generating-synthetic-surrogates

To replace PHI with synthetic surrogates, you use a de-identification process that substitutes detected identifiers with realistic, type-matched fake values. This preserves clinical note readability and structure, making the data safe for NLP training and demonstrations.

Yes, you can generate consistent fake identities by applying consistency controls to the replacement process. This ensures that repeated names or identifiers within a document map to the same surrogate identity, maintaining logical coherence in clinical notes.

You create reproducible and localized surrogate data by configuring seed values and locale settings during the replacement process. This generates repeatable surrogates appropriate for different regions and languages in multilingual clinical text.

Yes, you can register custom generators and providers to create organization-specific formats for clinical identifiers. This allows you to define specific formatting rules for MRNs, account numbers, and addresses when replacing PHI.

The best way to de-identify clinical notes while keeping structure intact is using synthetic surrogates that match the original data type. This method replaces protected health information without disrupting the natural language structure and readability of the document.

To securely retain mappings for reversible pseudonymization, you must configure secure handling of optional mapping files during the surrogate generation process. This allows you to link synthetic identities back to original PHI if reversibility is required.