Contents

• Background
• Qualifying Payment Amount
• Transparency in Coverage In-Network Rates
• Data and Methods Overview
• Datasets
• Methods
• Data Limitations
• Analytical Findings
• Aggregate Analysis
• State-Level Analysis
• Summary and Policy Implications

Summary

The No Surprises Act (NSA), part of the Consolidated Appropriations Act, 2021 (P.L. 116-260), established various consumer protections related to surprise billing—that is, circumstances in which individuals receive large, unexpected medical bills when they are unknowingly, and potentially unavoidably, treated by out-of-network (OON) providers.

The NSA also specifies a methodology to be used to determine how much insurers must pay OON providers for care in specified situations. Under the federal payment methodology, when an insurer and an OON provider cannot agree on the relevant payment amount, either party may initiate an independent dispute resolution (IDR) process before a private arbitrator. In a process commonly described as "baseball-style" arbitration, the insurer and provider each submit to the IDR entity an amount representing what the party believes the total price of care should be (i.e., a payment offer), and the IDR entity then selects between the parties' payment offers.

Since the IDR process became operational in April 2022 and through 2024, its utilization has exceeded initial expectations, with over 2 million initiated disputes. Providers have initiated nearly all disputes, and have been very successful in the IDR process; in 2023 and 2024, the IDR entity selected the provider's offer in approximately 80% and 85%, respectively, of the disputes that resulted in a payment determination.

Emergency service disputes represent the most common type of dispute. The median prevailing offers for emergency service disputes ranged from 2½ to 3 times the qualifying payment amount (QPA) in 2023 and 2024. The QPA is a statutory benchmark defined as an insurer's 2019 median in-network rate for a particular item or service furnished by a provider in the same/similar specialty and location, indexed for inflation.

The QPA is a significant part of the IDR process and has been a focus of interest groups during NSA implementation. Providers have expressed concern about the lack of QPA transparency and have asserted that insurers are incorrectly determining the QPA. Insurers have indicated that the QPA represents the market rate for services.

Outcomes from the IDR process affect health care costs through payments that insurers make to providers for OON care and can affect subsequent negotiation of in-network rates between insurers and providers. The Congressional Research Service (CRS) analyzed IDR process emergency service outcomes in 14 states against the in-network rates of IDR-participating providers to compare QPAs, insurer and provider offers, and prevailing offers.

For IDR emergency service disputes determined in the first half of 2024, (1) over 95% of disputes had a prevailing offer that was above the 2025 median in-network rate among IDR-participating providers, and (2) the median prevailing offer in the IDR process for emergency service was over three times the 2025 median in-network rate among IDR-participating providers. In addition, (3) the median QPA was slightly higher than the 2025 median in-network rate of IDR-participating providers.

Further analysis by state highlights that (1) in all 14 states assessed, the median prevailing offer for emergency service disputes determined in the first half of 2024 was greater than the 2025 median in-network rate of IDR-participating providers; however, (2) the degree to which the median prevailing offer exceeded the median in-network rate of IDR-participating providers varied by state. In Nevada and California, the median prevailing offer was slightly higher than the median in-network rate, whereas Colorado saw a median prevailing offer that was over six times the median in-network rate of Colorado IDR-participating providers. In total, nine states had a median prevailing offer that was more than three times the median in-network rate. Additionally, (3) the relationship between the QPA and median in-network rates varied across states. In six states, the median QPA was less than the median in-network rate; in eight states, it was greater than the median in-network rate.

The No Surprises Act (NSA), part of the Consolidated Appropriations Act, 2021 (P.L. 116-260), established various consumer protections related to surprise billing—that is, circumstances in which individuals receive large, unexpected medical bills when they are unknowingly, and potentially unavoidably, treated by out-of-network (OON) providers.¹ The law generally recognizes surprise billing circumstances to include OON emergency services, OON nonemergency services provided during a visit at an in-network facility, and OON air ambulance services. In those situations, the NSA generally limits the amount consumers pay for care and specifies a methodology to be used to determine how much insurers must pay OON providers for care.²

Under the federal payment methodology, when an insurer and an OON provider cannot agree on the relevant payment amount, either party may initiate an independent dispute resolution (IDR) process before a private arbitrator, referred to as an IDR entity.³ In a process commonly described as "baseball-style" arbitration,⁴ the insurer and provider each submit to the IDR entity an amount representing what the party believes the total price of care should be (i.e., a payment offer). The IDR entity selects between the parties' payment offers after considering a list of statutory factors, including the item's or service's qualifying payment amount (QPA). The QPA is defined as an insurer's 2019 median in-network rate for a particular item or service furnished by a provider in the same or similar specialty and location of service, indexed for inflation. Each participant pays a flat fee to participate in the IDR process, and the non-prevailing party also pays the IDR entity's fee.

