For much of his career as an anesthesiologist and researcher, Jake Sunshine, ’12, has been preoccupied with a question: What if the smart technology all around us — our speakers, phones, watches, etc. — could save our lives in an emergency?
A “once-in-a-generation investigator,” as one colleague put it, Sunshine began exploring options nearly a decade ago through his work at the University of Washington as an associate professor in the School of Medicine. After countless interdisciplinary partnerships, hours of research, trials and tribulations, he has this year debuted smartwatch technology that can detect a wearer’s cardiac arrest and call for help.
The need speaks for itself: Without intervention, out-of-hospital cardiac arrest is one of the world’s leading causes of death, carrying a 90% mortality rate for the approximate 350,000 Americans who experience it every year. Getting help there to administer CPR quickly can mean the difference between life and death.
What if?
Sunshine’s first attempt at answering his lifesaving technology question began with research examining how smart devices could detect another emergency frequently experienced without witnesses: overdose. Together with colleagues in computer science, he identified how a person’s smart device (smartphone or smart speaker) could be used to monitor for respiratory failure, the mechanism of how people die from opioid overdose, through inaudible tones that pinpoint breathing motion.
With a proof of principle that everyday tech could be used for health purposes beyond fitness, Sunshine and his team began exploring other ways to answer that big “what if?”. As it turns out, most people experiencing cardiac arrest also develop a particular respiratory problem called agonal breathing, classified by shallow, infrequent, irregular gasps that don’t deliver enough oxygen to the body. Because agonal breathing has such a distinct sound, Sunshine realized his team could use machine learning to train a smart device to identify the sound in the same way wake words like “Hey Siri” or “Alexa” signal devices to pay attention.
Sunshine wants to build smart-device features that average people can use to protect themselves in an emergency.
Agonal breathing, though, isn’t a word — and it’s not a sound that can be replicated in a lab. Instead, Sunshine partnered with Thomas Rea, UW professor of medicine and medical director for King County’s Emergency Medical Services Division, to gain access to agonal breathing samples collected in 911-call audio that the team could use in testing and training. With graduate student Justin Chan and professor Shyam Gollakota from the UW computer science department, Sunshine demonstrated that a smartphone and smart speaker could accurately classify agonal breathing audio arising from cardiac arrest.
Based on this research, Sunshine had all the reassurance he needed to understand that smart devices could play a role in saving lives and could be scalable. In 2018, he co-founded Sound Life Sciences (SLS), through the UW with the help of CoMotion, which has helped facilitate 2,160 licenses and 52 UW spinoffs in the last five years. SLS established its headquarters on campus in Startup Hall, one of UW CoMotion Labs’ incubator spaces for early-stage tech startups. CoMotion served a critical role in the company’s early success, connecting Sunshine and SLS with resources and guidance on licensing, funding and bringing an invention through the regulatory process.
“You can do research and write papers all the time, but something that takes the next step toward actually getting out into the world is a really, really, really long journey, and there are only a few examples of technological innovations that actually do that,” Sunshine says. “But this is a great example of the University being a multidisciplinary environment — people bringing different areas of expertise together to solve a problem. Using smart devices for detecting cardiopulmonary emergencies really got its start at the University of Washington.”
SLS received an FDA clearance for their smartphone breathing monitor in 2021. The following year, the company was acquired by Google. Sunshine stayed on to lead a team pursuing that important question: “What if?”
Sunshine says the UW’s collaborative environment helps him brainstorm and research with colleagues across campus.
Finally, bringing an idea to market
Last March, the Google Pixel Watch 3 launched the first publicly available version of a feature that builds on the idea that our everyday devices can be used to save someone’s life. The technology smartwatches already use to measure heart rate can be applied to detect when someone loses their pulse (when the heart stops beating in a life-sustaining fashion) and call for help.
Sunshine explains that watches are well-suited as sensing devices for emergencies because they’re worn consistently and have connective capabilities that allow them to call for help. Smartwatches can calculate wearers’ heart rates using rays of green light, a color absorbed by red blood cells, to measure blood flow — a process called photoplethysmography. Using this technology, Sunshine and his team developed a complex algorithm that can detect if a loss of pulse occurs, confirm if someone is not responsive (consistent with cardiac arrest) and call for help. All of this happens in around one minute. That speed is critical: Survival rate decreases by 10% every minute without treatment like CPR.
“This is a great example of the University being a multidisciplinary environment — people bringing different areas of expertise together to solve a problem.”
Sunshine adjusts his Google Pixel Watch, capable of detecting when a wearer‘s pulse stops.
Sunshine says the technology correctly identifies loss of pulse in about two out of every three occurrences. It’s a balancing act between sensitivity (a higher degree of accuracy) and specificity (which, if too low, could yield more false positives and overwhelm emergency services). “Status quo is that when people experience unwitnessed cardiac arrest, the survival rate hovers around zero,” Sunshine says. “So even if the watch catches two out of three, it could potentially be helpful and could lead to lives saved.” He hopes that performance will improve with additional research.
Sunshine’s loss-of-pulse research was published in the journal “Nature” on the same day the device was cleared by the FDA. Currently, the technology is only available on the Pixel Watch in 17 countries, but Sunshine is hopeful that it will be more widely available in the future. After all, he says, the use of everyday devices for passive detection of everyday emergencies is baked into the DNA of his original company — and that technology saves the most lives when it’s available to everyone.
Story by Chelsea Lin // Photos by Mark Stone
