When Apple Watch Series 10 is released in September 2024, Apple brings another attempt at health monitoring. Apple Watch Series 9, Ultra 2, SE 3 and newer models can detect abnormalities in the user’s breathing rhythm during sleep monitoring and alert the user of suspected nighttime apnea. Of course, this also includes the newly released Apple Watch Series 12 and Apple Watch Ultra 4.
On September 4, 2026, the “Mobile Sleep Apnea Sign Alert Software” was approved by the State Food and Drug Administration, and the function was officially launched in mainland China on October 14.
Why does Apple Watch develop this feature? What is the difference between it and previous sleep monitoring?
Why sleep apnea?
Who doesn’t know a few people who snore in their sleep?
Your family, roommates, the old man relaxing in the bathhouse, and your deskmate lying down in class may all make loud noises in their sleep – this is something we have discussed in “Snoring: What is the Night Thunder You Are Familiar with?” discussed in detail in the article: Snoring indicates that the upper respiratory tract is blocked during sleep. Snoring is caused by the vibration caused by the high-speed passage of airflow when the upper respiratory tract is narrowed.
We pay attention to snoring because snoring, as a phenomenon, may indicate a sleep-related breathing disorder – obstructive sleep apnea (OSA, obstructive sleep apnea). It is the most common type of sleep apnea. People with sleep apnea repeatedly stop and start breathing during sleep. If the pause is because the airway is blocked, such as being too fat, not growing well, drinking too much, etc., then it is obstructive; if the pause is because the brain suddenly no longer tells the respiratory muscles to exercise, it is central sleep apnea (CSA).
Regardless of the type or cause, sleep apnea will cause the patient’s blood oxygen level to drop suddenly. In order to survive, the brain can only wake the patient briefly, and the quality of sleep will inevitably decrease significantly. What follows is a series of problems such as daytime drowsiness, headache when getting up, and even secondary cardiovascular and cerebrovascular diseases. It can be said that OSA, or sleep-related breathing disorder, has brought huge health threats to people’s lives.
Can this disease be cured? Can be cured. As far as common OSA is concerned, treatment options have long been mature.
So what’s the problem? ——Confirmed.
The prerequisite for reasonable treatment is to be diagnosed with the disease. OSA has a fairly complete set of diagnostic methods, one gold standard and one silver standard, which we also introduced in detail in the article mentioned above. The gold standard is polysomnography (PSG), which has many parameters and can measure accurately. The disadvantage is that it requires an overnight stay in the hospital, and the patient may not be able to sleep. Moreover, the price of the test is high, which deters many potential patients. Silver standard home sleep monitoring allows patients to complete the diagnosis without going to the hospital, but it still requires a bunch of equipment including an actimeter. Sensitive patients may still be unable to sleep, and the data must be taken back to the hospital for secondary analysis.
This leads to two situations that often occur in patients with sleep-related breathing disorders: they have never been diagnosed, but they came to the conclusion of the disease based on reminders from others, and used positive pressure non-invasive ventilators with a strong sense of self-management; their own lives have been seriously affected, and even obvious complications have occurred, but they are completely unaware that these problems are caused by sleep apnea. It is both difficult to detect and diagnose, which makes the diagnosis rate and incidence rate of sleep apnea long-term disproportionate.
For this reason, our current research on sleep monitoring is focused on zero/low-load sleep monitoring equipment, striving to complete the diagnosis of OSA with the smallest possible equipment and as little impact on sleep as possible. But unfortunately, such equipment isNo diagnostic qualificationsYes, if you want to be diagnosed, you still have to go to the hospital.
What does the Apple Watch do?
Since smartphones and consumer-grade smart wearable devices are already a collection of various sensors, as long as there is a reasonable set of algorithms, it is not impossible to use them to gain a glimpse of the user’s sleep, even if it does not reach clinical-level accuracy.
For example, sleep monitoring apps before the Apple Watch era basically followed the same path: They would ask you to put your phone next to your pillow and leave the app in the foreground. At this time, the application will collect data from sensors such as gyroscopes and accelerometers, collect vibrations caused by the user’s movement on the bed, and determine the sleep phase of the person under test based on the developed algorithm. Some applications will also take samples to record snoring within a certain period of time.
