Medicaid represents nearly $1 out of every $5 spent on health care in the U.S. and is the major source of financing for states to provide health coverage and long-term care for low-income residents. Medicaid is administered by states within broad federal rules and jointly funded by states and the federal government through a federal matching program with no cap. With Donald Trump returning to the presidency and Republican control of the Senate and House, discussions are ramping up about major cuts to federal spending, including federal Medicaid spending, through a new government efficiency initiative and to help pay for an extension of tax cuts in the 2017 Tax Cuts and Jobs Act. Restricting federal Medicaid spending could leave states with tough choices about whether to offset reductions with state dollars, limit Medicaid utilization or benefits, or reduce coverage. Amid uncertainty for Medicaid financing at the federal level, states have also reported uncertainty regarding post-unwinding Medicaid spending and enrollment trends and recent shifts in state fiscal conditions. This brief examines the following key questions ahead of potential proposals to change Medicaid financing:

• How does Medicaid financing work?
• How much does Medicaid cost and how are funds spent?
• How does Medicaid relate to federal and state budgets?
• What factors affect Medicaid spending changes?

How does Medicaid financing work?

Medicaid financing is shared by states and the federal government with a guarantee to states for federal matching payments with no pre-set limit. The percentage of costs paid by the federal government (known as the federal medical assistance percentage or “FMAP”) varies across states, for specific services and types of enrollees, and depending on whether the costs are for medical care or program administration. The match rate has also been temporarily adjusted during economic downturns and most recently during the COVID-19 pandemic.

The FMAP for services used by people eligible through traditional Medicaid, which includes individuals who are eligible as children, low-income parents, because of disability, or because of age (65+), is determined by a formula set in statute. The formula is designed so that the federal government provides a match rate of at least 50% and provides a higher match rate for states with lower average per capita income. The resulting FMAP varies by state and ranged from 50% (the FMAP “floor”) in ten states (California, Colorado, Connecticut, Maryland, Massachusetts, New Hampshire, New Jersey, New York, Washington, and Wyoming) to 77% in Mississippi for federal fiscal year (FFY) 2026 (Figure 1).

To participate in Medicaid and receive federal matching dollars, states must meet core federal requirements. States must provide certain mandatory benefits (e.g., hospital, physician, and nursing home services) to core populations (e.g., low-income pregnant women, children, people with disabilities, and people ages 65 and older) without waiting lists or enrollment caps. States may also receive federal matching funds to cover “optional” services (e.g., adult dental care and home care, also known as home and community based services) or “optional” groups (e.g. people with income above the limits established for core populations). States also have discretion to determine how to purchase covered services (e.g., through fee-for-service or capitated managed care arrangements) and to establish provider payment methods and rates.

Medicaid also provides “disproportionate share hospital” (DSH) payments to hospitals that serve a large number of Medicaid and low-income uninsured patients to offset uncompensated care costs. DSH payments totaled over $17 billion in FFY 2023. Federal DSH spending is capped for each state and facility but within those limits, states have considerable discretion in determining the amount of DSH payments to each DSH hospital. The Affordable Care Act (ACA) called for a reduction in federal DSH allotments starting in FFY 2014 based on the assumption of reduced rates of uninsurance, but the cuts have been delayed several times (most recently delayed until April 1, 2025). DSH is one type of a broader set of “supplemental” payments that states make to supplement Medicaid “base” payment rates that often do not fully cover provider costs; unlike other supplemental payments, DSH payments can also be used to pay for unpaid costs of care for the uninsured.

There are special match rates for the ACA expansion group, administration, and other services. While the traditional FMAP applies to the vast majority of Medicaid spending, there are a few exceptions that provide higher match rates for specific services or populations, such as family planning and most notably people covered under the ACA Medicaid expansion. States that have implemented the expansion currently receive a 90% FMAP for adults covered through the ACA Medicaid expansion. States that had not adopted the expansion as of 2021 when the American Rescue Plan Act was enacted are eligible for a 5% increase in the state’s traditional FMAP for two years if they implement the expansion. Administrative costs incurred by states are usually matched by the federal government at a 50% rate, but some functions such as eligibility and enrollment systems receive higher match rates. Medicaid administrative costs are less than 4% of total Medicaid spending.

Unlike in the 50 states and D.C., annual federal funding for Medicaid in the U.S. territories is subject to a statutory cap and fixed matching rate. Once a territory exhausts its capped federal funds, it no longer receives federal financial support for its Medicaid program during that fiscal year. Over time, Congress has provided increases in federal funds for the territories broadly and in response to specific emergency events. Various pieces of legislation during the pandemic significantly increased the allotments for each of the territories and also raised the FMAP rates from the statutory level of 55% to 76% for Puerto Rico and 83% for the other territories. The 2023 Consolidated Appropriations Act extended the 76% FMAP for Puerto Rico through FFY 2027 and made the 83% match rate for other territories permanent.

Both the federal government and states are responsible for promoting program integrity. Program integrity broadly refers to the proper management and function of the Medicaid program to ensure it is providing quality and efficient care while using funds–taxpayer dollars–appropriately, with minimal waste. Improper payment reports in recent years have highlighted program integrity issues, and both the previous Trump Administration and the Biden Administration worked to advance program integrity, though through different means. Improper payments, which are often cited when discussing program integrity, are payments that do not meet Centers for Medicare and Medicaid Services (CMS) program requirements (they can be overpayments, underpayments, or payments with insufficient information to determine whether the payment was proper). Improper payments are not the same as criminal activities like fraud and abuse, which may be a subset of improper payments. CMS estimated the overall Medicaid improper payment rate was about 5% in 2024, the lowest rate since the COVID-19 pandemic began due, in part, to flexibilities granted during that time. Most improper payments (79%) were due to insufficient information (or missing administrative steps), not necessarily due to payments for ineligible enrollees, providers, or services (i.e., since they may have been payable if the missing information had been on the claim and/or the state had complied with requirements). The improper payment estimation process is not designed to detect or measure fraud. Program integrity efforts focused on areas identified as major contributors of improper payments, such as monitoring provider screening and enrollment for noncompliance, likely yield greater returns than focusing on reducing errors in eligibility determinations that could make it more difficult for eligible people to obtain and maintain coverage.

How much does Medicaid cost and how are funds spent?

Capitated payments to Medicaid managed care organizations (MCOs) accounted for more than half of Medicaid spending in FFY 2023 (Figure 2). Managed care and health plans accounted for the largest share (55%) of Medicaid spending, with capitated payments to comprehensive MCOs accounting for 52% of Medicaid spending in FFY 2023 and other Medicaid managed care (e.g., primary care case management (PCCM) arrangements or payments to specialty plans) accounting for another 3%. Smaller shares of total Medicaid spending in FFY 2023 were for fee-for-service acute care (21%), fee-for-service long-term care (19%), Medicaid spending for Medicare premiums on behalf of enrollees who also have Medicare (3%), and DSH payments (2%).

Enrollees eligible based on disability or age (65+) comprise 23% of all enrollees but account for over half of total spending due to higher per person costs (Figure 3). Children account for 34% of enrollees but only 14% of spending. Adult enrollees (those made eligible under the ACA Medicaid expansion, as well as low-income parents) account for 43% of all enrollees and 34% of spending. The disproportionate spending on certain eligibility groups stems from variation in spending per enrollee across the eligibility groups, reflecting differences in health care needs and utilization. Spending per enrollee for individuals eligible based on age (65+) and disability, the two groups with the highest per enrollee costs, is approximately six times higher than spending per enrollee for children, who had the lowest spending of any eligibility group. Those eligible on the basis of age or disability tend to have higher rates of chronic conditions, more complex health care needs and are more likely to utilize long-term care than other enrollees, contributing to higher spending.

Total spending per full-benefit enrollee ranged from a low of $3,750 in Tennessee to $12,425 in the District of Columbia in 2021 (Figure 4). Variation in spending across the states reflects considerable flexibility for states to design and administer their own programs – including what benefits are covered and how much providers are paid — and variation in the health and population characteristics of state residents. Within each state, there is also substantial variation in the average costs for each eligibility group and within each eligibility group, per enrollee costs may vary significantly. Overall, Medicaid spending has experienced slower cumulative growth since 2008 compared to Medicare and private insurance on a per-enrollee basis.

Overall, Medicaid spending totaled $880 billion in FFY 2023 with the federal government paying 69% ($606 billion) and states paying 31% ($274 billion). This share is slightly higher than historic shares due to the enhanced pandemic match rate, but there is variation across states (Figure 5). The combination of low per capita income and adopting the ACA expansion are both factors in determining the overall share of federal spending on Medicaid. While the federal share of Medicaid spending varies across states, so does the total amount of federal Medicaid dollars coming into the state. States with the largest populations (California, New York, Texas, Pennsylvania, Ohio) receive the most federal Medicaid funding.

How does Medicaid relate to federal and state budgets?

