For primary care groups

Know whether every referral reached the next step.

When a referral order leaves a primary care chart, the work has only started: honoring the patient's and provider's choice of specialist, sending complete records, reaching the patient, confirming the visit happened, and bringing the note back. Linear Health is operational AI that runs that closed loop on top of your existing EHR, with your staff owning every exception, so the chart can answer its own question.

95% completion at Aunt Martha's, a published FQHC customer
Primary care chartYour EHR, unchanged

Specialist referral ordered

Referral receipt

  • MatchPatient and provider choice honored
  • SentComplete, with records, in network
  • PatientReached, visit confirmed
  • ReturnedStatus and consult note filed in the chart
Illustrative workflow

Between the order and the visit sits a black box.

The EHR records that a referral was ordered. Everything after that happens outside the record. Closing the loop means running that outside work to completion and writing the result back in.

A referral loop is closed when the originating practice has confirmation that the patient was seen and the consult note or status is back in the chart. Closed-loop coordination is everything between those points: match, send, contact, schedule, track, and write back.

  1. The order leaves, preference first

    The patient's and provider's chosen specialist always comes first. Linear searches for an in-network, soonest-available alternative only when no preference is set or the choice is not workable, and your practice can keep booking in staff hands entirely. The full mechanics live on the outbound referral coordination page.

  2. Verified, then sent complete

    Eligibility, network status, and any required authorization are confirmed before the appointment is made, and the referral goes out structured with the records the specialist actually needs. Out-of-network surprises stop at this step instead of surfacing as denials months later.

  3. The patient is reached in minutes

    Outreach runs by text, email, and voice AI in English and Spanish, with first contact in about 5 minutes instead of the 3-7 days a manual queue takes. The visit lands on the specialist's calendar through scheduling that books real availability and protects the slot with reminders.

  4. A snag gets a named owner

    A patient unreachable after the full outreach sequence, incomplete records, an unclear eligibility answer, a declined referral: each pauses the automation and moves to a coordinator with the reason attached. Nothing exits the queue silently, and resolution puts the referral back on the automated path.

  5. Seen, and confirmed seen

    Linear follows the referral to the visit and confirms the patient was actually seen, so the answer to “did they go?” is on the screen as it happens, not a mystery until the next annual.

  6. The receipt docks back into the chart

    The consult note is pursued, captured, and filed in the right place in the originating record, next to the appointment status. The loop is closed where the next clinician will look for it.

Primary care chart
  • Specialist referral ordered
  • Patient reached, visit confirmed
  • Consult note returned and filed

Afterward, the black box stays open to inspection: every referral in flight has a state, an owner, and an age your team can query across sites, instead of living in one coordinator's memory.

Coordinators stop chasing every order and work the short exception list that genuinely needs a person.

Made for groups where referrals are a volume problem.

The loop earns its keep in multi-site primary care groups sending to broad external specialist networks, in centralized referral teams that should be working exceptions rather than chasing every order, and in organizations accountable for outcomes, where referral completion and documentation are operational requirements rather than courtesies.

Not the right fit: solo practices with a handful of referrals a week, or hospital enterprise deployments. The loop pays for itself on volume.

Measured at Aunt Martha's Health & Wellness

Evidence from the loop at scale.

Aunt Martha's is a 100-provider Federally Qualified Health Center and a published Linear Health customer. Its referral operation runs the same outbound loop this page describes, across 35 sites.

95%Referral completion, up from Aunt Martha's own 35% baseline
20 to 2Coordination FTEs, redeployed to patient-facing work
10,000+Referrals and coordination events per month

“Linear Health automated our referral operations end to end. We process tens of thousands of referrals every month across Athena and multiple payer portals without manual work.”

Leadership team, Aunt Martha's Health & Wellness

Scope, baseline, and method: the Aunt Martha's case study. The 35% starting point is Aunt Martha's own. Industry-typical referral completion runs 50 to 65%; the Institute for Healthcare Improvement and National Patient Safety Foundation report that up to half of referrals are never completed. Results reflect this customer's volume and staffing.

What changes, seat by seat.

Four views of the same returned receipt.

Referral coordinators

The chase queue becomes an exception queue. Instead of spreadsheets, phone tag, and asking offices to fax notes, coordinators resolve the short list of judgment calls and let the automation carry the rest.

Care teams

The chart answers its own question before the follow-up visit: appointment status and the consult note are where the next clinician expects them, without anyone re-keying a fax.

Practice leaders

Referral flow becomes measurable across sites: volume, completion state, exception reasons, and aging, with one shared definition of done instead of a per-site dialect.

Patients

Contact comes quickly, in their channel and language, and someone follows through until the visit happens. The referral stops being homework the patient has to manage alone.

Frequently asked questions

Direct answers for referral coordinators and medical directors evaluating the loop.

How does a primary care practice know whether the patient saw the specialist?

Linear Health follows each outbound referral past the send: it confirms the appointment, verifies the patient was seen, pursues the consult note, and returns the result to the originating chart. Coordinators see live status instead of discovering the outcome at the patient's next visit.

What does closed-loop referral coordination mean?

A referral loop is closed when the originating practice receives confirmation of the outcome: the patient was seen and the consult note or status is back in the chart. Closed-loop coordination is the operational work between those two points, matching, sending, contacting the patient, scheduling, tracking, and writing back, handled so no referral exits the workflow unresolved.

Who chooses the specialist, the AI or our providers and patients?

Preference comes first. Linear honors the specialist your provider or patient chose, and only searches for an in-network, soonest-available alternative when no preference is set or the chosen option is not workable. Your practice can also keep everything except booking and have staff place the appointment.

What happens when a patient cannot be reached or records are missing?

The referral does not vanish. It moves to an exception queue with a named owner and a visible reason: unreachable after the outreach sequence, missing records, an eligibility question, or a canceled appointment. Staff resolve the judgment call and the automation resumes from there.

Does referral status write back into our EHR?

Yes. Appointment status and the retrieved consult note are written back to the correct place in the originating chart. athenahealth, eClinicalWorks, NextGen, Greenway Health, Veradigm, and Epic are established integrations, connected through standard health-data interfaces.

Does this work for multi-site groups with a centralized referral team?

That is the operating model it is built for. All sites feed one coordination layer with shared status definitions, so a centralized team works one exception queue instead of per-site spreadsheets, and each site keeps visibility into its own patients.

How long does implementation take, and what does it cost?

Go-live is 4 weeks on top of your existing EHR, with no migration. Pricing is $1,000-$8,000/mo, usage-based, month-to-month, and our customers typically see a 3:1 return within 90 days, measured against their own baseline.

Send a referral in the demo. Watch it come back.

Bring your own workflow: your EHR, your specialist network, your payer mix. Go-live is 4 weeks, on the EHR you already run, and customers typically see a 3:1 return within 90 days.

Stay updated

Get the latest on AI healthcare coordination.