The first prior authorization metrics are public, and they are hard to compare. By March 31, 2026, the payers covered by CMS’s January 2024 interoperability and prior authorization rule, from Medicare Advantage organizations to issuers on the federal exchanges, had to post their first annual prior authorization metrics. For many of these plans, it is the first time any prior authorization data has been published at all.
We read a sample of the postings in the weeks after the deadline. Approval rates for standard requests in the reports we reviewed ran from about 92% to 99%. Beyond that headline, the reports diverge. Some include the counts behind every percentage and some do not, and decision times appear in hours in one report and in days in the next. Appeals show up as a share of all requests in some postings and as a share of appeals in others. Martin’s Point broke its results out by service area and vendor, which the rule does not require.
Our March piece argued that prior authorization metrics should be published by service, and the first postings are the test of that argument.
Key takeaways
- The data exists now. Plans posted 2025 metrics on approvals and decision times for standard and expedited requests.
- Approval rates are high. Standard approval rates in our sample ranged from about 92% to 99% at the contract or plan level.
- Formats differ in ways that matter. Plans vary in whether they show counts and in what they use as the denominator for appeals and extensions.
- Appeals are rare. In one large Medicare Advantage contract, about 2% of standard denials were appealed. In a Medicaid plan, about 13%.
- Medians hide the tail. Most Medicare Advantage medians were a day or less, but one plan’s median for medical requests was 8 days in 2025, above the 7-day standard that took effect in January 2026.
- Service detail is possible. Martin’s Point voluntarily reported results by service area, showing denial rates that differed widely between categories.
What the rule requires
The interoperability and prior authorization rule requires impacted payers to post, each year on a public website, a list of items and services that require prior authorization and a set of metrics for the prior year. The metrics, aggregated for all items and services, cover 5 measures.
- the percentage of standard requests approved and denied
- the percentage of standard requests approved after appeal
- the percentage of requests approved after the review timeframe was extended
- the percentage of expedited requests approved and denied
- the average and median time from request to decision, for standard and expedited requests
Drugs are excluded. The level of reporting depends on the program. According to Becker’s summary at the deadline, Medicare Advantage plans report at the contract level and exchange issuers at the issuer level. Medicaid and CHIP report by state for fee-for-service programs and by plan for managed care.
The rule’s reach is also limited by who it covers. Employer-sponsored plans, which cover most Americans with private insurance, are not impacted payers under the rule and do not have to post these metrics. Neither do plans sold outside the federal exchanges unless a state requires it. The first public metrics therefore describe public programs and federal exchange coverage, not the whole market.
The rule’s 2025 data predates its new decision deadlines. Starting in January 2026, Medicare Advantage plans and Medicaid and CHIP programs must decide standard requests within 7 calendar days and expedited requests within 72 hours. In 2025, the Medicare Advantage standard was 14 days. The first year’s metrics therefore show how plans performed under the old clock, which makes them a useful baseline for the new one.
What the first postings show
We reviewed postings from 5 plans, 4 of them Medicare Advantage contracts and 1 a Medicaid managed care plan. They are not a representative sample. We chose them because they illustrate the range of formats.
| Plan and contract | Market | Standard approved | Standard denied | Approved after appeal | Median standard decision |
|---|---|---|---|---|---|
| Humana CarePlus, H1019 | Medicare Advantage | 96.95% | 3.05% | 80.17% of appeals | 0 days |
| Wellpoint, H3240 | Medicare Advantage | 95.5% | 4.5% | 85% | 1 day |
| Blue Advantage (Louisiana), H6453 | Medicare Advantage | 98.99% | 1.01% | 84.21% | Under 1 day |
| Martin’s Point, H1365, medical requests | Medicare Advantage | 88% | 12% | 62% | 8 days |
| Aetna Better Health of Texas | Medicaid and CHIP | 92.30% | 7.70% | 36.16% of appeals | 3 days |
From each plan’s 2025 posting, linked in the first column.
The Humana contract is the largest in the sample, with 752,924 standard requests and 110,216 expedited requests in 2025. It approved 96.95% of standard requests and 95.19% of expedited ones, and reported a median standard decision time of 0 days and a median expedited time of 22 hours.
