Electronic lab reporting, eCR, and syndromic surveillance feeds

Reference article · 5 min read

A health department's ability to detect and investigate disease depends on a handful of data feeds that arrive continuously, mostly without human intervention, and mostly from systems the department does not control. Electronic laboratory reporting brings test results from clinical and commercial labs. Electronic case reporting brings case reports triggered out of provider EHRs. Syndromic surveillance brings near real time visit data from emergency departments. The disease surveillance system where epidemiologists work sits downstream of all of them, and it is where an incoming message becomes a case with an investigation attached.

When something in that chain breaks, the symptom epi staff report is generic: conditions arriving late, arriving with fields missing, landing in the wrong jurisdiction, or not arriving at all. The cause is usually specific and fixable. What follows maps the pipeline, the onboarding work behind it, and the failure points informatics staff spend most of their time on.

What each feed actually carries

ELR. HL7 version 2 result messages from hospital, commercial, and public health laboratories. Tests are identified with LOINC and coded results with SNOMED CT. For most jurisdictions this is the highest volume structured feed and the backbone of notifiable condition detection.
eCR. An electronic initial case report generated from a provider's EHR when trigger codes fire during a clinical encounter, routed through shared national infrastructure, with reportability determined by decision support rules the jurisdiction authors itself. A reportability response goes back to the sender. eCR carries clinical context a lab result cannot: symptoms, pregnancy status, travel, treatment, provider contact information.
Syndromic surveillance. Admission, discharge, and transfer messages from emergency departments and other facilities, generally processed through CDC's National Syndromic Surveillance Program platform and analyzed in ESSENCE. Chief complaint text and diagnosis codes support trend detection and situational awareness, not case confirmation.
The surveillance system. NBS, Maven, an Epi Trax deployment, or a state-built system, where messages become investigable cases and from which notifiable case notifications flow to CDC.

Reportable condition lists, reporting timeframes, and the authority behind them are set in state statute and administrative code and change over time. Confirm current requirements with your state's official rule source and your counsel rather than with a vendor summary.

Onboarding a facility

Onboarding is a project per facility, not a switch. A workable sequence:

Registration and intent. Record the facility, its EHR or LIS vendor and version, contacts, and which feeds it will send. Federal incentive programs have historically required providers to register intent to report and engage with public health, but program details shift, so verify the current rules.
Transport. Agree on the channel and credentials. Test connectivity before testing content.
Test messages. Validate structure against the applicable implementation guide using conformance tooling, then review content by hand. Structural validity does not mean the message says what the sender thinks it says.
Vocabulary mapping. Compare the sender's local test and result codes against LOINC and SNOMED. Expect gaps.
Parallel reporting. Run electronic alongside the existing fax or portal process until epi staff agree the electronic feed is complete and correct.
Production and monitoring. Move to production with a named owner on both sides and an agreed path for reporting problems.

Where messages break

Unmapped or wrong codes. Local codes with no LOINC equivalent, LOINC codes that do not match the test performed, and free text where a coded result belongs. These messages either fail processing or, worse, process into the wrong condition.
Address and geocoding failures. Jurisdictional assignment depends on patient residence, not facility location. Missing units, PO boxes, foreign addresses, and homeless or congregate settings all defeat automated geocoding and push cases into manual queues or the wrong health department.
Patient identity. The same person arrives from multiple labs and providers with different identifiers, name spellings, and dates of birth. Weak matching creates duplicate case records; aggressive matching merges two people. Both are serious, and the second is harder to detect.
Demographic completeness. Race, ethnicity, and occupation fields are frequently blank or coded as unknown, which limits equity analysis after the fact.
Feed overlap. The same case can arrive by ELR and eCR. The surveillance system needs deduplication logic that treats them as one event.
Silent stoppage. A facility upgrades its EHR and the interface quietly dies. Without volume monitoring per facility, nobody notices for weeks.

Surge

Outbreak volume exposes design assumptions. Queues back up, manual review of unmatched addresses becomes the bottleneck, and staff who normally investigate cases spend their time triaging messages. Plan for it before it happens: instrument throughput and backlog by stage, decide in advance which conditions get automated processing with lighter review, and know which steps can be temporarily relaxed without corrupting the case record.

What to ask vendors and require in the contract

Ask how vocabulary mapping is maintained over time, who does it, and what happens when a sender introduces a new local code.
Ask for the deduplication and matching approach in writing, including how a bad merge is undone.
Require that interface configurations, mappings, and transformation logic are documented and exportable, and that your agency owns them.
Require monitoring and alerting on per-facility message volume, rejection rates, and processing lag, visible to your staff rather than only to the vendor.
Require a defined onboarding runbook and a service commitment for facility onboarding, not an open-ended best effort.
Ask how the system handles reprocessing after a mapping fix, since retroactive correction is routine work.

How to start

Inventory what you have: every sending facility, every feed, last message received, and current rejection rate. Most departments find stopped interfaces they did not know about. Then pick the two highest volume error categories and fix those before adding new senders. Onboarding more facilities into a pipeline that already loses messages multiplies the problem.

HQT is a technology and software company serving United States state and local government agencies, with capabilities in data integration, analytics, and systems engineering.

Frequently asked questions

What is the difference between electronic lab reporting and electronic case reporting?

Electronic laboratory reporting sends structured test results from a laboratory to the health department, identifying tests with LOINC and coded results with SNOMED CT. Electronic case reporting is generated from a provider's EHR when trigger codes fire during an encounter, and it carries clinical context such as symptoms, treatment, and pregnancy status that a lab result does not include. Many cases arrive through both channels, so the surveillance system needs deduplication logic that recognizes them as one event.

Why are reportable conditions landing in the wrong jurisdiction?

Jurisdictional routing is normally based on the patient's residence address, not the reporting facility's location, so routing quality depends entirely on address quality and geocoding. Missing apartment numbers, PO boxes, out-of-state or foreign addresses, and congregate or unhoused situations commonly defeat automated geocoding. These records either fall into a manual review queue or get assigned to the wrong health department, which is why per-stage queue monitoring matters.

How long does it take to onboard a facility for ELR or eCR?

It varies widely by facility, EHR or LIS vendor, and how many local codes need mapping, so treat each onboarding as a small project rather than a configuration change. The sequence generally runs from registration and transport setup through test messages, vocabulary mapping, parallel reporting alongside the existing process, and finally production with ongoing monitoring. The parallel reporting phase is where most schedule risk lives, because that is when epi staff discover content problems that passed structural validation.

What should a health department require in a surveillance integration contract?

Require that interface configurations, vocabulary mappings, and transformation logic are documented, exportable, and owned by the agency rather than the vendor. Require monitoring and alerting on per-facility message volume, rejection rates, and processing lag that your own staff can see. Also require a documented onboarding runbook, a defined service commitment for facility onboarding, and a clear process for reprocessing messages after a mapping correction.

How do we detect an interface that has quietly stopped sending?

Monitor expected message volume per facility and per feed, and alert on unexpected drops rather than only on errors. A facility that upgrades its EHR or changes a network path often stops sending without generating any failure the health department sees. Reviewing last-message-received dates for every sender is usually the fastest way to find stopped interfaces in an existing pipeline.