Autonomy Health Care Platform Decodes Unstructured Data
Autonomy, a maker of enterprise information-management software, has developed a new platform that allows health professionals to search through the patterns of unstructured data in electronic health records, journals and textbooks.
With the Meaning-Based Healthcare software platform, announced on Nov. 15, Autonomy builds on its experience in archiving, compliance, eDiscovery, records management and private cloud computing, the company reports. The platform comprises two components: Auminence and Information Governance.
Using Bayesian Inference Technology, Auminence allows health care providers to make diagnoses based on contextual patterns within the data, according to Dr. Joseph Britto, Autonomy's head of medical technologies. The application also runs on a smartphone.
Meanwhile, Information Governance helps health care companies archive patient records and manage the large sets of data they contain while complying with federal regulations.
The software, which handles 106 languages, is available as an on-premise client, in the cloud or as a combination of the two, Britto told eWEEK.
With the application, Autonomy hopes to improve the accuracy of medical diagnoses and maintain security while lowering health care costs.
Meaning-based computing allows health care professionals to search through patient data without using a specific keyword, phrase or exact question. The technology uses probabilistic modeling and matches patterns to create a context for the unstructured data.
"Our real strength is finding the pattern within the data, whether it be textbooks or journals or whether it be the mass communication of e-mail," Britto said. When you perform a search using meaning-based computing, the software will call up e-mails or documents based on the concept rather than an exact word, Britto explained.
The software processes unstructured data such as e-mail, images and voice recordings. "It allows you to search by meaning and concept, up and down a clinical health record-think about the power of searching a group of electronic health records looking for patterns," Britto said.
By recognizing concepts and patterns in this data, the Autonomy software can create meaning and a checklist for the medical professional to explore, based on statistics, along with data from EMRs.
If you search for "fever" or "pyrexia," raised temperature may appear in results, Britto noted. "What you would get back are documents that are conceptually related," he said.
In another example, a search of "man," "dog" and "room" could return the conceptually related words "boy," "puppy" and "corridor," Britto said. "All of these are different descriptions of the same concept," he explained.
Autonomy has also applied these techniques in the financial, legal, pharmaceutical and retail fields.
Being able to process both structured and unstructured data in EMRs is key for health care IT according to one Autonomy customer.
"Health care providers that are able to better leverage unstructured and structured information stored within their health care applications will improve quality of care, reduce costs and treat far greater numbers of patients," Brian Harris, executive director for enterprise solutions at Loma Linda University Medical Center, in Loma Linda, Calif., said in a statement.
Earlier this year Autonomy also introduced Evolve, a document-management application for hospitals.
Although people hope EHRs can lead to better care, transferring the data into knowledge that's usable is their key benefit, Dr. Stephen Borowitz of the University of Virginia Medical Center said in a statement. "To do this, we need technologies that help us identify and characterize currently unseen patterns and meaning in the sea of data and help us transform those patterns into usable knowledge."
With meaning-based computing now available, Britto calls for a shift in the way people will search through medical records.
"Advanced visualization and analytics allow you to make sense of the unknown," Britto said. "In the increase in data, I argue that the next generation of search has to go beyond the keyword Boolean search engine."