off label.

Prior authorization

How often are prior-authorization denials wrong?

Medicare Advantage denied 6.4 percent of prior authorization requests in 2023, but 81.7 percent of the denials that were appealed got overturned, so most denials that patients actually challenged were wrong.

Question

The problem

Prior authorization has become a national utilization-management layer that shifts clinical time, administrative cost, and access risk onto hospitals, physicians, and patients. The strategic problem is not only denial volume, but weak accountability when denials are overturned, unappealed, or driven by documentation process rather than clinical appropriateness.

The recommendation

Reposition prior authorization as a targeted clinical-control tool rather than a default administrative hurdle. The recommended program is to require plan-level denial and overturn transparency, gold-card low-risk providers, standardize documentation, and audit service lines where overturned denials signal systematic overuse.

How often

Size the denial machine and its direction, the raw rate plus the absolute volume it hides.

53M
Medicare Advantage prior-authorization determinations in 2024
Up from 49.8 million in 2023.
4.1M
of those determinations were denied in 2024
A 7.7 percent denial rate across 53 million requests.
7.7%
Medicare Advantage prior-authorization denial rate in 2024
Up from 6.4 percent in 2023, reversing the prior-year dip.

Medicare Advantage prior authorization denial rate by year

Share of prior authorization requests denied by Medicare Advantage plans. Only Medicare Advantage is publicly reported at this granularity. Commercial and Medicaid managed-care series are not available. The 2024 point of 7.7 percent, from KFF's 2024 determination data, reverses the 2023 dip.

Read it this way Trace the line year by year: the denial rate climbed from 5.7 percent in 2018 to a peak of 7.4 percent in 2021 and 2022, dipped to 6.4 percent in 2023, then rose again to 7.7 percent in 2024. The single 2023 dip does not establish a lasting improvement, since the rate reversed the very next year. Use this chart to see whether the authorization burden is volume, correctness, appeal friction, or harm, and how that evidence supports targeted reform rather than blanket approval or blanket denial.

0.0% 2.5% 5.0% 7.5% 10.0% 2018201920202021202220232024 MA denial rate DENIAL RATE (%)
Payer typeYear range
⊞ data table⬇ CSV
YearMA denial rate %
20185.7
20196.2
20206.9
20217.4
20227.4
20236.4
20247.7

KFF, Medicare Advantage prior authorization determinations, 2023 and 2024 · 2018 to 2024 · source

Is it right?

Interrogate whether the no is correct, using two independent lenses: the appeal funnel, which is self-selected, and OIG's random-sample audit, which is not.

13%
of denied prior-authorization requests met Medicare coverage rules in OIG's random-sample audit
These would likely have been approved under traditional Medicare.
18%
of denied payment requests should have been paid, in the same OIG audit
Most stemmed from manual-review or claims-system errors, not clinical judgment.

What happens to 1,000 Medicare Advantage prior authorization requests

Modeled per 1,000 requests using published 2023 rates: 6.4 percent denied, 11.7 percent of denials appealed, 81.7 percent of appeals overturned. Most denials are never appealed.

Read it this way Follow the widths left to right: only 7.5 of every 64 denied requests are appealed, and of those, 6.1 are overturned. The chart shows the 56.5 unappealed denials, not the appealed 7.5, are where most denied requests end without a second look. Use this chart to see whether the authorization burden is volume, correctness, appeal friction, or harm, and how that evidence supports targeted reform rather than blanket approval or blanket denial.

Caveat A modeled illustration of real published rates, not a raw per-request count. The 81.7 percent overturn rate applies only to the small appealed slice, which is why unchallenged denials are the larger concern.

1,000 PA requests 1,000 Approved 936 Denied 64 Appealed 8 Not appealed 57 Overturned on appeal 6 Denial upheld 1
⊞ data table⬇ CSV
FromToRequests per 1,000
1,000 PA requestsApproved936
1,000 PA requestsDenied64
DeniedAppealed7.5
DeniedNot appealed56.5
AppealedOverturned on appeal6.1
AppealedDenial upheld1.4

KFF, Nearly 50 million MA prior authorization requests in 2023, and HHS OIG appeal-overturn data · 2023 · source

Share of Medicare Advantage denials that get appealed

Share of denied requests that were appealed. Reporting exists only for 2019, 2023, and 2024, so the line is drawn from three points and not interpolated. Points are evenly spaced, not scaled to the year gaps.