The addition of smart wearable devices makes these monitoring more accurate. For example, it can reduce the influence of bed partners, and can also more effectively monitor respiratory rhythm, heart rate and other data. Some applications that are more relevant to life have also been born, such as smart alarm clocks that adjust the ring or vibration time based on the sleep phase inferred by the algorithm, making it easier for users to wake up and less annoying to get up.
Sensors and algorithms are indispensable. For example, the way Apple Watch determines the current sleep phase of the wearer is to rely on the built-in three-axis accelerometer to capture subtle movements that are difficult to detect by the human eye, and then use these movement data to determine the algorithm based on a cycle every 30 seconds. The development of this set of algorithms came from a development trial that included 858 participants and 1171 nights of polysomnography (PSG) data, and a validation trial with an additional 166 participants and 299 nights. It ultimately allowed Apple Watch users to obtain a rough sleep phase distribution without using EEG.
Similarly, you may also notice that watchOS 11 introduces nighttime vital signs feature. It includes respiratory rate (beats/min), blood oxygen (%), heart rate (beats/min), wrist temperature (°C), sleep duration (min) and other data. Some rely more on new sensors, such as the temperature sensor introduced in Apple Watch Series 8; some rely on new self-developed algorithms to convert the data detected by the sensor into the value of a certain indicator.
This time the breathing disturbances data and sleep apnea notification function belong to the latter. The Apple Watch solution is still based on a built-in three-axis accelerometer. To find a relationship between a measure of breathing disorder (accelerometer changes) and actual sleep apnea, Apple launched a wider trial than previous trials. The algorithm development phase included 2160 participants with a total of 3936 nights of home/laboratory PSG data; the algorithm validation phase included an additional 2542 participants with a total of 7220 nights of data. The subjects included patients with sleep apnea of varying degrees and normal people. The resulting correlation can be represented by this graph:
It can be seen that there are regular patterns to be followed between the changes in accelerometer readings (ABC) and the changes in abdominal band action (D) and oral and nasal air flow (E) when apnea events occur in conventional PSG. The algorithm ultimately used by Apple Watch is to detect acceleration changes that conform to these laws and provide respiratory disorder data. It will conduct an evaluation every 30 days after the first measurement. If there are at least 10 days of valid data during the period, and 50% of the data is higher than the normal value, a reminder will be sent to the wearer, indicating that sleep apnea may exist.
What’s the point of this feature?
Contrary to common speculation after the press conference, Apple’s design of this feature is relatively conservative. The Apple Watch device itself is not a medical device. On the one hand, the monitoring accuracy and range of the sensor may be different from clinical medical-grade equipment. On the other hand, the dimensions of the data it collects are also relatively limited. It is imprecise and unscientific to draw a diagnosis directly from incomplete data. Therefore, Apple’s own design of this feature is limited to “danger reminders for people who have not been diagnosed with sleep apnea”, rather than “helping people at potential risk to self-diagnose sleep apnea.” It is also approved by the FDA as an “Over-The-Counter Device To Assess Risk Of Sleep Apnea” (Over-The-Counter Device To Assess Risk Of Sleep Apnea).
The app uses “disturbed breathing” as a measurement value. This is not a direct indicator of sleep apnea. When studying this disease, there is usually an index called the apnea-hypopnea index (AHI, apnea-hypopnea index), which represents the average number of apneas or hypopneas recorded per hour. It is generally considered that 5 to 15 times are mild sleep apnea, 15 to 30 times are moderate, and more than 30 times are severe. The “hypopnea state” in this concept is clearly defined, requiring a single airflow reduction event lasting more than 10 seconds and a drop in blood oxygen of 4% or more. (Suspected) respiratory disorders calculated solely from acceleration changes obviously fail to meet this requirement. In the test data provided by Apple, the obtained respiratory disorder data is 91.4% within the acceptable error range of the true AHI value. This value is not low, but it is not 100% after all.
So why develop a separate algorithm? Doesn’t Apple Watch already have breath detection and blood oxygen detection?
Indeed there is. But obviously, Apple Watch does not come with a breathing mask, so the breathing data of Apple Watch is an indirect value, which is calculated by the PPG sensor used by the heart rate detection function.
