What Prior Authorization Actually Costs a Practice in Staff Time

A number practices already have

The American Medical Association runs an annual survey of physicians on the administrative burden of prior authorization. In its most recent results, physicians and their staff completed an average of about 39 prior authorization requests per physician per week. Staff time spent on those requests averaged around 13 hours per physician per week.

Those two figures are usually cited as evidence that prior authorization is a burden. They are also enough to build a real cost estimate for one practice, using payroll figures the practice already has on hand.

The calculation

The AMA figure is a national average, but the arithmetic behind it applies to a practice of any size. The formula has four inputs:

• Hours of staff time spent on prior authorization per physician per week. The AMA average is 13; a practice can also time its own staff for a week or two to get a local number.
• Fully loaded hourly cost of the staff who do that work: wages plus payroll tax, benefits, and overhead, not just the wage printed on the paycheck.
• Number of physicians, or other prescribing providers, whose orders trigger prior authorization requests.
• Number of working weeks per year, typically 48 to 50 after holidays and time off.

Multiply the four together: hours per week, times hourly cost, times number of providers, times weeks per year.

As an illustration with round numbers only: 13 hours per week, a fully loaded staff cost of $28 per hour, four physicians, and 48 working weeks comes to 13 × $28 × 4 × 48, or about $69,888 a year.

That figure covers only the staff hours identified in the AMA survey. It leaves out the cost of denied claims, delayed care, and the physician time spent on peer-to-peer review calls, which the AMA tracks separately and which add to the total for many practices.

The same formula scales in either direction. A solo practitioner with one part-time referral coordinator handling prior authorization on the side is absorbing a smaller total, but likely a larger share of that person's week than the average suggests. A twenty-provider multispecialty group can multiply the same 13 hours across twenty providers and will often find the number large enough to justify a dedicated prior authorization role, rather than spreading the work across several front-desk and clinical staff who each lose a few hours to it.

Where the 13 hours actually go

The AMA's survey work breaks the weekly burden into a small set of repeating tasks, most of them performed by the same staff members for every payer, every plan, and often every renewal of a chronic prescription:

• Fax and portal submission. Requests go through payer portals or fax rather than a shared standard, so a staff member re-enters the same clinical information in a different format for each payer.
• Callbacks and status checks. Staff follow up by phone when a portal shows no status update, often more than once per request.
• Appeals of prior-authorization denials. Denials tied to prior authorization are up roughly 31% across both commercial and Medicare Advantage plans. Each denial that gets appealed adds another round of documentation gathering and submission on top of the original request.
• Resubmissions. Requests get returned for missing information, an incorrect code, or a documentation format the payer's system did not accept, and the same information has to be prepared and sent again.

None of these tasks require clinical judgment. They are procedural: find the form, find the fax number or portal, enter the data, wait, follow up. That is why the burden is measured in staff hours rather than physician hours, even though a physician's order is what starts the process.

Why the number keeps showing up

The AMA survey is not a one-time finding. Medical Economics, AJMC, and AHA News have covered the same rising burden through 2025 and 2026, with request volume and staff time holding steady or increasing rather than declining. The 31% rise in prior-authorization-related denials points in the same direction: payers are not only requiring more prior authorizations, they are approving a smaller share of them on first submission, which pushes more of the 13 weekly hours into appeals instead of original requests.

For a practice planning staffing or budgeting for revenue cycle work, this matters because it rules out waiting for the problem to shrink on its own. Three years of independent reporting show the trend flat or worse, not improving. A practice that measured its prior authorization cost last year and did nothing about it should expect this year's number to be the same or higher, not lower.

The staff time also does not move evenly across a practice's payer mix. A practice that tracks its prior authorization volume by payer usually finds that a small number of plans generate a disproportionate share of the requests and denials, which means the 13-hour average can hide a much larger burden concentrated on whichever staff member handles those specific payers.

What to baseline before evaluating any automation tool

A growing share of practices are looking at automation for this work. In MGMA Stat polling, over half of practices surveyed said they plan to apply AI tools to revenue-cycle tasks going forward, a category that commonly includes prior authorization. Before evaluating any tool in that category, a practice gets more useful vendor conversations by measuring its own baseline first, rather than relying on the AMA's national average. That means tracking, over at least four to six weeks:

• The actual count of prior authorization requests submitted per week, broken out by payer and by service line.
• The actual hours logged against prior authorization work, separated from other revenue-cycle tasks the same staff perform.
• The first-pass approval rate compared with the rate after appeal, so a vendor's claimed improvement has a real starting point to be measured against.
• The fully loaded hourly cost of the staff doing the work, calculated the same way as above.

Without those four numbers, a vendor's claimed time savings or denial-rate improvement has nothing local to be measured against. With them, the same calculation used earlier to estimate the current cost can be rerun after any process change, automated or not, to check whether the hours actually moved.