AUTOMATIC IDENTIFICATION AND EXTRACTION OF MEDICAL CONDITIONS AND EVIDENCES FROM ELECTRONIC HEALTH RECORDS

摘要

This document describes systems, methods, devices, and other techniques forautomatically identifying and extracting medical conditions and supportingevidences fromelectronic health records. In some implementations, formatted text extractedfrom anunstructured electronic health record is obtained. The formatted text issegmented into multipledocuments, wherein each document comprises a respective document type andrepresents arespective document encounter. Medical condition entities and supportingevidence entitiesreferenced in each of the multiple documents are extracted. Extractedsupporting evidenceentities within a same document are linked to respective extracted medicalcondition entitiesfrom the same document using one or more of i) medical ontologies, or ii) amedical knowledgebase. Output data representing linked supporting evidence entities and medicalcondition entitieswithin a same document is provided.