Social Security, Medicare, and Medicaid are the three main entitlement programs and accounted for 41% of all federal outlays in FFY 2024 (Figure 6). Of these three programs, Medicaid is smallest in terms of federal outlays, though it covers a larger number of people than Medicare or Social Security. Overall, federal spending on domestic and global health programs and services accounted for more than one-fourth of net federal outlays in FFY 2024, including spending on Medicare (12%), Medicaid and CHIP (8%), and other health spending (6%). It is expected that in 2025, the new Congress and Administration will leverage reconciliation to make major reductions in spending which can offset some of the costs of extending the expiring tax cuts. Budget reconciliation is a special legislative process used to make changes to taxes and mandatory spending that allows the Senate to pass legislation with only 50 votes rather than the customary 60 votes. To begin the reconciliation process, Congress must develop a budget resolution, which establishes spending levels including any large deficit or spending reduction targets. With President-elect Trump taking substantial cuts to Medicare and Social Security off the table, there will be significant pressure to reduce spending on Medicaid.

Medicaid is a spending item and at the same time the largest source of federal revenues for state budgets. As a result of the federal matching structure, Medicaid has a unique role in state budgets as both an expenditure item and a source of federal revenue for states. According to data from the National Association of State Budget Officers (NASBO), in state fiscal year (SFY) 2023, Medicaid accounted for 30% of total state spending for all items in the budget (Figure 7). Medicaid accounted for only 15% of expenditures from state funds (including state general funds and other state funds), second to K-12 education (23%). On the other hand, Medicaid accounted for 57% of all expenditures from federal funds. States have an incentive to control Medicaid spending because they pay a share of Medicaid costs. At the same time, research shows that federal matching dollars from Medicaid spending have positive effects for state economies. A number of studies show that states that have adopted the ACA Medicaid expansion have realized budget savings, revenue gains, overall economic growth as well as observed positive effects on the finances of hospitals and other health care providers.

States can use provider taxes and IGTs (intergovernmental transfers) to help finance the state share of Medicaid. States have flexibility in determining how to finance the non-federal share of state Medicaid payments, within certain limits. In addition to state general funds appropriated directly to the Medicaid program, most states use funding from local governments or revenue collected from provider taxes and fees to help finance the state share of Medicaid. While data are limited, the Government Accountability Office (GAO) estimated that provider taxes accounted for approximately 17% of the non-federal share of total Medicaid payments in SFY 2018 though the share varies by state. All states (except Alaska) have at least one provider tax in place and many states have more than three (Figure 8). The most common provider taxes are on nursing facilities (46 states) and hospitals (45 states). As of July 1, 2024, 48 states including DC also reported at least one provider tax that is above 3.5% of net patient revenues, and 38 states including DC also reported at least one provider tax that is above 5.5%, which is close to the maximum federal safe harbor or allowable threshold of 6%. Provider taxes and funds from local governments are often used by states to help finance DSH payments, non-DSH supplemental payments like upper payment limit (UPL) payments, and MCO state directed payments (uniform payment increases through managed care that are similar to supplemental payments). New Medicaid managed care rules finalized in 2024 permit states to pay hospitals and nursing facilities at the average commercial payment rate (ACR) when using directed payments, contributing to expected increases in federal Medicaid spending projections.

What factors affect Medicaid spending changes?

Medicaid enrollment and spending typically increase during recessions. Medicaid is a counter-cyclical program, meaning that more people become eligible and enroll during economic downturns. At the same time, states typically may face declines in revenues that make it difficult to finance the state share of funding for the program. As in past economic downturns, Congress enacted legislation during the recent pandemic-induced recession that temporarily increased the federal share of Medicaid spending to help states maintain their Medicaid programs. In exchange, states were required to keep individuals continuously enrolled in Medicaid, contributing to historic growth in Medicaid enrollment. High enrollment growth rates, tied first to the Great Recession, then ACA implementation, and later the pandemic, were the primary drivers of total Medicaid spending growth over the last decade (Figure 9). The continuous enrollment provision ended on March 31, 2023, and Medicaid enrollment growth has since declined and total spending growth has slowed.

Medicaid spending is also driven by other external factors and policy choices. Other external factors can include state fiscal conditions and the costs of providing health services and long-term care. In KFF’s latest annual budget survey, states reported inflation and workforce shortages were driving higher labor costs, resulting in pressure to increase provider rates. State policy choices include changes to eligibility, benefits, or provider reimbursement as well as delivery system reforms. Recent eligibility changes include the adoption of the ACA Medicaid expansion in seven states since 2020 as well as other expansions for children. KFF’s annual budget survey also found many states were implementing benefit enhancements, particularly for mental health and/or substance use disorder services, as well as authorizing a wide range of provider rate increases, particularly for direct care workers in nursing facilities and those providing home care to help address workforce shortages.

Virtually all enrollees in Medicare Advantage (99%) are required to obtain prior authorization for some services – most commonly, higher cost services, such as inpatient hospital stays, skilled nursing facility stays, and chemotherapy. This contrasts with traditional Medicare, where only a limited set of services, including certain outpatient hospital services, non-emergency ambulance transport, and durable medical equipment, require prior authorization (see Box 1).

Prior authorization requirements are intended to ensure that health care services are medically necessary by requiring approval before a service or other benefit will be covered. Medicare Advantage insurers typically use prior authorization, along with other tools, such as provider networks, to manage utilization and lower costs. This may contribute to their ability to offer extra benefits and reduced cost sharing, typically for no additional premium, while maintaining strong financial performance. At the same time, prior authorization processes and requirements, including the use of artificial intelligence to review requests, may result in administrative hassles for providers, delays for patients in receiving necessary care, and in some instances, denials of medically necessary services, such as post-acute care.

This analysis uses data submitted by Medicare Advantage insurers to the Centers for Medicare and Medicaid Services (CMS) to examine the trends in the number of requests for prior authorization determinations, denials, and appeals for 2019 through 2023, as well as differences across Medicare Advantage insurers. It does not include determinations or denials by type of service or plan because CMS does not collect or report this information, though such data could help inform consumers in choosing among plans. It also presents data from CMS about the use of prior authorization in traditional Medicare, including the number of reviews and denials for 2021 through 2023, and the share appealed and the outcome of the appeal for 2021 and 2022 (the 2023 data do not include this information).

Key Takeaways:

• Medicare Advantage insurers made nearly 50 million prior authorization determinations in 2023, reflecting steady year-over-year increases since 2021 (37 million) and 2022 (46 million) as the number of people enrolled in Medicare Advantage has grown. The determinations represent requests for approval that providers are required to submit before providing a service. Substantially fewer prior authorization reviews for traditional Medicare beneficiaries were submitted to CMS – just under 400,000 in fiscal year 2023 – though the number of people enrolled in Medicare Advantage and traditional Medicare were similar in these years.
• In 2023, there were nearly 2 prior authorization determinations on average per Medicare Advantage enrollee, similar to the amount in 2019. In contrast, in 2023, about 1 prior authorization review was submitted per 100 traditional Medicare beneficiaries – a rate of about 0.01 per person — which reflects the limited set of services subject to prior authorization in traditional Medicare.
• In 2023, insurers fully or partially denied 3.2 million prior authorization requests, which is a somewhat smaller share (6.4%) of all requests than in 2022 (7.4%). Though there were substantially fewer prior authorization reviews for traditional Medicare beneficiaries, a larger share was denied – 28.8% in 2023.  Denial rates varied across the limited set of services subject to prior authorization in traditional Medicare.
• A small share of denied prior authorization requests was appealed in Medicare Advantage (11.7% in 2023). That represents an increase since 2019, when 7.5% of denied prior authorization requests in Medicare Advantage were appealed. A relatively small share of denied prior authorization reviews was appealed in traditional Medicare (6.4% in 2022) as well.
• Though a small share of prior authorization denials were appealed to Medicare Advantage insurers, most appeals (81.7%) were partially or fully overturned in 2023. That compares to less than one-third (29%) of appeals overturned in traditional Medicare in 2022. These requests represent medical care that was ordered by a health care provider and ultimately deemed necessary but was potentially delayed because of the additional step of appealing the initial prior authorization decision. Such delays may have negative effects on a person’s health.

Prior authorization practices have gotten a fair amount of attention in recent years. During the Biden Administration, CMS finalized three rules related to the use of prior authorization in Medicare Advantage. Among other changes, the three rules clarify the criteria that may be used by Medicare Advantage plans to establish prior authorization policies, streamline the prior authorization process for Medicare Advantage and certain other insurers, and require Medicare Advantage plans to evaluate the effect of prior authorization policies on people with certain social risk factors. In December 2024, the outgoing Biden Administration proposed further changes, including clarifying coverage requirements in Medicare Advantage. The Trump Administration will have an opportunity to modify or finalize these proposed changes and may propose additional regulatory changes. Additionally, lawmakers in Congress have held hearings, requested detailed data from the largest Medicare Advantage insurers, and introduced several bills to improve transparency and reform other aspects of prior authorization (see Box 2). Despite bipartisan support encompassing a majority of members in both houses of Congress, legislation on the use of prior authorization has not been enacted.

