Prior Authorization for Cardiac Imaging Appointments

Prior authorization for cardiac imaging is usually described as a paperwork problem. In a cardiology practice it is a scheduling problem wearing paperwork’s clothes. The submission almost always goes out. What goes wrong is where the appointment sits relative to the window that comes back.

An authorization is not a permission slip. It is a date range attached to a specific service, and a study performed outside that range is unpaid work the practice already did.

The sequence that produces the failure is ordinary. A patient calls, a scheduler offers the first opening that fits the room and the technologist, the study gets booked, and the authorization request goes to the payer afterward. When the approval returns with a window that starts after the appointment date, or expires before it, nobody re-opens the appointment to check.

The same thing happens in reverse when a patient reschedules. The new appointment is not automatically attached to the existing authorization, and a practice that has not assigned someone to re-link them will not notice until the claim comes back.

Both failures are invisible on the schedule. The appointment looks completely normal.

The window, not the approval, is what gets missed

Practices that measure this find the leak is larger than they assumed, and that it is created by well-intentioned schedulers.

At a multi-site pain practice, a share of procedures were being scheduled outside the authorization window by human schedulers. Not because anyone was careless, but because the person booking the appointment was optimizing for the first available room and the person handling authorization was working a separate queue with its own backlog. Neither had the other’s information at the moment of the decision.

That is the structural point. Authorization and scheduling are two workflows owned by two teams, and the failure lives in the seam between them. Adding staff to either side does not close the seam.

The fix is to move the check to booking time. Before an appointment for a study that carries an authorization requirement is confirmed, the automation reads the existing authorization on the chart through GET /patients/{patientid}/referralauths, compares the approved date range to the slot being offered, and either confirms the fit or holds the booking for the authorization team.

Regulation set the clock, which makes the timing plannable

The timing side of this became more predictable rather than less, and most scheduling rules have not caught up.

Under the CMS Interoperability and Prior Authorization Final Rule, impacted payers must send prior authorization decisions within 72 hours for expedited requests and seven calendar days for standard ones, and must give a specific reason for denials. That is a planning input. A practice that knows the outer bound of a standard decision can set its booking offers against it instead of guessing.

The practical translation is a two-tier booking rule. A study that needs no authorization gets offered same-day or next-day availability. A study that does need authorization gets its earliest offer set far enough out that the decision can return before the appointment. Some practices generalize the rule further by plan type, because certain plan designs require authorization on nearly everything while others rarely do.

Rules like this are worth writing down explicitly, because staff get the underlying list wrong in both directions. Front desk teams have told patients a practice did not take their plan when it did, and told patients a referral was required when their plan did not require one. Both were caught only in call review.

The complication: rescheduling silently breaks the link

This is the failure practices are most surprised by, because nothing about it looks like an error.

When an appointment carrying an authorization is rescheduled, the new appointment is not automatically linked to the existing authorization. One multi-site practice solved it by assigning a human team to re-link them, which works and costs a headcount.

There is a second rule inside it that is easy to get backwards. If a patient reschedules an appointment whose authorization has not come back yet, the appointment can only be pushed later, never pulled earlier. There is no approved date range to schedule inside of, so an earlier slot is a guess that the payer may not honor.

An automation handling reschedules has to know both. On any move, re-read the authorization, check the new date against the approved range, and re-attach the link. When the authorization is still pending, offer only later dates and say why in plain language rather than presenting an arbitrary restriction.

The handoff belongs to the authorization team in three cases: the new date falls outside an approved range, the authorization has expired, or the payer approved a different service than the one now being booked. Each of those goes to a person with the authorization record and the appointment attached. None of them should be resolved by an automation choosing a date.

Expiring authorizations need a clock, and nobody owns it

The other half of the timing problem runs in the opposite direction.

For an ongoing condition, a specialist authorization might be valid for a year, and then someone has to notice that it lapsed and renew it. In most practices no one is assigned to that. It surfaces when a patient calls to book, the visit gets scheduled, and the claim is denied months later.

The authorization number also has to reach the patient, usually through the portal, and patients call about it when it does not arrive. That is administrative traffic generated entirely by a tracking gap.

Running this as a queue rather than a memory is straightforward. Read the authorizations attached to active patients, sort by expiration, and work the ones expiring inside a defined horizon before the patient calls. Renewal requests submitted ahead of a scheduled study are considerably cheaper than appeals filed after one.

The measurement that matters is not authorizations obtained. It is the share of authorization-requiring appointments whose approved window contains the appointment date, checked at the time of service rather than at the time of booking.

Where the phone time goes, and what that implies

There is a staffing argument here that usually settles the business case.

A March 10, 2026, MGMA Stat poll of 294 applicable responses asking practice leaders which phone tasks consume the most staff time put eligibility and prior authorization at 45%, ahead of scheduling at 31%, intake at 9%, and prescription refills at 6%. Nearly half the phone burden in a practice sits on the two workflows that decide whether a study gets paid.

Most of that time is not judgment. It is status checking: calling a payer portal or a phone line to find out whether a decision has come back, then recording the answer somewhere the scheduler will see it. That is exactly the kind of work that scales badly with headcount and well with automation, because it is repetitive, high volume, and has a clear correct answer.

What stays with people is the part that requires argument. Peer-to-peer conversations, appeals, and cases where the payer approved something other than what was ordered are handled by staff, with the automation supplying the assembled record rather than an opinion about it.

Key Takeaways

Treat prior authorization for cardiac imaging as a scheduling constraint. The appointment date has to fall inside the approved window, and that is checked at booking, not at submission.
Read the authorization on the chart before confirming a slot for any study that carries a requirement, and hold the booking when the dates do not fit.
Set a two-tier booking rule: no-authorization studies get near-term offers, authorization-required studies get an earliest offer that allows the decision to return.
Regulated decision timelines give you a planning bound. Set booking offsets against the standard decision window instead of guessing.
Re-check and re-attach the authorization on every reschedule. The new appointment does not inherit the link automatically.
When an authorization is still pending, only later dates can be offered. An earlier slot has no approved range to sit inside.
Run expiring authorizations as a sorted queue worked ahead of the patient’s call, not as something staff are expected to remember.
Measure the share of authorization-requiring appointments whose approved window contains the appointment date at time of service. Authorizations obtained is the wrong number.

Cardiac imaging concentrates a practice’s authorization risk into a small number of high-value studies, which is good news: the list of appointment types that need this discipline is short enough to automate precisely. The reason it usually is not automated is that the check has to happen at the moment of booking, inside a conversation, using information that lives in a different team’s queue. An AI team working inside athenaOne can read the authorization record before offering a slot, keep the appointment and the approved window aligned through every reschedule, work the expiring authorizations before patients call about them, and route the cases that need a human argument to the people who make it.

Sources