Combining unsupervised, supervised and rule-based learning: the case of detecting patient allergies in electronic health records

Table 1 A subset of the document types and database datatypes stored in the hospital’s enterprise EHR system

Data formats	Oracle datatypes	Document type examples
RTF	CLOB	Discharge summaries, surgery notes, journal notes
PDF	CLOB	Laboratory results, anesthesia and intensive care journals
Formflow	CLOB	Specialized forms, surgical nurse reports, medical certificates
TIFF	BLOB	Scanned documents, clinical curve data
XML	XML	Physician referrals, CAVE/drug reactions
HTML	VARCHAR	Physician referrals
Proprietary/legacy data formats	VARCHAR	Specialized forms, pre-operative assessments and planning forms
Text	VARCHAR	Older documents from converted legacy EHRs

ISSN: 1472-6947