Wellpoint’s contract reported an average standard decision time of 5 days against a median of 1 day. A gap that wide means that while most requests are decided quickly, a meaningful share take much longer, enough to pull the average up by several days. That is exactly the group of requests a patient worries about, and the median hides it.
Blue Advantage reported approval rates near 99% for both its HMO and PPO contracts, with median standard decisions under a day and median expedited decisions of about an hour. Those figures are consistent with a high share of requests approved automatically or near-automatically.
The Aetna Medicaid plan in Texas reported 41,416 standard requests, of which 92.30% were approved, with a median decision time of 3 days. It also reported that 4,027 standard requests had their review timeframe extended, and that 73.93% of those were eventually approved. By our calculation, nearly 1 in 10 standard requests had its review extended.
Where the formats diverge
The rule specifies the metrics but leaves much of the presentation to plans. The first postings show how much that matters.
Counts. All but Wellpoint posted the number of requests behind each percentage, and Wellpoint’s posting for its contract showed percentages only. Without counts, a reader cannot tell whether an approval rate rests on 500 requests or 500,000, or whether a difference between plans is meaningful.
Time units. Plans reported decision times in days or hours, sometimes with different units for different service areas in the same report. Wellpoint’s posting explained that a reported time of 1 day means up to 24 hours and includes requests approved in real time. Blue Advantage reported its standard median as less than 1 day. A table that says 1 day can therefore describe anything from a few seconds to a full day.
Appeal denominators. Humana reported approvals after appeal as 388 of 484 appeals, or 80.17%. Aetna Better Health reported 162 of 448 appeals, or 36.16%. Both are shares of appealed requests. As we found in the equity reports in February, other plans have read the same line as a share of all requests, which produces a number near 0%. A reader comparing a figure of 85% with a figure of 0.16% may be comparing 2 different measures.
Extension denominators. The same ambiguity applies to approvals after an extended review. Blue Advantage reported 0.03% for its HMO contract, apparently as a share of all requests. Aetna Better Health reported 73.93%, as a share of requests whose review was extended.
Breakouts. Martin’s Point went beyond the rule, reporting medical and behavioral health requests separately from the oncology and radiology requests a vendor manages. That choice makes its report the most informative in our sample, and also makes its numbers harder to set beside plans that reported a single figure.
A table that says 1 day can describe anything from a few seconds to a full day.
What the numbers say, and what they cannot
Even with those caveats, the first postings support a few observations.
Approvals dominate. Every plan in the sample approved more than 85% of standard requests, and most approved more than 95%. That is consistent with KFF’s analysis of 2024 Medicare Advantage data, which found that 7.7% of requests were denied in full or in part.
Few denials are appealed. Where plans posted counts, the share of denials that were appealed was small. Humana’s contract reported 22,927 denied standard requests and 484 appeals, which by our calculation means about 2% of standard denials were appealed. Aetna Better Health reported 3,332 denials across standard and expedited requests and 448 appeals, or about 13%. In both cases, a large share of appeals succeeded. The combination of high overturn rates and low appeal rates means that most denials are never tested.
Service areas differ widely. Martin’s Point’s breakout shows why aggregate rates can mislead. For contract H1365, medical requests had a standard denial rate of 12%, against 2.8% for oncology and radiology requests. Behavioral health requests had none. A single aggregate rate would have averaged those very different processes into one number.
Medians are not deadlines. Most Medicare Advantage medians in the sample were a day or less. But Martin’s Point reported a median of 8 days for standard medical requests in 2025, within the 14-day standard that applied then and above the 7-day standard that applies now. A median also says nothing about how many decisions took longer than the deadline, which is the number that matters to a patient waiting for care.
The numbers cannot answer the questions most readers bring to them. They do not show which services are denied most often, or how many decisions missed the required timeframe. They also cannot account for differences in what each plan chooses to require prior authorization for, which shapes approval rates as much as how plans decide.
How to read a prior authorization report
For anyone trying to use the postings now, a few checks get the most out of them.
- Find the count. A rate without the number of requests behind it cannot be weighed. Look for the total requests and the total appeals.
- Check the denominator for appeals. If the approved-after-appeal figure is near 0%, it is probably a share of all requests. If it is above 30%, it is probably a share of appeals. Only the second tells you how often appeals succeed.
- Check the unit of time. A median of 1 day may mean anything up to 24 hours. A large gap between the average and the median means some requests take much longer.
