TL;DR: A multi-specialty group should build HL7 interfaces only for its highest-volume labs and handle every other source with a centralized lab results ingestion workflow. That workflow captures outside results from fax, email, and portals in one queue, matches each to the patient and ordering provider, files it into the EHR, and routes it to the right specialty with abnormal values flagged. Centralizing intake across locations, rather than letting each site manage its own fax tray, is what keeps results from getting lost.
The problem: interfaces cover the big labs, not the long tail
Most multi-specialty groups have done the obvious work. Quest and Labcorp are interfaced. Orders go out electronically and results come back as discrete data. On paper, labs are solved.
Then you look at the fax queue.
Cardiology is getting lipid panels and troponins from the hospital where a patient was admitted. Nephrology is getting dialysis labs from three different dialysis units. Endocrinology is getting A1c results from the patient's PCP down the street. Oncology is getting CBCs from an infusion center and pathology from a hospital lab. GI is getting outside biopsy results. None of those sources are interfaced, and most never will be.
ONC's research on interoperability among office-based physicians has repeatedly shown that integrating outside information into the EHR without manual effort trails far behind just receiving it. For a multi-specialty group, that gap multiplies across every specialty and every location.
How do you decide which labs get an interface and which get ingested?
The answer comes from your own volume data, not a vendor pitch. Two steps get you there.
Step 1: Size your long tail before you decide anything
Before you pick an approach, get real numbers. Run a two-week count of every outside lab result that is not delivered by an existing interface.
Capture for each result:
- Source (which hospital, lab, dialysis unit, or practice)
- Channel (fax line, eFax, email, portal, paper)
- Receiving specialty and location
- Result type (chemistry, hematology, pathology, microbiology)
- Time from arrival to filed in the chart
You'll almost always find two patterns. First, a small number of sources account for a large share of volume. Second, the remaining sources are scattered: dozens of senders, each with a handful of results a month.
That split tells you what to do. If one hospital lab sends you hundreds of results a month, it may deserve an HL7 interface. Everything below that threshold belongs in an ingestion workflow.
Step 2: Decide where interfaces still make sense
Interfaces are still the gold standard for discrete, order-matched results. They're also projects. Each one requires coordination with the sending lab, configuration in your interface engine or EHR, testing, and ongoing maintenance when either side upgrades.
A reasonable rule for a multi-specialty group:
- Interface it if a source sends high, steady volume across multiple specialties and the sending organization will actually cooperate on a project.
- Ingest it if volume is modest, concentrated in one specialty, inconsistent, or the sender can't support an interface.
- Revisit annually, since referral patterns and hospital affiliations change.
The mistake is treating every new lab source as a potential interface project. Your IT team will never catch up, and results will keep arriving by fax in the meantime.
Why each specialty sees different outside results
A multi-specialty group isn't one lab workflow. It's several, and your ingestion rules need to reflect that.
Cardiology
Outside results are often hospital-generated: troponins, BNP, lipid panels, and coagulation studies tied to recent admissions or ED visits. Speed matters when a post-discharge visit is coming up, and INR values matter for anticoagulation management.
Nephrology
Dialysis units and outside labs send recurring panels (creatinine, eGFR, potassium, phosphorus, PTH). Volume is high and repetitive, which makes nephrology a strong candidate for discrete-value writing so values trend.
Endocrinology
A1c, thyroid panels, and lipids frequently come from the patient's primary care practice. These are high-value for trending and often for quality measures.
Oncology and hematology
CBCs and chemistries from infusion centers and outside labs arrive frequently and are time-sensitive relative to treatment decisions. Pathology reports may come from hospital labs.
Gastroenterology and urology
Outside pathology from biopsies, plus PSA and other markers in urology. These are less frequent but high-consequence if misfiled.
Your routing rules, document categories, and discrete-data priorities should differ by specialty. One generic "Outside Labs" folder for the whole group is how results end up in front of the wrong provider.
Should you centralize intake or let each location handle it?
Centralize. That's the short answer, and it's the right one for nearly every multi-specialty group with more than a couple of sites.
Site-by-site handling usually looks like this: each location has its own fax line, its own MA or front desk staff checking it, its own habits for labeling documents, and its own backlog. When someone is out sick, the tray fills up. When a hospital faxes to the wrong location, the result sits there.
A centralized model looks different:
- All fax lines, eFax accounts, and shared inboxes feed one digital intake queue, while keeping each location's existing fax numbers so senders don't have to change anything.
- Portal retrieval for hospital and lab portals is handled once for the whole group, not separately at each site.
