String Identifier in Multiple Medical Databases

In a distributed medical system, building cross-site records while maintaining appropriate patients anonymity is essential. The distributed databases contain information about the same individuals, often described by using the same variables, which do not fit quite frequently due to accidental distortions. In such cases, the record linkage methods are used to find records that correspond to the same individuals in order to create a consistent database. Our goal was to find a solution for this problem. In this paper, we propose an anonymous identifier, based on combinations of first two letters from the surname, name, date of birth and gender, which can allow a deidentifying merged dataset from multiple databases of a distributed medical system.