Amazon has launched a new machine learning platform capable of extracting relevant medical data from unstructured text.
Amazon Comprehend Medical can read clinical notes, prescriptions and audio transcripts and identify clinical information buried within them.
It does this through what Amazon terms medical language processing, a form of natural language processing that can “identify medical conditions, anatomic terms, medications, details of medical tests, treatments and procedures,” without the need to first train the system or apply “large numbers of custom rules”.
It is hoped Amazon Comprehend Medical will enable healthcare providers, suppliers and researchers make better sense of large swatches of unstructured medical data, ultimately helping support clinical decision-making, medical trials and population health initiatives.
“The majority of health and patient data is stored today as unstructured medical text, such as medical notes, prescriptions, audio interview transcripts, and pathology and radiology reports,” a post on Amazon’s machine learning blog read.
“Identifying this information today is a manual and time-consuming process, which either requires data entry by high skilled medical experts, or teams of developers writing custom code and rules to try and extract the information automatically.
“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.”
Amazon is working with the Fred Hutchinson Cancer Research Centre in Seattle, where Comprehend Medical is being used in medical trials to identify patients who may benefit from specific cancer therapies.
According to the tech firm, the platform has enabled the hospital to rapidly extract and index medical conditions, medications, and choice of cancer therapeutic options from millions of clinical notes.
Matthew Trunnell, CIO of Fred Hutchinson Cancer Research Center, said: “For cancer patients and the researchers dedicated to curing them, time is the limiting resource. The process of developing clinical trials and connecting them with the right patients requires research teams to sift through and label mountains of unstructured medical record data.
“Amazon Comprehend Medical will reduce this time burden from hours per record to seconds. This is a vital step toward getting researchers rapid access to the information they need when they need it so they can find actionable insights to advance lifesaving therapies for patients.”
Another AWS customer that has previewed the service is Roche Diagnostics.
Anish Kejariwal, the company’s director of software engineering for information solutions, said: “With petabytes of unstructured data being generated in hospital systems every day, our goal is to take this information and convert it into useful insights that can be efficiently accessed and understood
“Amazon Comprehend Medical provides the functionality to help us with quickly extracting and structuring information from medical documents, so that we can build a comprehensive, longitudinal view of patients, and enable both decision support and population analytics.”
17 December 2018 @ 08:48
Earlier this week, AWS also unveiled Amazon Comprehend Medical , new HIPAA-eligible machine learning tool, enabling developers to process unstructured medical text and spot specific data such as diagnosis, treatments, dosages, symptoms and more.
10 December 2018 @ 12:49
BTW Digital Health
Why does your software remove my carefully considered paragraphs breaks and white space from my comments – Making them less redabale?
10 December 2018 @ 12:47
i’m sure Natural Language Processing and Machine Learning have an important role to play in extracting valuable insights from unstructured health data, but I have some serious concerns.
Health data is unusually complex and the quality of health records is poor such that even a skilled reader can often not properly understand them.
the old adage ” rubbish in – rubbish out” remains a valid as ever and it would be a mistake to think NLP will extract meaning that’s not present in the source. The great danger is that it might make it look like it has.
If we want to harness the power of AI we still need to work to create properly structured records capable of conveying the meaning in the mind of the author to the reader (be that a person or an AI)
8 December 2018 @ 16:05
I hope it will understand the HPC when it’s been given in the local dialect of places such as Glasgow, Liverpool and Newcastle!