, and on an EKG when it started flipping out. My sister is a nurse, I guess somehow she talked someone in to wearing a holter monitor after she heard about my brother, and sure enough same issue. It's not exactly harmless, but not warranting medical treatment at the moment. It's good to know these things.
Should you depend on your watch to tell you this? No. But if it can do this for you, that's great news.
Considering the abnormal rhythm occurs way less than 3% of the time, if the monitoring app displays just a static image of "Ok" at all times, accuracy even higher than 97% is achieved.
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
You feel fine. You have a common "missed beat" or something. You never knew it. Now you know it. You worry. A positive feedback loop ensues, in which the dread, stress, and occasional panic induce more serious symptoms.
Maybe this wouldn't happen. Maybe it would. That's why we need some real studies. Until then, I'll just assume that leading a better lifestyle than my Dad \(who died at 82\) is OK. Statistically, I know it is. Any *proven* tech that enhances that is icing on the cake.
I'd love to have such an app, but I don't have an iPhone nor am I going to be buying an iPhone. The data plans are prohibitively expensive and I don't need mobile calling or texting at that cost, when I can get it with pay-as-you go for $5/mo on a flip phone. I do have an iPad with Wifi though, so if the Apple Watch ever works with those, this app would be enough to make me want to buy one.
Also, by how much is the "machine learning, neural network" approach better than simpler approaches? There's no point in shooting machine learning bullets at things that can be analyzed with much simpler means to a similar degree of sensitivity and specificity.
I bet that a much simpler algorithm could produce similar results. But nowadays it seems to be the latest fad to throw machine learning at fairly plain signal/data processing problems.
A crashed advertisement reveals the code of the facial recognition system used by a pizza shop in Oslo
A crashed advertisement reveals logs of a facial recognition system
Theres less than 3% of people having an abnormal heart rhythm. Are those that dont get detected?
had absolutely no idea they had heart arrythmias until it was discovered by accident.
And I'm willing to bet that after checking it out - nothing was done. Most arrhythmia are harmless.
It's not exactly harmless, but not warranting medical treatment at the moment.
There you go. Not warranting medical treatment at the moment - in a litigious society - means that his doctor is willing to bet that it IS harmless. Nothing will come of it. I'll keep my eye on it - we'll check every few years and monitor it and make sure it doesn't change... etc. This is something I happen to know quite a bit about, since I am both a physician AND I have heart disease and a pacemaker/defibrillator for MY arrhythmia...one that is NOT harmless. But usually you get those after a heart attack (I've had 5) unless you have some sort of anatomical defect (like WPW syndrome).
Should you depend on your watch to tell you this? No. But if it can do this for you, that's great news.
I believe strongly in patient autonomy. If people want their watch to do this then fine. The trouble comes with the people who think it replaces actual medical advice (despite the legalese disclaimers on the EULA which means nothing except to lawyers). Like all those people who "self diagnose" with WebMD - that's all well and good, but I have a real problem with the patient who delays treatment for something preventable because they were playing doctor on themselves or worse - on their kids.
Seven puppies were harmed during the making of this post.
Except there's no evidence that it improves survival rate at all. 97% accuracy is easy to achieve when the chances of this condition are so incredibly low. It's also not impressive when you consider there are devices that get much closer to 100% with almost no error. What do you want, the thing that maybe works, or the thing that virtually always works?
FFT even an audio signal of the beat, or light-level through the finger or whatever.
Produce some simple stat on the regularity and speed of the heartbeat from that data.
Use that number to establishment a limit, to use as a diagnostic against those who are medically diagnosed with such conditions.
Apply that limit to Yes/No answer.
If it got WORSE than 97% accuracy, I'd be surprised.
It it took more than a handful of code coupled with an audio/camera and FFT library, I'd be amazed.
The processing power required would be pathetic.
Whether or not it would actually BE USEFUL as a diagnostic device with that kind of accuracy? That's another matter entirely.
An irregular rhythm detection device already exists that is $50 and works all the time. What monopoly is being broken with the crapple watch? What improvement is being made?
5 heart attacks? Wow. That's some scary shit. I have vasovagal/neurocardiogenic syncope. I've had symptoms that are similar to heart attack symptoms that were scary enough to put me in the ER. I couldn't imagine going through an actual one 5 times. I bet after that it's hard to listen to people complain about "having a bad cold".
...when paired with an AI-based algorithm
Sounds to me like it's the AI-based algorithm that's doing the detecting here. Not the watch.
