Home / Insights / Connectivity
HL7 v2 vs FHIR for lab results: which should your lab use?
Every lab connecting to an EHR hears the same question: HL7 v2 or FHIR? The honest answer is that the receiving system usually decides, and many labs need both. This guide explains how each standard carries lab orders and results, where each is strong, and how to choose.
Written and reviewed by the Catenix team. How these guides are written and checked.
In brief
- HL7 v2 is the long-established standard for lab messaging. Orders travel as ORM or OML, results as ORU, usually over MLLP on a private network.
- FHIR is the newer, web-based standard. Orders are ServiceRequest resources; results are DiagnosticReport and Observation resources, exchanged over HTTPS APIs.
- Most hospital and ambulatory EHRs still accept lab results over HL7 v2; many cloud EHRs and patient apps prefer FHIR or their own APIs.
- The meaning matters more than the syntax: codes, units, ranges, flags, corrections and patient matching must be right in either standard.
- Catenix sends results as HL7 v2 or FHIR, so the choice follows your receiving system.
Take it with you: POCT connectivity buyer's guide · POCT readiness checklist (PDF, no form).
How HL7 v2 carries lab results
HL7 v2 messages are text, made of segments separated by line breaks and fields separated by pipes. A result message (ORU^R01) carries an MSH header, a PID patient segment, ORC and OBR order segments, one OBX segment per result with code, value, units, range, flag and status, and NTE comments. Messages are acknowledged with an ACK. They usually travel over TCP with MLLP framing across a VPN or private network.
Strengths: almost universal support in hospital and ambulatory systems, decades of practice, predictable behaviour. Weaknesses: many local variations, so every interface needs its own specification; network set-up (VPNs, fixed IPs) can be slow; not designed for apps or the open web. Our HL7 field guide shows real segments.
How FHIR carries lab results
FHIR represents data as resources exchanged as JSON or XML over HTTPS REST APIs. For labs:
| FHIR resource | Role |
|---|---|
| ServiceRequest | The order for one or more tests |
| Specimen | The sample, with collection time and type |
| Observation | One result, with code, value, unit, reference range and interpretation |
| DiagnosticReport | The report grouping the observations, with status and conclusion text |
| Patient | The patient the results belong to |
Strengths: web-native, easier for modern developers, works naturally with apps and patient-facing services, consistent resource model. Weaknesses: each EHR implements FHIR with its own profiles, permissions and app registration process; write access for outside lab results is not always offered; and version and profile differences mean "we support FHIR" does not guarantee a working interface.
Side by side
| HL7 v2 | FHIR | |
|---|---|---|
| Format | Pipe-delimited text segments | JSON or XML resources |
| Transport | MLLP over TCP, often via VPN | HTTPS REST APIs |
| Orders | ORM^O01, OML^O21 | ServiceRequest |
| Results | ORU^R01 | DiagnosticReport and Observation |
| EHR support for outside labs | Very broad | Growing, varies by vendor |
| Patient apps and portals | Rare | Common |
| Set-up effort | Network and per-interface specification | App registration, permissions and profiles |
The same result in both standards
One TSH result as an HL7 v2.5.1 ORU^R01 message:
MSH|^~\&|MIDDLEWARE|LAB|EHR|CLINIC|20261118101530||ORU^R01|MSG000417|P|2.5.1 PID|1||MRN12345^^^CLINIC^MR||TEST^PATIENT||19850214|F ORC|RE|ORD7781|2611180042 OBR|1|ORD7781|2611180042|3016-3^TSH^LN|||20261118094500|||||||||||||||20261118101530|||F OBX|1|NM|3016-3^TSH^LN||2.15|mIU/L|0.27-4.20|N|||F|||20261118101500
The same result as a FHIR R4 Observation:
{
"resourceType": "Observation",
"status": "final",
"category": [{"coding": [{"system": "http://terminology.hl7.org/CodeSystem/observation-category", "code": "laboratory"}]}],
"code": {"coding": [{"system": "http://loinc.org", "code": "3016-3", "display": "TSH"}]},
"subject": {"reference": "Patient/example"},
"effectiveDateTime": "2026-11-18T09:45:00Z",
"valueQuantity": {"value": 2.15, "unit": "mIU/L", "system": "http://unitsofmeasure.org", "code": "m[IU]/L"},
"referenceRange": [{"low": {"value": 0.27, "unit": "mIU/L"}, "high": {"value": 4.20, "unit": "mIU/L"}}]
}
The facts are identical: a LOINC code, a value, units, a reference range, a time and a final status. HL7 v2 puts them in numbered fields of the OBX segment; FHIR names them. In FHIR the Observation would normally sit inside a DiagnosticReport, and the patient and order would be references to resources the EHR already holds. The units in FHIR also carry a UCUM code (m[IU]/L), which receiving systems use to compare values reliably. All identifiers here are invented for the example.
What this means in practice
In the US, most ambulatory EHRs accept lab results from an outside lab system over HL7 v2, sometimes through the EHR vendor's own interface programme; some also accept results through their API. In the UK, routes into GP systems such as SystmOne and EMIS and into private clinic EHRs vary, and personal health record and NHS App routes increasingly use FHIR. In both countries, patient-facing apps lean toward FHIR.
So the practical rule is: use whatever the receiving system supports best for an outside lab, and pick a lab system that can do both. A lab serving several receivers often runs both at once: HL7 v2 to one EHR, FHIR to another, and a signed PDF to a third. See how lab orders and results interfaces work and our EHR integration pages.
Getting it right in either standard
- Codes: map tests to the codes the receiver expects, such as LOINC in the US or SNOMED CT in the NHS. See LOINC and SNOMED coding.