The Congressional Budget Office (CBO) has indicated that outcomes from the IDR process can affect health care costs in at least two ways. First, the IDR process directly determines payments that insurers make to providers for OON care. Second, in general and more broadly, IDR outcomes can affect costs through subsequent negotiation of in-network rates between insurers and providers, where insurers or providers can leverage the threat of the provider being OON and having the ability to use the IDR process to obtain lower or higher in-network rates.⁵ Both sets of costs have implications for premiums and, subsequently, the federal budget.

Given the potential implications of the IDR process on health care costs, researchers have sought to analyze and contextualize IDR outcomes through several approaches. Studies, for example, have examined IDR outcomes relative to QPAs and concluded that the median IDR outcome for emergency services generally ranged from roughly 2½ to 3 times QPA in each quarter in 2023 and 2024.⁶ Another series of studies compared IDR outcomes to Medicare rates and found that the mean IDR decision for emergency services in the second half of 2023 was four times the price Medicare would pay.⁷

This CRS report contributes to the existing literature by using a more contemporary and market-based benchmark of IDR-participating providers' in-network rates, offering a distinct perspective on the extent to which IDR determinations for services provided OON diverge from the negotiated in-network rates of providers that participate in the IDR process. As such, this report analyzes a subset of IDR outcomes for OON emergency services in selected states relative to certain available data on in-network emergency services rates of IDR-participating providers. Specifically, this report uses the in-network rates of providers from selected states that (1) CRS identified as having participated in at least one IDR emergency service dispute in the first half of 2024, (2) had in-network rates that were reported by health insurers under the Transparency in Coverage requirements in May 2025, and (3) had in-network rates that were included in a database maintained by Payerset, an aggregator of transparency data. There were 14 states included in this analysis, which were selected to minimize the effects of geographic variation between underlying datasets.

This report begins with a brief background on the IDR process and an overview of the data sources, data limitations, and methods used for this research before describing the analytical findings. It concludes with a summary and potential policy implications. A more thorough description of the methods can be found in the Appendix A, and a table describing different selected characteristics of the IDR disputes analyzed in this report can be found in Appendix B.

Background

From when the IDR process became operational on April 15, 2022, through 2024, over 2.3 million disputes were initiated, far exceeding initial expectations.⁸ Providers initiated nearly all of these disputes and have been very successful in the IDR process. In 2023 and 2024, the IDR entity selected a provider's offer in approximately 80% and 85%, respectively, of the OON emergency and nonemergency service disputes that resulted in a payment determination.⁹

Dispute prevalence and outcomes vary by specialty. Emergency service disputes represented the most common type of dispute (over 50% of all determinations in 2023 and roughly 45% of all determinations in 2024). Median prevailing offers for emergency service disputes ranged from 2½ to 3 times the QPA in each quarter in 2023 and 2024 and were 4 times Medicare rates in the second half of 2023.¹⁰

In the IDR process, the Centers for Medicare & Medicaid Services (CMS) has indicated that providers often benchmark their offers to historical OON payment amounts with the disputing insurer and historical in-network rates with the disputing insurer or a different insurer in that state.¹¹ Insurers often benchmark their offers to their QPA for the particular item or service.¹²

When selecting between insurer and OON provider payment offers, IDR entities are allowed to consider a list of statutory factors, including the QPA, and information related to specified "additional circumstances." These additional circumstances may include the provider's level of training, experience, and quality and outcome measurements; the market shares held by the insurer and the provider; the individual's acuity; the teaching status, case mix, and score of furnished services; demonstration of good faith efforts made by the insurer and the provider to enter into network agreements; contracted rates from the previous four years between the insurer and the provider; and additional information submitted by parties regarding their offer.¹³

