What a good prior authorization decision looks like from every seat
Key takeaways
- Meeting turnaround times does not show that a prior authorization decision was clinically sound or clearly explained
- Only 24% of physicians in the AMA's 2025 survey said qualified clinicians consistently reviewed medical-necessity denials
- CMS-0057-F public metrics show patterns, not causes. Appeal and overturn rates are especially easy to misread
- Make administrative preparation and clinical authority explicit, and keep evidence, criteria and rationale tied to each decision
- Start with a small sample of appealed cases, then pilot one bounded workflow and measure speed, effort and clinical quality together
Ask five people what a good prior authorization decision looks like and you will hear five answers. The provider office wants a fast response it can act on. The intake team wants a complete request the first time. The medical director wants the right evidence in front of the right reviewer. The member wants to know what happens next. The compliance lead wants a record that holds up when someone asks how the decision was made.
Each of them is right. Problems start when a plan optimizes for one of those answers and assumes the others will follow.
Turnaround time has become the default measure, and it matters. Patients should not wait longer than necessary. Still, a decision that arrives on time and is later overturned on appeal can cost everyone more than one that took a little longer and held up. My view: plans get the most out of prior authorization when they judge every decision from all five seats at once.
Five seats, five definitions of good
No single team can deliver all five. That is why prior authorization is an operating design question as much as a clinical or technology one.
What appeal data says about decision quality
Medicare Advantage insurers made nearly 53 million prior authorization determinations in 2024 and fully or partially denied 7.7% of them. Only 11.5% of those denials were appealed, and four in five appealed denials (80.7%) were overturned (KFF).
A reversal does not prove the first decision was wrong. It may follow new documentation, a different clinical assessment or a change in the case. Yet every reversal is paid for at every seat. The provider resubmits. The clinical team reviews the case twice. The member waits. And the result now shows up in numbers anyone can read.
Three measures, read together
Since January 1, 2026, the Medicare Advantage, Medicaid and CHIP programs covered by CMS-0057-F have had to decide expedited requests within 72 hours and standard requests within seven calendar days. Impacted payers also posted their first public prior authorization metrics by March 31, 2026 (CMS fact sheet).
Those figures show how fast a plan decides and how often it approves. They cannot show why a pattern looks the way it does. For that, I would track three measures side by side:
- Timeliness: how long each stage takes, from receipt to a complete case to decision.
- Administrative effort: touches, follow-ups and resubmissions per case.
- Decision quality: whether the evidence was available at review, the criteria and authority were documented, and the notice explained the next step.
Improving one of these while another slips is easy to miss when you only report the first.
Where AI fits
AI-supported workflow tools can capture information, classify requests, organize records and flag missing documentation before a case reaches review. That is real preparation work, and it frees clinical time.
Clinical judgment stays with qualified reviewers. Any potential adverse determination has to follow the qualifications, authority and escalation rules that apply to the program and jurisdiction. The design question for every plan is simple to state: the tools prepare the case, a qualified clinician makes the call, and the record shows both.
A test you can run this quarter
Pull a small sample of recently appealed cases. Read each one from every seat:
- Provider: Did the original notice give a clear reason and next step?
- Intake: Was the request complete and correctly classified when it arrived?
- Clinical reviewer: Was the relevant evidence in front of the reviewer, with the criteria easy to find?
- Member: How long did the member wait from first request to final answer?
- Compliance: Could someone outside the team reconstruct why the decision was made?
Then look at what changed between the original determination and the appeal. A small sample will not describe your whole operation. It will tell you which seat to start with.
The plans that lead
The plans that lead on prior authorization will be the ones that can explain any decision from any seat: fast enough for the provider, clear enough for the member, grounded enough for the medical director and documented well enough for anyone who asks.
Our whitepaper, Prior Authorization Under Public Scrutiny: A Practical Guide to Timely, Clinically Grounded Decisions, gives you the tools to start: a regulatory timeline, a responsibility map, a diagnostic for locating avoidable work and a pilot scorecard for evaluating change.
FAQ
Why is faster prior authorization turnaround not enough to rebuild provider trust?
Speed shows that a plan met the clock. It does not show whether the decision rested on relevant evidence, followed the right clinical authority or came with an explanation the provider can act on. In the AMA's 2025 survey, only 24% of physicians said appropriately qualified clinicians consistently reviewed medical-necessity denials.
What prior authorization metrics does CMS-0057-F require payers to publish?
CMS-0057-F requires impacted payers to publicly report specified prior authorization metrics for medical items and services, excluding drugs. Operating requirements and API obligations follow separate timelines, and applicability varies by payer type.
Why are prior authorization appeal and overturn rates easy to misread?
A reversal can follow new documentation, a different clinical assessment or another change in the case, so an overturn does not prove the original decision was wrong. Cross-plan comparisons also need care because samples, weighting and reporting practices differ.
How should health plans divide prior authorization work between workflow tools and clinical reviewers?
Workflow tools can support information capture, request classification, record organization and flagging missing documents. Clinical assessment and potential adverse determinations must follow the qualification, authority and escalation rules for the relevant program and jurisdiction. Evidence, criteria, review authority and rationale should stay connected to the decision.
How can a health plan test its prior authorization process without a full overhaul?
Review a small sample of recently appealed or repeatedly handled cases. Check whether information was available at clinical review, criteria and evidence were easy to find, authority and escalation were clear, and the notice explained the reason and next step. Then pilot one bounded workflow and measure timeliness, administrative effort and clinical process quality together.
%20Thumbnail%20Images/Article2_408x213.png?updatedAt=1792078849371&tr=w-412,h-232,c-maintain_ratio,q-80)