HealthcareITNews
Linguamatics, RealHealthData to mine patient data with natural language processing
By: Mike Miliard
Linguamatics, whose AI and machine learning algorithms fuel natural language processing technology for text mining, is joining forces with RealHealthData, which works with medical transcription companies to populate its database of narrative medical records.
Cambridge, UK-based Linguamatics' I2E technology is able to pore through a variety of text resources, such as EHR data and clinical trials information. Santa Cruz, California-based RealHealthData maintains a detailed database of provider notes. Together they'll use NLP to better understand an array of clinical data.
One aim is to give pharmaceutical and biotech companies better insight into the real-world impact of patient therapies, rather than having to rely clinical trial data, according to the companies.
"We believe this partnership will enhance the value we can provide our life science customers for health economics and outcomes research, epidemiology, and medical affairs,” said Jane Reed, head of life science strategy at Linguamatics.
Medical records offer a key trove of such data, which could can inform drug development and beyond, but it's often unstructured. Linguamatics I2E can extract key facts from the narratives in RealHealthData's database, which covers every medical specialty, by using specific ontologies and queries for better decision making.
Unstructured EHR text offers a level of detail that's not available in the structured fields that life science companies are used to. RealHealthData's trove of patient records include narrative information such as patient social status and clinical notes such as comorbidities, complications, co-medications, lab values, adherence or switching issues.
Manuel Prado, CEO of RealHealthData, said customers of the companies "can now access the unique and valuable insights in the database using a first-in-class, healthcare-specific natural language processing platform."
I2E technology can mine large amounts of unstructured data – incorporating machine learning to directly specific patients, such as diabetics who smoke and are overweight, said David Milward, chief technology of Linguamatics, using longitudinal data to look at outcomes or behavior over time.
