Continuing its aggressive push into healthcare, Amazon announced this week that it’s launching a machine learning service that will aim to mine data from electronic health records (EHRs).
According to company officials in a blog post earlier this week, the software, Amazon Comprehend Medical, is “a 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, and more.”
They continued, “Comprehend Medical helps healthcare providers, insurers, researchers, and clinical trial investigators as well as healthcare IT, biotech, and pharmaceutical companies to improve clinical decision support, streamline revenue cycle and clinical trials management, and better address data privacy and protected health information (PHI) requirements.”
As Amazon officials noted, a core issue in healthcare and health IT today is that a large amount of critical data is stored as unstructured medical text, such as medical notes, prescriptions, audio interview transcripts, and pathology and radiology reports. “This means that being able to identify this information can be 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,” Amazon officials outlined. “In both cases this undifferentiated heavy lifting takes material resources away from efforts to improve patient outcomes through technology,” they added.
But what Amazon Comprehend Medical will aim to specifically allow developers “to identify the key common types of medical information automatically, with high accuracy, and without the need for large numbers of custom rules. Comprehend Medical can identify medical conditions, anatomic terms, medications, details of medical tests, treatments and procedures. 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,” according to officials.
Indeed, Amazon executives are quite bullish on the machine learning capabilities of the new software. Taha Kass-Hout, a senior leader with Amazon’s healthcare and AI efforts told the Wall Street Journal that “During testing, the software performed on par or better than other published efforts, and can extract data on patients’ diseases, prescriptions, lab orders and procedures,” according to a report this week.
Officials noted that since there will be no servers to provision or manage, developers will only need to provide unstructured medical text to Comprehend Medical. The service will “read” the text and then identify and return the medical information contained within it, they explained.
Importantly, as pointed out in a CNBC report on the Amazon initiative, Christina Farr noted that “It looks unlikely at this point that Amazon will compete directly with medical records companies like Allscripts and Cerner, as there are plenty of money-making opportunities to work with those vendors and provide services like population health and clinical trial support.”
Rather, Farr continued, “Amazon is most directly taking on UnitedHealth Group's Optum, which is already in the space, as well as technology rivals Apple and Alphabet.” To this point, according to the WSJ report, “The market for storing and analyzing health information is worth more than $7 billion a year, according to research firm Grand View Research, a business in which International Business Machines Corp.’s Watson Health and UnitedHealth Group Inc.’s Optum already compete.
Amazon Pushes Further into Healthcare
It’s been quite the year of convergence for Amazon and the healthcare industry. In January, Amazon, Berkshire Hathaway, and JPMorgan Chase & Co announced they were teaming up on an initiative to improve satisfaction and reduce costs for their companies’ employees. Although not many details are known about this collaboration, the organizations named Atul Gawande, M.D., as CEO of the initiative, back in June.
Meanwhile, in August, Amazon said it would be part of another endeavor related to healthcare—to remove interoperability barriers and to make progress on adoption of health data standards. For this, Amazon will be teaming up with Microsoft, Google, IBM, and others to jointly commit to support healthcare interoperability by advancing healthcare standards such as HL7 (Health Level Seven International), FHIR (Fast Healthcare Interoperability Resources), and the Argonaut Project.
Also this summer, Amazon acquired PillPack, a Boston-based online pharmacy startup.
Liam Bouchier, principal with the Illinois-based consulting firm Impact Advisors, says that this latest initiative is simply another example of Amazon “doing what it does best”—working to analyze large data sets as a means to gain meaningful insights into the consumer through a variety of different ways. “This concept does make some in the healthcare industry uncomfortable, which is understandable given the amount of regulations, including HIPAA, that are a challenge to manage daily,” he says, importantly adding that EHRs were not the norm even 10 years ago. “Now, the [EHR] market is saturated with almost every health system and healthcare provider partaking,” he points out.
To this point, as explained in the WSJ article, “Amazon Web Services won’t see the data processed by its algorithms, which will be encrypted and unlocked by customers who have the key, Dr. Kass-Hout said. Its service is designed to conform with privacy rules laid out in the federal Health Insurance Portability and Accountability Act,” he said, per that report.
As Bouchier notes, in healthcare there are companies that provide this type of NLP (natural language processing) service to varying degrees of success, many of whom use Amazon Web Services as their infrastructure. “Is this really that big a leap for Amazon?” he asks. “The analytics market is full of vendors that provide a variety of analytics tools using both claims and clinical data primarily. The difference is these vendors started with healthcare in mind, yet few can scale and provide their services in a cost-effective manner to the customer and to the health system.”
But a key difference, he asserts, is that “they almost most assuredly don’t have the depth of knowledge and know-how to execute successfully in the same manner along with the depth of data and knowledge Amazon possess on the world population today.”
In the same announcement this week, Amazon also said that it was working closely with Seattle’s Fred Hutchinson Cancer Research Center in an effort to identify patients for clinical trials who may benefit from specific cancer therapies. Fred Hutch was able to evaluate millions of clinical notes to extract and index medical conditions, medications, and choice of cancer therapeutic options, reducing the time to process each document from hours, to seconds, according to officials.
“Curing cancer is, inherently, an issue of time,” Matthew Trunnell, CIO, Fred Hutchinson Cancer Research Center, said in a statement. “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.”
