Laboratory

Datasets

US Core 6.1.0 maps each USCDI element to FHIR.

Dataset	US Core 6.1.0 target profile(s)
`labs`	US Core Laboratory Result Observation, US Core Specimen

labs

One row per analyte result. A panel flattens: each member analyte is its own row, grouped back together by a shared diagnostic_report_id. The FHIR category is fixed to laboratory — non-lab test results belong in clinical_observations.

Column	Required	Format / values	Example
`record_id`	Yes	your stable key for this result	`LAB-0001`
`patient_identifier`	Yes	patient key	`MRN-4471903`
`status`	Yes	`registered`, `preliminary`, `final`, `amended`, `corrected`, `cancelled`, `entered-in-error`, `unknown` observation-status	`final`
`loinc_code`	Yes	LOINC us-core-laboratory-test-codes	`2339-0` glucose [mass/volume] in blood
`value_quantity`	If numeric	decimal	`104`
`unit`	If `value_quantity`	UCUM ucum-common	`mg/dL`
`value_comparator`	If censored	`<`, `<=`, `>=`, `>` quantity-comparator
`value_string`	If non-numeric	text	`POSITIVE`
`value_code`	If coded	SNOMED CT code, with `value_system` (SNOMED CT assumed when empty)
`data_absent_reason`	If no value	`unknown`, `not-performed`, `error`, etc. data-absent-reason
`interpretation`	If available	`H` high, `L` low, `N` normal, `A` abnormal observation-interpretation	`H`
`reference_range_low`	If available	decimal	`70`
`reference_range_high`	If available	decimal	`99`
`effective_datetime`	Recommended	datetime, at least to the day	`2026-04-18T08:40:00-04:00`
`specimen_type_code`	If available	SNOMED CT specimen code Specimen Type	`119297000` blood specimen
`diagnostic_report_id`	If available	`record_id` of the `diagnostic_reports` row	`DR-771`
`performer_npi`	If available	10 digits	`9999999991`
`is_deleted`	If retracting	`true` retracts this row	`true`

loinc_code is the USCDI Tests element and the one coded field with no fallback: a row without it cannot become an Observation. The binding accepts any laboratory LOINC code (http://loinc.org assumed).
Every row needs a result or a reason it is missing. Send exactly one of value_quantity (+ unit), value_string, or value_code; when the source has none — the test was cancelled, the specimen was unsatisfactory — data_absent_reason becomes required. A row with neither is rejected.
unit is not free text: US Core requires UCUM for quantity results. value_code should be SNOMED CT, e.g. 260385009 Negative for a coded qualitative result.
A censored numeric result — below or above the assay's detection limit — stays structured: <0.5 is value_quantity 0.5 with value_comparator <, not a value_string.
interpretation and the reference_range_* pair go beyond the USCDI v3.1 floor (they joined USCDI in v4), but pass them through whenever your source has them — they are what makes a bare number readable.
specimen_type_code is the USCDI Specimen Type element; it becomes a companion US Core Specimen resource the Observation points at, not a separate dataset you deliver.
diagnostic_report_id groups the analytes of one panel under a report in the diagnostic_reports dataset. A lab row is complete without it.
performer_npi is the resulting lab or clinician and must match a row in practitioners or organizations.

These resources are served by Patient Access, Provider Access, and Payer-to-Payer.