The fall detection feature that went live with the latest Apple Watch Series 4 in one such, even though the ability to take an ECG from your wrist captured most of the attention.
Can technology save lives? And I ask this question not in the context of technology that we find in hospitals, but in gadgets that most of us use. The answer is slowly but surely moving towards the affirmative. This is because, over the past few years, personal technology has been putting a lot of emphasis on wellness, stress and, of late, more serious stuff like blood pressure monitoring. While Apple showcased ECG functionality with its latest Apple Watch, Samsung’s new Galaxy Watch now comes with blood pressure monitoring. Meanwhile, having sold over 90 million devices, a company like Fitbit is sitting on so much user data, which when combined with the best of medical knowledge could help change lives like never before.
Apple seems to be way ahead of others when it comes to adding features that improve accessibility and safety for those wearing its devices. The fall detection feature that went live with the latest Apple Watch Series 4 in one such, even though the ability to take an ECG from your wrist captured most of the attention. While fall detection might seem like a relatively easy technology to execute, there is quite a bit of computing that goes into getting this right.
What is fall detection? The Apple Watch can detect a hard fall based on a bunch of algorithms and by analysing wrist trajectory and impact acceleration. If the Watch senses the person has been immobile for 60 seconds after a hard fall was detected, it will automatically call emergency services and send a message along with location to emergency contacts in the language chosen by the user beforehand.
Accurate fall detection has become possible with Watch Series 4, which is capable of measuring up to 32 g-forces, twice that of its predecessor. The accelerometer and gyroscope are up to eight times faster than previous generations of Apple Watches. All of these together help the Watch collect the higher fidelity motion data needed for fall detection.
But creating algorithms needed was not easy either, as fall detection had never before been put on a watch and hence the kind of data needed was not available. Apple got this training data over a long-term study working with real people to ensure the data was “true”. With 250,000 days of data from 2,500 participants, Apple’s team looked at the most common types of falls—slipping and tripping then bracing, falling off a ladder, tripping on an uneven curb—to hone in on/and recognise the angle/time of impact.
The study taught them that the types of falls for elderly are often different to falls that occur during recreational activity or contact sports. This is significant because a large chunk of older patients fail to tell their caregivers about falls, making it difficult for doctors to understand the origin or underlying causes, such as balance, movement disorders, reduced leg strength, or worsened hand-eye coordination. So, it was critical for falls to be documented so as to provide the patient, caretaker and physician with more insights. Fall detection is hence on by default for users who are above the age of 65. Look at this from the business side, and you understand how big an opportunity it is to cater to the elderly.
From its study of the fall data, Apple learnt about specific motion patterns associated with different types of falls—someone tripping, pitches forward and tries to brace with their hands, or someone slipping naturally swings their arms upwards in a windmill motion. The Apple Watch Series 4 can clearly recognise these fall patterns.
Interestingly, some events like a car crash can lead to the fall detection being triggered and there have been instances where the Watch has alerted emergency service even when it was not a fall per se. The Watch filters out a lot of regular trips and falls, and Apple has data on what these false positives usually are. But when a hard fall is detected, the Watch sends the user an alert and she has the option to either confirm she’s fine or to call for emergency services. If the user is immobile for a minute, the Watch will start emergency call automatically. The user can cancel this call, if needed. This also helps prevent overburdening of emergency services.
But if you thought this technology will get better over time as millions of people use the latest Apple Watch, then you are wrong. Apple does not collect user data from the watches, so the algorithm is based on original dataset only. The new data stays locally on user’s watches and hence is not used to change the algorithm itself. However, some local customisations do gradually happen on the Watch itself.
This is what technology is supposed to do—change lives for the better. And this year you can rest assured that across brands there will be a technology that can save more lives.