Scanwatch SpO2 Accuracy: Oxygen Saturation Test (Review, Withings)
ฝัง
- เผยแพร่เมื่อ 7 ก.ค. 2024
- I tested the SpO2 / Oxygen Saturation accuracy of the Withings Scanwatch.
Instagram: / quantified_scientist
Twitter: @QuantifiedRob
Linkedin: / rob-ter-horst-90756a4a
Website: www.robterhorst.com
📹Fitbit Sense videos📹
First look review and test: • Fitbit Sense Accuracy ...
Heart rate: • Fitbit Sense Heart Rat...
📹Withings Scanwatch videos📹
Unboxing and first review: • Withings Scanwatch Rev...
Sleep test: • Withings Scanwatch Sle...
Heart rate test: • Withings Scanwatch Hea...
Q&A with Withings: • Scanwatch Q&A: Answers...
⏱️This comparison is part of a larger project where I spend over 11 hours a week tracking everything in my life. If you'd like to know more:
• This mad scientist sca... ⏱️
Timestamps:
00:00 Intro
00:58 Experimental Setup
02:42 Results: Range of SpO2 values
04:12 Results: Consistency
05:07 Results: Accuracy
06:08 Implications for Health
06:45 Results: Hypoxia Clinical Trial
07:27 Should You Buy the Scanwatch?
Since the release of the Scanwatch I have taken 457 oxygen saturation measurements and at the same time also taken measurements with a dedicated SpO2 monitor.
In this video I compare the results, take a brief look at the clinical trial that Withings ran for the SpO2 sensor and finally put the results in the context of health tracking.
For the test between the scanwatch and SpO2 monitor I want to look at three things:
(1) Are the results I get for the Scanwatch in the same range of values as the SpO2 monitor
(2) If I take 4 measurements in a row, are the results consistent
(3) Is there an association between the measurements taken with the Scanwatch and those that were taken with the dedicated monitor.
There are of course several limitations to the experiment I did, they most important are:
I assume that the finger oximeter I used is correct
I only tested the Scanwatch on myself
I only tested the Scanwatch in a “normal” range of SpO2 values, and not in hypoxic conditions.
See you in the next video!
🎵Music🎵
(1) Not Without the Rest by Twin Musicom is licensed under a Creative Commons Attribution license (creativecommons.org/licenses/...)
Artist: www.twinmusicom.org/
(2) Wonders by JVNA
(3) Can’t Sleep (TH-cam Audio Library)
(4) Classique (TH-cam Audio Library)
(5) Blurry Vision (TH-cam Audio Library)
Video by Pressmaster from Pexels
Video by Yerson Retamal from Pixabay
Video by Taco Fleur from Pexels
Video by Anna Shvets from Pexels
Video by Tima Miroshnichenko from Pexels - วิทยาศาสตร์และเทคโนโลยี
![CAMPปลิ้น | EP.77[2/2] "รับทราบ" แก๊งเพื่อนรัก 10 ปี ที่ไปไหนก็มีตำนาน!](http://i.ytimg.com/vi/ifc0ofD51B0/mqdefault.jpg)
Instagram: instagram.com/quantified_...
Twitter: @QuantifiedRob
Linkedin: www.linkedin.com/in/rob-ter-h...
Website: www.robterhorst.com
Please make a review of Whoop 3.0 sleep accuracy
Every time I watch one of your videos, I feel like an incredibly bright light has illuminated an incredibly dark, murky, and confusing landscape. Thank you for your work Rob.
Once again, brilliant video! Thanks for such useful and thorough tests, also for putting the results into perspective (like that, for most people with consistently high values of SpO2, their accuracy doesn't really matter).
That's sobering! Again: thank you for your valuable work!
My pleasure! Have a great day Jörg!
Great video once again!
