Winning Cardiology Referrals by Being Easy to Refer To

Cardiology referrals rarely move because of a relationship between physicians. They move because a coordinator in a primary care office has four practices she can send to and one of them makes her day easier. She is not comparing outcomes. She is comparing how long it takes to get a patient booked and how often she has to call to find out what happened.

Most cardiology groups compete on the wrong thing here. They invest in physician outreach and lunches while the actual decision gets made by someone who has never met the cardiologist and is working from a note that says which practice calls back.

The referral is a work object, not a document

The mental model that fixes most of this is treating an inbound referral as a piece of work with a defined end state rather than a document to be filed.

A referral is complete when the patient has an appointment on the calendar, the required records are attached, coverage is verified, and the referring office knows all of that. Filing the PDF into the chart accomplishes none of it, and most practices measure the filing.

That framing also tells you what to automate. Every step between arrival and booking is administrative, repetitive, and has an unambiguous definition of done. Match the referral to an existing chart or create one, read what the referring office sent, identify what is missing, request it, verify coverage, call the patient, book, and report back. None of that requires clinical judgment and all of it currently competes for a coordinator’s attention with the phone.

Cardiology books in pairs, and nothing enforces the pairing

Here is the complication that makes cardiology referral intake harder than it looks, and it is the one most likely to be handled by convention rather than by any rule in the system.

A large share of cardiology visits are an imaging study plus a provider visit, booked together with specific spacing between them. Nothing in the scheduling template enforces the pairing. It exists as practice knowledge, and the patient calls asking to see the doctor without knowing there are two appointments involved.

The failure mode is booking one leg and not the other. The second half silently does not exist until a human catches it, usually in a schedule review the day before, sometimes never. The patient arrives for a study with no follow-up booked, or for a consult without the study that would make the visit useful.

There is a second trap sitting directly on top of it. Native reminders commonly fire on the chronologically first appointment only. A patient with a 9:30 imaging slot and a 10:00 provider visit gets reminded about one of them and misses the other. The practice records a no-show and reads it as a patient reliability problem when it is a reminder configuration problem.

Automation handles both cleanly because both are rule shaped. Book the pair as a pair, refuse to consider the referral booked when only one leg exists, and confirm both appointments explicitly in every reminder rather than trusting the default.

Speed to first contact is the whole game

Whatever else a practice fixes, the interval that decides referral share is how long the patient waits for someone to call them.

Access pressure is real and uneven. A July 2025 MGMA poll found that 66% of medical groups reported new-patient wait times had stayed the same at 40% or shortened at 26% compared with the prior year, while 31% said wait times had increased. The same analysis flagged internal referral leakage as a common failure in multi-specialty settings, where available slots get absorbed by easier-to-schedule follow-up visits and true new patients get crowded out.

That is worth sitting with, because it describes a practice losing referral volume through its own scheduling behavior rather than through demand. If new-patient capacity is not protected, referral marketing just increases the number of people waiting.

Someone calling the patient beats the patient calling you

The other half is who does the calling. The ASPN referral study, following 776 referred patients across 81 practices, identified scheduling of the specialty appointment by the referring physician or staff as a positive predictor of the patient completing the referral.

Whichever side does it, a practice that calls the patient converts more referrals than one that hands out a phone number. For a referred cardiology patient who has just been told they should see a specialist, the difference between being called and being expected to call is often the difference between a booked appointment and a referral that quietly expires.

What the referring office should never have to do

Three things drive coordinators to a competitor, and all three are inside your control.

Retyping information you already have. If the referral arrived with demographics, do not ask for them again. Pull what exists and ask only for the gaps.

Guessing what you need. Publish the requirement per referral type: which records, which imaging, what coverage information. A coordinator who knows the list can send a complete referral the first time, which is cheaper for both offices than a chase.

Calling to find out what happened. Push status back without being asked. This is the single most valuable item on the list and it costs almost nothing once the intake is instrumented.

Where a person takes over

A referral that does not match what the practice does goes to a person. The automation checks it against the practice’s own list of accepted services and provider assignments, and flags a mismatch with the reason attached. It does not assess clinical appropriateness and should not be built to.

Anything the referring office marked urgent goes to clinical staff immediately, without the automation interpreting what urgent means in that context. The AI moves the paperwork and the booking. Clinicians decide the care.

A patient who cannot be reached after a defined number of attempts becomes a person’s problem, and the referring office should be told rather than left assuming their patient was seen. Unreachable is information. Silence is not.

Key Takeaways

Treat the referral as a work object with an end state: booked, records attached, coverage verified, referrer informed. Filing the PDF is not completion, and most practices measure the filing.
Book paired imaging and provider visits as a pair, and refuse to call the referral booked when only one leg exists. Convention does not survive a busy Monday.
Confirm both appointments in every reminder. Native reminders that fire on the first appointment only manufacture no-shows that get blamed on patients.
Protect new-patient capacity explicitly. Referral growth against unprotected slots just lengthens the queue.
Call the patient rather than handing them a number. Practice-side scheduling is a documented predictor of referral completion.
Publish your requirements per referral type so a coordinator can get it right the first time, and pull the data you already have instead of asking twice.
Specialty mismatches, anything marked urgent, and unreachable patients all go to a person, and the referring office gets told.

The practice that wins cardiology referrals is usually not the one with the best outreach program. It is the one where a coordinator across town knows the patient will get called within a day, both appointments will be booked correctly, and she will not have to chase anybody to find out what happened. That reputation is built out of unglamorous administrative work, and it compounds quietly in a way physician relationships alone never do.

Sources