Practice Operations
Vascular Prior Authorization and the Appointment Window
Vascular prior authorization fails on dates, not paperwork. A procedure booked outside its window is a denial the practice created weeks before the claim.
Vascular prior authorization is usually described as a paperwork problem, and that description is why practices keep losing money to it. The authorization is rarely missing. It exists, it was approved, and the procedure was still booked on a date it does not cover. The denial arrives six weeks later, addressed to the billing team, for a mistake that was made at the front desk on a Tuesday.
A procedure-heavy specialty has a scheduling problem that most practices treat as a revenue cycle problem, and the gap between those two views is where the money goes.
At one multi-site procedural practice, roughly 6% to 8% of procedures were being scheduled outside their authorization window by human schedulers. Nothing about that is carelessness. A scheduler is holding a patient on the phone, an open slot, a provider preference and a set of dates in a different system, and the dates are the part that loses.
The internal framing that helps is to think about revenue cycle backwards. Get the appointment to land inside the authorized dates at the moment of booking, and the claim works at checkout. Fix it afterward and you are working a denial, which is the most expensive form of the same task.
MGMA’s January 2026 poll put denials and appeals at 48% of the biggest revenue cycle leaks, ahead of front end issues at 23%, billing and collections at 14%, coding at 13% and charge posting at 2%. A large share of that top category is not adjudication. It is dates that did not line up.
The window is a scheduling constraint, so it belongs in the scheduling step
Authorizations have start dates, end dates and unit counts. Appointments have dates. Everything hard about this specialty’s revenue cycle lives in the join between those two things, and in most practices that join is a person’s memory.
GET /patients/{patientid}/referralauths is where the authorized dates live, and reading them before offering a time is the entire intervention. The times offered to the patient are the times inside the window, and no others. That is not a clever workflow. It is just refusing to display dates that will not be paid.
Coverage detail matters alongside it. GET /patients/{patientid}/insurances/{insuranceid}/benefitdetails answers whether this plan requires authorization for this service in the first place, which practices get wrong in both directions. One practice’s staff told a patient with a preferred provider plan that a referral from their primary care physician was required when the plan did not require one. Another told a patient the practice did not accept their coverage when it did. Both were found in call review.
The rule that experienced practices settle on is a booking lead time keyed to plan type. Where authorization is not needed, book near term. Where it is, hold the earliest offered date far enough out that the authorization team can actually work it. Managed care plans get the delayed treatment as a class, because they essentially always require it.
A reschedule silently breaks the link
This is the failure that surprises people, and it is worth checking whether it is happening in your own practice this week.
When an appointment carrying an authorization is rescheduled, the new appointment is not automatically linked to the existing authorization. One practice’s answer was a dedicated human team whose job is re-linking. That is a real cost, paid every day, for something a system should carry.
There is a second rule underneath it that schedulers have to know. If a patient reschedules an appointment whose authorization has not come back yet, the appointment can only be pushed later, never pulled earlier, because there is no authorized date range to schedule inside of.
PUT /patients/{patientid}/referralauths/{referralauthid} is how the link gets maintained rather than rebuilt by hand. The behavior that matters is simple to describe: when the date moves, re-check the window, re-link the authorization, and if the new date falls outside it, do not complete the reschedule quietly. Flag it to the authorization team before the patient hangs up, while an extension is still a phone call rather than an appeal.
Authorizations expire, and expiry has no owner
The other date nobody watches is the far one.
Referral authorizations for an ongoing condition are frequently good for a year, and then somebody has to notice. In practice nobody does, because the expiry lands in the middle of a quiet period between visits and there is no queue that surfaces it. The patient turns up for a follow-up procedure and the coverage that was fine in March is not fine in April.
This is the cleanest possible automation target. It is a date, in a field, with a known consequence, and the remedy is a renewal request submitted before the date passes. Reading the authorization list on a schedule and acting on what is about to lapse costs nothing and prevents a category of denial entirely.
The adjacent case is the multi-appointment structure this specialty runs on. At one vascular group close to 90% of visits are an imaging study paired with a provider visit, and nothing in the scheduling template enforces the pairing. The rule existed only as convention, visible in the booking history. On a live test call, an agent booked both halves for one body region but the follow-up task that should have been created was not, so the second half of the workflow silently did not exist until staff caught it in review.
When a linked pair carries an authorization, both halves have to sit inside the same window. Booking one and losing the other is how a partially authorized episode gets billed.
The payer clock is getting shorter, which helps
The regulatory direction of travel favors practices that have their dates organized, and it is worth planning around.
Under the CMS interoperability and prior authorization final rule, impacted payers are required to send prior authorization decisions within 72 hours for expedited requests and seven calendar days for standard requests, and beginning in 2026 they must give a specific reason for a denial.
A shorter and more predictable payer response changes what a practice can commit to at booking. A two-week buffer built for an unpredictable payer is capacity the practice was holding back, and part of it can be released once the response time is dependable.
The specific denial reason matters even more. A reason code that names the problem is a machine-readable input, which means the resubmission or the appeal can be prepared without a person reading a letter and guessing. Practices that are already reading authorization data structurally will get value from that on day one. Practices working from a spreadsheet will not.
What the authorization team keeps
Nothing described here submits an authorization or argues one. That distinction is the whole scope boundary and it is easy to audit.
The automation reads authorized date ranges before offering appointment times and refuses to offer dates outside them. It re-links the authorization when a visit moves, and flags a move that lands outside the window before it is confirmed. It watches for authorizations approaching expiry, queues the renewal, checks whether a plan requires authorization at all, and keeps both halves of a linked visit pair inside the same window.
The authorization team still owns the submission, the clinical documentation that supports it, the peer conversation when a payer pushes back, and every appeal. Those are judgment calls with a licensed professional behind them.
What changes is the volume. Most of what an authorization coordinator does in a procedural practice is not submission. It is chasing dates across two systems, and that is the part a practice should stop paying people to do.
Key Takeaways
- Read the authorized date range before offering appointment times, so the schedule cannot produce a date the payer will not pay for.
- Set booking lead time by plan type, holding the earliest offered date far enough out for the authorization team where authorization is required.
- Re-link the authorization every time an appointment moves, and stop the reschedule when the new date falls outside the window.
- Remember that an appointment whose authorization has not returned can only be pushed later, never pulled earlier.
- Queue renewals against authorization expiry dates, since a lapse between visits has no natural owner in the practice.
- Keep both halves of a linked imaging and provider visit inside the same authorized window, because the pairing is convention rather than an enforced rule.
- Expect payer decisions within 72 hours expedited and seven calendar days standard, and release some of the buffer capacity that unpredictability used to require.
A vascular practice does not lose authorization revenue in the billing office. It loses it at the moment of booking, on a date that looked fine to everyone involved. Putting the authorized window in front of the person picking the date is the cheapest denial prevention available.
Related reading
- linked visits and the pairing nothing enforces
- referrals arriving across five different inboxes
- the prior authorization letter packet
Sources
Ready to See It in Action?
See how PGA keeps procedures inside their authorization window and re-links them after a reschedule
Schedule a Demo →Written by Kevin Henrikson