The laboratory information system is where clinical results become clinically real. A pathogen identification result that never reaches the LIS, or that reaches it as an unstructured text attachment, is operationally invisible to the ordering physician and the care team. Integrating sequencing results into the clinical LIS requires solving a vocabulary problem (how do you express a metagenomic pathogen call in HL7 result semantics?) and an interface problem (how does a real-time analysis pipeline talk to a LIS that expects discrete, formatted messages?). Neither is hard in isolation. Together, they require careful engineering.
The HL7 v2 ORU message as the clinical result carrier
HL7 version 2.x — specifically the ORU^R01 (Unsolicited Observation Result) message — remains the dominant result transmission format between laboratory analyzers, middleware, and LIS systems in US hospital environments. Despite the ongoing migration toward FHIR, the vast majority of clinical LIS interfaces in production today receive lab results over HL7 v2. A sequencing pipeline that only exports FHIR R4 will not integrate with most hospital LIS deployments without a translation layer.
An ORU message for a conventional microbiology result has a well-understood structure: OBR segment (order information), OBX segments (individual result observations). A culture result might have: one OBX for organism identification, one OBX for colony count, and one OBX per antibiotic tested in the susceptibility panel. The LIS knows exactly how to display this because it has been mapping microbiology culture results to this structure for decades.
A metagenomic sequencing result doesn't map cleanly to this structure. The result contains: a primary pathogen identification (or multiple, for polymicrobial specimens), confidence scores, AMR gene findings, sequencing QC metrics, and potentially a free-text interpretation. Mapping all of this into OBX segments requires decisions about what each OBX represents, what result code is used, and how the interpretive narrative is structured.
LOINC and SNOMED coding for genomic results
LIS systems expect result observations to be coded with standard vocabularies. For microbiology results, LOINC codes identify what was measured (e.g., LOINC 634-6 = Bacteria identified in specimen by culture) and SNOMED CT codes or organism codes identify what was found (the organism name). Sequencing-based identification should use LOINC codes appropriate for molecular methods rather than culture — for metagenomic identification, relevant LOINC categories fall under nucleic acid-based organism identification.
In practice, many clinical LIS systems have limited LOINC code sets that were built for culture-based microbiology and don't include codes for metagenomics-specific observations. This is a real integration challenge: the receiving LIS may not have a code for "metagenomic organism identification, nanopore sequencing" in its result catalog, requiring either LIS configuration changes (adding new order codes and result codes) or creative mapping to existing codes with interpretive comments carrying the methodological context.
The sequencing software should output result data in a format that makes standard vocabulary lookup straightforward — organism names in standardized nomenclature (not abbreviated, not genus-only when species is available), AMR gene names in standard nomenclature (CARD gene symbols), and QC metrics with defined normal ranges. The integration engineer shouldn't have to interpret or reformat analytical output to perform the LOINC mapping.
HL7 FHIR R4 DiagnosticReport for genomic results
FHIR R4 provides a richer data model for genomic results than HL7 v2. The DiagnosticReport resource can reference Observation resources for individual findings, MolecularSequence resources for underlying sequence data references, and structured coded results with explicit confidence values as extensions. For EHR systems with FHIR R4 capability (the major US EHR vendors have published SMART on FHIR and bulk data APIs), FHIR-native result delivery avoids the vocabulary compression required by HL7 v2 encoding.
A FHIR DiagnosticReport for a metagenomics result would typically include: the DiagnosticReport resource (result category, status, subject reference, effective datetime, issued datetime, performer reference); one or more Observation resources for identified organisms (SNOMED coded, with valueQuantity or interpretation extension for confidence); additional Observation resources for detected AMR genes; and a narrative text component for clinical interpretation.
The practical challenge for FHIR integration in hospital settings is that even EHRs with FHIR APIs often receive lab results through their HL7 v2 interface engine — the FHIR API is read-facing (for patient portals, third-party apps), not write-facing for inbound lab results. The write path for results into the EHR commonly still goes through the LIS over HL7 v2. Understanding the actual inbound integration architecture at the target facility — not the vendor's stated FHIR capability — is a prerequisite for designing the integration.
Middleware and interface engines
Most hospital integration environments include a middleware layer (interface engine) between laboratory analyzers and the LIS. Products like Mirth Connect, Rhapsody, or Epic's integration framework process inbound HL7 messages, transform them to LIS-specific formats, and route results to the appropriate result record. For a sequencing pipeline, the interface engine is both an opportunity (message transformation and enrichment can happen there) and a dependency (the sequencing software must be registered as a sender, the inbound channel must be configured, and the message format must pass validation).
The configuration process for adding a new analyzer or result source to the interface engine is typically a hospital IT project with its own ticket queue and change management review. Timeline estimates for LIS integration projects at large hospitals range from weeks to months. Laboratories evaluating sequencing software for clinical deployment should include interface engine configuration in the implementation timeline and loop in hospital IT and LIS team early — not as a post-procurement integration task.
Result display and physician-facing presentation
Even with a technically correct HL7 message arriving in the LIS, the physician-facing display depends on how the LIS is configured to present that result. A well-structured ORU message with five OBX segments for organism identification, confidence, AMR genes, coverage metrics, and interpretation text will display as five result rows in many LIS systems — which may be confusing to a physician accustomed to seeing a simple culture result.
Designing the result structure for physician interpretability requires collaboration between the integration engineer, the laboratory medical director, and the LIS configuration team. The goal is a result that is clinically complete but not overwhelming: primary organism identification prominently displayed, resistance flags immediately visible, confidence metric available but not requiring expertise to interpret, and detailed methodological information available in the report comments rather than in the primary result view.
We're not saying LIS integration is the sequencing software vendor's sole responsibility — the hospital IT and LIS teams own the receiving infrastructure. We're saying that sequencing software which delivers clean, well-structured, vocabulary-coded output makes the integration project substantially simpler than software that delivers unstructured text or proprietary JSON that requires custom transformation. The integration complexity is real; the software can either amplify or reduce it.
A practical integration checklist for the laboratory team
Before beginning LIS integration of a clinical sequencing pipeline, the laboratory should have documented answers to:
- What HL7 version does our LIS support for inbound results? (v2.3, v2.5.1, v2.6+?)
- Do we have an interface engine, and who manages it?
- What LOINC codes exist in our LIS for molecular/sequencing-based organism identification?
- How long does adding a new analyzer interface typically take through IT change management?
- Does the EHR receive results from the LIS via HL7 v2 forwarding, or does it have a separate direct inbound interface?
- What result display format does our LIS use for microbiology culture reports, and is there an equivalent for molecular results?
These are IT infrastructure questions, not bioinformatics questions. The laboratory medical director and IT liaison need to be at the same table early in the sequencing deployment project. The analytical validation and the integration are parallel workstreams, and the timeline often compresses when integration is treated as a final step rather than a parallel track.