HL7 history: from v2 (1987) to FHIR (2014–today)

HL7's history spans nearly 40 years and can be examined through four broad eras. From v2's delimiter-based messages, through the RIM-based v3 family and CDA, to FHIR resources and web-based exchange mechanisms, each era reflects a different design scope. This article also distinguishes evidence that is publicly traceable in Vietnam from areas where no nationwide adoption data are available.

Quick summary

  • 1987–1996: HL7 v2 emerges to support exchange among hospital systems; messages use delimiters and are commonly transported over MLLP.
  • 1996–2005: the v2 family continues to expand; HL7 launches the RIM in 1997 as the reference model for the v3 family.
  • 2005–2014: HL7 v3 and CDA Release 2 are published; v3 messaging and CDA have different deployment models, tooling, and scopes.
  • 2014–today: FHIR brings Resources, JSON/XML, and multiple exchange paradigms into one specification; the 2020 ONC rule applies an R4.0.1 requirement to a specific US API certification criterion.
  • Vietnam: no public dataset comprehensively describes HL7 v2 use before 2024; traceable milestones include the hl7vn/vn-core-ig repo in 2024 and the trial-use hl7.org.vn initiative in 2026.

On this page

  1. Context — what did US healthcare need in the 1980s?
  2. 1987–1996: HL7 v2 formation and early deployment
  3. 1996–2005: v2.x expands and the RIM is launched
  4. 2005–2014: v3 messaging and CDA Release 2
  5. 2014–today: FHIR and web-based exchange models
  6. Vietnam in the HL7 timeline
  7. Lessons for Vietnam
  8. References and further reading

1. Context — what did US healthcare need in the 1980s?

HL7 was founded in 1987 as hospitals needed to exchange data among hospital information, laboratory, and radiology systems. Without a shared specification, each pair of systems typically required its own data format and interface agreement.

In the 1980s, US health-information vendors used multiple proprietary messaging formats. Connecting a Hospital Information System (HIS) to a Laboratory Information System (LIS) or Radiology Information System (RIS) therefore depended heavily on custom interfaces and bilateral agreements.

The name "Health Level Seven" refers to layer 7 — the application layer — of the OSI model. HL7's scope focuses on exchange content and rules at that layer rather than specifying the entire underlying network stack.

2. 1987–1996: HL7 v2 formation and early deployment

HL7 was established in 1987. HL7 v2.0 was released in September 1988, and the organization became an ANSI-accredited Standards Development Organization in 1994.

Early v2 releases followed with v2.1 in March 1990, v2.2 in December 1994, and v2.3 in March 1997. Installed systems may use different v2 versions for ADT (Admission–Discharge–Transfer), laboratory-result, or medication-order flows. HL7 v2 uses characters such as pipes and carets as delimiters; MLLP over TCP is a commonly used transport mechanism, but it is not the only transport an implementation may use. 

MSH|^~\&|HIS|HOSP|LIS|LAB|202604301430||ORM^O01|MSG001|P|2.3
PID|1||MRN0001||NGUYEN^THI LAN||19850315|F

The delimiter syntax allows many v2 messages to be inspected directly. Production interoperability still depends on the version, implementation guide, code systems, optional fields, and local agreements; syntactic readability does not make two systems automatically compatible.

3. 1996–2005: v2.x expands and the RIM is launched

During this period, v2.3 (1997), v2.4 (2000), and v2.5 (2003) continued to expand the standard family. Later releases through v2.9 in 2019 show that v2 remained maintained alongside newer standard families. Versions and implementation scope vary by jurisdiction, program, and interface.

HL7 v2 defines message structures, segments, and fields; compatibility issues more often arise from extensive optionality and differences among local implementation guides. Z-segments are a valid extension mechanism, but they need shared specification and governance. V2 primarily supports event-driven messaging rather than every query style or clinical data model.

In 1997, HL7 launched the Reference Information Model (RIM) as the reference model for the v3 family. The RIM defines root classes such as Act, Entity, Role, Participation, and ActRelationship, providing a common modeling foundation for v3 artifacts.

