Lab result routing so abnormal values reach a clinician first

Results arrive from reference labs and in-house analyzers, and the workflow sorts them to the ordering clinician or covering provider, puts abnormal and critical values at the top, drafts a patient-friendly summary for the clinician to edit, and tracks that every result is acknowledged.

01the problem and who owns it

Provider inboxes mix lab results, refill requests, and messages in one long list. Results for a provider on vacation sit unread, faxed reports from outside labs get scanned without being matched to an order, and abnormal values look like everything else.

The medical director owns result follow-up as a patient safety process. Under information blocking rules, patients may see results in the portal promptly, sometimes before their clinician has reviewed them, which raises the cost of slow routing.

02what the AI does, step by step

1. Receive and parse resultsStructured results arrive as HL7 v2 ORU messages or FHIR DiagnosticReport and Observation resources. Faxed or PDF reports go through extraction and are matched to an open order.
2. Match to patient and orderEach result is linked to the patient and the originating order. Results that cannot be matched with confidence go to a reconciliation queue, never to a guess.
3. Apply flagging rulesLab-reported abnormal and critical flags, plus thresholds your clinicians define, set priority. The model does not decide clinical significance; it applies written rules and the lab's own flags.
4. Route to the right personResults go to the ordering clinician, or to the covering provider when the schedule shows the orderer is out. Critical values also trigger the practice's existing phone notification policy.
5. Draft the patient messageFor normal results, the model drafts a plain-language note using clinician-approved templates. The clinician edits and sends it, or chooses to call instead.
6. Track acknowledgmentEvery result has an owner and a due time. Unacknowledged results escalate to the covering provider and then the medical director.

03systems it connects to

• Lab interfaces. Reference lab connections through an interface engine such as Mirth (NextGen Connect) or the EHR's built-in lab integration.
• EHR. Results inboxes and orders in Epic, athenahealth, eClinicalWorks, or similar.
• Fax and document intake. Cloud fax and document extraction for reports that arrive outside the interface.
• Provider schedule. To know who is covering when the ordering clinician is away.

04human checkpoints

• Clinician reviews every result. Interpretation and next steps are always a licensed clinician's decision; the workflow only sorts and drafts.
• Flagging rules approval. Thresholds and priority rules are set and periodically reviewed by the medical director.
• Unmatched results. Medical records staff resolve results that cannot be confidently matched to a patient and order.

05what to measure

• Time to acknowledgment. Separately for critical, abnormal, and normal results.
• Unacknowledged results past due. Count and age, reviewed daily.
• Unmatched result volume. Results needing manual reconciliation, by source lab.
• Patient notification time. From result received to patient informed.

06risks and guardrails

• Missed critical values. No automation replaces your critical value callback policy. Keep the phone path, and alert on any interface outage so results do not silently stop.
• PHI and access. Results are among the most sensitive data a clinic holds. Use BAAs with every vendor, role-based access to queues, minimum necessary data in drafts, and audit logs on every view.
• Over-reliance on drafts. A well-written draft can make a clinician skim. Keep drafts short and show the source values beside them.

07build vs buy

Your EHR's results inbox and routing rules are the first stop, and many clinics only need them configured properly, with coverage rules kept current.

Custom development pays off when results arrive from many labs and formats, when faxed reports are a large share, or when you need acknowledgment tracking and escalation the EHR inbox does not provide.

08related playbooks

Browse every healthcare clinic operations playbook or the full library.