COMPARING HOSPITALS FROM DEPERSONALIZED HEALTH DATABASES WITHOUT APPARENT QUASI-IDENTIFIERS

摘要

FIELD: data processing devices of various types.;SUBSTANCE: method involves combining two or more depersonalized health databases (X, Y). Each depersonalized health database has personal data which are not identified, including by replacing medical institutions with aggregates for medical institutions. When combining databases, statistical distribution of patient characteristics is calculated for aggregates for medical institutions in anonymous health databases. Fillers for medical institutions in different impersonal health databases are compared by matching the corresponding statistical distribution of patient characteristics for the respective aggregates for medical institutions. Patients in different depersonalized health databases are compared by comparing the matched pairs of aggregates for the respective medical institutions, which increases efficiency of calculations. Forming a database matching table (16) identifying the matched patients, or forming a combined depersonalized health database, which combines patient characteristics in different impersonal health databases for each mapped patient into a single patient record taking into account medical facilities.;EFFECT: technical result is faster processing and security of data.;20 cl, 4 dwg