Amazon Web Services Inc. today updated its natural language processing service for healthcare providers with a new feature that lets them link to information from medical ontologies.

AWS Comprehend Medical is a machine learning service that models topics, detects language, conducts sentiment analysis and extracts phrases from unstructured medical texts such as doctors’ notes, clinical trial reports and patient health records.

The service analyzes these texts in order to help medical professionals to get a better understanding of the complex relationships between things such as drug dosage and their impact on health conditions, for example. The new feature will help professionals to better detect medical information from unstructured clinical texts and then link those findings to the ICD-10-CM and RxNorm medical ontologies.

For the uninitiated, medical ontologies are meant to enable more efficient communication of information between the numerous different computer systems that healthcare providers use. Comprehend Medical helps with this because it can link detected entities, like symptoms and diagnoses, from medical records to specific concepts in ontologies.

Amazon said that linking AWS Comprehend Medical to the ICD-10-CM and RxNorm medical ontologies would improve medical outcomes and care, and also reduce the cost, time and effort involved in processing large amounts of unstructured medical text.

“Using Amazon Comprehend Medical ICD-10-CM and RXNorm Ontology Linking APIs, developers can quickly and accurately extract codes from a variety of data sources, such as doctor’s notes or patient health records,” AWS Technical Evangelist Danilo Poccia wrote in a blog post. “Our deep learning approach to ontology linking provides much higher accuracy than existing rules-based systems by understanding the context each entity is found in.”

Photo: SIM USA/Flickr