Use of Prior Authorization in Medicare Advantage

As part of its oversight of Medicare Advantage plans, CMS requires Medicare Advantage insurers to submit data for each Medicare Advantage contract (which usually includes multiple plans) that includes the number of prior authorization determinations made during a year, and whether the request was approved. Insurers are additionally required to indicate the number of initial decisions that were appealed (reconsiderations) and the outcome of that process, including whether the initial decision was affirmed, partially overturned, or fully overturned. These data are useful for assessing overall trends and variations across insurers, but do not contain the information necessary to understand how the use of prior authorization varies by type of service or type of plan.

In 2023, Medicare Advantage insurers made nearly 50 million prior authorization determinations.

After dropping in 2020 amid the initial phase of the COVID-19 pandemic, prior authorization determinations increased steadily between 2021 and 2023 (Figure 1). The decline in 2020 was likely due to both a decline in utilization, as well as the option for insurers to temporarily pause prior authorization requirements during the public health emergency.

The recent increase in the total number of prior authorization determinations since 2020 corresponds to an increase in Medicare Advantage enrollment. Between 2019 and 2023, the number of Medicare Advantage enrollees rose from 22 million people to 31 million people. In 2019, there were approximately 1.7 prior authorization determinations per Medicare Advantage enrollee. That number dropped at the onset of the COVID-19 pandemic to 1.4 in 2020 and 1.5 in 2021, before returning to the pre-pandemic level of 1.7 determinations per enrollee in 2022 and rising slightly to 1.8 in 2023 (Figure 2).

Medicare Advantage insurers denied 3.2 million (6.4%) prior authorization requests in 2023.

Of the 49.8 million prior authorization determinations in 2023, more than 90% (46.6 million) were fully favorable, meaning the requested item or service was approved in full. However, the remaining 3.2 million prior authorization determinations (6.4%) were unfavorable, meaning they were denied in full or in part by Medicare Advantage insurers. This is slightly lower than the 7.4% of requests that were denied in 2022 (which amounted to 3.4 million denials) (Figure 3). Both the share and number of requests denied was higher in 2023 than in 2019. Across all years, most denials (81% in 2023, data not shown) were denied in full, while a minority of denials were determined to be partially favorable, meaning that only part of the request was approved. For example, the insurer may have approved 10 of 14 requested therapy sessions.

Just 11.7% of denied prior authorization requests were appealed to Medicare Advantage insurers in 2023.

The majority of the 3.2 million denied prior authorization requests were not appealed, similar to previous years. In 2019, just 7.5% of all denials were appealed. That share increased somewhat in 2020 to 10.2% and was relatively stable in 2021 (10.6%) and 2022 (9.9%) (Figure 4). These include appeals of determinations that were both fully and partially denied.

The vast majority of denied prior authorization requests that were appealed were subsequently overturned by Medicare Advantage insurers.

From 2019 through 2023, more than eight in ten (81.7%) denied prior authorization requests that were appealed were overturned (Figure 5). This raises questions about whether the initial request should have been approved, although it could also indicate that the initial request was missing the required documentation to justify the service. In either case, patients potentially faced delays in obtaining services that were ultimately approved because of the prior authorization process.

Variation in Use of Prior Authorization Across Medicare Advantage Insurers in 2023

In 2023, the volume of prior authorization determinations varied across Medicare Advantage insurers, as did the share of requests that were denied, the share of denials that were appealed, and the share of decisions that were overturned upon appeal, meaning people may have different experiences depending on the Medicare Advantage plan in which they enroll.

Across most insurers, a higher number of prior authorization determinations per enrollee was correlated with a smaller share of requests being denied and vice versa. For example, prior authorization determinations for UnitedHealthcare and Humana, the two largest Medicare Advantage insurers, were among the highest (Humana, 3.1 determinations per enrollee) and lowest (UnitedHealthcare, 1.0 determinations per enrollee) observed, and correspondingly, denial rates were below average (Humana, 3.5%) and above average (UnitedHealthcare, 9.1%) for these insurers.

While all Medicare Advantage insurers require prior authorization for at least some services, there is variation across insurers and plans in the specific services that are subject to these requirements. In addition, some insurers waive prior authorization requirements for certain providers, for example, as part of risk-based contracts or through “gold carding” programs that exempt providers with a history of complying with the insurer’s prior authorization policies.

Prior authorization determinations were most common among Humana and Anthem plans.

The number of prior authorization determinations per enrollee ranged from a low of 0.5 determinations per enrollee in Kaiser Permanente plans to a high of 3.1 determinations per enrollee in Humana and Anthem plans (Figure 6). Kaiser Permanente is atypical among insurers in that it generally operates its own hospitals and contracts with an affiliated medical group. Looking across insurers that are more similar, the low end of the range was 1.0 determinations per enrollee in UnitedHealthcare plans. Differences across Medicare Advantage insurers in the number of prior authorization determinations per enrollee likely reflect some combination of differences in the services subject to prior authorization requirements, the frequency with which contracted providers are exempted from those requirements (which may be related to the extent to which providers are affiliated with the insurer), how onerous the prior authorization process is for a particular insurer relative to others, and differences in enrollees’ health conditions and the health care services they use.

Centene denied the highest share or prior authorization requests while Humana denied the fewest.

The denial rate ranged from 3.5% of prior authorization requests for Humana plans to 13.6% of prior authorization requests for Centene plans (Figure 7). The overall denial rate includes requests that were both fully and partially denied (adverse and partially favorable determinations, respectively).

Most insurers that had more prior authorization determinations per enrollee than average denied a smaller share of  requests than average and vice versa. Centene was an exception with both a relatively high number of prior authorization determinations (2.4 per enrollee) and the highest denial rate (13.6%).

Across all insurers, a small share of denials was appealed.

Fewer than one in five denied prior authorization requests were appealed across all firms. The shares ranged from 1.7% for Kaiser Permanente to 18.0% for Cigna (Figure 8).

Across most firms, at least two-thirds of appeals were successful.

Even though most denials were not appealed, when they were, most of the initial decisions were partially or fully overturned. The share of appeals that resulted in favorable decisions was lowest for Kaiser Permanente (42.4%) and highest for Centene (93.6%), which also had the highest share of requests initially denied (Figure 9).

The Use of Prior Authorization in Traditional Medicare

The use of prior authorization is relatively new to traditional Medicare and only used for a limited set of services, including certain outpatient hospital services, non-emergency ambulance transport, and durable medical equipment (see Box 1). The prior authorization process does not change any documentation requirements that are not already necessary for receiving Medicare payment – they are just required earlier in the process. CMS has recently published two reports presenting data on the use of prior authorization in traditional Medicare for fiscal years 2021, 2022, and 2023. These reports include information on the number of reviews completed and the number and share of reviews that were affirmed. For 2021 and 2022 only, the data also include information on appeals and the outcome of the appeal.

Just under 400,000 prior authorization reviews were completed by CMS for traditional Medicare in 2023.

Across the three categories of services that required prior authorization for certain services, there were 216,571 reviews completed in 2021, 260,986 reviews completed in 2022, and 393,749 reviews completed in 2023 (Figure 10). This translates to about 1 prior authorization review per 100 traditional Medicare beneficiaries in 2023 – a rate of about 0.01 per person.

About one-quarter of prior authorization reviews in traditional Medicare denied coverage of the service.

CMS approved (or affirmed) coverage in the majority of prior authorization reviews it completed. CMS reported that 24.8% of requests were denied (or non-affirmed) in 2021, 27.6% of requests were denied in 2022, and 28.8% of requests were denied in 2023 (Figure 11). This reflects 53,680 denied requests in 2021, 72,029 denied requests in 2022, and 113,448 denied request in 2023.

A small share of denied prior authorization requests in traditional Medicare was appealed to CMS.

In both 2021 and 2022, fewer than 5,000 denied prior authorization requests were appealed to the first level. As a share of all denied requests that translates into 8.7% appealed in 2021 and 6.4% appealed in 2022 (Figure 12). Appeals data are presented differently across the two CMS reports. Specifically, the 2023 report includes a separate claims and appeals section for each category of service, which appears to include a broader universe of reviews by Medicare Administrative Contractors, including those for payment of services rendered, than what is presented in the report for 2021 and 2022. Given the differences in the data reported, we present total appeals for 2021 and 2022 only.

Just over one-quarter of appeals to CMS overturned the denied prior authorization request in traditional Medicare.

In 2021, 1,365 prior authorization denials were overturned upon appeal. In 2022, 1,323 prior authorization denials were overturned upon appeal. That represents 28.7% of all first level appeals in 2022 (Figure 13).

The share of prior authorization reviews that were denied, appealed, and overturned varied across service type in traditional Medicare.

In 2022, the largest number of prior authorization reviews in traditional Medicare were for certain hospital outpatient department services (132,565), followed by reviews for certain durable medical equipment, prosthetics, orthotics and supplies (97,334), which are both applied nationwide. Repetitive, scheduled non-emergent ambulance transport had the fewest number of requests (31,087), which in part may reflect that it became a nationwide program after expanding to additional jurisdictions in 2022, as well as the ability to approve multiple ambulance rides through a single request (Figure 14).

