Protected health information that excludes 16 specific direct identifiers but may contain dates and geographic information for permitted purposes.
A limited data set (LDS) is a category of protected health information from which 16 specific direct identifiers have been removed, but which may still contain certain indirect identifiers such as dates of service, dates of birth, and geographic information at the city or zip code level. Unlike fully de-identified data, which is not considered PHI at all, limited data sets remain protected health information and are subject to HIPAA requirements. However, they can be shared without individual authorization for specific permitted purposes, making them a valuable middle ground between fully identified PHI and completely de-identified data.
Limited data sets may only be used or disclosed for three purposes: research, public health activities, and health care operations. Before sharing a limited data set, the covered entity must enter into a data use agreement (DUA) with the recipient. The DUA must establish the permitted uses and disclosures, specify who is permitted to use or receive the data, require the recipient to use appropriate safeguards to prevent misuse, require the recipient to report any known violations, ensure the recipient does not attempt to identify or contact individuals, and require the recipient to extend these restrictions to any agents or subcontractors.
It is important to distinguish between limited data sets and de-identified information. De-identified information has had all 18 HIPAA identifiers removed (or has been certified by an expert as having very low re-identification risk) and is no longer considered PHI, meaning it can be used without restriction. Limited data sets retain some identifying elements, particularly dates and geographic data, which makes them more useful for research and analysis but also means they remain subject to HIPAA protections. Organizations should use the least identifiable form of data that meets their needs: de-identified data when dates and geography are not required, and limited data sets when these elements are essential for the intended purpose.