Only registered and logged in users can download files from this category.
Computable Phenotype Lecture
This recorded lecture discusses the art of creating a "computable phenotype," or electronic health record data-driven case definition. The content is a high-level overview of the concept, and is intended for clinical researchers who may not be informaticians themselves, but wish to work with informaticians or analysts to define patient cohorts electronically.
CLARK! is a machine-learning classifier developed through a collaboration between NC TraCS and CoVar Applied Technologies, a Durham, NC-based machine learning company. CLARK! is designed to take clinical notes as input, and classify those notes based on features defined by the user as regular expressions. To try CLARK! for yourself, download all files in this directory, and open the README file for a brief overview on getting started. Fuller documentation of CLARK! (in addition to a user-friendly manual) will be available in the near future, but in the meantime, reach out to epfaff at email dot unc dot edu with any questions.