Atrius Health, one of the leading and longest-standing accountable care organizations in the country, will roll out natural language process technology from Linguamatics Health to help it extract hard-to-reach information hidden within unstructured clinical data.

Boston-based Atrius Health, which serves 740,000 patients in Massachusetts, implementing Linguamatics' enterprise NLP platform, which runs on I2E text mining technology, to make better use of patient records and help with its advance quality care initiatives – especially identifying at-risk patients to minimize care gaps.

"A vast amount of critical clinical data exists as unstructured text which is difficult to access and analyze," said Joe Kimura, MD, Atrius Health’s chief medical officer, in a statement. "We are leveraging the power of NLP to replace the manual, inefficient data extraction processes that many healthcare organizations struggle with, in order to advance our quality care initiatives more rapidly."

[Also: Athenahealth partners with NoteSwift on AI-powered EHR documentation]

Beyond closing gaps in care, he said, Linguamatics' technology can help Atrius enhance clinical documentation for chronic care management, reduce litigation risks and streamline Medicare ACO quality reporting.

Atrius Health plans to broaden its use of the I2E technology to help with other pop health and value-based care initiatives, such as identifying social determinants of health to improve available interventions supporting outreach for behavioral health, he said.

"By operationalizing our use of Linguamatics NLP we can more efficiently support our value-based care initiatives,” added Craig Monsen, MD, Atrius Health's medical director of analytics.

"Clinicians and patients benefit from more accurate problem lists for chronic disease without adding to existing documentation burden, the ACO reporting process can be simplified, and Medicare Advantage coding is more complete," he said.

For example, the ACO is ramping up its safety-net efforts because the new technology can help it detect those patients with specific conditions that are hard to track, such as pulmonary nodules, said Monsen: "By responding early to harbingers of lung cancer or other conditions for which we have safe, effective preventative interventions, we aim to improve our patient outcomes.”

Simon Beaulah, Linguamatics' senior director of healthcare, said the move toward value-based care requires deeper understanding of individual patients. "With 80 percent of EHR data in unstructured form, the use of NLP to fill in those gaps in the jigsaw puzzle is expanding rapidly,” he said.

Twitter: @MikeMiliardHITN
Email the writer: mike.miliard@himssmedia.com