Research Question: (RQ09):
Can we identify vulnerable groups from users of the UK’s largest mental health service provider?
Robert Stewart, Honghan Wu; South London and Maudsley; HDR UK Scotland, Edinburgh
Research Question Narrative:
Use natural language processing to identify patients with vulnerable physical underline conditions on near real time SLaM EHR data.
Intended Research Outcomes:
An automated pipeline to identify most vulnerable people from EHRs.
SAGE Priority Area:
Direct Impact; Other conditions
Progress So Far:
33 phenotypes have been defined including various autoimmune disorders. A set of multi-morbidity NLP models have been adapted to run on the whole set of free-text SLaM EHRs.
PPIE Activities So Far:
hub for data sharing and interoperability
HDR UK community members with skill & will to contribute:
Original content: https://web.www.healthdatagateway.org/project/190009