How fall detection is moving beyond the pendant

Digital health innovators look to the wrist, the ears and the wall for new ways to keep seniors safe.
By Jonah Comstock
04:22 pm
Share

For the over-65 population, falling can be a serious health risk. Three million older people each year are treated in emergency rooms for fall-related injuries, according to the CDC, and one in four older adults falls each year.

But the options for guarding against falls are limited. For a long time, the standard of care has been a “personal emergency response system” or PERS device — a pendant worn around the neck with a button that an older person can press to contact a call center or emergency services. Newer PERS devices, called mPERS, can be worn outside the house and some can even detect a fall automatically in case someone can’t press the button.

Yet they’re still far from an ideal solution, not the least because of the stigma attached to them. Most people, elderly adults included, associate these devices with the elderly woman in a famous 1987 LifeCall commercial who helplessly intones “I’ve fallen, and I can’t get up!”

“No one likes to admit that they might have a risk of falling,” Noga Barpal, a business development associate at Vayyar, a company working on a new fall detection technology, told MobiHealthNews. “…And what happens when you wear something is you’re kind of calling out to everyone ‘Hey I might be falling, so I’m wearing this pendant.’ And that’s not nice for anyone.”

Recently, the digital health world has stepped up its effort to create an alternative that’s not only stigma-free but also proactive, detecting and even preventing falls.

The Apple Watch

Perhaps the most high-profile recent foray into fall detection came from Apple, which announced the addition of a fall detection feature to the Apple Watch Series 4 last year. The announcement was somewhat overshadowed by the announcement at the same event of the Watch’s ECG functionality.

"It’s interesting, identifying a fall may seem like a straightforward problem but it requires a tremendous amount of data and analysis," Apple COO Jeff Williams said at the time. "We collected data on thousands of people and captured data on real-world falls.”

Specifically, the company worked with a combination of movement disorder clinics and friends and families of employees to collect 250,000 days worth of data on 2,500 people before launching the feature, according to Apple.

The feature is automatically turned on for anyone who tells the Apple Watch during the setup process that they are over 65. When it detects a fall, it first displays a notification on the device. In the event of a false alarm or a fall that the user can easily get up from, they can dismiss the alert. But if they don’t, the Watch automatically dials emergency services. Then, when emergency services pick up, it reads aloud an audio message that includes the user’s location.

This was no easy feat for a device that works anywhere in the world, the company told MobiHealthNews, and it’s even designed to read that information in the local language based on what emergency department is being contacted.

According to Apple, the technology to do fall detection well — both the high-fidelity sensors and the computer processing power to have those sensors and algorithms running constantly — really hasn’t existed until recently, which is part of the reason fall detection didn’t debut until the Series 4. The other part of that reason is the time it took the company to perfect the algorithm, which isn’t fooled by a stunt fall or by movements in certain sports, like tennis, that can resemble a fall to sensors.

“When you trip, your body will naturally pitch forward and your arms will go forward to brace yourself,” Williams said at the event last year. “However, if you slip, there’s a natural upward motion in the arms. These are motions Series 4 is ideally suited to recognize. With a new accelerator and gyroscope, the Watch analyzes wrist trajectory and impact acceleration to determine when a fall occurs.”

FallCall Solutions

Even before Apple announced its fall detection offering a startup company called FallCall Solutions was working on creating a fall detection software for the Apple Watch.

“FallCall Solutions was born out of the frustration that I was feeling as a trauma surgeon seeing that the majority of our population coming into the trauma room were people over the age of 65 who fell and were on anticoagulation, like blood thinners, for their atrial fibrillation and a variety of other conditions,” Shea Gregg, FallCall’s CEO, told MobiHealthNews. “…This was right around the time that the smartwatches started coming out, 2015, and the smartwatch that really caught my attention at that time was Apple Watch.”

Gregg’s vision was to build an app for the Apple Watch that could distinguish between high-mechanism and low-mechanism falls. His team was able to develop an algorithm approaching 90 percent accurate fall detection for high mechanism falls. But they ran into a problem.

“We basically hit a brick wall when it came to WatchiOS, when it came to being able to run fall detection 24-7 on a device,” Gregg said. “And the reason for that is when the Watch app is open in the foreground, it can run continuously and it was a very effective fall detector. When you close the app, your app shares the operating system’s processing time with Apple Music and phone calls and all this other stuff. We decided not to release it for that reason: that it was not going to give continuous fall detection.”

Instead, the team developed an app for iPhone and Apple Watch that included fall response features, but without the fall detection. Gregg says it has a leg up on Apple’s built-in system because it allows people to set levels of escalation, calling friends and family members or a designated call center rather than immediately calling emergency services.