4. 2005–2014: v3 messaging and CDA Release 2

In 2005, HL7 published the Normative Edition of HL7 v3 and CDA Release 2 (Clinical Document Architecture). The RIM was subsequently published as ISO/HL7 21731. Although both derive from the RIM, v3 messaging and CDA address different exchange patterns and should be evaluated separately.

HL7 v3 messaging uses XML and RIM-derived models, with more formal datatype, vocabulary, and structural constraints than v2. That formality can improve semantic consistency while increasing requirements for tooling, conformance artifacts, and modeling expertise.

V3 messaging deployment varied across programs and domains. A single program cannot establish that the entire v3 family "succeeded" or "failed"; assessments should distinguish v3 messaging, CDA, and other RIM-derived artifacts and examine each program's actual conformance scope.

CDA R2 defines an XML clinical document with a header and body and preserves document context. CDA can participate in storage or digital-signature workflows, but the specification alone does not determine a signature's legal validity. IHE XDS provides infrastructure for sharing multiple document types and can carry CDA documents; XDS does not restrict its document payload to CDA. CDA remains in use in some national and regional programs, with varying scope.

Experience from this period shows that a formal semantic model is only one component of interoperability. Tooling, implementation guidance, version governance, conformance testing, and operational capability also affect outcomes.

5. 2014–today: FHIR and web-based exchange models

In 2011, Grahame Grieve published the "Resources for Health" proposal; HL7 developed that approach into FHIR — Fast Healthcare Interoperability Resources. FHIR organizes data into Resources such as Patient, Observation, and MedicationRequest, provides JSON/XML representations, and supports multiple exchange paradigms, including RESTful APIs, documents, and messages. FHIR therefore does not require every implementation to use REST.

The first published milestone was FHIR DSTU1 on 30/09/2014. FHIR architecture describes the "80/20 rule" as a design principle: prioritize common requirements and use profiles and extensions for specific contexts. It is a reuse heuristic, not evidence that FHIR covers exactly 80% of all use cases; extensions also do not replace governance and conformance testing.

DSTU2 was published in 2015, STU3 in 2017, and Release 4 on 27/12/2018; technical correction 4.0.1 followed in 2019. R4 introduced FHIR's first Normative content while other artifacts remained Trial Use. Standards status is assigned per artifact and, in some cases, to content within an artifact; R4 is therefore described as a mix of Normative and STU content, not as wholly Normative.

In 2020, the US ONC Cures Act Final Rule established a requirement for certification criterion §170.315(g)(10): certified API technology had to support standardized access based on FHIR R4.0.1 and related implementation guides, with a compliance date of 31/12/2022. Its scope is health IT modules in the certification program, not every EHR or every healthcare API in the United States.

In 2022, HL7 released FHIR R4B (4.3.0) — a Trial Use release focused on changes in specific areas, not a new Normative edition. In 2023, FHIR R5 (5.0.0) was published through the Trial Use process. R6 is currently a ballot/CI build, not a stable release. The current releases of US Core, JP Core, KR Core, CH Core, and AU Core reviewed by this project are based on R4; that does not make R4 the default for every National IG.

By the mid-2020s, many major health platforms and cloud services support FHIR for interoperability use cases. Adoption depth, version, and conformance scope still vary by country, program, and vendor, so this does not imply that every healthcare system already exposes usable FHIR data.

6. Vietnam in the HL7 timeline

No public national dataset establishes the first HL7 v2 production date, number of facilities, or versions currently in operation in Vietnam. Evidence from individual projects is insufficient to estimate nationwide adoption, so this article does not use an estimated year as a national milestone.

Output data for BHYT claim review and payment developed through a separate series of XML specifications. Decision 4210/QĐ-BYT is one milestone in that lineage; Decisions 130, 4750, and 3176/QĐ-BYT (29/10/2024) subsequently amended or replaced related requirements. These formats are not HL7 v2, CDA, or FHIR; mapping them to FHIR requires a separately versioned and tested mapping layer.

