Sifting through and assessing medical records is undeniably time-consuming, a task that limits the potential of administrators, doctors and patients to not only make connections but also completely grasp data.
Nonetheless, Amazon, a globally recognized e-commerce giant, recently revealed that it intends to commence the selling of software that will sort through medical records to get information that can be utilized in not only improving patient treatment but also minimizing costs.
Amazon is hoping to simplify the sorting through and analysis of medical records with Amazon Comprehend Medical, its novel machine learning service.
The service leverages machine language processing in decoding the information contained in unstructured writing like doctor’s notes or even medical records.
Recently rolled out during Amazon Web Services’ re: Invent conference, the new machine learning service (Amazon Comprehend Medical) has already conducted tests with its partners including leading pharmaceutical company Roche and Seattle-based Fred Hutchinson Cancer Research Center.
Recent reports from The Wall Street Journal revealed that Amazon Web Services has been selling the same text analysis software, particularly for use in supply-chain management, travel booking, and customer support.
“The service will ‘read’ the text and then identify and return the medical information contained within it,” said Amazon in a recent blog post.
A blog post by AWS said that the program could “identify medical conditions, anatomic terms, medications, details of medical tests, treatments and procedures” in numerous documents.
A user keys in the unstructured into Amazon Comprehend Medical that, in turn, “reads” and mines relevant patient health information as well as medical terms.
At Fred Hutch, Amazon Comprehend Medical was utilized in identifying patients for clinical trials who can benefit from various therapies.
According to Fred Hutch’s Chief Information Officer Matthew Trunnell, claimed that the program reduced the processing time to minutes from hours.
Machine learning is an element of artificial intelligence that is loosely modeled, specifically on the brain.
For example, after training a neural network using real-world data, it can learn to recognize a cat, transcribe spoken words into text or identify a spam email.
“Time is the most valuable resource in cancer,” Fred Hutch told GeekWire. “Fred Hutch is utilizing technology that allows us to spend more time doing breakthrough research. Eliminating cancer and related diseases requires an open and collaborative approach across industries and disciplines which is why Fred Hutch is partnering with companies like Amazon Web Services to increase our research capabilities and gather insights that allow us to get cures to patients faster.”
Although the unstructured medical text has been a hindrance for most algorithms, Amazon’s software utilizes machine learning in identifying the ways that physicians make their notes.
According to Amazon, the software has the potential to mine data such as dosages, treatments, and diagnoses, amongst other essential data points.
The e-commerce giant also added that the software is expected to better address both privacy and protected health information (PHI) requirements.
“No data processed by the service is stored or used for training,” said Amazon.
Amazon’s recent move marks an exciting step for the company as it ventures into the healthcare space.