“So, as opposed to burdening the 911 system we’ve employed that electronic connection to a central monitor and we encourage people to sort of test it and get used to it, as opposed to being in that situation where ‘Oh my god I contacted 911 and I have the police on the phone.’ It can be very intimidating, especially for an elderly person or a caregiver.”

Gregg says that Apple’s announcement was both frustrating and validating, but his team is pushing forward with the launch of a fall-detection app — not just for the Apple Watch, but also for Android devices on WearOS. And the app will integrate with Siri, so users can program the Watch to respond to a phrase like “Hey Siri, I fell”.

Gregg believes that the wrist is the right place for fall detection, and that the best chance of overcoming the PERS stigma is to build it into existing, trendy devices — which has the added benefit of not overwhelming an older adult with a lot of disparate pieces of tech.

“There are devices out there that can do one thing very well — mobile PERS, they can call for help — but why would you limit yourself and have to carry all these different devices when you can get everything in one on your wrist?” he said. “It’s automated, it’s smart, and I believe that that’s going to be the dominate solution going into the future.”

Beyond the wrist

While everyone interviewed for this story agreed that the pendant’s day should be over, not everyone sees the wrist as the best spot. Vayyar, for instance, has created a wall-mounted device that uses radar-based 3D imaging to monitor for falls without invading a user’s privacy.

Currently, the device, called Walabot Home, is on sale for use in the bathroom, where Barpal says 80 percent of falls occur — and where a wearable device, whether it’s a pendant or Apple Watch — might fall short.

“A lot of times if you’re getting up to go to the bathroom at night, you might be at risk of falling then but you didn’t remember to wear your pendant,” she said.

Vayyar is working on not only deploying the device to other rooms of the house, but also to expand its capabilities beyond falls to other kinds of behavioral monitoring that could yield health insights.

Starkey Hearing Technologies, a well-established player in the hearing aid market, has recently begun incorporating fall detection into its high-end offerings, like the Livio AI connected hearing aid.

“The closer that we can put our sensors to the ear, the closer you are going to be to the actual vestibular system,” Starkey CIO Dr. David Fabry told MobiHealthNews. “So we think that the ear is not only a good idea, it’s one of the best places because you’re close to the end organ of balance. Specifically for fall detection, having two [sensors] and having them close to the ear is the best location.”

Hearing aids are a device already used by many elderly folks, and those users are less likely to accidentally leave them behind than they would be with an Apple Watch, since they need them to hear well. But on the other side of the coin, hearing aids have the same adoption and stigma problem as PERS pendants.

Fabry thinks building additional functionality into hearing aids, like fall detection, will actually help address this problem, as well as help to make the devices more affordable.

“What we see is while cosmetics are one angle to work with hearing aids, the other is related to the functionality and this functional utility,” he said. “We see that as we can reduce the stigma by improving the performance [and] improving the applications for people with hearing loss disorders, higher adoption rates will lead to higher numbers of people using them and then the economies of scale kick in. We’ll reduce stigma, we’ll make hearing aids cool and in turn there will be more of them in the market which will address the cost.”

Like Apple, Starkey has been waiting for the technology to catch up with this vision. Miniaturizing the processing and power requirements for either a hearing aid or a fall detection device would be a tall order; fitting both into one package was an even loftier technical challenge.

From fall detection to fall prevention

As fall detection is moving forward into new territory, the holy grail of fall prevention still remains out of reach. But in a world of artificial intelligence and big data, it’s not out of sight. As devices like the ones described above collect more and more data on what telltale signs precede falls, many believe we’re moving to the point where a device could intervene somehow and prevent a fall from happening.

Fabry thinks a hearing aid could accomplish this by offering a warning or even coaching in the moment.

“We know that many people who are at risk for falling start to lose muscle control, so that if they’re simply going to sit in a chair, they have less control over that slow deceleration from a standing to a sitting position than people who don’t experience those balance disorders,” he said. “And what we can do is start thinking about not only measuring that motion, but trying to coach them the way a physical therapist or an audiologist might do in the course of vestibular rehabilitation. … That’s where we’re beginning to see the movement between fall detection and fall prevention.”

Apple declined to comment on whether it was looking into prevention. But the company would be at a disadvantage in some ways because of the hard line it’s taken on data privacy. Apple isn’t collecting any kind of training data from its fall detection features, even in aggregate, the company said. It’s all stored locally, for the benefit of that particular user and, if they choose to share the data, their caregiver.

Over the years, fall prevention efforts have included everything from an augmented reality app to smart shoes and socks. Who knows what that technology will eventually look like, but hopefully we are headed for a future where no one will need to say “I’ve fallen, and I can’t get up!” ever again.

Consumerization of Healthcare

In April, we'll look at the consumerization of healthcare from a variety of angles, including how to treat patients as customers.

Share