In 2024, the Department of Health Information Technology under the Ministry of Health (the unit's name at publication) released the hl7vn/vn-core-ig repository, a Vietnam Core IG based on FHIR R4 with canonical http://fhir.ehealth.gov.vn/core/. The mirror at fhir.chiaseyhoc.vn made a built copy publicly accessible; its maintenance and adoption scope must be assessed separately and cannot be inferred from the website's existence alone.

Circular 13/2025/TT-BYT (issued 06/06/2025, effective 21/07/2025) defines the implementation schedule for electronic medical records. Law 91/2025/QH15 and Decree 356/2025/NĐ-CP establish personal-data protection duties, with health data classified as sensitive. These instruments create digitization, interoperability, and data-control requirements, but they do not mandate FHIR; a FHIR choice must follow the use case, conformance scope, and security architecture.

In 2026, Omi HealthTech initiated the trial-use VN Core project at hl7.org.vn, published under CC-BY-4.0 with canonical http://fhir.hl7.org.vn/core/. It uses an open, Working Group-oriented model and is developing its review and technical-proposal process; it is not an approved national standard or an HL7 Affiliate recognized by HL7 International.

7. Lessons for Vietnam

Forty years of HL7 history give Vietnam several very concrete lessons as it builds the VN Core IG.

One: publish artifacts through tested iterations. FHIR's 80/20 principle is a design heuristic for prioritizing common requirements, not a quantitative coverage target. Each VN Core profile or extension needs a use case, scope, examples, and explicit conformance criteria before expansion.

Two: use multi-stakeholder governance and traceable decisions. Clinical SMEs, healthcare facilities, vendors, payers, and regulators need a transparent review channel. HL7 Affiliate status should be stated only after formal recognition; until then, the project must accurately describe its trial-use status and emerging Working Group model.

Three: retain VN Core's published R4 baseline. This choice follows the package dependencies, validators, and reference IGs the project has tested. R5 is the latest published major release but remains Trial Use, while R6 is still at the CI/ballot stage; a version upgrade therefore requires a dedicated compatibility assessment rather than an assumption about global popularity.

Four: govern terminology alongside profiles. Interoperability depends on bindings, versions, and authoritative publication sources. Use of ICD-10 VN, SNOMED CT, LOINC, medical-service catalogs, or administrative-division codes must identify authority, license, version, and scope; a local translation or CodeSystem must not be presented as an official release by default.

Five: maintain legal traceability without exceeding the source. The IG should identify the relevant instrument, affected data element, and technical constraint selected by the project. A requirement for EMR, data protection, or interoperability does not automatically mean that the instrument mandates FHIR.

Historical note. Some Vietnamese sources have incorrectly recorded the FHIR DSTU1 release date as 21/02/2014, or claimed that R4 is fully Normative. This article reflects the official HL7 history page and the current R4 spec: DSTU1 was released on 30/09/2014, and R4 is Mixed Normative + STU.

8. References and further reading

Primary references

Relevant Vietnamese legal documents

  • Circular 13/2025/TT-BYT — Electronic medical records (issued 06/06/2025, effective 21/07/2025).
  • Law 91/2025/QH15 — Personal Data Protection Law (effective 01/01/2026).
  • Decree 356/2025/NĐ-CP — Implementing guidance for the Personal Data Protection Law (effective 01/01/2026).
  • Decision 3176/QĐ-BYT — Output data standard for healthcare services (29/10/2024).
  • Resolution 202/2025/QH15 — Reorganization of provincial-level administrative units, from 63 down to 34 (operational from 01/07/2025).

More from the knowledge hub

→ What is HL7? Overview of the organization and its standards → What is FHIR? Resources, REST API, and the 80% rule → HL7 v2 vs v3 vs FHIR — which standard for a new project → FHIR versions: DSTU1, DSTU2, STU3, R4, R4B, R5 → FHIR in Vietnam timeline — milestone-by-milestone details