This is a fairly common sensor, and a number of algorithms have been developed based on its readings. In “Is the “heart rate detection” on the wrist reliable? What data can smart watches help you understand? 》In the article, we briefly introduced the principle and use of PPG sensors. The strapless blood pressure measurement function of Samsung Galaxy Watch relies on the PPG sensor. Similarly, blood oxygen detection also relies on a PPG sensor, but it uses a red LED. As early as October last year, South Korea’s Ministry of Food and Drug Safety approved an application for Galaxy Watch 5 and above models to use the Samsung Health Monitor app to detect sleep apnea symptoms. The standard it uses is blood oxygen value changes. Users only need to monitor sleep for more than four hours twice within ten days to obtain the results.
In fact, the blood oxygen test commonly used in hospitals is also an “indirect” test based on PPG. After all, you cannot draw blood every few seconds and run a blood gas analysis. The specific principles are as follows in “Popular Science | What is the use of blood oxygen monitoring on Apple Watch Series 6?” “Introduced in “. Compared with professional fingertip oximeters, the fit degree, fitting position, and fixation method of the back of the smart watch to the wrist are worse than those of a clip with thinner skin and more stable fixation on the fingertip. This is still a common problem faced by all similar wearable devices. In the future, with the evolution of algorithms and watch fixation methods, blood oxygen may also be integrated into the determination of sleep apnea.
The solution chosen for Apple Watch will of course also be affected by posture, nighttime activities and wear. Apple’s adjustment plan is to reduce the false positive rate as much as possible. In the clinical validation study of the algorithm, the weighted sensitivity of the reminders issued by the algorithm for moderate to severe patients with clear evidence (sickness and receipt of reminders) reached 66.3%, exceeding the pre-designed 50%; the weighted specificity for mild patients/normal people (normal and no reminders/mild illness and receipt of reminders) reached 98.5%, also exceeding the pre-designed 85%.
In the report, Apple emphasized that the specificity for normal subjects reached 100%, which means that among a total of 1,499 participants, those who received the notification had at least mild sleep apnea. This means that for this function of Apple Watch, the priority is to ensure that those who receive the reminder should check it out, but it does not guarantee that those who do not receive the reminder will be fine.
Then, the significance of developing this feature for Apple Watch is obvious. Unlike the previous atrial fibrillation detection, the greatest significance of this function is not “condition monitoring”, but “finding problems”. The premise that PSG testing is troublesome and expensive will not change for a while. In other words, most potential patients still need to rely on reminders from others and perception of their status during the day to be aware of their condition. The Apple Watch’s reminders can, on the one hand, help people living alone to discover their sleep status without being reminded. On the other hand, they can sound an alarm for potential patients who know they have risky symptoms but do not get checked or consult a doctor for various reasons, so that they can make up for the situation before it develops and affects other aspects of their health.
Related reading
- Epstein LJ, Kristo D, Strollo PJ Jr, Friedman N, Malhotra A, Patil SP, Ramar K, Rogers R, Schwab RJ, Weaver EM, Weinstein MD; Adult Obstructive Sleep Apnea Task Force of the American Academy of Sleep Medicine. Clinical guideline for the evaluation, management and long-term care of obstructive sleep apnea in adults. J Clin Sleep Med. 2009 Jun 15;5(3):263-76. PMID: 19960649; PMCID: PMC2699173.
- Gottlieb DJ, Punjabi NM. Diagnosis and Management of Obstructive Sleep Apnea: A Review. JAMA. 2020;323(14):1389–1400. doi:10.1001/jama.2020.3514
- Abbasi A, Gupta SS, Sabharwal N, Meghrajani V, Sharma S, Kamholz S, Kupfer Y. A comprehensive review of obstructive sleep apnea. Sleep Sci. 2021 Apr-Jun;14(2):142-154. doi: 10.5935/1984-0063.20200056. PMID: 34381578; PMCID: PMC8340897.
- Estimating Sleep Stages from Apple Watch – Apple
- Estimating Breathing Disturbances and Sleep Apnea Risk from Apple Watch – Apple
- Obstructive Sleep Apnea – John Hopkins Medicine
- Popular Science | What is the use of blood oxygen monitoring on Apple Watch Series 6?
- Snoring: What is the night thunder you are familiar with?
- Is the “heart rate detection” on your wrist reliable? What data can smart watches help you understand?
> Follow Minority Little Red Book and experience the wonderful digital life🍃
> Practical and easy-to-use genuine software, minority group presents it to you🚀