Atrial Fib is absolutely not harmless. GET YOURSELF A BETTER DOCTOR. The statistics are very clear that it is a significant early mortality predictor. If you have more than one Atrial Fib event every six months you should be on an anti arrhythmia medication and a blood thinner.
Heart arrhythmia (missed beats) is not all that uncommon and is not something that one needs to be overly concerned about. Sugar intake can cause it in some people, but their heart will return to normal beating after the sugar is processed by the body. I know someone with this problem and the only thing they have to watch is to not get too much sugar at a time. The watch is basically worthless concerning this.
Fair enough. Thing is, I just spent 2 weeks in hospital due to exactly this. Paroxysmal Fibrillation. If the watch can give me warning of impending problems before they happen I could avoid that. I know what medical equipment looks like, I just spent 2 weeks attached to it. An Apple watch would be a perfect device to keep an eye on things in the next few years instead of carrying a bloody EKG with me all the time.
Sure it is not perfect. But since I am not going to actually wear real medical equipment for a few years it is waaaaay better than nothing. You only notice arrythmia when the ticker is already going bonkers and damage has been done. If I pick pick up early signs as a tripwire the watch will have dont its job just nicely, thank you. Risk reduction, not elimination is the key here.
The dangers of excessive individualism are nothing compared to the oppressiveness of excessive collectivism
Except for me, who just spent time in hospital due to precisely this. If the chances are low for you it does not mean it is low for everyone. I would be happy to wear the device as a tripwire trigger. The fitbit I have did not cut the bacon, it did not pick up the problem.
The dangers of excessive individualism are nothing compared to the oppressiveness of excessive collectivism
Where do I buy it? Does it fit on my wrist?
The dangers of excessive individualism are nothing compared to the oppressiveness of excessive collectivism
That's not how 97% accuracy works. This isn't saying "it will detect whether you have an abnormal heart rate at this instant with 97% accuracy", it's saying "it will detect whether an individual suffers from this condition with 97% accuracy".
I had AF a few weeks ago and my FitBit did not pick it up, except that the pulse was a bit high. The problem is that a AF is not very visible if you only measure the pulse at the wrist. You need to check the impulses on the chest, the wrist is a bit indirect. If Apple has an algo that gives a better accuracy on that with a neural net I am all for it. No, trust you health to a iToy is not good, but then, most people trust their health to a $5 thermometer somewhere in their life.
Or even better, just feel the pulse with the hand and count in the head. $0. Not accurate.
The dangers of excessive individualism are nothing compared to the oppressiveness of excessive collectivism
Go ahead and try to write an algorithm that can measure heart rate purely in software. We'll wait.
Except there's no evidence that it improves survival rate at all.
That wasn't the goal of this study. The point was to find out if it could detect the condition.
97% accuracy is easy to achieve when the chances of this condition are so incredibly low. It's also not impressive when you consider there are devices that get much closer to 100% with almost no error. What do you want, the thing that maybe works, or the thing that virtually always works?
Yes, but they are expensive, Rx only in most cases, and somewhat inconvenient for people to wear all day every day. Having something that people already own that is capable of having close to the accuracy of those devices is valuable. No doctor will use it as a diagnostic, that's not the point. The point is to make the person wearing it aware that something might be wrong and that they should see a doctor to have it properly diagnosed.
Also, I think you are under-estimating the prevalence arrhythmias. They are not what I would consider to be "incredibly low". According to the CDC the prevalence of atrial fibrillation is about 2% of people under 65, and 9% over 65.
I browse on +1 so AC's need not respond, I won't see it.
Who said anything about trusting it? It's just a handy feature to have. "We have detected an unusual heartbeat rhythm. We suggest you go in for a checkup."
If I read the article correctly and the Blog post, they took the readings from the Apple Watch and processed them separately, during which they were able to to detect the abnormal rhythms.
This is a far cry from the Apple Watch continually analyzing your pulse and providing real time warnings of impending cardiac events. It seems to merely point out that that Apple Watch is an adequate data collection device.
When Fascism comes to America, it will call itself Anti-Fascism, and tell you to give up your guns.
Same here, it almost killed me.
I hate having any kind watch but if it could have saved me from going through this it would almost be worth joining the fruity cult.
The Apple Watch detected a life threatening, rapid heartbeat in my husband. The ER personnel compared the Apple Watch readout to their expensive ER equipment. They agreed to within 3 beats per minute. That's close enough to distinguish between calling 9-1-1 and scheduling an appointment for later the same day.
