SMART SUPPRESSION USING RE-IDENTIFICATION RISK MEASUREMENT

摘要

System and method to produce an anonymized cohort, members of the cohort having less than a predetermined risk of re-identification. The method includes receiving a data query of requested traits to request in an anonymized cohort, querying a data source to find records that possess at least some of the traits, forming a dataset from at least some of the records, and calculating an anonymity histogram of the dataset. For each patient record within the dataset, the method anonymizes the dataset by calculating using a threshold selector whether a predetermined patient profile within the dataset should be perturbed, calculating using a value selector whether a value within the indicated patient profile should be perturbed, and suppressing an indicated value within the indicated patient profile. The anonymized dataset then is returned.

主权项

1. A method to produce an anonymized cohort, members of the cohort having less than a predetermined risk of re-identification, comprising:
receiving a data query via a user-facing communication channel to request an anonymized cohort, the data query comprising requested traits to include in members of the cohort; querying a data source, using a data query transmitted via a data source-facing communication channel, to find records that possess at least some of the traits; forming a dataset from at least some of the records; calculating, by a processor coupled to the user-facing communication channel and the data source-facing communication channel, an anonymity histogram of the dataset; for each patient record within the dataset, anonymizing the dataset by performing the steps of:
calculating, by the processor using a threshold selector, whether a patient profile within the dataset should be perturbed;calculating, by the processor using a value selector, whether a value within the indicated patient profile should be perturbed; andsuppressing, by the processor, an indicated value within the indicated patient profile; and providing, via a user-facing communication channel, the anonymized dataset.