Read it this way The three plotted points show the appeal share rising from 7.5 percent in 2019 to 11.7 percent in 2023, then easing slightly to 11.5 percent in 2024. Because the x-axis spaces these three years evenly rather than to scale, the slope should not be read as a steady year-over-year climb, and the missing 2020 to 2022 years are not shown at all. Use this chart to see whether the authorization burden is volume, correctness, appeal friction, or harm, and how that evidence supports targeted reform rather than blanket approval or blanket denial.

Caveat The 2020 to 2022 values are null in the source and are not shown. Appeals rose from 7.5 percent in 2019 to 11.7 percent in 2023, but only about one denial in nine is ever contested.

0.0% 5.0% 10.0% 15.0% 20.0% 201920232024 Share of denials appealed PERCENT OF DENIALS APPEALED
⊞ data table⬇ CSV
YearDenials appealed %
20197.5
202311.7
202411.5

KFF and CMS, Medicare Advantage appeal data · 2019 to 2024 · source

Denial overturn rate on appeal, by service type

Share of appealed denials overturned, by service type. A high overturn rate means the original denial was reversed on appeal. Post-acute care for the sickest patients is overturned far more often than the all-denials average.

Read it this way Rank the bars by service type: skilled nursing facility denials are overturned 95 to 97 percent of the time on appeal, far above the 81.7 percent all-denials average, while inpatient rehab (43 percent) and long-term care hospital stays (36 percent) sit below it. The 43 percent inpatient rehab figure is only an average; individual plans in the underlying review ranged from 14 to 86 percent overturned. Use this chart to see whether the authorization burden is volume, correctness, appeal friction, or harm, and how that evidence supports targeted reform rather than blanket approval or blanket denial.

Caveat The inpatient rehab figure of 43 percent is an average. Across the 19 MA plans reviewed, IRF overturn rates ranged from 14 to 86 percent. naviHealth, a delegated contractor, processed about half of the SNF reviews.

0.0% 25.0% 50.0% 75.0% 100.0% SNF admission (naviHealth contractor) 97.0% SNF admission (all MA plans) 95.0% All MA denials (2023) 81.7% Inpatient rehab (IRF) 43.0% Long-term care hospital (LTCH) 36.0%
⊞ data table⬇ CSV
Service typeOverturned on appeal %
SNF admission (naviHealth contractor)97
SNF admission (all MA plans)95
All MA denials (2023)81.7
Inpatient rehab (IRF)43
Long-term care hospital (LTCH)36

HHS OIG post-acute prior-authorization reviews (OEI-09-24-00330, OEI-09-24-00331) and KFF 2023 appeal data · 2023 to 2024 · source

Why, and who pays

The reasons behind denials and the human cost they land on: patient harm and physician administrative burden.

94%
of physicians report that prior authorization delays necessary care
Share of physicians reporting at least one such delay, not a patient-incidence rate.
45
prior-authorization requests per physician per week, on average
Rising to 64 in rheumatology and 58 in oncology.
14 hrs
of staff time per physician per week spent on prior authorization
Just over a third of practices use staff who work on it exclusively.

Why prior authorizations get denied

Share of denials by reason. Step therapy and missing-documentation defects are process barriers, not clinical judgments, and together account for 35 percent of denials.

Read it this way Read the bars by share of denials: medical necessity (31 percent) and step therapy (22 percent) are the two largest categories. Step therapy and missing documentation together add up to 35 percent of denials, process barriers rather than clinical judgment, but the chart cannot show whether any individual denial was ultimately correct. Use this chart to see whether the authorization burden is volume, correctness, appeal friction, or harm, and how that evidence supports targeted reform rather than blanket approval or blanket denial.

Caveat A synthesized categorization hand-curated from KFF, ProPublica, and AMA sources, not insurer-reported counts from a single dataset.

0% 13% 25% 38% 50% Medical necessity 31% Step therapy (fail first) 22% Site of care 14% Missing documentation 13% Coding or formulary 11% Quantity or duration limits 9%
⊞ data table⬇ CSV
ReasonShare of denials %
Medical necessity31
Step therapy (fail first)22
Site of care14
Missing documentation13
Coding or formulary11
Quantity or duration limits9

AMA and ProPublica, synthesized denial-reason taxonomy · 2023 · source

Physician-reported patient harm from prior-authorization delays

Physician-reported harm from prior-authorization delays, from most to least common. The gradient runs from delays that nearly all physicians report to catastrophic outcomes a smaller share report.