IDR entities are prohibited from considering usual and customary charges; the provider's billed charges; and Medicare, Medicaid, TRICARE, and other public payer rates for the disputed service.¹⁴

Qualifying Payment Amount

The QPA plays a significant role in the IDR process. In addition to serving as a statutorily defined factor considered by IDR entities and a common benchmark for insurers in the IDR process, it serves as the amount used to determine consumer cost sharing in surprise billing situations.¹⁵

The QPA is an insurer-specific cost measure for a particular service. The insurer generally determines the QPA as the median contracted rate for a service among all 2019 contracts with providers in the same or similar specialty within each applicable insurance market and in the same geographic region, as adjusted for inflation according to the Consumer Price Index for all Urban Consumers (CPI-U).¹⁶ Insurers provide the QPA and selected QPA information to providers with an initial payment (or denial of payment) for OON billing services, a step that precedes the IDR process in the federal methodology.¹⁷

Providers and insurers have expressed differing opinions on the QPA.¹⁸ Providers have expressed concern about a lack of QPA transparency and have pointed to instances of particularly low QPA rates that are below Medicare rates as evidence of potential inaccurate calculation of QPA by insurers.¹⁹ In addition, providers have argued that certain implementing methodological features artificially lower the QPA.²⁰ Insurers, by contrast, have indicated that the QPA represents the market rate for services.²¹ Insurers also are required to share certain information with OON providers about the QPA upon an initial OON payment and upon a provider's request, and IDR-related rulemaking (and proposed rulemaking) around these disclosures has focused on finding a balance between ensuring meaningful, transparent QPA reporting and minimizing administrative burden.²²

CMS is statutorily required to conduct QPA audits.²³ To date, it has released one audit, where it found the insurer was incorrectly determining air ambulance QPAs.²⁴ CMS has conducted additional audits as part of its investigations of complaints but has yet to release findings.²⁵

Transparency in Coverage In-Network Rates

In 2020, the Departments of Health and Human Services (HHS), Labor, and the Treasury (hereinafter, the Departments) issued the Transparency in Coverage final rule. The rule mandated that insurers publish in-network negotiated rates with providers for all covered services and items (excluding prescription drugs under a fee-for-service arrangement) on a publicly accessible website for each offered plan, make the files available free of charge and without conditions, and update them monthly.²⁶

Some private vendors have emerged to aggregate, clean, standardize, and analyze Transparency in Coverage data. One such vendor is Payerset, which provided in-network rates for this analysis under a contract with CRS.

Data and Methods Overview

This CRS report looks to identify how IDR outcomes, provider offers, insurer offers, and QPAs for emergency services compare with the in-network emergency service rates of the providers participating in the IDR process. To do this analysis, this report uses selected data from the Departments' IDR and health insurer transparency data to calculate, for each OON emergency service dispute, the IDR outcome, provider offer, insurer offer, and QPA as a percentage of the applicable median in-network rate for the same service, among all IDR-participating providers in the same state, and whether provided by an individual provider or a facility.

This section briefly describes the datasets and methods used to conduct this analysis and selected data limitations. A more in-depth discussion of the datasets and methods appears in Appendix A.

Datasets

Three datasets were used to conduct the analysis. Two of these datasets are available in public-use files (PUFs) released by the Departments containing IDR dispute data,²⁷ and the third is insurer price transparency data provided by Payerset.

The first PUF dataset is OON Emergency and Nonemergency data (i.e., the Details File) for Q1 and Q2 of 2024. This dataset contains specific characteristic details for each party to an IDR dispute, such as provider National Provider Identifier (NPI), provider name, insurer name, service code, place of service code, and type of service code, but does not have dollar values for the applicable QPAs, insurer offers, provider offers, and prevailing offers. Instead, a dispute's applicable QPA, insurer offer, provider offer, and prevailing offer are presented as a percentage of other benchmarks (e.g., the prevailing offer is presented as a percentage of the QPA). This file includes the state where the disputed service was provided.

The second PUF dataset is the QPA and Offers data (i.e., the Outcomes File) for Q1 and Q2 of 2024. This second dataset contains dollar-specific QPAs, insurer and provider offers, the prevailing offers, service code, place of service code, and initiating party for each dispute but does not contain specific, identifying information about the parties participating in each dispute. This dataset includes the metropolitan statistical area (MSA) where the service was provided.

