BMC Medical Research Methodology

Table 2 Types of PHI and other data detected by the de-identification systems

From: Automatic de-identification of textual documents in the electronic health record: a review of recent research

De-identification system	PHI							Clinical data
	Person names	Ages > 89	Geographical locations	Hospitals/HC org.	Dates	Contact information	IDs
Aramaki	P+D	✔	✔	✔	✔	✔	✔	None
Beckwith	P+D	✔	✔	✔	✔	✔	✔	None
Berman	✸	✸	✸	✸	✸	✸	✸	UMLS
Fielstein	P+D	-	✔	✔	✔	✔	✔	None
Friedlin	P+D	✔	✔	✔	✔	✔	✔	None
Gardner	P	✔	-	-	✔	-	✔	None
Guo	P+D	✔	✔	✔	✔	✔	✔	None
Gupta	P+D	✔	✔	✔	✔	✔	✔	None
Hara	P+D	✔	✔	✔	✔	✔	✔	None
Morrison	✸	✸	✸	✸	✸	✸	✸	MedLEE
Neamatullah	P+D	✔	✔	✔	✔	✔	✔	None
Ruch	P+D	-	-	✔	✔	✔	✔	MEDTAG
Sweeney	P+D	✔	✔	✔	✔	✔	✔	None
Szarvas	P+D	✔	✔	✔	✔	✔	✔	None
Taira	P	-	-	-	-	-	-	None
Thomas	P+D	-	-	-	-	-	-	None
Uzuner	P+D	-	✔	✔	✔	✔	✔	None
Wellner	P+D	✔	✔	✔	✔	✔	✔	None

✸ Only extracted concepts (i.e. UMLS or other clinical concepts) are retained.
P+D = Patient and healthcare provider names; P = Patient name

Back to article page

ISSN: 1471-2288

Contact us

General enquiries: journalsubmissions@springernature.com