De-Identification, Anonymization, Redaction Toolkit Services (DARTS) Implementation Guide

This project will examine the use of a services to de-identify FHIR data to receive line-level (patient level) data based on applicable regulations, program requirements, and policies. The generation and exchange of FHIR data continues to grow exponentially, driven by the increasing use of data for various purposes such as reporting to government agencies, sharing data with research, and training artificial intelligence (AI) models.

There are use cases when there is a need to provide data without protected health information (PHI) or personally identifiable information (PII) to meet federal, state, and local policies and jurisdictional policies and regulations that may restrict PHI and PII.

To share line-level de-identified data, a set of services are needed which can consume identifiable FHIR data and provide de-identified or anonymized data. This set of services will be called DART services. DART services that are currently being considered are a de-identification service and an anonymization service. These DART services will transform FHIR data containing PHI/PII to FHIR data that does not have PHI/PII.

To achieve these objectives, we propose DeIdentifcation, Anonymization, Redaction Toolkit Services (DARTS) Implemenation Guide. that will define a set of services for de-identification and anonymization that accept FHIR data as an input along with one or more policy designators to either transform or remove PHI/PII data.

Policy designator examples include "HHS Safe Harbor Rules," "General Data Protection Regulation (GDPR)," or specific research institute policies.

We intend the IG to be balloted in the May 2026 cycle.

This project reuses work performed under the UDS+ IG (https://www.fhir.org/guides/hrsa/uds-plus/) and the bulk import work stream in Argonaut.