Best channel for this kind of stuff. Thx alot!
getting to be my fave channel cause i cherish knowing the accuracy of tech
Love your content and subscribed. Was torn between The Sense and the Scanwatch. I've had the scanwatch for 2 days now, and made the decision after watching your videos. I use it alongside my Fitbit Charge 3 and have found the step count on the Fitbit to be almost double. I really like the simplicity of both, without distracting or asking much from the user as with Smartwatch type trackers. I do believe that body temperature tracking would have been a great additional feature to the Scanwatch and would really have completed the package. Still hope Withings listens to feedback and allows the user to choose features that sacrifice battery life for continuous monitoring.
Dude I really appreciate your videos reviewing various products in depth! I was looking for a good smartwatch to track my sleep apnea and irregular heart activity. So far you are the only TH-camr who's actively reviewing smartwatches' features by scientifically testing/comparing the products with proper tools in the lab. Thanks for your videos. Keep it up!
Great video as always! As a small suggestion for this and other videos, a little bit more info of the relevance of the range would be super helpful i.e. how important is each point of deviation from the actual value for SpO2, heart rate, etc. or for something like sleep tracking is having high levels of accuracy important vs. a broadly indicative view of readings. Perhaps tying the level of accuracy that's required to draw different types of conclusions about your health would be a way to ground this type of info in a relatable context (you do this a little in this vid, but a bit more would make it even more relateable 🙂)
Thanks for the suggestion Daniel! Indeed, this is something I need to consider. I find it difficult to judge what others find most valuable in my content, so this really helps! Have a great day :-)
Your work adds a lot of info in the knowledge of these devices and it's very useful!
Hope your numbers will increase quickly!
Thank you very much Filippo!!
Sweet content; keep at it! When are the Apple Watch studies coming out
Hi Nigel! My order has been delayed, but I should hopefully receive my Apple Watch this week, so a video in ~ 2 weeks or so 😄 . Have a great day!
Thanks again, great video. The scan watch doesn't seem to be quite matching up to its potential sadly. Had hoped to pick one up but not so sure now. The withings trust pilot score is very poor currently too. Disappointing as the scan watch is exactly the kind of smartwatch I wanted.
Very interesting review of all the products. I like the detailed info you give for us to look at. I am wondering which one is the best fitness/smartwatch with ECG, SpO2 and others features from your point of view? I am considering one to give to seniors that have a heart attack history.
Nice work! Since you already found that Scanwatch had some measurements completely off, it would have been interesting instead of the median value of the 4 measurements, then just to throw out the most inaccurate measurement and see if the deviation of the remaining 3 tended to be within a small discrepancy. This would be somewhat similar to what you did for the heartbeat tracking where you found some initial measurements were way off but the remaining measurements were quite accurate.
Overall, considering all the measurements available, would you still recommend the scanwatch if you do not necessarily need 100% accurate results?
Great video!! Will you be reviewing garmin products. Such as the venu sq or the forerunner 745?
Results as expected. WHen I had the scanwatch, I was disappointed with the accuracy compared to my finger spo2 monitor. Garmin was the same though. Reflective Wrist based devices shouldn't be used if you've any expectation of accuracy.
Thanks, which Garmin did you use for the comparison?
Thank you.
Hello Rob, Your videos are great and you imply scientific methods for evaluating the products. My opinion for SpO2 is that because normal SpO2 measurements have a narrow range, one cannot easily conclude about reliability of a SpO2 device if hypoxia cases are not included/examined. Of course consistency/repeatabillity is something important but in any case if 97 is the correct and the watch gives 92 and 98 as next value it's not a big deal medically speaking. It is important to detect hypoxia measurement and to know how the device behaves in the low range. Is there a way of finding that?
