A de-identification tool for removing personal identifiable information from clinical free text narratives. The current version include the knowledge-driven tool for processing clinical letters.

Available NERs:

• AGE,
• CITY-TVD,
• COUNTRY,
• COUNTY,
• DOB,
• FAX,
• HOSPITAL,
• ID,
• MEDICALSTAFF,
• PATIENT,
• PHONE,
• POSTCODE,
• PROFESSION,
• STREET,
• EMAIL, and
• URL.

The ChristieDEID is a simple command line tool. The source code should be straight forward to disinsect and integrate (see Controller.java) into your own application.

1. Build maven project: mvn package
2. Basic usage:

Use case 1: Print help screen. java -jar target/christie-deid-0.1.jar -h

Use case 2: Process a set of plain text letters in directory inputdir/ and save results in outputdir/. The output should include de-identified text. java -jar target/cDeid-0.1.jar --letters --dtext inputdir/ outputdir/