- Units and ranges: send units explicitly and use the lab's reference ranges.
- Flags: map abnormal and critical flags consistently.
- Status: send preliminary, final and corrected statuses correctly, and test that corrections replace rather than duplicate.
- Identity: carry the receiver's patient and order identifiers exactly.
- Delivery: log every message, handle acknowledgements or API errors, retry and alert on failure.
Which one to use, case by case
| Receiving system | Usually the best route |
|---|---|
| Hospital EPR or LIS with an integration engine | HL7 v2 ORU over MLLP, to the trust's or hospital's specification |
| US ambulatory EHR with a lab interface programme | The vendor's HL7 v2 lab results interface, or its API where offered for outside labs |
| Cloud EHR with a documented FHIR API that accepts lab results | FHIR DiagnosticReport and Observation |
| Patient app or personal health record | FHIR, or the platform's own API |
| A system that accepts neither | A signed PDF report filed against the patient, while a structured route is agreed |
Common mistakes in either standard
- Assuming "we support FHIR" means "we accept your results". Many EHRs offer read access only. Ask specifically whether an outside lab system can write results.
- Mismatched versions and profiles. FHIR R4 with US Core is not the same as an older version or another country's profile.
- Units without UCUM in FHIR. Free-text units make values hard to compare and chart.
- Status mapping errors. Preliminary, final, corrected and cancelled must map correctly in both directions.
- Skipping acknowledgements. In HL7 v2, an unacknowledged message must be retried or flagged; in FHIR, an API error must not be ignored.
Questions to ask the receiving system's vendor
- Can an outside lab system send results to you, and by which route?
- Which HL7 v2 version or which FHIR version and profiles do you require?
- Which code system do you expect on each result?
- How are corrected and cancelled results handled?
- Is there a test environment, an approval process and a fee?
Our integrations pages describe the route for each record system, and the open connectivity dataset lists both analysers and record systems. You can also ask inside ChatGPT or Claude through our AI assistant connector.
Transport and security
HL7 v2 messages travel over TCP with MLLP framing. Across the internet, that connection should run inside a VPN or be wrapped in TLS, and the two ends usually agree fixed addresses and ports. Some cloud EHRs accept HL7 v2 files over SFTP instead. FHIR runs over HTTPS. A lab system writing results to an EHR typically authenticates as a system rather than as a user, often with OAuth 2.0 client credentials or SMART Backend Services, and is limited to the permissions the EHR grants it. Whichever route you use, ask how credentials are stored and rotated, and how the connection is monitored.
What to test before go-live
- A single result, a panel and a result with a comment.
- An abnormal and a critical result, with flags displayed as intended.
- A corrected result replacing the original.
- A result for a patient the receiver cannot match, and where it goes.
- A rejected message or API error, with the retry and the alert.
- Special characters and long names.
Where Catenix fits
Catenix receives analyser results in ASTM E1394, HL7 v2, POCT1-A2 or proprietary formats, applies QC and competency rules, and delivers results to LIS and EHR systems as HL7 v2 messages or through FHIR and vendor APIs, with acknowledgements, retries and a full audit trail. The choice of standard follows your receiving system. Analyser not on the list? We build the driver, usually within a couple of days. Catenix is connectivity, workflow and record-keeping software; it does not interpret results and is not marketed as a medical device.
Questions people ask
Is FHIR replacing HL7 v2 for lab results?
Gradually, and mainly for apps and newer cloud systems. HL7 v2 remains the most widely supported route into EHRs for lab results, and will be for years.
Which should a new lab choose?
Whichever each receiving system supports best for an outside lab. Choose lab software that can send both.
Can FHIR be used for analyser connectivity?
Rarely. Analysers mostly speak ASTM E1394, HL7 v2 or POCT1-A2. Middleware converts their output to FHIR where needed.
Does Catenix support FHIR?
Yes. Catenix can deliver results as HL7 v2 or through FHIR and vendor APIs.
Which FHIR version do EHRs use for lab results?
Most current EHR APIs are based on FHIR R4, often with national or vendor profiles such as US Core in the US or UK Core in England. Ask which version and profiles the receiving system requires.
How is a corrected lab result sent in FHIR?
The Observation and DiagnosticReport are updated with status corrected or amended, keeping the same resource identity, so the receiving system replaces the earlier value rather than adding a second one.
Is FHIR more secure than HL7 v2?
Neither is secure or insecure on its own. Security depends on the transport and access control: HL7 v2 over a VPN or TLS, FHIR over HTTPS with proper authorisation. Ask how each connection is encrypted, authenticated and monitored.
Sources and further reading
Read next
More for coordinators.
All Catenix integrations
Plain-language guides to connecting point-of-care analysers to your LIS or EHR, and the messaging standards they use: HL7, POCT1-A2, ASTM.
Read →How lab orders and results interfaces work
How lab orders reach a lab system and results return to the EHR: HL7 ORM and ORU, FHIR, specimen IDs, acknowledgements, corrections and who builds...
Read →ASTM to HL7 conversion explained
How lab middleware converts analyser output in ASTM E1394 into HL7 v2 ORU messages or FHIR: record mapping, test codes, units, flags, QC and...
Read →Take it with you
The buyer's guide, in your inbox.
Nine pages on what POCT connectivity is, the standards in plain English, the cost of doing it by hand and eight questions to ask any supplier.
Get the POCT connectivity buyer's guide
Nine pages, vendor-neutral: what connectivity is, the standards in plain English, the cost of doing it by hand and eight questions to ask any supplier. The PDF also opens directly from the downloads page; leave an email only if you want it in your inbox.
See it done in software.
A 30-minute walkthrough on a live tenant, with your analysers and your allowable-error limits.