Read it this way The bars run from the most-reported harm to the least: 94 percent of physicians report delays and 78 percent report care abandonment, while the more severe outcomes, hospitalization (25 percent), permanent harm (19 percent), and death (7 percent), are reported by far fewer. These are physician-reported shares of at least one occurrence, not the rate at which any given patient experiences harm. Use this chart to see whether the authorization burden is volume, correctness, appeal friction, or harm, and how that evidence supports targeted reform rather than blanket approval or blanket denial.

Caveat These are shares of physicians who report at least one such event, not the rate at which patients experience them.

0% 25% 50% 75% 100% Delays in care 94% Care abandonment 78% Serious adverse event 33% Patient hospitalized 25% Permanent harm 19% Patient death 7%
⊞ data table⬇ CSV
Reported harmPhysicians reporting %
Delays in care94
Care abandonment78
Serious adverse event33
Patient hospitalized25
Permanent harm19
Patient death7

AMA, 2023 Prior Authorization Physician Survey · 2023 · source

Weekly prior authorization requests per physician, by specialty

Reference line is the all-physician average of 45 requests per week. Specialty burden is highest where treatments are expensive and chronic.

Read it this way Compare each specialty's bar to the 45-per-week reference line: rheumatology (64) and oncology (58) carry far more prior-authorization volume than average, while dermatology (28) carries much less. The pattern tracks specialties with expensive, chronic treatments, but the chart doesn't show whether higher-volume specialties also face higher denial rates. Use this chart to see whether the authorization burden is volume, correctness, appeal friction, or harm, and how that evidence supports targeted reform rather than blanket approval or blanket denial.

0 25 50 75 100 Rheumatology 64 Oncology 58 Endocrinology 53 Cardiology 47 Gastroenterology 44 Primary care 41 Psychiatry 36 Dermatology 28 All-physician average, 45/week
Service
⊞ data table⬇ CSV
SpecialtyWeekly PA requests per physician
Rheumatology64
Oncology58
Endocrinology53
Cardiology47
Gastroenterology44
Primary care41
Psychiatry36
Dermatology28

AMA, 2023 Prior Authorization Physician Survey · 2023 · source

Why this matters

Two independent sources agree the denials themselves are often wrong. HHS OIG's stratified random-sample audit, not self-selected the way appeals are, found 13 percent of prior-authorization denials met Medicare's own coverage rules and 18 percent of payment denials should have been paid. Overturn rates on appeal vary sharply by service type: skilled-nursing-facility admissions are overturned 95 to 97 percent of the time, versus 43 percent for inpatient rehab and 36 percent for long-term care hospital stays, suggesting some service lines are denied far more aggressively than the evidence supports. The most common stated reasons, medical necessity (31 percent) and step therapy (22 percent), leave room for dispute, but missing documentation alone accounts for 13 percent of denials, a pure process failure with no clinical basis.

Recommended actions

  • Require per-plan public reporting of denial and overturn rates by service type, so the SNF-level 95 to 97 percent overturn rate is visible to regulators and beneficiaries before they file an appeal.
  • Extend gold-carding, exempting providers with historically low denial rates from prior authorization, to the service lines with the highest overturn rates first: skilled nursing, inpatient rehab, and long-term care hospital stays.
  • Apply OIG's random-sample audit methodology plan-wide on a recurring basis, since self-selected appeals undercount how often initial denials are wrong.
  • Target missing-documentation denials specifically with standardized submission templates, since this is a process fix with no clinical tradeoff.
  • Track the appeal rate itself as a metric requiring intervention, since only 11.7 percent of denials were even challenged in 2023.

The recommendation

Therefore, reposition prior authorization as a targeted clinical-control tool rather than a default administrative hurdle. The recommended program is to require plan-level denial and overturn transparency, gold-card low-risk providers, standardize documentation, and audit service lines where overturned denials signal systematic overuse.

Demographic slice none. CMS prior-auth reporting and the AMA survey are payer or service-level.

Sources