Marijuana use is increasing in the U.S. due to increased legalization for medical use and may be used by patients with pain. The development and refinement of search terms is a necessary step preceding use of informatics tools to identify marijuana documentation in the free text of patient charts. We used low-level text processing tasks to compare terminology used in a different sample. Documentation of any positive-use term was present in 56% of pulled charts.

Learning Objective: Learn about the refinement of a preliminary lexicon used to develop Natural Language Processing-enabled identification of substance users from free text.


Termeh Feinberg (Presenter)
Yale University

Joseph Goulet, Yale University
Amy Justice, Yale University
Sameh Fodeh, Yale University
Lori Bastian, Yale University
Qing Zeng, George Washington University
Cynthia Brandt, Yale University

Presentation Materials: