Amazon Comprehend Medical Makes Sense of Scattered Healthcare Data

Amazon . com on Tuesday announced Amazon . com Comprehend Medical, a language processing service that lets users gather information -- like a patient's medical problem and medicine dosage, strength and frequency -- from sources including doctors' notes, medical trial reports, hospital admission notes and patient health records.



Most health insurance and patient data -- for example medical notes, prescriptions, audio interview transcripts, and pathology and radiology reports -- presently is stored as unstructured medical text. Identifying these details either requires skilled medical coders to deal with data entry, or groups of developers to create custom code and rules to extract it instantly.

salesforce commerce cloud

Blessing and Curse

Healthcare providers and insurers presently need to write and keep some customized medical coding classification rules for natural language software. A big change one classification code name may have a ripple impact on a large number of hard-coded rules, which then should be updated, or data is going to be missed or classified incorrectly.

Amazon . com Comprehend Medical lets users create mixers reliably appreciate this medical information, vastly improving medical coding.

"The quantity of regulation and regulatory needs in prescription medication is both a blessing along with a curse," noted Ray Wang, principal analyst at Constellation Research.

"The curse is handling the disparate teams of information that aren't always complete or joined correctly," he told TechNewsWorld. "The blessing is, these details may lead to numerous insights for organizations to organize the following best action. Including identifying medical coding errors, potential dangerous drug interactions, epidemiological patterns, possibilities to avoid new illnesses, and billing malpractice and quality assurance."

Data interoperability lengthy is a problem for that medical industry. Amazon . com, Alphabet, IBM, Microsoft, Oracle and Salesforce this summer time announced support for that Fast Healthcare Interoperability Foundation (FHIR), which aims to produce a common group of standards for that exchange of healthcare information .

Automating the procedure

Amazon . com Comprehend Medical lets developers find out the key common kinds of medical information instantly, rich in precision, and without getting to create plenty of custom rules. It may identify health conditions, anatomic terms, medications, and information on tests, treatments and operations.

"The opportunity to use Comprehend to extract and classify unstructured medical information is significant, since it also enables private information to become removed," stated Rebecca Wettemann, VP of research at Nucleus Research.

"Researchers and physicians can share and evaluate data sets which were not practical to evaluate as unstructured patient-specific notes," she told TechNewsWorld. "More information is better when understanding diagnoses, trying new drugs or treatments, or evaluating existing treatments."

Amazon . com Comprehend Medical uses advanced machine learning models. It provides two APIs developers can integrate into existing workflows and applications having a couple of lines of code. The APIs are Medical Named Entity and Relationship Extraction (NERe) and guarded Health Information Data Extraction and Identification (PHId).

Partners range from the following organizations:

  • The Fred Hutchinson Cancer Research Center
  • Roche Diagnostics
  • Pricewaterhouse Coopers
  • Deloitte (ConvergeHEALTH)


Data Security and privacy

Amazon . com Comprehend Medical is HIPAA-qualified. It rapidly can identify protected health information (PHI) for example name, age and permanent medical record number, so you can use it to produce applications that safely process, maintain and transmit PHI, based on Amazon . com. It adheres to General Data Protection Regulation (GDPR) standards.

The services are covered under Amazon . com Web Services' HIPAA eligibility and BAA.

Developers can use data security and privacy solutions by removing after which identifying relevant patient identifiers as described in HIPAA's Safe Harbor approach to de-identification. Further, the service doesn't store or save any customer data.

However, data security is one central consideration, noted Michael Jude, program manager at Stratecast/Frost & Sullivan.

"Data moving is way less secure than data resting," he told TechNewsWorld. "To use these Amazon . com capabiities, data should be moved. Amazon . com need to take pains to guarantee the security associated with a patient data they get access to."

Exactly what the Service Will Definitely Cost

Amazon . com Comprehend Medical expenditure is levied per block of 100 figures of examined text.

NERe API users pay one (US) cent per block and PHId API users pay 14 cents per block. Users only pay for which they will use, and you will find no minimum charges or upfront commitments.

Amazon . com Comprehend Medical provides a free tier for 25K units of text, or 25,000 figures, for that first three several weeks when customers use both APIs.