YouTubeMeetupFacebookLinkedInTwitterInstagramrss

Predicting Health Changes With AI And Intel



By Clinical Informatics News Staff

October 11, 2018 | After successfully deploying in Chile, AccuHealth is bringing its AI-empowered healthcare model—with Intel processors—to Columbia and New Jersey.

AccuHealth offers patient monitoring services to hospital systems. Using advanced analytics and artificial intelligence, AccuHealth delivers better care for patients with chronic ailments, by predicting changes in health status, and intervening before things get worse.

“When you have a disease that’s a ‘slow motion’ disease like diabetes, everything is going very slow motion… you get bored. But suddenly it goes fast forward. For a person who has diabetes, it would seem everything goes well until it doesn’t,” Xavier Urtubey, co-founder and CEO of AccuHealth, explained.

The turning point in chronic illnesses often means patients end up in the emergency room with significant health impacts. But early intervention and monitoring can avoid an ER trip, keeping patients healthier and reduce costs to insurers by up to 50% per patient.

But how? It was a question Urtubey wanted to tackle. “If we had a way to really look at these slow-motion moments, and look under the radar of a regular physician, we could see that there are some clues that show you that, ‘Hey, this may seem like slow motion, but under the radar it’s moving faster and faster.’”

Seeing under the surface of accumulated health data—whether from devices, sensors, or lab values—is not easy. AccuBrain, AccuHealth’s AI and advanced analytics engine, combines patient biometric, self-assessment, behavior, and demographic data with machine learning algorithms and predictive models based on population data. And the model updates constantly based on new inputs from individual patients and population health on the whole. “We are gathering the information that is specifically aiming to create an impact for the patient before it gets out of hand,” Urtubey says.

The AccuHealth team does not see patients in a physical location. Instead, AccuHealth physicians, nurses, dieticians, and behavioral analysts work with patients and caregivers via sensors, questionnaires, home-based tests, or telehealth to be able to react quickly once a downward health trend is identified.

AccuHealth preparing to launch phase one of a US pilot project, Urtubey says, with a hospital system in New Jersey. Up to 3,000 chronic disease patients will be constantly under AccuHealth services; another approximately 9,000 patients will be part of the population management approach, with less intensive monitoring.

For patients with chronic diseases, “We are their safety network to gather the actionable data,” Urtubey says.

Powering An AI Solution

“The real value is the data plus the insights you’re able to generate from them with the AccuBrain algorithm,” says Jennifer Esposito, General Manager, Health and Life Sciences Group at Intel. The AccuHealth data center runs Intel Xeon Scalable processors.

“Intel has been working with Xavier and his team for a few years now,” Esposito says. “When we first started, it was about the longer-term roadmap or picture of where [AccuHealth was] going to go with the business and how we, from a technology perspective, could future-proof what was going to happen.

“Most recently we’ve been working on optimization of the architecture that runs in their back office to support the AccuBrain. What we’re really trying to do—that we do with a lot of the partners we work with—is leverage the largest and greatest technology to make sure that the solution is really running as best as it possibly can and, in some cases, bringing with it additional efficiencies and speed as well as [decreasing] total cost of ownership.”

For others who want to experiment with AI, Esposito encourages them not to start from scratch. “Particularly when it comes to artificial intelligence, people think they need to invest in a lot of new specialty hardware to start doing AI. What we were able to do with AccuHealth, was work with them on the existing infrastructure that they already had, and do some upgrades into our Xeon scalable platform, which allowed them to continue to ramp and will give them a really good runway into the future.

“This has always been a really compelling solution to us,” Esposito says, “because it’s a way to really look at how to bring artificial intelligence all the way to the edge—to the patient.”

0 Comments

Add Comment

Text Only 2000 character limit

Page 1 of 1