What's model of the finger spo2 oximeter you are using can you share the Amazon link
I'd love to see similar test using apple watch 6. I'm curious if apple is any better. Thanks for the video, it was very insightful :) i was wondering why my garmin detects v low levels of spo2 but i guess its just probably the same story as with the scanwatch
Hallo, großes Lob für ihren Kanal, sehr informative und professionelle Tests. Ich suche schon seit Wochen nach einem Gesundheits Tracker. Es kommen ja auch immer mehr auf dem Markt. Jetzt bin ich auf den Viatom Checkme Pro gestoßen der ja nicht wirklich im vergleich zu einer Smartwatch ein Wearebal ist und schon einen sehr professionellen Eindruck macht. Vieleicht wäre das ja auch mal für Sie interessant darüber ein Testvideo zu machen. Ich finde das Gerät auf jeden Fall höchst interessant und auch ein Video ihrer Qualität wert.
Have you or has anyone looked into how Withings' ScanWatch compares to the Apple Watch 6 when it comes to SpO2 accuracy and consistency (e.g. for tracking sleep apnea during the night)?
I found the 42mm is more accurate in compare with the 38 mm I do own both of them and at the end I return the 38mm because the problems with the heart rate accuracy.
May i know what's the brand name of your oximeter?
wie immer ein super beitrag:-) hab die scanwatch, den oura ring und möchte meine schlafapnoe überblicken, hoffe die grafische darstellung der nächtlichen apnoen bei scanwatch wird bald freigegeben... mal sehen wie gut die ist. hast du bei der spo2 messung auch erfahrungen mit der neuen apple watch? lg gerald
Can you recommend a good SPO2 monitor, not a watch but a traditional finger type one. I have emphysema and had bowel cancer. Would love a dedicated spo2 monitor.
Hello, I have some questions:
1) Are the dedicated spo2 monitor made in China . I think that correct compare to use spo2 monitor made in Europe or USA .)
2) I saw the spo2 from brand Ihealth . What do you saw about this device?
3)For futures videos compere cardiocore from brand qardio . I think it is interesting device .
Thanks for your work 😎😏
Would be cool if you could do that for the Huawei Watch Fit too :)
Thanks for the video. Can you please be more clear about which brand and model SpO2 monitor you recommend? Ideally, also please provide an Amazon link. Thank you!
Sounds like the Scanwatch might think you have sleep apnea based on a low SpO2 reading that isn't real. Or do the low reading is sees not trigger that kind of feedback, since they are low, but still in the safe/normal range?
Great content. Personally interested in amazfit and Huawei smartwatch, but I'm pretty certain sleep and spo2 are very bad. You already killed my interest in whitings/nokia.
Hi, thanks for your work. Could you please share the oximeter you used for this?.
Also I recommend sharing that in the description using affiliate links from amazon or something similar that will help support the channel 👍👍.
Cheers!
Hi Fernando! Great idea! So, I have tried Beurer, iHealth and Arystone SpO2 sensors. To my surprise the Arystone (to me an unknown Chinese brand) gave the most consistent results, especially when tracking my SpO2 during a flight. So based on this I continued to use it, though I expect it does not have official FDA/CE clearance.
Can you do a sleep and heart monitor test with the apple 6 watch?
What do you think is the best wearable? Fitbit, Withings etc
What is the small hole by the winder? can't find any information about it.
Thanks for the video! I am wondering how does the sleep apnea detection work, does it measure SpO2 every second continuously throughout the sleep?
At least it checks your pulse very rarely at night, was it only once every ten minutes!? And it should detect atrial fibrallation...
Great content! Thank you!
Just one suggestion: IMHO when you are showing the distribution of readings over their value space like in 4:15 it would be more clear if you also visualized them in a distribution plot
At 6:00 it seems an odd conclusion to say there is no correlation between the two measurements, considering they are always within 2% of each other. The problem is that the measurements are all essentially the same. You could put a tattoo reading 98 on your wrist and it would be pretty accurate for most people most of the time. You really need some abnormal results to draw any useful conclusion. I'm trying to track my sleep apnea with a fitness tracker pulseox. It would be great to test the accuracy under those conditions, if you could find someone with the condition to test on.
Please can you share the name of the SPO2 monitor that you said you consider reliable? Also, do you think skin colour has an effect on the reliability of these devices, due to the method used? I'm not Caucasian, and so I wonder if they might be less accurate for me.