Across the three service categories, denials were most common for repetitive, scheduled non-emergent ambulance transport (36.8%), followed by certain durable medical equipment, prosthetics, orthotics and supplies (33.1%), and least common for certain hospital outpatient department services (21.4%). Denied requests for the ambulance transport services were also most often appealed (19.6%); just 6.6% of denials for certain hospital outpatient services and 1.5% of certain durable medical equipment, prosthetics, orthotics and supplies requests were denied (Figure 14).

The share of appeals that resulted in overturning the initial decision also varied widely. Nearly two-thirds (63.9%) of appeals for durable medical equipment, prosthetics, orthotics and other supplies were successful. That compares to 26.3% of appeals for the ambulance transport services and 22.2% of appeals for certain hospital outpatient services (Figure 14).

Note: This brief was updated on January 29, 2025 to incorporate new data and policy changes on emergency contraception.

Emergency Contraception

Emergency contraception (EC) is a form of backup birth control that can be used up to several days after unprotected intercourse or contraceptive failure and still prevent a pregnancy. In 1999, Plan B was the first oral product approved for use in the U.S. as an EC by the Food and Drug Administration (FDA). Since then, more EC products have been approved. Many have confused EC pills with mifepristone, the “abortion pill,” but EC pills do not cause abortion. EC pills prevent pregnancy by delaying or inhibiting ovulation and will not work if the pregnancy is established. This fact sheet reviews the methods of EC, known mechanisms of action, use of EC, and current national and state policies affecting EC access.

What is EC?

EC is used as a back-up birth control method to prevent pregnancy in the event of unprotected sex, sexual assault, or a contraceptive failure, such as a condom breaking. ECs do not terminate a pregnancy, stop the implantation of a fertilized egg, or affect a developing embryo. There are several methods of EC that are available in the U.S. including progestin-based pills, ulipristal acetate, and intrauterine devices (IUDs) (Table 1). The copper and levonorgestrel IUDs can be used after unprotected sex and as ongoing regular contraceptive methods. EC pills are only to be used after unprotected sex and are not intended for use as a regular method.

Progestin-Based Pills

• Plan B was the first oral form of EC to be made available in the U.S. as a pre-packaged dose of pills containing the progestin, levonorgestrel. Now, there are generic alternatives available as well. Progestin-based EC pills use the same hormones found in daily oral contraceptives and are the most widely used form of EC. EC pills are marketed today under the brand name Plan B One-Step and generic names and are available over-the-counter, without a prescription/ (Table 1).
• Progestin-based EC pills do not interrupt or adversely impact an established pregnancy, nor are they medical abortion drugs like mifepristone or methotrexate that end an established pregnancy. Plan B One-Step and the generic versions prevent pregnancy by inhibiting or delaying ovulation or by making it harder for sperm to reach an egg.
• Progestin-based EC is to be taken within 72 hours of unprotected sex in order to be most effective and reduce the likelihood of pregnancy by 81% to 90% when taken in this timeframe.
• There are no known serious side effects associated with progestin-based EC; side effects may include bleeding or spotting, headaches, nausea, and breast tenderness.
• Some research has suggested that efficacy of progestin-based EC is lower among women with Body Mass Index (BMI) levels greater than 25. However, in May 2016 the FDA announced that it had reviewed the available scientific data regarding the effectiveness of EC pills in overweight and obese women, and that the data are inconclusive and did not recommend a labeling change.
• Progestin-based pills can be stored for several years, with Plan B One Step having a shelf-life of four years when stored as directed. Consumers should always check the expiration date on the packaging prior to taking the pill.

Ulipristal acetate: ella

• Ulipristal acetate, marketed as ella, was approved by the FDA in 2010 for sale and use in the U.S.
• ella is a single-dose (30 mg) pill that is effective in preventing pregnancy up to five days after unprotected intercourse, giving women a longer timeframe to prevent unintended pregnancy than Plan B. Its mechanism of action is similar to that of progestin-based EC.
• Study findings show that side effects for ella are comparable to those for Plan B, and some research suggests that its effectiveness appears to diminish at BMI thresholds above 35.
• A 2025 study based on a small sample of women in Mexico suggests that when paired with misoprostol, one of the two drugs used in the FDA’s approved abortion medication regimen, higher doses of ulipristal acetate (60 mg) can terminate a pregnancy. It has already been demonstrated that misoprostol alone terminates early pregnancies. Prior research has shown that a 30 mg dose of ulipristal acetate alone, the dose that is used for emergency contraception in the U.S., does not disrupt an established pregnancy.

Combined/Combination Pills

• Certain daily oral contraceptive pills can also act as EC when taken in doses four or five times higher than the daily dose, although they are not specifically sold as emergency contraception. Oral contraceptive pills contain progestin and estrogen and are taken in two doses 12 hours apart to be effective as EC.
• Combined pills have been found to be safe and effective for preventing pregnancy within 5 days of intercourse.

Copper-T IUD: Paragard

• Available to women since the 1970s, copper-T IUDs are the most effective forms of EC, reducing the risk of pregnancy by more than 99% when inserted within 5 days of unprotected intercourse. IUDs are inserted into the uterus by a health care provider and require a visit to a clinic or provider’s office. They also can be used to effectively prevent subsequent pregnancy for up to 10 years.
• The hormone-free copper-T IUD works by interfering with egg fertilization by preventing sperm from reaching the egg. Previous research suggests the copper IUD inhibits implantation of a fertilized egg, but this mechanism of action has not been conclusively proven.
• Efficacy of copper IUDs does not diminish in women who are overweight or obese.

Levonorgestrel IUD: Mirena and Liletta

• Levonorgestrel IUDs are a type of hormonal IUD that contain the progestin levonorgestrel. To date, only levonorgestrel IUDs that contain 52 mg of levonorgestrel (LNG 52 mg IUDs) have been studied as a form of EC. LNG 52 mg IUDs are marketed under the brand names Mirena and Liletta.
• Like copper-T IUDs, LNG IUDs are inserted into the uterus by a clinician and require a visit to a clinic. LNG IUDs can subsequently be used as a regular form of contraception for up to seven years.
• Levonorgestrel IUDs recently started being used as a method of EC, and research has found that LNG 52 mg IUDs can be as effective as copper-T IUDs when inserted within five days after unprotected intercourse or contraceptive failure. The hormonal IUD prevents egg fertilization by making the cervical mucus impenetrable to sperm.
• Efficacy of levonorgestrel IUDs does not diminish in women who are overweight or obese.

Women’s Use of EC Pills

There have been numerous public health and educational initiatives to increase use of EC. Use of EC pills has increased over the past 20 years. Between 2022 and 2023, the most recent years for which data are currently available, 33% of women ages 15 to 49 who have ever had sex with a male reported they had used EC pills at least once in their lives, an increase from 22% in 2015-2017 (Figure 1).

Younger women are more likely to report that they have ever used EC (Figure 2). More than four in ten women ages 15-24 (44%) and women ages 25-34 (40%) say they have taken EC pills, compared to 25% of women ages 35-49. One in four Hispanic women (40%) and approximately one in three Black and White women (30% and 33%, respectively) report ever taking EC.

Access and Availability

At least one form of oral EC has been available in the U.S. for over a decade and there have been several efforts to broaden women’s access to EC, particularly since its effectiveness window is time-limited.

Over the Counter Access of EC Pills

• Prior to 2006, a prescription was needed for all individuals seeking EC pills. Between 2006 and February 2014, Plan B and its generic equivalent were available without a prescription for men and women 17 and older, but adolescents under 17 needed a prescription.
• In 2014, the FDA removed point-of-sale age requirements for Plan B obtained over the counter (OTC) and generic versions began to enter the OTC market. Currently, several generic brands of EC pills, including Next Choice One Dose, My Way, Fall Back Solo, Take Action, levonorgestrel tablet, and Aftera are available OTC to women of all ages.
• A prescription is still required for ella for women of all ages.

Cost and Coverage

• The Affordable Care Act (ACA) requires most new private health plans (individual and group) to cover without cost-sharing all FDA-approved contraceptive drugs and devices as prescribed, including emergency contraception. This means that although OTC formulations are available without a prescription, the pills are only required to be covered by insurance if they were prescribed. Private insurance plans must also cover the cost of IUDs, as well as services related to insertion, follow up and removal, without cost sharing. However, there have been numerous reports of individuals having to pay out-of-pocket for contraception that they believe should have been fully covered, including for services related to the insertion and removal of IUDs.
• Seven states require private health insurance plans to cover some over-the-counter contraceptives, without any cost-sharing, including EC. For employer-sponsored plans, state coverage requirements generally apply only to fully-insured plans and not to self-funded plans.
• Family planning services are a required benefit under Medicaid. The coverage requirements under Medicaid are different for states that have expanded eligibility under the ACA. These programs must cover all prescribed, FDA-approved contraceptives, meaning that they must cover EC pills if a woman has a prescription. States have discretion in deciding whether they include EC in their traditional full-scope Medicaid programs or family planning expansion programs. A 2021 survey of state Medicaid programs found that while all responding states cover at least one form of EC in their traditional Medicaid programs, some states impose quantity limits. To increase access to contraception, eight states have opted to use state-only funds to cover at least some methods of OTC contraception for their Medicaid enrollees, primarily emergency contraception, without a prescription.
• Without a prescription or insurance, women in most states accessing EC OTC must pay the retail price. Plan B pills and the generic versions average can range from $10 to $50 when purchased OTC. Without insurance, the cost of IUDs can range from $500 to $1,300.
• Individuals without health insurance or who are ineligible for Medicaid may be able to obtain free or low-cost EC (in addition to other reproductive health care services) through a clinic or community health center participating in the federal Title X family planning program. The Title X program may also be an option for lower-income individuals covered by employer-sponsored health insurance plans that are not required to cover contraception because they existed prior to the ACA (known as “grandfathered” plans) or because their employer excludes coverage of emergency contraception based on a religious or moral objection to it. Regardless of the pathway, eligibility for the program is based on the individual’s household income and costs are based on a sliding fee scale.