- One set of classification and patient-matching rules applies everywhere.
- Routing logic sends each filed result to the right specialty and ordering provider, regardless of which fax number it arrived on.
- One exception team (often a small central medical records group) clears anything the automation can't file confidently.
This is also where MSO-backed groups gain the most. A shared services team can run lab intake for every affiliated practice on one standard.
How to set routing rules by specialty and provider
Once results are in one queue, routing determines whether they help or just move the pile.
Practical rules that work:
- Ordering provider first. If the result names an ordering provider in your group, route to that provider's inbox.
- Specialty fallback. If there's no ordering provider (common with hospital labs), route by result type and the patient's active specialties. A dialysis panel for a nephrology patient goes to nephrology, not to the patient's PCP.
- Location fallback. If both of the above fail, route to the patient's home location's clinical lead.
- Abnormal flagging. Abnormal values sort to the top of the provider's queue.
- Critical values follow your group's critical result policy, with direct notification to the provider or on-call clinician.
This matters for patient safety. A study in Archives of Internal Medicine found practices failed to inform patients, or to document informing them, about clinically significant abnormal results about 7% of the time, and outside results with no clear owner are an easy place for that to happen.
Governance: who owns abnormal outside results?
Automation handles filing and routing. It doesn't decide clinical ownership. Your group needs a written policy that answers:
- Who is responsible for reviewing an outside result that arrived without an order from your group?
- How quickly must abnormal outside results be acknowledged, by specialty?
- Who covers a provider's inbox during PTO?
- How are unmatched results (patients not yet registered) tracked so they don't disappear?
Write the policy down, get sign-off from each specialty lead, and revisit it when you add a location or a new specialty joins the group. Review exception-queue reports monthly. Recurring problems, such as one hospital that always omits date of birth, are usually fixable with a phone call.
What should a multi-specialty group measure?
A centralized lab results ingestion workflow should be run like any other operational service, with a short set of metrics reviewed monthly by location and by specialty.
- Time-to-chart. Median hours from when a result arrives to when it's filed. Break this out by channel, since portal results and fax results often behave differently.
- Unfiled backlog at end of day. The single best early-warning signal. If it rises, find out why before it compounds.
- Automation rate. The share of outside results filed without human touch. This should climb over the first few months as rules are tuned.
- Exception reasons. Unmatched patient, unreadable document, duplicate, missing ordering provider. The mix tells you what to fix upstream.
- Misfile rate. From periodic audits of a random sample of filed results. This should be close to zero; anything higher means patient-matching rules need tightening.
- Abnormal acknowledgment time. How long it takes a provider to acknowledge an abnormal outside result, by specialty.
Share these numbers with specialty leads. Cardiology and nephrology will care about different things, and giving each specialty visibility into its own outside-lab flow builds support for keeping the central model running. It also gives you the data to decide whether a high-volume source has earned an HL7 interface.
Where Honey Health fits
Honey Health builds AI agents for back-office work in multi-specialty groups, specialty practices, and MSOs. For non-interfaced labs, two agents cover the workflow above.
The fax triage agent reads every inbound fax across all your locations, splits multi-document packets, recognizes lab results, matches the patient, and files the result into the EHR under the right specialty-specific category with routing to the ordering provider. The data fetching agent logs into hospital and lab portals on a schedule and pulls new results into the same flow. Neither requires a new interface project for each lab, and both run alongside the HL7 interfaces you already have.
Frequently Asked Questions
When should a multi-specialty group build an HL7 interface for a lab?
When the lab sends high, steady volume across multiple specialties and will cooperate on the project. For most groups that's the main reference labs and perhaps one or two affiliated hospitals. Lower-volume or inconsistent sources are better handled with a lab results ingestion workflow.
How do you route an outside lab result with no ordering provider?
Use fallback rules. Route by result type and the patient's active specialties first, then to the patient's home location's clinical lead. For example, a dialysis panel for an established nephrology patient should go to nephrology, not default to primary care.
Should each specialty have its own document categories for outside labs?
Yes. Specialty-specific categories make results easier to find and route. A single "Outside Labs" folder across the group tends to bury important results and send them to the wrong inbox. Keep the list short and consistent across locations.
Can centralized lab intake keep our existing fax numbers?
It should. Senders shouldn't have to change anything. Each location's existing fax numbers can feed a central digital queue, and routing logic decides where each result goes based on content, not on which number received it.
What's the biggest risk in automating outside lab intake?
Misfiling a result under the wrong patient. Require multiple matching identifiers before automatic filing, route anything uncertain to a human, and audit a sample of filed results regularly during the first few months.
