For most clinical labs, LOINC is the boring layer that quietly decides whether a result lands cleanly in a state public health database, a payer’s quality measure, or another health system’s chart. When the mapping is right, nobody notices. When it is wrong — a wrong specimen code, a generic “glucose” mapped instead of the method-specific code, a panel sent as a component — results get rejected at the trading-partner gateway, reportable conditions miss reporting deadlines, and the lab finds out from someone else.
This guide walks through what LOINC actually encodes, where mappings break in real labs, what tooling exists, and how to keep a mapping table current as LOINC itself evolves.
What LOINC is, in one paragraph
LOINC — Logical Observation Identifiers Names and Codes — is a universal coding system for laboratory and clinical observations maintained by the Regenstrief Institute since 1994. The set covers laboratory tests, vital signs, clinical documents, survey instruments, and other observations. A LOINC code is a numeric identifier (up to seven digits) followed by a hyphen and a single check digit — for example, 2345-7 for “Glucose [Mass/volume] in Serum or Plasma”. Each code is paired with a structured “Fully Specified Name” that describes the observation along six axes.
The six axes of a LOINC term
Every LOINC term is defined by six parts, and “mapping a test to LOINC” really means choosing each one correctly:
- Component — what is being measured (e.g. Glucose, Hemoglobin A1c, SARS-CoV-2 RNA).
- Property — the kind of quantity (Mass concentration, Substance concentration, Number, Ratio, Presence/absence).
- Time aspect — point-in-time (
Pt) for a single measurement, or a duration (24H,1H) for timed collections. - System (specimen) — the source — Serum, Plasma, Whole blood, Urine, CSF, Bronchoalveolar lavage, Stool, and so on.
- Scale — quantitative (
Qn), ordinal (Ord), nominal (Nom), narrative (Nar), or document (Doc). - Method — the analytical method, when method-dependent (immunoassay, PCR, mass spec, etc.). Many codes leave method unspecified, indicating any method is acceptable.
Two tests that look identical on the requisition can require different codes. “Vitamin D, 25-hydroxy” can be 1,25-dihydroxy versus 25-hydroxy versus the total — three different components. “TSH” can be measured in serum or plasma — different systems. “Influenza A” can be detected by PCR, antigen, or culture — different methods, which the receiving public health agency cares about because it affects how the result is interpreted.
This six-part structure is the reason LOINC mapping is harder than mapping CPT or ICD. It is not a vocabulary lookup; it is a small structured-data exercise per test.
Why LOINC mapping matters in 2026
Three forces have pushed LOINC from a nice-to-have to a hard requirement for most clinical labs.
Electronic Laboratory Reporting (ELR). State and local public health agencies require reportable conditions to be sent via HL7 v2.5.1 ORU messages with LOINC in the OBX-3 field. The CDC’s HL7 v2.5.1 Implementation Guide for ELR is the reference specification most states adapt. A non-LOINC code in OBX-3 typically fails validation at the public health gateway and the message is rejected. For COVID-era reporting that requirement turned into an enforcement mechanism: labs that wanted to report at all had to clean up their LOINC tables. The public health LIMS ELR workflow guide walks through where those gateway rejections surface in the reportable-condition lifecycle and how the LIS reconciles them.
USCDI and certified EHR exchange. The United States Core Data for Interoperability names LOINC as a required vocabulary for the Laboratory data class. ONC-certified EHRs must consume and exchange lab data with LOINC codes, which means LIS output must include them. For labs that send results to multiple EHRs, the cleanest implementation is “LOINC in the OBX-3 of every result”, not “LOINC only when the receiver asks”.
FHIR Observation resources. In FHIR, the Observation.code element is bound to LOINC for laboratory observations under the US Core profile. Patient apps, payer APIs, and bulk-FHIR pulls all expect LOINC, not the lab’s internal test code. Even labs that have no FHIR endpoint of their own are reached by FHIR-shaped queries through their EHR integration.
The net effect is that for any clinical lab that integrates with anything outside its own four walls, “LOINC in the LIS” is no longer optional.
What goes into a LOINC mapping in practice
A mapping in the LIS is a row that ties an internal test code — the code the lab uses on its catalog, requisitions, and instrument worklists — to a LOINC code. Most LIS systems store this at the result level, not the order level, because a single order code can produce multiple results that each need their own LOINC. A “Comprehensive Metabolic Panel” order generates fourteen separate OBX segments, each with its own component LOINC.
A typical row carries:
- The internal test code (lab catalog ID).
- The internal result code (analyte ID for one OBX line).
- The LOINC code.
- The specimen system the LOINC code expects (the lab catalog’s specimen type must match — Serum and Plasma are different terms).
- The unit of measure, ideally in UCUM format (
mg/dL,10*9/L). - The method when method-dependent.
- Effective date and version.
- The user who last verified the mapping.
For instrument-driven assays — particularly molecular and immunoassay platforms — the manufacturer’s LIVD (LOINC/LIVD digital format) catalog publishes the manufacturer-recommended LOINC for each test on the analyzer. Pulling from LIVD is faster than mapping from scratch and reduces the risk of method-dimension errors, because the manufacturer has already encoded their method. The immunology LIS guide covers the immunoassay analyzer interfacing and structured result capture that LOINC-coded autoimmune, allergy, and serology panels depend on.
Where LOINC mapping breaks
These are the failure modes that surface most often when a lab does an honest review of its mapping table.
Specimen drift. The lab’s catalog says “Blood”, the LOINC says “Serum”, the actual collection is “Plasma”. The result goes out under a code that does not describe the specimen that was tested. Downstream consumers cannot reconcile because they assume the LOINC system axis is authoritative. This is the single most common mapping defect in audits.