Online Contraceptive Platforms

• New markets have emerged, particularly since the start of the COVID-19 pandemic, that allow people to obtain health care services without the need for an in-person visit, including several online reproductive health services platforms and smartphone apps where an individual can obtain hormonal contraceptives, including EC pills, outside of traditional health care settings. A 2023 KFF study on telecontraception found that 9% of survey respondents report getting a prescription for emergency contraception.
• These online platforms and apps typically employ licensed medical professionals to determine a client’s eligibility and to prescribe contraception in a similar manner to a clinic. The patient is then either mailed their contraceptives or their prescription is sent to a nearby pharmacy for pick-up.
• Some, but not all, of these platforms accept private insurance and/or Medicaid, and there is considerable variation in out-of-pocket costs for those who self-pay.

Provision of EC in Health Care Settings

• Several major medical and public health organizations, such as the American Academy of Pediatrics, American College of Obstetricians and Gynecologists, and the American Public Health Association, endorse the use of EC and advocate for broader access to EC.
• Counseling and coverage of EC is included as a standard of care in the federal requirements for providing Quality Family Planning Services (QFP). Providers are encouraged to discuss EC with their patients, inform them of its availability, and provide them with an advanced supply of EC pills if the patient requests them. Advance provision of EC can increase the chances that someone will have EC on hand and use it when needed, but research has found that few clinicians regularly provide advance provision of EC pills.
• There have been ongoing efforts to make EC more readily available to survivors of sexual assault. Currently, 21 states and the District of Columbia require that emergency room staff provide EC to women after sexual assault (Figure 3). However, some studies have documented that a sizable share of clinicians are not fully informed about EC options, safety, and efficacy. Additionally, some community health centers have reported challenges stocking IUDs due to high costs.
• In 2015, Indian Health Service (IHS) clarified its policy on access to the OC pill for Native American women, including that a prescription or age verification to access Plan B is not required.

Availability and Access in Pharmacies

• Nine states have laws that allow pharmacists to directly prescribe and dispense EC to women of all ages without obtaining a clinician’s prescription.
• Four states have measures that require pharmacies or pharmacists to fill all valid prescriptions. These policies have been enacted, in part, in response to reports of pharmacists refusing to fill prescriptions for EC pills because they oppose its use on moral or religious grounds.
• Ten states have laws allowing pharmacies or pharmacists to refuse to dispense EC pills on the basis of moral or religious objections.
• Pharmacies are not required to stock EC pills and some studies show that EC pills that may be sold OTC are not consistently stocked on store shelves and are sometimes kept behind the counter or in a locked display due to the high cost of the product. This report also documented misinformation regrading age and ID requirements among pharmacy staff and consumers.

As President Trump begins his second term, the public’s trust in health information from key health agencies has fallen over the past 18 months, continuing a decline that began during the COVID-19 pandemic, finds the new KFF Tracking Poll on Health Information and Trust. 

Just over half (53%) of the public now says they trust the Food and Drug Administration (FDA) to make the right recommendations on health issues at least “a fair amount,” down from nearly two-thirds (65%) in June 2023. The share who says they trust their state and local public health officials fell a similar amount (to 54% now from 64% in 2023). Trust in health recommendations from the Centers for Disease Control and Prevention (CDC) also dipped slightly from 66% in 2023 to 61% now. 

When asked about other health institutions, two-thirds (66%) of the public say they trust scientists at the National Institutes of Health, and just over half trust the Department of Health and Human Services (HHS) (55%) and the Centers for Medicare and Medicaid Services (CMS) (54%) to make the right recommendations on health issues. 

Republicans are far less trustful of each of these health institutions than Democrats, with independents in the middle.

An even larger partisan divide in the opposite direction exists in trust in health information from President Trump and Robert F. Kennedy Jr., his nominee to be HHS Secretary.

Less than half the public overall trust President Trump (42%) and Kennedy (43%) to make the right recommendations on health issues. Among Republicans, however, similar shares say they trust President Trump (84%) and Kennedy (81%) as say they trust their own doctors (84%). 

The public overall continues to place the most trust in their own doctors, though the share who say they trust their own doctors’ health recommendations “a great deal” or “a fair amount” is down from 93% in 2023 to 85% now, mostly due to declining trust among independents and Republicans.

Attitudes Toward Childhood Vaccines Remain Positive, Though Eroding Among Republicans

While large shares of the public continue to express positive attitudes toward childhood vaccines and school vaccination requirements, the poll reveals some erosion in support among Republicans and parents.

About eight in ten (82%) parents of children under age 18 now say they normally keep their child up to date with recommended childhood vaccines such as the one for measles, mumps, and rubella (MMR), down 8 percentage points from 2023.

About one in six (17%) now report delaying or skipping some shots, up from 10% in 2023. The shift is most pronounced among Republican parents: About one in four (26%) now report skipping or delaying some vaccines for their children, up from 13% in 2023.

A large majority of the public continues to believe that the benefits of the MMR vaccines outweigh the risks (80%), while about one in five (18%) say that the risks outweigh the benefits. Among parents, about seven in 10 (72%) say the vaccines’ benefits outweigh the risks, while a quarter (25%) say the risks outweigh the benefits. Republican-leaning parents are twice as likely as Democratic-leaning parents to say that the vaccines’ risks outweigh the benefits (33% vs. 15%), though large majorities of each group continue to say the benefits outweigh the risks.During his campaign, President Trump had vowed to cut federal funding to schools with vaccine mandates, though the poll suggests large majorities of the public and of parents across political parties support such requirements.

Among the public overall, 83% say public schools should require some vaccines for students, allowing for health and religious exceptions. This includes large majorities of Democrats (93%), independents (85%) and Republicans (75%).

Among parents, three-quarters (76%) say public schools should require vaccines, while one in four (24%) say they should not. Two-thirds (66%) of Republican and Republican-leaning parents favor schools requiring vaccines, while a third (34%) say that schools should not require any vaccines.

Other findings include:

• About two-thirds (63%) of adults overall and parents (67%) say they have heard the false claim that the MMR vaccines have been proven to cause autism in children. Just 3% of adults say this false claim is “definitely true.” One-third (33%) say it is “definitely false.” The rest are somewhat uncertain what to believe, saying the claim is either “probably true” (20%) or “probably false” (41%).
• Parents who say that the false claim that MMR vaccines are proven to cause autism is definitely or probably true are much more likely to say they have delayed or skipped some vaccines for their children than parents who say it definitely or probably false (37% and 8%, respectively).
• When asked about the H5N1 bird flu in the U.S., about one in three (34%) say they are “very” or “somewhat” concerned that they or a family member will get sick. About four in 10 (44%) say they are “very” or “somewhat concerned” that there will be a widespread outbreak in the U.S.
• Looking toward the future, a larger share of adults say the U.S. government is now more prepared (40%) rather than less prepared (26%) to deal with a pandemic or widespread health crisis than it was in 2020. Similar shares of adults across partisanship, age, race, and ethnicity say the government is more prepared than it was before.

Designed and analyzed by public opinion researchers at KFF, the survey was conducted Jan. 7-14, 2025, online and by telephone, among a nationally representative sample of 1,310 U.S. adults in English and in Spanish. The margin of sampling error is plus or minus 3 percentage points for the full sample. For results based on other subgroups, the margin of sampling error may be higher.

HealthCare.gov insurers denied nearly one out of every five claims (19%) submitted for in-network services and an even larger share (37%) share of claims for out-of-network services in 2023, a new KFF analysis finds.

The analysis examines the main source of publicly available data on claims denials and appeals for individual-market plans available through the federal HealthCare.gov marketplace. The Affordable Care Act requires certain entities to report data about claims denials and appeals to encourage transparency about how insurance coverage works for enrollees. The publicly available data does not include marketplace plans sold on state-based marketplaces or employer health plans, the nation’s primary source of private health coverage.

The analysis finds a huge variation across HealthCare.gov insurers, which had in-network denial rates as low as 1% and as high as 54% in 2023 in some states. Nationwide, high-volume insurers with higher in-network denial rates across HealthCare.gov states included Blue Cross Blue Shield of Alabama (35% for its 12 plans in that state), UnitedHealth Group (33% across 274 plans in 20 states), Health Care Service Corporation (29% across 915 plans in four states), Molina Healthcare (26% across 72 plans in nine states), and Elevance Health (23% across 154 plans in seven states).

