Over 100 Utah Auctions End Today - Bid Now
Over 350 Total Lots Up For Auction at Three Locations - MD 02/08, TX 02/10, Ontario 02/14

Amazon Comprehend Medical to bring natural language processing to healthcare

by Thomas Dworetzky, Contributing Reporter | December 05, 2018
Health IT
Alexa may one day help patients keep their daily med routine on track – and more.

Online giant Amazon is developing a new service to provide better clinical decision support to health care providers, insurers, researchers, and clinical trial investigators as well as healthcare IT, biotech, and pharmaceutical companies.

“We are excited to announce Amazon Comprehend Medical, a new HIPAA-eligible machine learning service that allows developers to process unstructured medical text and identify information such as patient diagnosis, treatments, dosages, symptoms and signs,” Dr. Taha A. Kass-Hout and Dr. Matt Wood wrote on the Amazon Web Services (AWS) Machine Learning Blog.

More than just a better way to ensure patient compliance with drug regimens, the system is expected to help streamline revenue cycle and clinical trials management, while more efficiently addressing data privacy and protected health information requirements. These capabilities are predicted to make it a top competitor with solutions produced by other organizations such as IBM’s Watson Health and Optum from United Health Group, according to an AMA report.

Its creation furthers the objective to turn health records into “big data”, thereby doing away with the challenge of converting most health and patient information today, which comprises so-called “unstructured medical text”, into AI usable form.

Converting this information, which consists of doctor's notes, prescriptions, transcripts of interviews, and pathology and radiology reports, is too laborious to do by hand – a task made even more costly by the need to employ skilled medical experts to enter it into a system – or create custom software to tackle it.

Amazon Comprehend Medical spots key information “automatically, with high accuracy, and without the need for large numbers of custom rules,” said Kass-Hout. “Ultimately, this richness of information may be able to one day help consumers with managing their own health, including medication management, proactively scheduling care visits, or empowering them to make informed decisions about their health and eligibility," she said, along with Wood.

In addition, the system needs no onsite servers, and can read and return texts of medical information sent to it, requiring no models for training or machine learning experience from users. Another plus is privacy – no data is kept on the service or used for training.

You Must Be Logged In To Post A Comment