De-ID Software

De-ID software uses advanced heuristics and ‘machine reading’ to understand lexical and syntactic context to identify PHI from non-PHI in narrative and create a separate, de-identified version of a note or free text field.  De-ID not only removes PHI, it also replaces it with proxies (e.g., NAME; PLACE; offset dates) that retain the narrative integrity of the patients story.  The software understands words in context, and can differentiate between a when nouns are used as clinical terms (e.g., Hodgkins Disease; not PHI) and as names (e.g., Dr. Hodgkins; PHI) and permits “localization" of inclusion and exclusion rules.

De-ID can be set to automatically remove all 18 HIPAA Safe Harbor elements of PHI, or can be configured to process data based on a Limited Data Set of one or more of the elements.  In addition, there are software features that support patient identification via an Honest Broker, a feature than has been used for clinical trial recruitment, and also supports clustering of de-identified records associated with the same patient.

Originally developed and tested at the University of Pittsburgh (ref 4), De-ID software has been used by the Federal government, major academic medical centers and health systems, transcription and computer-assisted coding companies and consulting companies to remove “protected health information” (PHI) from discharge summaries, pathology reports, surgical/procedure and radiology notes and free text fields in electronic health record systems under HIPAA compliance programs utilizing both the Safe Harbor and ‘expert determination’ methods. (Ref 1, 2, 3)

1.   Danciu, I et al Secondary use of clinical data: The Vanderbilt approach. J Biomed Inform. 2014 Feb 14
2.   Crowley RS et al, caTIES: a grid based system for coding and retrieval of surgical pathology reports and tissue specimens in support of translational research
.  J Am Med Inform  Assoc 2010;17:253-264
3.   Roden MD et al   Development of a Large-Scale De-Identified DNA Biobank to Enable Personalized Medicine 
Clinical Pharmacology & Therapeutics 2008; 84:362-369  
4.   Gupta D, Saul M, Gilbertson M. Evaluation of a Deidentification (De-Id) Software Engine to Share Pathology Reports and Clinical Documents
for Research Am J Clin Pathol 2004;121:176-186


© 2014 De-ID Data Corp, LLC