Of limited information available on in-network claims denial reasons, the most common reason for denial was a general “other” reason (34%), while 16% involved excluded services, 9% involved lack of prior authorization or referral, and 6% involved medical necessity. Other common reasons for denials included administrative issues (18%) and exceeding benefit limits (12%).

Consumers appealed about 1% of denied in-network claims in 2023. Following those appeals, insurers often upheld their initial denials (56%), and consumers rarely took the next step to file an external appeal.

The analysis examines data from the Centers for Medicare and Medicaid Services on 425 million claims submitted to 175 insurers selling marketplace coverage in 2023, the most recent year available. Additional data files with insurer- and state-specific information are available at kff.org.

The Affordable Care Act (ACA) requires insurers to report transparency data for all non-grandfathered health plans sold on and off the Marketplace, including fully-insured and self-insured employer group health plans. The law requires data to be available to federal and state insurance regulators and to the public. However, federal implementation of this requirement has so far been limited to qualified health plans (QHP) offered on the federally facilitated Marketplace (HealthCare.gov) and does not include QHPs offered on state-based Marketplaces or group health plans. This brief analyzes federal transparency data released by the Centers for Medicare and Medicaid Services (CMS) on claims denials and appeals for non-group qualified health plans (QHPs) offered on HealthCare.gov in 2023. A downloadable working file based on CMS’s public use file is available on the right-hand side of this brief.

Key Takeaways

• Insurers of qualified health plans (QHPs) sold on HealthCare.gov denied 19% of in-network claims in 2023 and 37% of out-of-network claims for a combined average of 20% of all claims.
• The in-network denial rate ranged from 1% to 54%. There was significant variation by insurer and by state.
• Of limited information available on in-network claims denial reasons, the most common reason cited by insurers was “Other” at 34% followed by administrative reasons (18%), excluded service (16%), lack of prior authorization or referral (9%), and only 6% based on lack of medical necessity.
• Consumers rarely appeal denied claims (fewer than 1% of denied claims were appealed) and when they do, insurers usually uphold their original decision (56% of appeals were upheld).
• Marketplace enrollees filed 5,000 external appeals in 2023, or 3% of all upheld internal appeals. Due to the suppression of small values, the rate at which external appeals were upheld could not be calculated.

Introduction

The impact of claims denial is widely recognized by enrollees. The 2023 KFF Survey of Consumer Experiences with Health Insurance found that 58% of insured adults said they have experienced a problem using their health insurance, including denied claims. Four in ten (39%) of those who reported having trouble paying medical bills said that denied claims contributed to their problem.

As a part of the annual QHP certification process, issuers (referred to as insurers in this brief) must report certain denied claims information to CMS for plans that were offered in the previous year that they want to offer in the upcoming year. Data does not include information about denied requests for prior authorization (a claim decision made before a service is provided). The dataset only includes information about claims for benefits (medical and prescription drugs combined) made after a service was provided (post-service claims).

Insurers participating in the Marketplace in 2025 reported aggregated data on all HealthCare.gov QHPs they offered in 2023. Additionally, plan-level data from 2023 are reported for plans returning in 2025, including the number of in- and out-of-network claims submitted and denied, and reasons for claims denials. Among insurers participating in HealthCare.gov states in 2023, 43 are not participating in 2025 so they did not provide claims denial information. Among returning insurers, such denial information was only reported for 69% of their claims (the share of claims attributable to returning plans), as not all plans offered in 2025 were also offered in 2023.  Additionally, only 40% of plans in the dataset were offered in 2023 and are included in the plan-level reporting for denial reasons. See the Methods and Data Limitations section for more details.

Claims Denials and Appeals in 2023

Insurer-level Claims Denials Data

Insurers reported receiving 425 million claims in 2023, with 92% (392 million claims) filed for in-network services. Of these in-network claims, 73 million were ultimately denied, resulting in an average in-network denial rate of 19% (Figure 1). Out-of-network claims totaled 33 million, with an overall higher denial rate of 37%. Claims that were initially denied then subsequently resubmitted and paid are not included as denied claims in the denial rate.

Although the composition of HealthCare.gov states has continued to change since the inception of transparency reporting, the overall in-network denial rate in 2023 is similar to those from other analyses conducted by KFF (Figure 2).

Insurer denial rates for in-network claims received in 2023 varied widely, ranging from 1% to 54%. Twenty-two of the 175 reporting insurers had an in-network denial rate of less than 10% while twenty-nine insurers had a denial rate of 30% or more (Figure 3).

Denial rates also varied geographically, as shown in Figure 4. The state with the highest average in-network denial rate for HealthCare.gov insurers was 34%, in Alabama, and the lowest was 6%, in South Dakota. Average denial rates have the potential to obscure variation. For example, while the average denial rate for insurers in Florida (16%) was slightly below the national average (19%), denial rates for insurers in Florida had more variability than another other state included in this analysis, ranging from 8% to 54% (the highest single insurer-level denial rate in the country).

Limited ACA transparency data collected by the federal government continue to show wide disparities in the rate at which Marketplace plans pay claims. While HealthCare.gov insurers denied an average of 19% of in-network claims in 2023, some insurers reported denying a much higher share. Table 1 shows denial rates for claims filed by parent companies that received more than 5 million claims within HealthCare.gov states in 2023. For in-network claims processed by these parent companies, the average in-network denial rate was 19%, ranging from 13% to 35% by parent company. (Blue Cross and Blue Shield parent companies from different states are separated in this table because they operate independently.)

Plan-level Claims Denial Data

In all, insurers reported on 49 million denied in-network claims at the plan level for the 2023 coverage year. Denial rates varied only slightly between most plan metal levels. On average, in 2023, HealthCare.gov insurers denied 19% of in-network claims in their bronze plans, 18% in silver plans, 18% in gold plans, 15% in platinum plans, and 27% in catastrophic plans (Figure 5).

CMS requires HealthCare.gov insurers to report the reasons for in-network claims denials at the plan level. Specified denial reason categories include:

• Denials due to lack of prior authorization or referral
• Denials due to an out-of-network provider
• Denials due to an exclusion of a service
• Denials based on medical necessity (reported separately for behavioral health and other services)
• Denials due to enrollee benefit reached
• Denials due to a member not being covered
• Denials due to investigational, experimental, or cosmetic procedure
• Denials for administrative reasons (which include claims that were duplicate, missing information, untimely, for an unapproved provider, or that met other criteria)
• Denials for all other reasons not specified above.

Denials due to enrollee benefit reached (such as a limit on the number of physical therapy visits allowed per year); member not being covered at the time of service; investigational, experimental, or cosmetic procedure; and administrative reasons were reported on for the first time for 2022 data (in filings for application for the 2024 plan year).

A claim might be denied for more than one reason and on more than one submission. For example, if the initial submission of a claim misspelled a patient’s name and was denied because the patient could not be identified, the claim may be denied again after being corrected and resubmitted if the claim were for a service that was not covered. Additionally, denial reasons are also reported for claims that are ultimately paid if they are resubmitted to correct the deficit or are successfully appealed. Insurers reported about 71 million denial reasons for in-network claims that were denied at some point in the adjudication process. The adjudication process employed by the insurer may affect how denial reasons are reported. Although publicly reported data allow for multiple reasons throughout the life of a claim, in practice, insurers may file denial reasons sequentially and not capture all applicable reasons for denying claims, such as denying claims from an unidentifiable enrollee before determining whether the claim was for a medically necessary procedure.

The distribution of in-network denials by reason is shown in Table 2. Of in-network claims, about 16% of denials were because the claim was for an excluded service, 9% due to lack of prior authorization or referral, and only about 6% based on medical necessity. The share of denial reasons related to administrative reasons is 18%, the most common reason aside from “other” (34%). The share of denial reasons attributed to “other” reasons in 2023 is significantly smaller compared to the 2021 data due to the reporting of new specific denial reasons, notably administrative reasons. Among all in-network claims filed, 6% required a resubmission (not necessarily for administrative reasons though). A resubmission may occur when the original claim was incomplete, contained errors, or was rejected for non-compliance with billing guidelines.

Insurers also had wide variability in their use of denial reasons. While about 6% of all in-network claims denials by HealthCare.gov plans were based on medical necessity, several plans reported much higher shares for medical necessity reasons. For example, 30% of denial reasons for Cigna HealthCare of North Carolina were due to medical necessity. Similarly, while about 9% of all in-network denials by HealthCare.gov plans were based on lack of prior authorization or referral, some plans reported a much larger share. For example, 97% of denial reasons for Blue Cross Blue Shield of Arizona were for lack of prior authorization or referral.

Plans may apply utilization review techniques differently. For example, individual insurer policies and practices may affect the balance between denials for failure to obtain referral/prior authorization and medical necessity denials, as greater use of prior authorization would shift utilization review to before a service is provided and possibly decrease the number of denials due to medical necessity. However, without more detail on the types of claims subject to these denials, it is not possible to discern the possible implications for patients. Recent federal regulations may provide further insight into the prior authorization process and what services typically require prior authorization for Healthcare.gov plans. Furthermore, denials captured in this data do not reflect the share or types of services covered by insurers.