Said oximeter on the thumbnail
Have you tester the Withings body cardio squale?
Have you ever tested Garmin venu? Or vivosport 4?
May I suggest you try Oculus Quest 2, $399, fibre optic, wifi 6, 5G, and use real fencing foil, boxing glove and fitness wearable? It might be similar to Olympic sports with one real athlete vs one virtual. Or, two real, both remote.
Hi Rob, great video. However, I think there are some problems with your analysis. Maybe you could shortly address them or make a follow-up video?
1. I think there is a problem with the accuracy of both methods (finger pulseoximeter and scanwatch). Even the best finger pulseoximeters have an error of 2 %. This 2 % error can be seen for example for Nonin pulseoximeter where values are compared to a real gold standard (arterial gas measurement). So assuming your finger pulseoximeter is equal to some of the best, you could see an error of 2 % on both measurements. So a measurement of 97 % of the finger could result in a measurement of 93 % on the Scanwatch, but both can still be just as close to the potentially actual value of 95 %.
2. You first need to validate you gold standard, the finger pulseoximeter you use has not been validated using arterial gas measurement. Even though it recorded a low reading on an airplane as expected, there could still be a systematic bias to for example high values. Indeed, you show that there is not always a clear relation between the values (you still expect this with a systematic bias) there is no way of knowing exactly what value is correct and what to compare against without proper validation of the pulseoximeter. To see the problem with finger pulseoximeters as a gold standard: th-cam.com/video/puVqZaOOGdg/w-d-xo.html
3. As you indicate especially the low values (with a gold standard measurement!) are interesting and a comparison of high values is less useful. Because of your connections, getting access to a rebreathing mask should be doable. To know if the measurements of the watch are useful it would be nice to get those measurements in.
I have been comparing my Fenix 6 watch with the Nonin Onyx Vantage and the results vary wildly, especially when the watch is not exactly straight on my skin... Just for fun I could send you some of my data. BTW I have also compared a very similar finger pulseoximeter to yours (at least the same software in terms of display), and the values between this one and the nonin sometimes vary with 3 %.
Hi! Thanks for your very thorough (!) feedback , and I agree with most of what you are saying. The test I did has inherent limitations, probably even more than my usual tests. As I mentioned at the end of the video, I put the most important limitations in the description. Indeed addressing that I am assuming my finger oximeter is correct.
1. For sure the absolute values could be off, that is something I could have mentioned better. However, hopefully the deviation between the 4 consecutive measurements does provide some insights. You might just be interested in changes in SpO2, and not absolute values. Also, by taking the median of 4 measurements I was hoping to at least reduce some of the noise. I had some contact with Withings, and they also mentioned that their goal is less so to be accurate in the "high" range of SpO2: "The differences in the 95-100% range are not very significant anyway. What is interesting is to identify falls under 95% and under 90%. Especially associated with symptoms."
2. However, indeed as you mention in point 2. , with a systematic bias, I would still expect some relationship. And yes, I would have loves to measure SpO2 using arterial gas measurements, but unfortunately this is not a device I have access to (nor someone to operate it). But you are right that my "golden standard" cannot really be considered to be the best.
3. The idea of a rebreathing mask sounds interesting! Let me look into it.
In short, I agree with almost everything you are saying, and you worded it very well. I still hope my video ads some value for some people though, since most videos in TH-cam do not go to this detail.
Yes, having a look at your data sounds great! Could you send me a message via "www.robterhorst.com/contact", and I can see if I can include it in a future video?
Thanks so much for the feedback, really useful!
Have a great day!
@@TheQuantifiedScientist Indeed, I agree you already did a pretty good job of underlining the limitations. Overall this video is WAY better than any other video I have seen that discusses smart watches capable of recording SpO2.
However, my concern is that this research is not capable of answering your ultimate research questions. I guess that is: Can you improve my life (i.e., health and general well-being) using these kinds of metrics?