Method-specific tests mapped to method-unspecified codes. “SARS-CoV-2 RNA by PCR” mapped to a generic “SARS-CoV-2” code that does not encode the PCR method. Public health agencies that branch reporting logic on method (PCR vs antigen) end up routing the result the wrong way.
Panels mapped instead of components. “Lipid panel” carries a LOINC at the panel level for ordering, but each component (cholesterol, HDL, LDL, triglycerides) needs its own LOINC for the result OBX segments. Mapping the panel code into the result causes the EHR to receive five copies of the same code.
Calculated results. LDL calculated by Friedewald is a different LOINC than LDL measured directly. eGFR has separate LOINCs by formula (MDRD vs CKD-EPI vs the 2021 race-free CKD-EPI). The calculation method is part of the term, and the lab must know which one its LIS is doing.
In-house assays and laboratory-developed tests (LDTs). Some LDTs do not have an existing LOINC term. The lab can submit a new term request to Regenstrief, or use the closest existing term with the method axis disclosed in the OBX-6 (range comment) or OBX-15 (producer ID) until a precise term exists.
Stale mappings after LOINC version updates. LOINC releases twice a year — typically in June and December — and existing codes can be marked DEPRECATED or DISCOURAGED with a recommended replacement. A lab that never refreshes its mapping ages out of the active set without realizing it. ELR validators do flag deprecated codes, but only if the lab is paying attention to the warnings.
Tools labs actually use
A short, factual list — there are more, but these four show up in nearly every implementation.
- RELMA (Regenstrief LOINC Mapping Assistant). Free desktop tool from Regenstrief that lets a lab load its internal test catalog and search the LOINC database with structured filters across the six axes. Still the workhorse for first-pass mapping.
- LOINC search on loinc.org. Browser-based search of the current LOINC release, useful for one-off lookups and for cross-checking a RELMA candidate against the live database.
- Manufacturer LIVD catalogs. Most major IVD vendors publish a LIVD file per analyzer, in HL7 v2 or FHIR format. The lab matches its instrument and assay to the manufacturer’s row and inherits the recommended LOINC.
- CDC PHIN VADS. The CDC’s vocabulary repository publishes the LOINC value sets that public health agencies expect for specific reportable-condition messages. When a state ELR rejects a code, PHIN VADS is the source of truth for what the value set actually contains.
Mapping governance — keeping the table honest
LOINC mapping is not a one-time project. A working governance pattern looks like this:
- Initial mapping done by an informatics analyst with sign-off from the laboratory director. The director’s role is to confirm specimen, method, and calculation choices that affect clinical interpretation.
- A change ticket for any new test, instrument change, or method change. The mapping is validated against test orders before the test goes live.
- A semi-annual LOINC version review following each LOINC release. The lab compares its active LOINCs against the release notes, retires deprecated codes, and accepts vendor LIVD updates for any analyzer that issued one.
- An ELR rejection log monitored monthly. Public health gateway rejections are the canary — they catch mapping defects before payers and EHRs do.
- Annual sample audit as part of routine CAP / CLIA preparation. Pull a representative slice of recent results and check that the LOINC sent matches the actual specimen, method, and measurement performed.
The audit trail behind a mapping change — who changed it, when, why, against what LOINC version — is part of the LIS validation evidence. Inspectors increasingly look for this when they review interoperability and reporting.
Why this matters
When the LOINC layer is clean, the downstream effects are quiet and cumulative. Reportable conditions hit state public health on time. EHRs across the region show results under codes their clinicians recognize. Patient app integrations through the EHR’s FHIR endpoint return structured, interpretable lab data. Payer quality measures (HEDIS, MIPS, value-based contracts) pick up the lab’s contribution because the codes match the measure denominators.
When the LOINC layer is stale, the failures are diffuse. The lab loses a few dollars per result on payer denials. Public health complains, intermittently, about messages that were rejected. The lab is told by an outside system that a test was “missing” when it was sent under the wrong code. The connection between any single failure and the underlying mapping defect is usually invisible — which is what makes routine governance, not heroic remapping projects, the part that actually works.
Where LIMS IQ fits
LIMS IQ stores LOINC at both the test catalog and result-component levels, with UCUM units, method, and effective-date on every mapping row. The same fields populate the outbound HL7 OBX-3 / OBX-6 in ORU result messages, the FHIR Observation.code and valueQuantity.code in FHIR Observation resources, and the ELR submissions that route through the electronic laboratory reporting pipeline. LOINC version refresh and deprecated-code review are scheduled work, not a one-time setup task. Want to see how the mapping layer is exposed in the LIS — including the audit trail, the LIVD import, and the LOINC version history? Request a demo.
This article is part of the clinical laboratory informatics series — the companion guides cover failed-QC review, critical-value callbacks, ELR reporting, and lab KPI dashboards.
Sources
- LOINC — About — Regenstrief Institute, history and structure of LOINC.
- LOINC Users’ Guide — Major Parts of a LOINC Term — definition of the six LOINC axes.
- LOINC — Releases — release cadence and version history.
- RELMA — Regenstrief LOINC Mapping Assistant — official tool documentation.
- CDC — Electronic Laboratory Reporting (ELR) — HL7 v2.5.1 ELR Implementation Guide and reporting requirements.
- HealthIT.gov — United States Core Data for Interoperability (USCDI) — Laboratory data class and required vocabularies.
- HL7 US Core — Observation Lab Profile — FHIR LOINC binding for laboratory observations.
- FDA — In Vitro Diagnostics and LIVD — LIVD digital format for manufacturer LOINC catalogs.
- UCUM — Unified Code for Units of Measure — companion unit-of-measure standard referenced alongside LOINC.
- CDC PHIN VADS — public health vocabulary value sets that constrain LOINC use in ELR.