Appeals Data

CMS requires insurers to report the total number of denied and internally appealed claims at the insurer level. Internal appeal is a process that allows a consumer to challenge a denied claim made by their health insurer. As in KFF’s previous analysis of federal claims denial data, we find that consumers rarely appeal denied claims and when they do, insurers usually uphold their original decision.

Appeal to Insurer (Internal Appeal). Of the 73 million in-network denied claims in 2023, HealthCare.gov consumers appealed 376,527 – an appeal rate of less than 1%. Insurers upheld 211,393 (56%) denials on appeal. Relatedly, the 2023 KFF Survey of Consumer Experiences with Health Insurance found that only one in ten insured adults who reported experiencing a problem with their insurance in the past year had filed a formal appeal.

Appeal to Third Party (External Appeal). Consumers whose denial is upheld at internal appeal may have the right to an independent external appeal (also called external review) for certain types of claims. Among insurers that reported at least 10 external appeals in 2023, Marketplace enrollees externally appealed at least 5,000 claims in 2023 (CMS suppresses reporting of observations lower than 10 so the number of externally appealed claims could be higher). Among insurers that reported at least 10 external appeals in 2023, 3% of upheld appeals were externally appealed. Due to the suppression of small values, the rate at which external appeals were upheld could not be calculated.

It is not well known that consumers can appeal claims denials through an external appeal process. KFF’s 2023 consumer survey found that just 40% of consumers believed they have a legal right to appeal to a government agency or independent medical expert, while 51% said they were unsure if they had appeal rights, and 9% did not believe they had this right. Furthermore, Marketplace enrollees (34%) were less likely to know they had external appeal rights compared to those with Medicare (58%) and Medicaid (45%).

Other Data Sources

Absent data on how often insurers in other markets deny claims, it is difficult to put ACA transparency data in context. Below are other sources of claims denial data.

Covered California

California requires insurers to report data on claims received and denied each year for both in- and out-of-network services, in a manner similar to HealthCare.gov insurers. Among insurers submitting complete 2023 claims data to Covered California, the in-network denial rate was 21%, similar to HealthCare.gov insurers. One insurer had a denial rate of 87%. When excluding this insurer from the analysis, the overall claims denial rate among Covered CA insurers was 19%.

Specified denial reason categories are the same for both Marketplaces. At the plan-level, about 14% of in-network denials were due to lack of prior authorization or a referral, followed by about 6% due to administrative reasons, and about 1% for lacking medical necessity.

The appeal rate for Covered CA insurers (1%) was similar to HealthCare.gov insurers. Among all Covered CA insurers with complete data, about 40% of internal appeals and 47% of external appeals filed were upheld, substantially lower than HealthCare.gov insurers. Like denials, one insurer also represented a large share of the appeals data reported. When excluding that insurer, the rate of internal appeals upheld by Covered CA insurers was 61%.

Connecticut Health Insurance Report Card

Connecticut law requires private health insurers in all market segments with at least 1,000 enrollees to report annual data on claims payment practices, prior authorization requests and denials, claims denial reasons, and several other metrics (Table 3). The state insurance department publishes the aggregated data at the insurer level in its annual Consumer Report Card, which includes data from the largest insurer and is intended to inform consumer decision-making. Claims denial data include the total number of claims received and the total number of claims denied by reason.

Insurers in Connecticut reported receiving more than 11.8 million claims and denying more than 2.7 million claims in 2023, for an overall denial rate of 23%. In 2023, the largest shares of claims denials were for reasons related to the benefit not being covered (11.4% of denials) and for other reasons not specified (62.4%).

Connecticut’s claims denial data are not directly comparable to those reported by Covered CA or HealthCare.gov insurers for several reasons, including that Connecticut’s data includes group health plans, denial reasons are reported at the insurer level rather than the plan level, and claims data in Connecticut are not separated by network status.

National Association of Insurance Commissioners

The National Association of Insurance Commissioners (NAIC), via the Market Conduct Annual Statement (MCAS), collects uniform data annually on claims denials, prior authorization requests, appeals, and more from many insurers in the individual and group markets in nearly every U.S. state. MCAS data are intended to help state insurance regulators monitor the market conduct of insurance companies, and insurers can use this information to identify areas to improve performance. However, full MCAS health insurance data are shared with state regulators only, not the general public or CMS. A limited national summary published by the NAIC shows that the average claims denial rate for both in- and out-of-network claims (excluding pharmacy) in 2023 was about 16%.

Medicare Advantage and Medicaid Managed Care

Medicare Advantage plans have come under scrutiny in recent years over concerns about policies and processes related to claims and prior authorization denials. According to a 2024 KFF analysis of federal data, Medicare Advantage plans denied (fully or partially) 3.4 million prior authorization requests for health care services in 2022, for an overall denial rate of about 7%, a share that has increased over the past few years. (Prior authorization is a process used by health insurers that requires providers to obtain approval before a service or other benefit is covered.) Additionally, a 2018 federal report found that 8% of claims and prior authorization requests (combined) submitted to Medicare Advantage plans between 2014 and 2016 were denied by insurers, which was less than half the denial rate reported, on average, by HealthCare.gov insurers during that period.

Medicaid managed care organizations (MCOs) also may require prior authorization. A 2023 federal report found that Medicaid MCOs denied more than 2 million prior authorization requests in 2019 for an overall prior authorization denial rate of nearly 13%–more than 2 times higher than the Medicare Advantage rate. However, these data are not directly comparable to the HealthCare.gov data being analyzed for this report, as the former pertains specifically to prior authorization denials while the latter covers post-service claims denials.

Looking Forward

Although research and investigations into health insurer practices have garnered attention from lawmakers and patient advocates over the past several years, the December 2024 killing of UnitedHealthcare’s CEO ignited broad public outrage over insurer claims denials. According to a January 2025 KFF public opinion poll asking about certain health care priorities for Congress and the Trump administration, most people (55%) say more closely regulating insurers’ decisions to approve or deny claims for health services or prescription drugs should be a “top priority.” While prospects for significant changes in response to the public outrage may be limited, interest in providing the public with more transparency about how insurer claims review and appeals operate could, in the same way as providing more accurate price transparency information, better enable consumers and employers to make more informed choices when purchasing private coverage. Efforts might include:

Including more specific information in existing datasets.

Using the current data, the proportion of claims denied for a given reason cannot be calculated. For example, it is not possible to know the share of services that were denied due to a lack of medical necessity. Federal reporting on denials could be more useful when presented as claims ever denied for a given reason, instead of tallying the total reasons. Also, reporting that includes denial information about all claims from all insurers in the previous year, and not just those attributable to plans that are returning to the Marketplace next year, could be useful. Additionally, information about the types of services approved and denied (e.g., specialty of service and type of prescription drug) would give a more comprehensive picture of insurer practices and what type of care was actually covered by an insurer or employer. Information about appeals, especially external appeals, could provide insight into how this consumer protection mechanism is working for patients. Information about what services required prior authorization and how often the prior authorization itself is approved and denied is another data element not included in the CMS Marketplace public use file but is included in NAIC MCAS data not available to the public.

Providing claims denial information about employer coverage.

Employer-sponsored insurance covered 154 million people under 65 in 2023. Since most Americans have employer-sponsored coverage, efforts to provide more information to this group may be a way to begin to address concerns about insurer denials. A proposed regulation from 2016 that was never finalized would have added a claims denial metric to reporting required under the Employee Retirement Income Security Act of 1974 (ERISA). Also, in 2024, some members of Congress, urged the Department of Labor (DOL) to collect information on claims and claims denial, citing reports of “widespread denials of health benefits.” An outside advisory panel to DOL also recently issued recommendations for increased data collection, among other claims and appeal reforms in this area. Also, federal mental health parity regulations updated in 2024 will require employer plans (and non-group plans) to collect and evaluate certain data, including the number and percentage of certain claims denials.

State-level initiatives.

There has been some activity at the state level to provide more transparency into claims denials and prior authorization requests. For example, in addition to California’s and Connecticut’s requirements for reporting claims denial data, Vermont requires insurers of state-regulated plans to report certain pre- and post-service claims denial data to the state, including breakdowns by mental health, substance use disorder services, and prescription drugs. Insurers in Oregon are required to report to the state claims denial and appeals data for behavioral health services compared to certain medical and surgical services. Additionally, Washington state requires insurers to report certain data related to prior authorization requests to the state, issue prior authorization determinations within certain timeframes, and use a standardized and streamlined prior authorization process. All of these states make at least some of this information available to the public annually. Going forward, more states may act to enact similar initiatives at the state level. These state laws, however, do not apply to self-insured health plans sponsored by private employers, which cover most insured Americans under age 65. Absent more uniform and complete data at a national level, efforts to fully understand and address issues related to health insurance claims denials will remain limited.

