Using nlp-insights with QuickUMLS

QuickUMLS is a service designed for fast, unsupervised concept extraction from medical text. The code base and documentation is located here. Another great article on the technology can be found here.

The nlp-insights service has been designed to interact with QuickUMLS for detecting medical concepts within FHIR resources.

Prereqs

• You must have access to a deployed QuickUMLS service to complete this tutorial. Instructions to start a server on your local machine are described here:
• You must have a container runtime installed on your machine
• You must have a python 3.9 and pip distribution
• This tutorial uses curl to submit REST requests to the service

Start the nlp-insights service

If the nlp-insights service has not been started, start the service in a local container by following the instructions here.

Configure nlp-insights to use quickumls for NLP

The nlp-insights service must be configured to use QuickUMLS prior to using the service to obtain insights. The steps to configure the service are described here.

Enrich FHIR resources with additional codings

The nlp-insights service can use QuickUMLS to derive additional coding values in FHIR resources. Learn how here

Derive new FHIR resources from unstructured content

The nlp-insights service can use QuickUMLS to derive new FHIR resources from clinical notes embedded in other FHIR resources. Learn how here

FHIR Integration

The nlp-insights service is designed to enrich a bundle prior to posting that bundle to a FHIR server. Learn how to to work with derived data that is stored in a FHIR server here You will need a viewer for jupyter-notebooks to view the tutorial.