The anatomy of a prior authorization cost

Line one: the visible cost

The AMA 2023 Prior Authorization Physician Survey reported that a physician and their staff spend, on average, twelve to thirteen hours a week completing prior authorizations. The average physician handles forty-three prior auths a week. Translated into operating cost at a fully loaded medical-assistant rate of roughly thirty-eight dollars an hour, that is between four hundred fifty and five hundred dollars in MA labor per physician per week, before management overhead.

At the practice level, that is twenty-three thousand dollars a year in pure prior-auth labor per physician. A ten-physician group is carrying a quarter of a million dollars in PA labor cost annually, before any of the second-order costs start. This is the line every administrator sees first, and it is the smallest one.

Line two: the patient cost

The AMA survey also reported that seventy-eight percent of physicians say prior authorization sometimes or often leads patients to abandon their recommended course of treatment. Twenty-four percent reported that prior auth has led to a serious adverse event in a patient they treat. Ninety-four percent reported that PA delays patient access to necessary care.

These numbers are not a marketing line. They are the clinical cost of a workflow that was built to control utilization but in practice controls access. For an independent practice, patient abandonment shows up downstream as no-shows, dropped care episodes, and ultimately patients who choose a different practice the next time they need care. It is a churn line item, and it does not show up on the P&L as PA-related, because by the time it shows up, the patient is gone.

Line three: the clinical cost

Pulling clinical staff into administrative work is the most expensive form of headcount substitution a practice can run. A medical assistant or registered nurse doing prior-auth submission is not doing rooming, vitals, patient education, or in-basket triage. The hour they spend on a PA is an hour the clinical encounter does not get. Across a ten-physician practice running thirteen hours of PA labor per physician per week, that is one hundred thirty clinical-staff hours redirected out of patient-facing work every week. The opportunity cost is the encounter volume, throughput, and clinical-quality improvement that does not happen.

This is the cost line that compounds. Practices that staff up to absorb the PA burden eventually run out of room in the building. They cannot hire more rooms. They cannot hire more clinicians. They can only hire more administrative bodies, and after a certain point those administrative bodies start to require their own management overhead. The work has reorganized the practice around itself.

Line four: the downstream revenue cost

A prior authorization that is delayed past the point of clinical urgency does not just cost the patient access; it costs the practice the revenue tied to the service. If the patient drops the therapy, the practice loses the visit revenue, the follow-up revenue, and any procedural or infusion revenue downstream. If the PA is denied and never appealed, the service is never billed. If the appeal succeeds but the service was already rendered without the auth in place, the claim is denied for missing-auth and enters the denial queue, where roughly forty percent of practice denials are never reworked at all.

MGMA cost-per-rework data puts a single denial recovery at roughly twenty-five dollars in operating cost. A PA-driven denial is almost always more expensive than that because the rework typically requires clinical documentation review, payer escalation, and sometimes a peer-to-peer call. Forty to sixty dollars per worked PA denial is closer to the real number, and that is on top of the upstream PA labor that produced the denial in the first place.

The mechanism, not the symptom

The reason this cost compounds is that prior authorization is a workflow problem, not a documentation problem. The default response, hire more MAs, build a PA tracker, contract a point software tool, treats it as a documentation problem. It adds capacity to the symptom and leaves the workflow alone. The cost ratio improves marginally and then stalls.

The change that actually moves the cost line is removing PA from the clinical staff's workload entirely. A dedicated, full-time team that owns the queue, knows the payer-specific quirks, runs the appeals, escalates the stuck cases, and reports the numbers weekly is operating on a different unit-cost curve than a rotating MA who handles PA when they have a free hour. The dedicated team is cheaper per PA, faster per PA, and produces denials at a lower rate. More importantly, it gives the clinical staff their hours back.

This is what Premier Medsolutions runs as a program. The PA team is not a software product. It is staffed operators who work inside the practice's EHR, own the queue under named accountability, and report turnaround and overturn rates the practice can read every week.