Section 1115 Medicaid demonstration waivers offer states an avenue to test new approaches in Medicaid that differ from what is required by federal statute, if [in the HHS Secretary’s view] the approach is likely to “promote the objectives of the Medicaid program.” They can provide states additional flexibility in how they operate their programs, beyond the considerable flexibility that is available under current law. Waivers generally reflect priorities identified by states as well as changing priorities from one presidential administration to another. Nearly all states have at least one active Section 1115 waiver and some states have multiple 1115 waivers. This brief explains what Section 1115 waivers are and how they are used, summarizes key waiver requirements, and outlines the application and approval process.

What are waivers and how are they used?

Authority & Purpose. Under Section 1115 of the Social Security Act, the Secretary of Health and Human Services (HHS) can waive certain federal Medicaid requirements.¹  In addition, the Secretary may permit states to use federal Medicaid funds in ways that are not otherwise allowed. Each administration has some discretion over which waivers to approve and encourage (Table 1). While the Secretary’s waiver authority is broad, it is not unlimited. Section 1115 waivers have been challenged in court. The Secretary does not have authority to waive some elements of the program, such as the federal matching payment system for states, or requirements that are rooted in the Constitution, such as the right to a fair hearing.² 

Waiver Scope/Use. Waivers have been used to expand coverage or benefits, change policies for existing Medicaid populations (e.g., testing premiums or other eligibility requirements), modify delivery systems, restructure financing or authorize new payments (e.g., supplemental payments or incentive-based payments), as well as make other program changes. Waivers vary in size and scope. States can obtain “comprehensive” Section 1115 waivers that make broad program changes or narrow waivers focused on a specific population. Some policies introduced through 1115 waivers can only be implemented through Section 1115 authority while others could be implemented under other authorities (e.g., State Plan authority or 1915(c)). MACPAC analysis found about half of all Medicaid spending (in FY 2019) was authorized under Section 1115 demonstrations, but most of that spending could have been covered without an 1115 waiver. States may seek to include some populations or services in Section 1115 waivers that could be covered under other authorities to capture “budget neutrality” savings (discussed in more detail below). For example, although states can implement mandatory managed care for most populations under other authorities (e.g., State Plan or 1915(b)), states may implement managed care under 1115 authority to show budget neutrality savings, which can be used to finance other waiver costs that are not otherwise covered / allowed by Medicaid. In addition, many states have comprehensive waivers that make broad and intertwined program changes and may include both provisions that require 1115 authority as well as provisions that could be implemented without a waiver.

What are the rules about waiver financing?

Financing. Under long-standing policy and practice (although not required by statute), waivers must be “budget neutral” to the federal government over the course of the waiver. In other words, federal costs under an 1115 waiver may not exceed what they would have been for that state without the waiver. Typically, budget neutrality calculations are determined on a per enrollee basis—so, per enrollee spending over the course of the waiver cannot exceed the projected per enrollee spending calculated in the “without-waiver baseline” (putting states at risk for the costs per individual but not for the number of individuals enrolled). Waiver budget neutrality—measured against the estimated without-waiver baseline over the entire demonstration period—is not the same as a federal per enrollee limit on spending set at rates lower than expected under current law to generate federal savings. Budget neutrality calculations and the use of “budget neutrality savings” (to fund the federal share of costs not otherwise allowed) are negotiated between states and CMS (and the Office of Management and Budget (OMB)).

Because Section 1115 budget neutrality is not defined in statute or regulations, CMS agency policy and guidance to states has changed over time. For example, the Trump administration made changes to 1115 waiver budget neutrality policy in 2018, limiting the amount of federal funds that could be used for waiver spending. Later, the Biden administration made changes to Section 1115 budget neutrality policies that could provide greater flexibility for states to design and implement 1115 demonstration programs, including health-related social needs initiatives.

What are waiver timelines and processes?

Waiver Timeframe. Section 1115 waivers generally are approved for an initial five-year period and can be renewed, typically for three-to-five-year periods. Some waivers have been continually renewed over many periods, allowing waiver operations to continue for many years. Under the Trump administration, in a departure from prior policy, CMS approved waiver extension requests for up to 10 years.

Incoming administrations may let waivers expire, choosing not to renew certain waiver provisions if they don’t align with the administration’s waiver priorities or if they determine the provisions do not promote the objectives of the Medicaid program. Additionally, outlined in waiver approval terms and conditions, CMS reserves the right to withdraw Section 1115 waiver or expenditure authorities at any time (including those already in operation under an active/approved waiver). The Biden administration withdrew Medicaid work requirement waivers in all states that had approvals, concluding that the provisions do not promote the objectives of the Medicaid program. States can appeal withdrawal decisions to the HHS Department Appeals Board and/or challenge recissions in court.

Transparency, Public Input, and Evaluation. The Affordable Care Act (ACA) made Section 1115 waivers subject to new rules about transparency, public input, and evaluation.³  Regulations require public notice and comment periods to occur at the state and federal levels before CMS approves new Section 1115 waivers and extensions of existing waivers. Although the final regulations on public notice do not require a state-level public comment period for amendments to existing/ongoing demonstrations, CMS has historically applied these regulations to amendments as well. The Trump administration did not enforce state-level public notice and comment procedures maintained by previous administrations for certain 1115 waiver requests, including waivers that proposed significant changes.

The ACA also implemented new evaluation requirements for Section 1115 waivers, including that states must have a publicly available, CMS-approved evaluation strategy. States have traditionally also been required to submit quarterly reports as well as an annual report to HHS that describes the changes occurring under the waiver and their impact on access, quality, and outcomes.

Waiver Application, Monitoring, and Evaluation Process. Medicaid policy changes, including through Section 1115 waivers, may require state legislative action or may be authorized at the direction of the governor. Once proposed policy changes have been formulated, a demonstration waiver proposal must be drafted. Key steps in the waiver process include (Figure 1):

• State public notice and comment.⁴  Prior to submitting an 1115 waiver application (or extension request) for official federal review, states must provide a 30-day public notice and comment period and must hold at least two public hearings, sharing sufficient detail about the proposed waiver to allow for meaningful public input. The state must share waiver proposal materials on its website. Federal rules also require tribal consultation (with federally recognized tribes) prior to application submission.
• Waiver application submission. State waiver applications must contain specific components including a comprehensive description of the demonstration, enrollment estimates (including for each category of beneficiaries impacted by the demonstration), a list of specific requested waiver and expenditure authorities, research hypotheses, and written documentation of the state’s compliance with public notice requirements, with a report of the issues raised and how the state considered those comments when developing the application.
• Federal public notice and comment.⁵  The federal government conducts a review for application completeness and sends the state a notice of receipt, indicating the start date of a 30-day federal comment period. CMS will publish the waiver application on its website and must make the comments received publicly available. Rules require CMS to review and consider all comments submitted by the deadline.
• Federal review and negotiation. CMS reviews the waiver application sometimes with the involvement of other HHS agencies and the Office of Management and Budget (which reviews the budget neutrality component). Significant negotiation may occur between the state and HHS.
• Approval. If a waiver is approved, CMS issues an award letter to the state (also published on Medicaid.gov), listing the specific sections of the Social Security Act and applicable regulations that are being waived or modified and the types of expenditures allowed as well as the “terms and conditions” of approval, including a budget neutrality agreement. There has been significant variation in the length of time it takes to get final approval of a waiver. NAMD has noted the typical negotiation / approval timeframe ranges from 6 months to 2 years.
• Implementation plans and protocols. For some waiver initiatives, CMS may require states to submit detailed implementation plans or protocol documents for review and approval.
• Monitoring & Evaluation.^6 ,7  Because Section 1115 authority is intended for research and demonstration purposes, states must have an approved evaluation strategy in place that is publicly available. States are required to submit an interim evaluation report (one year before a waiver’s expiration or with a renewal application) and a summative evaluation (due 18 months after a waiver period ends). States have traditionally also been required to submit quarterly reports as well as an annual report to HHS that describes the changes occurring under the waiver and their impact on access, quality, and outcomes. States must also hold public forums to solicit feedback following waiver approval / implementation.
• Amendments & Renewals. States may submit “amendment” requests to CMS to alter existing / ongoing 1115 demonstrations. To ensure public transparency, CMS has historically required states to follow public notice and comment rules (even though final regulations left open the applicability of public notice requirements to proposed amendments). Extension / waiver renewal requests must contain specific components enumerated by CMS, including evaluation results.

*State waiver applications (including amendment and renewal requests), CMS issued approval documents, required implementation plans and protocols, and monitoring and evaluation reports are made publicly available on Medicaid.gov (search by state).

1. 42 U.S.C. § 1315. ↩︎
2. The Secretary’s waiver authority is limited to the provisions of 42 U.S.C. § 1396a, provided that waivers are demonstration projects that further Medicaid program objectives. 42 U.S.C. § 1315. ↩︎
3. [3] §10201(i) of P.L. 111-148 added a new subsection (d) to Section 1115 of the Social Security Act. CMS issued final regulations implementing these provisions of the ACA (42 CFR Part 431 where a new Subpart G is added). ↩︎
4. § 431.408 ↩︎
5. §431.416 ↩︎
6. §431.424 ↩︎
7. §431.428 ↩︎