The Google Pixel Watch 3 ‘Loss of Pulse’ feature is now available in some European countries.
If you’re not already familiar with it, you can read more about the feature here, and watch the keynote. The high-level overview is that it’s a new medical feature exclusively on the Google Pixel Watch 3 that *can* detect a ‘loss of pulse’ and dial emergency services. It is the most targeted off-the-shelf technology applicable to cryonics monitoring thus far. The feature was rolled out in the last week of September 2024, and I will publish a 'Review' article in the Reviews section of the website once I'm able to gather more data on its usability in the real world.
From an engineering and practicality standpoint, two key questions arise: What is the reliability of the feature, and will it be available to people in the U.S. in the future? Both questions are crucial, given the importance of minimizing delays in cryopreservation and the fact that most cryonicists are based in the U.S.
Sensitivity - 69.3% (Missed 30.7% of Cases in Clinical Trial)
With regards to reliability, Google’s ‘Loss of Pulse’ feature is regulated under the Medical Device Regulation (MDR) (EU) 2017/745, and the European Database on Medical Devices entry for the product can be found here. MDR obligates manufacturers and their Authorized Representatives to maintain transparency regarding the safety and efficacy of medical device features. Specifically, Annex I, Chapter III, Section 23.4 of the MDR requires manufacturers to include clinical performance data and expected outcomes in the Instructions for Use (IFU).
As a result of MDR regulation, and the feature's classification as a Class IIa medical device, even though Google is not required to include performance information in the press release or general promotional material for the product, they do have to provide an Instructions for Use (IFU) document with Clinical Study information. That IFU document can be found here. The relevant section is on page 11 where it is stated that in the clinical trial 'Loss of Pulse' detection had a sensitivity of 69.3%. That is to say, if in the study 100 participants experienced cardiac arrest or other loss of pulse events, 30 of those would have gone undetected by the 'Loss of Pulse' system. A sensitivity of 69.3% means the system had a false negative rate of 30.7%.
Why Isn't Sensitivity Higher?
Some readers may be surprised that the sensitivity is not closer to 100%. We’ve discussed this topic at length in Cryonics Monitoring posts and media, but the high-level summary is that the PPG (Photoplethysmography) heart rate sensor encounters an unfavorable signal-to-noise ratio when the pulse is low, weak, or absent; when the room is bright; or when the watch is loose, etc. Sensor noise comes from several sources, such as ambient light, electromagnetic interference, and motion. Furthermore, the noise can be periodic, and cleaning and filtering algorithms can sometimes exacerbate, rather than reduce, this problem. The result is heartbeats on chicken drumsticks. What? If you’re not yet familiar, see a humorous, Cryonics Monitoring depiction of the issue here.
It’s clear from the clinical trial data above that, despite Google’s focused efforts to improve accuracy in detecting a lack of pulse—leveraging cutting-edge AI, significant funding, and top-tier talent—the feature still faces inherent limitations due to the photoplethysmography (PPG) sensor mechanism, which operates by emitting light into the skin to measure blood flow. This technology has fundamental limitations when the pulse is weak, low, or absent, and its effective detection range is constrained to heart rates between approximately 30 and 220 bpm based on the underlying optical principles. The persistent challenges Google faces highlight the inherent engineering constraints encountered when using this sensor for this purpose—constraints that even extensive resources and innovation could not overcome.
U.S. Availability - Soon or Not At All
Regulations for medical devices in the EU have recently become more stringent, narrowing the regulatory gap between the EU and FDA on medical devices, and the EU Class IIa regulatory requirements may now be comparably strict to those of FDA Class II devices. Both regulatory frameworks typically require comprehensive premarket data and specific regulatory controls to ensure the safety and effectiveness of medical devices. While the FDA designates these additional requirements as Special Controls, the EU achieves similar objectives through its Essential Requirements and conformity assessment processes under the Medical Device Regulation (MDR) 2017/745.
While the 'Loss of Pulse' feature is classified as a Class IIa medical device in the EU, there is a possibility it will be classified as a Class III device in the U.S. due to its critical role in detecting life-threatening situations, though this classification is unlikely. Class III devices are subject to the most stringent regulatory controls and require Premarket Approval (PMA), involving extensive clinical trials to demonstrate safety and effectiveness.
However, even if classified as Class II by the FDA, given the feature's current accuracy level and its critical application, Google may still face challenges in meeting the FDA's requirements, and may ultimately not be able to activate the feature for U.S. Pixel Watch 3 customers. It’s not necessarily a bad sign that it has been released in Europe but not yet in the U.S., as the FDA's approval process can be more time-consuming—in particular if the device is classified under the De Novo pathway, which is designated for novel devices without an existing predicate. However, based on average timelines for medical device approval process in the U.S., as well as average timelines for the De Novo route, and considering that Google already has experience in the Class II regulatory process (having gone through it for atrial fibrillation detection on Fitbit), I would estimate that if we don't see a U.S. release in the next 6-12 months, we likely won't see one at all.
Key Takeaways and Recommendations
1. Sensitivity of 69% The Google Pixel Watch 3 'Loss of Pulse' is an important new tool for cryonics monitoring in Europe and has been shown in clinical trial data to have a sensitivity of about 69%. The feature is designed to call emergency services promptly if it detects a problem.
2. Emergency Services But Not Contacts One feature lacking is the notification of contacts; the device currently only contacts emergency services, so be sure to wear your cryonics bracelet. The device does have a separate check-in feature that contacts emergency contacts, but it is currently not tied to ‘Loss of Pulse’. It also cannot be set on a repetitive basis (e.g., every day at 6pm). Other check-in apps are available within the cryonics space, from Cryonics Institute and Alcor.
3. FDA Approval Soon Or Likely Not At All Class II FDA approval can be more difficult and lengthy than approval in Europe. However, if we don’t see activation of the feature in the U.S. within the next 6-12 months, we are unlikely to see it at all.
4. Backup Plan for High-Risk Cryonicists By way of clinical trial data submitted by Google to the EU under MDR regulation, we can estimate the feature would fail to detect approximately 30% of loss of pulse events. High-risk cryonicists should establish a backup system for monitoring, such as a check-in app, for monitoring coverage in the event that the 'Loss of Pulse' system fails to detect you are having a problem.
European Availability:
Loss of Pulse Detection is currently available in Austria, Denmark, France, Ireland, Netherlands, Norway, Sweden, Switzerland, and the United Kingdom.
For European users interested in additional performance information, given the test data published so far is rather limited and invites speculation as to what conditions are favorable for accurate detection, I have been pursuing additional test data for the feature. Through consumer protections and MDR, citizens of the EU are likely entitled to greater transparency on test and performance of the feature upon request.
There would be various ways to go about obtaining this data, including by a request made to Google, requests made to the Notified Body that performed the conformity assessment, and the Authorized Representative of the feature in the EU. I have written formally to these entities to inquire further and request additional data. Such data may ultimately also appear here, with the ClinicalTrials.gov Study ID of NCT06430216, if FDA approval is achieved.
I will keep readers updated with a subsequent research article if I acquire any additional relevant data regarding testing and accuracy.