Being able to distinguish afib from other forms if trachycardia is important in determining stroke risk. If this app helps at all, it may save lives.
How does this apply to their competitors??
Accuracy of a test can be deceiving when the base rate is lower than the inaccuracy of the test. In other words, if the accuracy of this test is 97% and the base rate of arrythmia is 2.5% (wikipedia) then false positives will outnumber true positives meaning that if your phone says you have arrythmia, there's about a 55% chance it's right, not a 97% chance. Take 350,000,000 people. 2.5% or 8,750,000 have arrythmia. 8,487,500 (97%) will recieve a correct positive reading. 262,500 (3%) will have a false negative reading. Of those without arrythmia, 341,250,000 people, 97% (341,250,000 people) will recieve a correct negative reading. 3% (10,237,500 people) will receive a false positive diagnosis. The false positives outnumber the correct positives by about 1.2:1.
Atrial fibrillation is not the most common arrhythmia either. You are the first person to mention atrial fibrillation here. We're talking about the most common stuff - like the odd extrasystole yeah? I could just bring up ventricular fibrillation too, BUT YOU DON'T NEED A WATCH TO TELL YOU YOU JUST DIED.
Seven puppies were harmed during the making of this post.
I'm an Afib patient, so I thought I'd weigh in:
1. Any device that measures your pulse can be used detect an arrhythmia, but it takes an electrocardiogram and a cardiologist to tell what kind of arrhythmia it is. Since your pulse comes from the left ventricle, I don't see how it can distinguish the various types of superventricular (originating above the ventricles) arrhythmias. I see this app as mostly useful to monitor people with previously diagnosed atrial fibrillation, on the assumption that they are unlikely to have a second type of arrhythmia.
2. The stroke risk comes from blood clots forming in the heart, because atrial fibrillation reduces the ability of the atria to pump blood, allowing it to stagnate and form a clot, which can break loose and travel into the arteries. The general rule of thumb seems to be that Afib attacks lasting less than 48 hours don't pose much of a stroke risk, but longer spells require anticoagulant drugs.
3. The electrocardiogram machines in my local hospital's ER use a diagnostic algorithm to identify different types of arrhythmias, but the ER doctors generally take that as a suggestion, not a diagnosis.
4. There are (relatively) inexpensive electrocardiogram machines available on Amazon, but I don't know how useful they are. I've never yet seen a cardiologist or electrophysiologist who would show me an electrocardiogram printout and say "this is what your heart is doing." They all seem to think that understanding an ECG is beyond the ability of most patients.
Well the first time I was told I was too young to have a heart attack, even though my cardiac enzymes were elevated. So after 12 hours or so in the ER in observation they sent me home. I was 26.
The second time I stayed for the weekend. I even ended up getting iv lidocaine for an arrhythmia I developed. And another idiot cardiologist came and saw me Monday morning and again told me I was too young to have a heart attack, and sent me home. I was 30. He told me to come back for a stress test later that week though.
Later that week, 2 hours after the stress test, I felt really bad. Got driven to the hospital doing a Carrie Fisher - puking up all over my dad's car. I'm in the ER for 2 mins explaining what I felt to the doc when I did a ventricular fibrillation. That's when they kind of realized that I wasn't too young to have a heart attack... 2 weeks later I was getting quintuple bypass surgery.
In 2010 I had a small heart attack that required 3 stents. Despite my (arterial) bypass grafts being fine, my disease progresses and my coronary arteries are slowly clogging up on the other side of my bypasses. So those got opened up. A week after that I had a massive heart attack on the right side - probably one of the new stents got suddenly blocked, it happens. About 80% of the motility of the right side of my heart was gone. That earned me a couple months in the hospital, my defibrillator, and more stents (I have 9 now, on top of my bypasses).
Since then I've been ok with the odd episodes of angina. I get a cardiac catheterization done every few years and I get to see technology progress. The last one 16 months ago was my first one with intravascular ultrasound. They found a blocked section and opened it up, and I've been fine since. I'm grateful because I still have a decent ejection fraction and am only NYHA II on the heart failure scale so I can still do basic stuff like walk my dogs (on flat ground).
Seven puppies were harmed during the making of this post.
97% of Apple Watch wearers have abnormal heart activity.
Heart racing... breathing rapid... alertness heightened... bloodflow to extremities increased... it must be product announcement season at One Infinite Loop again!