The Payerset data (i.e., Transparency Data) contains all May 2025 in-network emergency service rates for a curated list of providers that were identified by CRS in the PUF Details File to have participated in at least one IDR emergency service dispute in the first half of 2024 and whose rates are within the Payerset database. Rates are associated with over 40 different insurers, including Aetna, Blue Cross Blue Shield, Kaiser, and United. Data reported by plans under Transparency in Coverage requirements do not include geographic variables (e.g., state).²⁸

Methods

To conduct this analysis, the PUF Details File was used to identify providers that furnished an applicable OON emergency service that was subject to an IDR dispute. That provider list was used to extract corresponding negotiated in-network rates from the Transparency Data, which was then trimmed to better align the data with the PUF datasets. The PUF Details File then was linked to the Transparency Data using the provider NPI variable, allowing the provider state reported in the Detail File to be assigned to the negotiated rates in the Transparency Data.²⁹

Transparency Data were then used to calculate state level median in-network rates for a given service across all insurers and providers. More specifically, median rates for a given service were calculated based on each unique combination of billing code, state, and whether the provider was an individual provider or a facility; the median was calculated across all negotiated rates between insurers and the providers in the state that participated in an emergency service dispute that was resolved in the first half of 2024.³⁰ For example, the median in-network rate among all Texas IDR-participating providers for a physician's billing of a low-level emergency room visit (CPT code 99283) was $117.62.³¹

To prepare a trimmed version of the PUF Outcomes File (consisting of applicable emergency service disputes) to link with the Transparency Data, a state was assigned to each IDR dispute in the trimmed PUF Outcomes File based on the MSA. There are differences in the level of geographic detail between the PUF Details File, which is used to assign a provider's state in the Transparency Data, and the PUF Outcomes File—the PUF Details File lists geographies at the state level, and the PUF Outcomes File lists geographies at the MSA level. Since MSAs can span state lines, the analysis was restricted to include only states in which at least 95% of a state's trimmed emergency service disputes occurred in an MSA that was entirely encompassed within that state.³² The following 14 states met this criterion: Alabama, Arizona, California, Colorado, Connecticut, Florida, Idaho, Louisiana, Michigan, New Mexico, Nevada, Texas, Utah, and Washington. Among disputes within these states, in instances where a dispute occurred in an MSA that was located entirely within a state, that state was assigned to the dispute in the PUF Outcomes File. In instances where a dispute occurred in an MSA that spanned state lines, the dispute was assigned the state that was both (1) included in the MSA and (2) included in this analysis.³³

Transparency Data median rates were linked with the applicable emergency service disputes in the PUF Outcomes File using the following variables: assigned state, dispute service code, and whether the rate/dispute related to an individual provider or a facility. After linking, each dispute items' QPA, insurer offer, provider offer, and prevailing offer were then calculated as percentages of the applicable median in-network rate by dividing each of those values by the applicable median in-network rate.

In total, over 82,000 disputes were analyzed for this report, which represented about half of the total sample of emergency services disputes.

Data Limitations

Data collection timeline differences, in-network rate data challenges, missing billing code modifiers, and geographic variable differences created certain limitations with this analysis.

First, the PUF Files and the Transparency Data are from different time periods. The PUF Files are for disputes resolved during the first half of 2024 and include services provided in 2022 and 2023. The Transparency Data used for this analysis were applicable in roughly May 2025.³⁴ Since the in-network rates were collected in 2025, it is possible that rates differed from those applicable at the time the disputed service was furnished. Furthermore (and potentially more problematic), the 2024 IDR outcomes may have influenced the 2025 in-network rates. For example, insurers and in-network providers could have renegotiated rates as a result of their experiences with the IDR process. Similarly, IDR outcomes may have affected provider and insurer decisions to contract with each other and, correspondingly, the in-network rates included (or excluded) in this analysis.³⁵

Second, the Transparency Data have limitations associated with the Transparency in Coverage data reporting.³⁶ For example, certain insurers negotiate rates that are a percentage of the provider's billed charges, which are displayed in the transparency data as a percentage (and not a dollar amount). Without the amounts that providers charge, which are not reported under the Transparency in Coverage requirements, these rates cannot be compared with rates expressed as dollar amounts. Although this represented a small percentage of total rates, the exclusion of these rates could affect the results of this analysis in the selection of benchmark median in-network rates. Similarly, it could affect provider and insurer decisions in regard to the IDR process and in negotiations outside of the IDR process.