- Check the scope. See whether the report covers one contract or several, and which services require prior authorization at all. A plan that reviews fewer, harder services will look stricter than one that reviews many routine ones.
- Look for breakouts. Results by service area, where a plan provides them, are more useful than the totals.
- Follow the hard cases. Look at expedited denial rates and extensions. A high share of extended reviews can signal missing documentation or slow processes.
Read this way, a single posting can still be informative. Read as a league table, the first year’s postings are more likely to mislead than to inform.
What a better second year would look like
The second round of postings, due in March 2027, could be far more useful without new legislation or a new rule, if plans and CMS converge on a few practices.
- Show counts with every percentage. Every rate should carry the number of requests behind it. Most plans in our sample already did.
- State the denominator. Approvals after appeal should be reported as a share of appeals, with the number of appeals shown, and the share of denials appealed should be reported alongside it.
- Use hours for time. Decision times reported in hours, with the share of decisions that exceeded the required timeframe, would describe performance far better than a median in days.
- Break out by service. Results by service category, as Martin’s Point did and as Minnesota already requires for commercial plans, would separate what a plan reviews from how it decides.
- Publish a data file. A simple machine-readable file alongside the PDF would let researchers and regulators compare plans without retyping tables.
None of these steps requires plans to collect anything new. Each requires only that they publish what their systems already record.
Where Redlake stands
We build tools health plans use in utilization review, and we think the first public metrics are a real step that plans should build on rather than treat as a compliance exercise. A plan that publishes counts and service-level results will be easier to understand and easier to trust than one that publishes the minimum.
Our rules make that kind of reporting straightforward. Every decision we help record carries the criteria applied and the reason for the outcome, each with a time stamp. A report built on those records reconciles from the single decision up to the public table. The same records answer the harder questions the public metrics leave open, such as how many decisions missed the deadline and which services are denied for missing documentation rather than clinical reasons.
The metrics also point to operational priorities. A service with thousands of requests and an approval rate near 100% is a candidate to stop reviewing, and a median decision time above the new 7-day standard needs to change now. A plan where few denials are appealed but most appeals succeed should ask what its first decisions are missing, because the denials nobody appeals are probably not all correct.
What to watch in 2026
- Compliance gaps. Whether any required plans failed to post, and how CMS responds.
- Analyses by researchers. Whether researchers or state regulators compile the postings into comparable data sets, and what they find.
- CMS guidance. Whether CMS issues more specific reporting instructions, such as required counts or denominators, before the second round.
- Decision times under the new deadlines. Whether plans whose 2025 medians exceeded 7 days report faster decisions for 2026.
- Service-level reporting. Whether more plans follow Martin’s Point and break out results by service without being required to.
About this analysis
Basis. We reviewed 2025 prior authorization metrics posted by the 5 plans named in the table. We also drew on the CMS fact sheet for the interoperability and prior authorization rule and on KFF’s analysis of 2024 Medicare Advantage data.
Data notes. Figures are as the plans posted them. Martin’s Point figures in the table are for its medical requests only. The shares of denials appealed and of requests extended are our calculations from posted counts. The 5 plans are not a representative sample, and differences in reporting conventions mean the figures should not be used to rank plans.
Sources. CMS, interoperability and prior authorization final rule fact sheet (January 2024). Becker’s Payer Issues (March 2026). Postings from the 5 plans in the table. KFF (January 2026).
Frequently asked questions
What prior authorization metrics did payers have to post in March 2026?
Impacted payers had to post 2025 approval and denial rates for standard and expedited requests, aggregated for all items and services. The postings also had to show how many requests were approved after appeal or after an extended review, along with decision times. A list of services requiring prior authorization was required too.
Which plans had to post them?
Medicare Advantage organizations report at the contract level and exchange issuers at the issuer level. Medicaid and CHIP report by state for fee-for-service programs and by plan for managed care. Employer plans are not covered.
Do the metrics include drugs?
No. The rule’s prior authorization requirements, including the public metrics, exclude drugs.
Can the metrics be used to compare plans?
Only with care. Plans differ in whether they show counts and in how they define appeal rates, and the metrics combine all services into single rates.
What is the next reporting deadline?
Metrics for 2026 are due by the end of March 2027 and every year after that.