Series on Exploiting the Intersection of NLP, Phenotyping, and Ontologies in Electronic Health Records (Ongoing)

Posted On 2020-12-31 14:37:00

The Series on “Exploiting the Intersection of NLP, Phenotyping, and Ontologies in Electronic Health Records” is edited by Dr. Daniel B. Hier from University of Illinois at Chicago, IL, USA.

Daniel B. Hier, MD, MBA
Department of Computer and Electrical Engineering, Missouri University of Science and Technology and Department of Neurology and Rehabilitation, University of Illinois at Chicago, IL, USA

Daniel B. Hier is neurologist with interests in stroke, dementia, machine learning, medical ontologies, and electronic health records. He holds an MD degree from Harvard Medical School and MBA from University of Illinois at Chicago. He completed a neurology residency and fellowship at Massachusetts General Hospital. He is emeritus Professor of Neurology at University of Illinois at Chicago and adjunct Professor of Electrical and Computer Engineering at Missouri University of Science and Technology. His current interests include machine learning in neurology and applications of ontologies to electronic health records.

The main purpose of this series:
Electronic health records (EHRS) hold patient data.
Although some patient data is structured and computable (such as laboratory data), many patient data (signs, symptoms, imaging findings, electrophysiological findings, nursing observations, etc.) are held as free text and are neither searchable nor computable. In order to unlock the value of patient data in electronic health records, free text and narrative must be converted into a computable form.

Natural language processing (NLP) and ontologies provide a mechanism to convert free text into clinical concepts (each with a machine-readable code). Patients can then be converted to phenotypes (collections of signs and symptoms that have been mapped to clinical concepts from ontologies). Algorithms are needed to calculate distances between patients and match patients to canonical phenotypes in large phenotype repositories.

This special series of the Journal of Hospital Management and Health Policy (JHMHP) is devoted to exploring how clinical concepts in ontologies combined with improvements in natural language processing can be used to unlock patient data held in electronic health records by making this data more computable. Improved patient similarity calculations and improved patient phenotyping are waystations on the road to improved precision medicine.

Series outline:

  1. The Intersection of NLP, phenotping, and electronic health records
  2. Simplifying documentation in EHRs with medical ontologies
  3. Using NLP to extract information from EHRs to support clinical decision support
  4. Ontology Guided Extraction of Clinical Concepts from Medical Texts
  5. Strategies for using NLP to extract clinical concepts from EHRs
  6. Deriving Patient Similarity Networks from EHRs
  7. Measuring Phenotype similarity in Electronic Health Records
  8. Finding phenotypes in electronic health records
  9. Current Status: Phenome Wide Association Studies (PheWAS)
  10. Deriving information from EHRs by NLP to improve management of chronic disease
  11. Approaches to Identifying Cohorts of Phenotypes in EHRs
  12. Using Ontologies to Encode Phenotypes in EHRs
  13. Update on the Human Phenotype Ontology
  14. Recognizing clinical concepts in clinical narratives
  15. Advances in Patient Similarity Algorithms for Electronic Health Records

The series “Exploiting the Intersection of NLP, Phenotyping, and Ontologies in Electronic Health Records” was commissioned by the editorial office, Journal of Hospital Management and Health Policy without any sponsorship or funding. Daniel B. Hier is serving as the unpaid Guest Editor for the series.