A clinician says "get me the LOINC for hemoglobin." A developer sits at the terminology tool and stares. LOINC has three dozen entries that match that request depending on specimen, panel, and unit, and the clinician does not want to have that conversation. Bridging the language gap is the first and hardest task of any LOINC integration.

Naming the entry pathway is the discipline. For related walkthroughs, the FHIR terminology desk collects the surrounding material.

The Clinical Phrase Is Almost Never Enough

"Hemoglobin" is not a single measurement in LOINC. It can be hemoglobin mass concentration in blood, hemoglobin A1c percent, or hemoglobin electrophoresis panel. Each has its own code, its own units, and its own clinical context.

The bridge starts by asking clarifying questions the requester probably has answers to: what specimen, what method, what units. A pass through the site's LOINC picker narrows the field by common vitals and labs so the conversation starts closer to the answer.

The Four Clarifiers That Matter

Four clarifiers narrow almost every LOINC search:

• Specimen: blood, urine, serum, plasma, sputum.
• Property: mass concentration, count, ratio, percent.
• Time aspect: point in time or over a duration.
• Method: manual, automated, immunoassay, chromatography.

Each clarifier maps to a LOINC axis. For the parts model that formalizes these axes, the LOINC parts model: property, system, and method at a glance covers the shape.

The Panel vs Component Question

The requester's phrase might refer to a single measurement or to a panel of measurements. "CBC" is a panel with sixty-plus component codes; "hemoglobin" alone is one component of the CBC.

Determining panel versus component early prevents scope thrash. For the specific framing, distinguishing lab tests, panels, and observations in LOINC covers the shape.

When No LOINC Match Exists

Some clinical measurements do not have a LOINC code yet. Rare tests, new methods, and highly-local protocols may need a local code and a proposal to LOINC. For the proposal-side framing, when a local code has no LOINC match and you need to propose one covers the workflow.

Absence of a LOINC match is not a failure of the search; it is a signal to consider a proposal or a local code with a documented mapping plan.

Start Broad, Narrow Deliberately

The reliable search pattern is broad to narrow:

• Start with the clinical phrase in the long common name search.
• Filter by specimen (blood, urine).
• Filter by property (mass concentration, ratio, percent).
• Confirm the units against the LOINC parts.

Each filter cuts the candidate list. Deployments that jump to a single guess produce inconsistent choices across searches.

The Language of Confirmation

Before adopting a LOINC code, confirm the choice with the requester using the LOINC long common name, not the code. Clinicians recognize the long common name; developers recognize the code. Confirming with the wrong side of the language gap invites disagreement two months later.

Documenting the Choice

Every adopted LOINC code should ship with a documented rationale: which clarifiers narrowed the choice, which alternative was rejected, which reference material informed the decision. The rationale becomes the audit trail future readers use to reconstruct the choice.

Rationale that lives in tribal knowledge decays; rationale in a shared document survives. Every LOINC integration that lands well maintains the rationale explicitly.

The Payoff Is Speed

The initial search cost is high; every subsequent search benefits from the established clarifier vocabulary. Six months into a program, most searches are minutes rather than hours because the clarifier habits are cached in the team.

Starting a LOINC search from a clinical phrase is not glamorous. It is where every terminology-driven integration begins.

Sources

• LOINC canonical getting-started guide from Regenstrief - LOINC canonical getting-started guide from Regenstrief Institute, evergreen reference