J. Tod Fetherling

CEO / Data Geek

18 CARE Prediction Models Now Available in Amazon's SageMaker Machine Learning Marketplace

We have been working for the last year developing and refining multiple disease state prediction models for everything from opioid addiction to lung cancer. Originally launched at AWS re:Invent 2018 in Las Vegas, the first of many CARE™ Prediction Models are available through Amazon's SageMaker machine learning marketplace within AWS.

Chris McLaurin, Chief Technology Officer at Perception Health said: "With a simple extract of a patient's medical history, our algorithms help doctors by providing a score that identifies early indicators of a particular disease or the need for a surgical procedure, in the next 360 days.”

These proprietary CARE algorithms, provided by Perception Health, can be licensed on a per patient basis to run on Amazon SageMaker. This fully-managed platform gives developers and data scientists the power to quickly and easily build, train, and deploy machine learning models at any scale.

“CentraState partnered with Perception Health two years ago to improve the ecosystem of our local referral network,” said John Ulett, Chief Information Officer, CentraState Healthcare System, Inc. “The disease predictions provided by Perception Health will allow us to use advanced analytics to predict the onset of serious diseases for the communities we serve. We believe this is the next evolution of care management and CentraState is at the forefront.”

“Perception Health already brings advanced disease prediction models to the industry,” said Garth Fort, Director, AWS Marketplace, Amazon Web Services, Inc. “Now, with Perception Health’s CARE algorithms running on Amazon SageMaker, Perception Health can rapidly deploy disease models that utilize AI in the selection and screening of patients through the prediction models.”

The following is a list of the first eighteen CARE Disease State Prediction Models from Perception Health available on Amazon SageMaker and there are many more to come.

Original Press Release