Abstract: We analyzed a year's worth of patient admissions at our institutions to determine the risk of re-identification based
on unique combinations of ICD-9 and ICD-10 diagnosis codes in a de-identified data set. More than 70% of patients
were uniquely identifiable. The percentage approached 100% in certain subpopulations. Our findings highlight the
risk of re-identification and the factors influencing it.
Learning Objective 1: Recognize patients at risk of re-identification based on their unique combination of ICD-9 and ICD-10 codes
Elizabeth Mulvey (Presenter)
Medical University of South Carolina
Jihad Obeid, Medical University of South Carolina