Imaging Before the Consult: Neurosurgery Referral Intake

Neurosurgery referral intake has one job that outranks all the others: get the images in the chart before the patient sits down. A consult without the actual study is a conversation about scheduling another consult. Everybody in the practice knows it, and it still happens every week, because the report arrives with the referral and the images arrive whenever the sending facility gets around to it.

The referral itself is usually complete enough. A letter, a report, sometimes a portal message. What is missing is the part that takes a phone call to a records department at a hospital that has never heard of your practice.

So the packet sits in a pending bucket while somebody means to call. The patient calls first, asking when they will be seen, and now your coordinator is handling an access complaint instead of chasing a study. The consult eventually gets booked without the images, because saying no to a waiting patient is harder than hoping.

The channel is part of the problem. In 2025, 40% of hospitals reported often using mail or fax to send a summary of care record, and 35% reported often using it to receive one, according to ONC tracking of interoperable exchange methods. Records movement between organizations is still largely a paper-shaped process, which means it is still largely a phone-shaped process.

The report is not the study

Intake coordinators learn this distinction in their first month. No system enforces it. A referral arrives with a written read attached, the chart looks populated, and the case gets marked ready. The surgeon opens it the morning of and there is nothing to look at.

athenaOne gives you a place to be precise about this. Documents carry classes, so an outside read and an imaging study are different objects in the chart rather than two attachments in the same pile. Intake can require both before a consult appointment type becomes bookable.

The automation reads the inbound referral, files each document under its class, and checks whether the study itself is present. When it is not, it opens a follow-up task, calls the sending facility’s records line, and keeps calling on a schedule rather than when somebody remembers. The patient’s consult stays in a pending state that is visible instead of implied.

Callers ask for one surgeon by name

In neurosurgery the referring physician and the patient often have a specific surgeon in mind, and that surgeon is frequently the one with the longest wait. This is a conversation-design problem rather than a flag on a chart.

A caller asking for a named surgeon needs language that offers the providers who actually have availability without sounding like a brush-off. Done badly it reads as a downgrade and the patient calls somewhere else. Done well it is a straightforward statement of what is open, from whom, and when.

Provider groups and appointment types in athenaOne already carry the answer. The AI knows which providers are bookable for that visit type in that department, offers the real options in one call, and books the requested surgeon when the patient prefers to wait. What it never does is guess who the patient should see.

Route on what the referral says, not what it implies

A neurosurgery practice usually runs several distinct service lines with different templates and different lead times. The referral names the requested service, and the referring physician has already made that call.

The routing rule is narrow on purpose. Read the requested service off the referral, match it to the appointment type and provider group configured for it, and book against that. Anything that does not map cleanly, or that names two services, or that arrives with no requested service at all, goes to a person.

This is the same boundary that governs referral intake from pulmonology and oncology in a thoracic practice. Automation moves paperwork against written instructions. It does not interpret them, and a vendor who tells you otherwise is describing a compliance problem rather than a feature.

A pending consult is not a booked consult

The tracking failure is treating a referral as handled once it has been entered. It has been entered. It has not been scheduled, the records are not in, and nobody has told the referring office anything.

More than three medical groups in four (76%) manage referrals in their EHR (66%) or in referral management software (10%), while about one in five (21%) still rely on manual tracking, according to MGMA Stat polling. Owning the tool is not the same as working the queue, and the queue is where consults go to age.

What closes the loop is status moving outward without a person typing it. The sending office learns the referral was received, learns what is still missing, and learns the date when there is one. That is the difference between a practice referring physicians trust and one they use when their first choice is full.

Where the intake work hands back

The AI does not decide who needs surgery, which service line fits a patient, or how urgent a referral is. It builds the chart, files the documents, chases the outside study, offers real dates, and reports status back to the referring office.

The exceptions are specific and they go to named people. A records department refusing the request because the authorization form is wrong goes to your records clerk. A referral naming a service your practice does not offer goes to your intake lead. A patient who wants the surgeon with a fourteen-week wait and needs to hear that plainly goes to a human, because that conversation should be had by one.

Your coordinator’s day changes shape. Less opening PDFs and dialing hospital records lines, more time on the handful of referrals that genuinely need a person.

Key Takeaways

Require the imaging study itself, not just the outside read, before a consult appointment type becomes bookable.
File inbound referral documents under athenaOne document classes so a report and a study are distinguishable objects rather than two attachments.
Run the outside-records chase on a schedule with a follow-up task attached, because it will never happen on memory.
Give callers who ask for one surgeon by name a scripted set of real options instead of an apology or a long wait.
Route referrals on the service the referring physician requested, and send anything ambiguous to a person.

Neurosurgery consults fail on logistics far more often than on medicine. The referral is complete, the study is somewhere else, and the phone call that would fix it competes with every other phone call. Separate the read from the study, chase the study automatically, offer honest dates when a caller asks for one surgeon, and push status back to the referring office without anyone typing it. The consult happens once instead of twice.

Sources