Keeping loved ones safe while preserving independence with Coral and Care.ai

Chakri Toleti was at his office in Florida in January 2018 when he got the call from India alerting him that his mother had fallen.

Only hours after the fact did he learn that his mother had lain on her bathroom floor for a good 20 to 30 minutes before a live-in caregiver discovered her. He felt helpless that he didn’t know she was in trouble until much later and therefore couldn’t get her help. He knew that many other people must face similar struggles. So he did what any good serial entrepreneur does. He started a company to solve what he perceived as a major challenge: How can a caretaker keep a loved one or patient safe while protecting the patient’s privacy and allowing that person a large measure of independence?

Within just a few months, he had formed a company, Care.ai, deploying computer vision and AI from Coral to detect falls and other dangers for the elderly and other at-risk individuals. By late 2019, the company had grown to 40 employees and was running trials in nursing homes across the United States. “It’s [a] HIPAA-compliant solution that uses AI at the edge,” Toleti says.

A vital factor in its success: Coral's Iocal AI platform, which keeps processing in the room with patients, preserving their privacy.

A growing need

The global population is aging. In the United States, for example, people over the age of 65 will outnumber children, for the first time in the country’s history, by 2035. With an aging population comes a growing need for care—at home, in hospitals, and in nursing facilities.

The demographic shift is contributing to a looming shortage of healthcare workers. Industry advisory firm Mercer projects a deficit of 446,300 home health aids alone in the U.S. by 2025. One clear area of need is patient monitoring. Even if nurses, aids, and family members are available, they can’t be everywhere at once. They often are delayed in responding to falls and other problems.

The solution at the heart of care.ai’s monitoring device: optical sensors that feed data to a Coral tensor processing unit (TPU) inside. The TPU interprets the data to determine whether someone in the room is at risk, and if so, sends alerts to caregivers. No video gets transmitted, preserving privacy.

Monitoring without watching

Care.ai’s monitoring system includes a sensor package about the size of a hardback book that care providers can mount on a wall. Optical sensors feed data to the onboard Coral Mini PCIe Accelerator. The processor runs AI algorithms trained on an extensive library of behavioral data from healthcare settings.

The AI can detect when a patient falls, when someone in a weakened state risks a fall by getting out of bed before they’ve fully recovered, when a dementia patient wanders out of their room, and other potentially dangerous conditions. It can even verify that care providers entering rooms use handwashing stations. The AI then sends alerts via its power-over-ethernet connection or Wi-Fi straight to nursing stations, the cell phone of a loved-one, or another specified device.

The result is what Toleti calls a self-aware room. “Imagine a Tesla car sitting in a room,” Toleti suggests. In other words, the same kind of processing power and reaction time that lets a self-driving car navigate highways goes to work in a room ensuring patient safety, thanks to care.ai. None of which would be possible without being freed by Coral of the latency burden imposed by cloud-based processing.

“What the team at care.ai has put together is truly ground-breaking,” says Mark Crandall, CIO at Consulate Health Care, which provides services to seniors at 140 centers around the U.S. "This is the first and only autonomous monitoring solution in healthcare that can truly transform an ordinary room into a self-aware room."

At Consulate Health Care, the system helps prevent patients from falling, from developing pressure ulcers, from wandering off, and more. Crandall likes that the system doesn’t impose any burdens on patients. It requires no action from them, and they don’t even have to wear a monitoring device. Every other system he's seen requires one or the other, curtailing independence. Compliance with U.S. healthcare privacy regulations is another key selling point for Crandall. “Care.ai’s architecture was built from the ground up with the security of our patient data in mind,” he says.

A future of smart healthcare

Along with the pilot project at Consulate, Toleti says care.ai is working with several other Hospitals and post-acute care pilot customers. He'd like to see care.ai bring a new standard of care for at-risk patients in hospitals, residential care centers, and homes around the world.

Fortunately, Toleti’s mother sustained only minor injuries from her fall. But Toleti hopes the scare that inspired him to found care.ai will soon help tens of thousands of others stay safe while remaining independent and maintaining their privacy, even as it gives peace of mind to loved ones both near and far.

If you're interested in how Coral can help you build your next product please contact us.