Written and maintained by CASRAI Editorial Board
Published
On this page
IHE XDS.b (Cross-Enterprise Document Sharing) is a foundational IT Infrastructure (ITI) integration profile specified by Integrating the Healthcare Enterprise (IHE) for registering, indexing, discovering, and retrieving longitudinal clinical documents across unaffiliated healthcare organizations, regional health information exchanges (HIEs), and multi-center clinical research networks.
Unlike point-to-point messaging where a sender must know every downstream recipient in advance, XDS.b establishes a federated, patient-centric architecture that strictly separates searchable document metadata (maintained in a central Document Registry) from the actual clinical document byte streams (stored across distributed Document Repositories).
What Is IHE XDS.b and How Is It Content-Neutral?
Introduced to supersede the earlier XDS.a profile by adopting OASIS ebXML Registry 3.0 together with SOAP 1.2 and MTOM/XOP binary attachment transport, IHE XDS.b operates within a shared governance boundary called an XDS Affinity Domain. Participating institutions in an Affinity Domain agree on a common patient identity domain, security policy, and master value sets for document classification.
Crucially, XDS.b is content-neutral: while health information exchanges most commonly populate XDS.b repositories with structured HL7 C-CDA continuity-of-care summaries, CDA Release 2 reports, or PDF clinical notes, the transport and metadata layer can index any MIME-typed payload, including DICOM Kos (Key Object Selection) manifests inside the radiological extension profile XDS-I.b (Cross-Enterprise Document Sharing for Imaging).
How Do Document Source, Repository, Registry, and Consumer Actors Interact?
The XDS.b technical framework coordinates five distinct system actors through numbered IHE ITI transactions:
| IHE Transaction | Initiating Actor | Responding Actor | Protocol & Operational Role |
|---|---|---|---|
| ITI-41 (Provide and Register Document Set-b) | Document Source | Document Repository | SOAP 1.2 + MTOM/XOP; transmits both the clinical document byte payload and the ebXML submission metadata set. |
| ITI-42 (Register Document Set-b) | Document Repository | Document Registry | SOAP 1.2 + ebXML RS 3.0; forwards only the metadata plus the repository’s repositoryUniqueId, SHA-1 hash, and byte size. |
| ITI-18 (Registry Stored Query) | Document Consumer | Document Registry | Queries ebXML metadata by patient ID, classCode, typeCode, practiceSettingCode, and service time window. |
| ITI-43 (Retrieve Document Set) | Document Consumer | Document Repository | Retrieves the actual document payload directly from the holding repository using repositoryUniqueId and documentUniqueId. |
| ITI-8 / ITI-44 (Patient Identity Feed) | Patient Identity Source | Document Registry | HL7 v2 ADT or HL7 v3; synchronizes Affinity Domain patient identifiers and merges duplicate master patient index records. |
The Three Core ebXML Metadata Objects
Inside the XDS Document Registry, metadata is organized around three ebXML Registry Information Model (RIM) objects:
DocumentEntry: Represents an individual clinical document, storing standardized classification attributes (classCode,typeCode,formatCode,healthcareFacilityTypeCode,practiceSettingCode,confidentialityCode), patient identifier, service start/stop timestamps, cryptographic hash, and pointers to its host repository.SubmissionSet: An immutable audit envelope created during each ITI-41/ITI-42 publication event that groups one or moreDocumentEntryobjects with author, source institution, and submission timestamp metadata.Folder: An optional logical container used to group related documents across multiple submission events for a shared care episode or clinical study protocol.
Document lifecycle updates are managed without overwriting historical files in place: a Document Source submits a new DocumentEntry connected to the prior entry via an ebXML association of type RPLC (Replaces), APND (Appends), or XFRM (Transforms), causing the Registry to transition the older entry’s availability status from Approved to Deprecated.
How Does IHE MHD Bridge XDS.b to HL7 FHIR?
Because parsing SOAP 1.2 envelopes, MTOM multipart streams, and verbose ebXML RIM XML slots imposes heavy complexity on lightweight web applications, mobile health tools, and modern research data pipelines, IHE developed two major companion profiles:
- IHE XCA (Cross-Community Access): Extends XDS.b across multiple independent Affinity Domains via Responding and Initiating Gateways (transactions ITI-38 Cross-Gateway Query and ITI-39 Cross-Gateway Retrieve), enabling national federated queries such as the eHealth Exchange and Carequality networks.
- IHE MHD (Mobile access to Health Documents): Provides a RESTful HL7 FHIR API facade over an existing XDS.b or XCA infrastructure. In MHD, an XDS
DocumentEntrymaps directly to a FHIRDocumentReferenceresource, an XDSSubmissionSetorFoldermaps to a FHIRListresource, and the clinical document payload is exchanged as a FHIRBinaryresource.
With an MHD Document Responder sitting in front of an XDS.b Affinity Domain, a research data pipeline or EHR client can execute standard FHIR REST searches (ITI-67 Find Document References) and HTTP GET retrievals (ITI-68 Retrieve Document) without constructing SOAP or ebXML payloads, and can subsequently pass retrieved C-CDA XML files into a C-CDA to FHIR transformation pipeline.
Frequently Asked Questions
What is the difference between a Document Registry and a Document Repository in IHE XDS.b?
A Document Repository stores the actual clinical document files (such as C-CDA XML or PDF byte streams) and computes their size and SHA-1 integrity hash, but does not support multi-parameter metadata search. A Document Registry stores zero clinical document files; instead, it indexes standardized ebXML DocumentEntry metadata from one or more Repositories so Document Consumers can search across the entire Affinity Domain before retrieving selected files from their respective Repositories.
How does IHE XDS-I.b handle large medical imaging studies without duplicating DICOM files?
Rather than copying multi-gigabyte CT or MRI pixel datasets into a general XDS Document Repository, IHE XDS-I.b publishes a compact DICOM Key Object Selection (KOS) manifest into the XDS Repository/Registry. The KOS manifest lists the Study, Series, and SOP Instance UIDs along with WADO-RS or DICOMweb retrieval endpoints pointing back to the originating PACS or vendor-neutral archive.
How are patient identifiers reconciled across institutions in an XDS Affinity Domain?
An XDS Document Registry accepts only patient identifiers registered within the Affinity Domain’s assigning authority via the Patient Identity Feed (ITI-8/ITI-44), typically backed by an enterprise Master Patient Index (eMPI) or IHE PIX/PDQ (Patient Identifier Cross-Referencing / Patient Demographics Query) services that link local medical record numbers to a shared cross-enterprise identifier.