BTW maybe you should look into the Nonin finger pulseoximeters, they are a bit expensive, but people with COPD or reduced lung function tend to use them. Furthermore, it is the only brand I know of that validates their measurements with arterial blood gas.
And you showed your Dutchness ;). A very cool article was written about the misuse of "golden standard":
Claassen, J. A. (2005). The gold standard: not a golden standard. Bmj, 330(7500), 1121.
@HighlandRegal Thanks so much! Yes indeed, these measurements are only valuable for people if they can improve their lives. And, yes, I cannot escape my Dutch-ness 😁
BTW Are you involved in some kind of research yourself too?
@@TheQuantifiedScientist Yes, I am a postdoc at a large pharmaceutical company. Project on machine/deep learning to predict small molecule behavior.
Was very happy to find your channel, the most scientific channel out there to review these measurements. Nearly all reviews of fit/sport watches and their measurements are "useless". It seems like people do not take the effort or think that people just do not care about the accuracy of these things.
I'm not sure of you're done testing the scanwatch or if you still have more tests to run, but given what you've discovered so far in this video and your other videos on the scanwatch would you recommend it as a good product?
What I have seen here, and especially read from the Withings forum... Never!
@@GreenlifeFin if you read Garmin and fitbit forums, all reports of these watches are not accurate. All steps on watches are all different, etc. The moral of the story is non of these watches, including withings are accurate enough in the real world. They may show trends, which you could go by over time but dont rely on them for real world accuracy
@@3barazi1 not ture. Have had fitbit, samsung smart watch, garmin and all vary and none are properly accurate. The first thing I do is as I walked with the watches, walked 250 steps and counted, and all gave different results.
Withings is the worst, and stel counts are well down
Will you try smart scale?
How accurate is the SpO2 of the Fitbit Sense? It asks me to sleep one night for a reading. So it should have time to do many more than 4 short readings. Theoretically it could be very accurate. Not so sure about reality though... Greetings from Carinthia!
Please test Samsung galaxy 3 HR
Tiếc là không bán tại Việt Nam
Thanks a lot brother. Fuck these overly expensive shits.
Why didnt you test ecg ?
Can you test the scanwatch 2
Hi Rob! @6:55 you claim, that in order to measure a low oxygen level, you would have to go to a high altitude or have low oxygen air. However there is a third option: You could use the Wim Hof breathing method (th-cam.com/video/tybOi4hjZFQ/w-d-xo.html). Basically you blow out lots of CO2 first, then you stop breathing. As the CO2 concentration steers your reflex to breath, you can now stop breathing for a longer time. During this time your oxygen level will drop. In my own experience, I dropped it down to 60%. So plenty of opportunities to compare the devices...
BTW, you have created a really nice channel!
Have garmin vivomove luxe and not great ether. Being totally accurate is not important. Its trends you need to take notice of.
Ref
You cannot compare to differing technologies and say that one is not very good. You would not wear an oximeter on your wrist. If you did then the Withings wins hands down.
Withings has such nice designs but their instruments are such a let down...
There is something wrong with the way you have calculated and presented the correlation analysis. It would be best to ask a statistician's advice. Of course there is a strong correlation between the two devices since in the vast majority of cases they both showed values within the normal range. You have probably taken only the values between 95 to 99% to test for correlation but this is inaccurate because the range can be much broader. Based on a 5% interval analysis where most measurements are around 97% you cannot draw the conclusion that there is no correlation.
first ?
All your tests prove not to rely on any smart or fitness watches, non are accurate and only give ballpark results. Garmin is pretty poor also when doing oxygen blood levels.
All fitness watches are not accurate and even within the same brand vary with the results.
I think these watches give people false security and taken them to seriously.
Think rather than going by accuracy people should be looking at tends.
Its like scale accuracy comes into play also, maybe not accurate but do show a trend if gaining or losing weight.
So a watch may show trends, but overall inaccurate.
Your regression doesn't make much sense!