In this guide
How big is the PA problem, really#
The AMA's 2023 prior authorization physician survey put the volume at ~43 PA requests per provider per week, consuming an average of ~12 hours of physician + 14 hours of staff time. 94% of physicians reported PA-related care delays, and 24% reported a serious adverse event attributable to PA. 1 The problem is measurable and growing — PA volume has roughly doubled since 2014. 3
From a pure economics standpoint, the cost is not the denied claims (about 80% of denials are overturned on appeal, per KFF). 2 The cost is the 25+ FTE-hours of clinical and administrative time per provider per week — that is the cost a vendor's ROI math has to beat.
What automation actually does#
Modern PA automation falls in three buckets: 1. Payer-portal RPA / browser bots — fill the payer's web form from EHR data, log the case ID, monitor status. Coverage is broad but brittle; portal changes break workflows weekly. 2. ePA integrations (Surescripts, CoverMyMeds, Availity, MyAvanza) — direct payer API where supported. Coverage is narrower (varies by payer and code) but reliable when it works. 3. AI-assisted denial appeals — drafts appeal letters from chart context, citing payer policy. The 80% appeal-overturn rate is the lever. 2
Realistic touch-time reduction is 40–70% on the codes and payers the tool covers. The honest variable is coverage. A tool that covers 90% of your PA volume produces step-change savings; a tool that covers 40% adds a parallel workflow your staff still hates.
When the ROI math works#
Reference math for a 5-provider practice at $0.85 fully-loaded labor cost per minute: • PA volume: 5 × 43 = 215 PAs/week → ~11,200/year. • Average touch-time at baseline: ~28 minutes per PA → ~5,200 staff-hours/year → ~$265K labor cost. • 50% reduction with automation on 75% of volume = ~$100K/year labor recovered. • Vendor cost: $30K–$80K/year for mid-market tools, depending on volume and EHR. • Net: $20K–$70K/year, plus 3–7 day faster time-to-decision (revenue acceleration) and material provider time freed.
Below ~$120K of annual PA labor cost — roughly 2.5 staff FTE on PA — most vendor tools don't pay back inside 12 months once you factor implementation and integration. Above ~$200K of PA labor cost, it's almost always the highest-ROI ops investment available.
Where automation doesn't pay back#
Three patterns: (1) practices on EHRs the vendor doesn't integrate with deeply — manual data re-entry erases the savings; (2) specialties whose PA load is concentrated in 2–3 unusual codes the vendor doesn't cover well (ask for the exact code/payer coverage map before signing); (3) practices that haven't first done the free wins — a payer-by-payer PA-rule cheat sheet, a denial-pattern review, and reassigning PA work from clinicians to a single dedicated coordinator.
The CMS Interoperability and Prior Authorization Final Rule (CMS-0057-F, effective in waves through 2027) will require Medicare Advantage, Medicaid managed care, and ACA-marketplace plans to support API-based ePA and publish PA metrics publicly. This will compress vendor differentiation over the next 18–24 months — factor a shorter useful-life into any multi-year contract. 4
How to run a real evaluation#
Six steps: (1) baseline your current PA volume by payer and code for 30 days; (2) ask each vendor for a written coverage map against that list; (3) pilot one payer or one service line for 60 days before signing a multi-year contract; (4) measure touch-time reduction, time-to-decision, and overturn rate, not just denial rate; (5) require a clean-exit clause; (6) reassess against the CMS-0057-F rollout timeline annually.
For a clinic-specific projection of PA labor cost and vendor ROI, run the Staff utilization report.
Run this for your clinic
Reading is free. The dollar figure for your practice requires a qualified diagnostic report.
Qualified full audits start at $1,500 per location. Refunded if quantified opportunity is under five times the fee; opportunities.
Written by
Jack Gierlich· Founder, Clear Median
Founder of Clear Median. Has built and reviewed diagnostic reports for independent primary care, specialty, and aesthetic clinics across the US. Background in operations and revenue analytics for ambulatory practices.
Reviewed by
Clear Median Analyst Team
Multi-specialty analyst team that builds the underlying benchmarks and stress-tests every recommendation against MGMA, CMS, and internal peer data before delivery.
How we write these →Frequently asked
- Will the CMS interoperability rule make this problem go away?
- It will reduce friction on the payers it covers (MA, Medicaid managed care, ACA marketplace), starting with API standards in 2026 and full effect in 2027. Commercial and traditional Medicare are not currently covered. Expect 30–50% volume relief on covered payers, not a full solution. 4
- What about AI tools that promise to handle the entire PA?
- AI-assisted drafting and appeal generation works well for the 80% of denials that are reversible. End-to-end submission with no human review is still risky for medical-necessity edge cases; mandate a human reviewer in the loop and audit a 5% sample monthly.
- Is gold-carding worth pursuing?
- If your top payer offers it (UnitedHealthcare's program is the most widely deployed), absolutely — it removes PA for high-performing providers on specific codes. The application work is real but one-time, and the workflow savings compound. 5
Sources & references
5 sources
- 12023 AMA Prior Authorization Physician Survey· American Medical Association
- 2Use of Prior Authorization in Medicare Advantage 2023· Kaiser Family Foundation
- 3Prior Authorization Trends in Medicare Advantage· Health Affairs
- 4CMS Interoperability and Prior Authorization Final Rule (CMS-0057-F)· CMS
- 5UnitedHealthcare Gold Card Program· UnitedHealthcare
Reimbursement figures cite CMS-published rates at the time of writing and are not locality-adjusted. Clear Median guides are educational and not legal, billing, or medical advice. Confirm any code or rate against your contracts and current CMS fee schedules before acting.