As another example, Transparency Data indicate whether each particular rate applies to an individual provider, a facility, or both. The Kaiser Family Foundation (KFF) examined Transparency in Coverage data and found instances where rates in these datasets may seem unlikely based on the individual provider/facility label.³⁷ In one example, rates for an individual practitioner were labeled as a facility rate in the Transparency in Coverage data. For purposes of this report, CRS was unable to use PUF Details File information to identify whether an IDR-participating provider was going through the IDR process as an individual provider or a facility,³⁸ and therefore could not use such information to verify and/or filter out rates for providers in the Transparency Data that did not have matching provider/facility labels.³⁹ Without this verification and/or filter, additional rates could be included in this analysis that could affect the calculation of median in-network rates, since median rates are calculated for each combination of assigned state, dispute service code, and whether the rate related to an individual provider or a facility. Future research may be able to address this limitation using a new variable added to 2025 IDR dispute data that indicates whether a dispute in the PUF Details File was initiated by an individual provider or initiated by a facility.

Third, the PUF Outcomes File does not include billing code modifiers that may apply to a disputed service. Billing code modifiers can be attached to a billing code to indicate that the performance of a medical service has been altered by a specific circumstance but not changed. For example, a billing code modifier could indicate that a service was provided via telehealth. In-network rates can vary by billing code modifier, and if a disputed service in the PUF Outcomes File had a billing code modifier, this could create discrepancies in the analysis.

Finally, geographic variable differences between the PUF Details File and the PUF Outcomes File limited the specificity of the analysis. QPAs generally are determined at the MSA level.⁴⁰ Because the PUF Details File indicates the state in which the disputed service was provided and is used to assign in-network rates to states, in-network rate summary statistics do not account for variation across in-network rates within a state. This could contribute to distinctions between the median QPA in the PUF Outcomes File and the median in-network rate, particularly if there are states where disputes are spread across multiple MSAs within a state and such MSAs have large variations in rates. This may be less consequential in states that have emergency service disputes that are largely concentrated within one MSA, such as Idaho and Nevada.

Analytical Findings

This section first analyzes—at the multi-state level—how QPAs, insurer and OON provider offers, and outcomes of selected IDR emergency service disputes resolved in the first half of 2024 relate to the 2025 median in-network emergency service rates of IDR-participating providers.⁴¹ It then discusses the extent to which these outcomes vary at the state level.

The findings are reported as a relative percentage of in network rates. For example, 100% means the IDR outcome equals the median in-network rate for that particular service in the same state and as provided by an individual provider or facility. Above 100% means the IDR outcome exceeds the corresponding median in-network rate. Below 100% means the IDR outcome is lower than the corresponding median in-network rate.

Aggregate Analysis

Across all 14 states examined, for IDR emergency service disputes resolved in the first half of 2024, the median prevailing offer was over three times the median in-network rate of IDR-participating providers (312% of median in-network rate) for that service, and over 95% of disputes had a prevailing offer that was above the median in-network rate (see Figure 1, where the median in-network rate of IDR-participating providers is depicted by the dotted vertical line at 100%). For these disputes, the median provider offer was similarly above the median in-network rate (321% of median in-network rate).

Figure 1. Distributions of IDR Emergency Service QPA, Offers, and Outcomes Relative to Median In-Network Rates

Source: CRS analysis of Payerset Transparency Data for May 2025 and Departments of Health and Human Services, Labor, and the Treasury, Federal Independent Dispute Resolution (IDR) Public Use File (PUF) for 2024, Q1, Federal IDR PUF for 2024, Q2. Notes: Disputes resolved in Q1 and Q2 of 2024 can pertain to services provided prior to 2024. Median in-network rate is the 2025 median rate for a given service among all providers in the state that participated in an emergency service dispute that was resolved in the first half of 2024 and has negotiated rates included in Payerset Transparency Data. It is based on each unique combination of billing code, state, and whether the provider was an individual provider or a facility. The median in-network rate of IDR-participating providers is depicted by the dotted vertical line at 100%. In each box plot, the vertical black line represents the median, the ends of the box represent the 25th percentile and the 75th percentile, and the ends of the tails represent the 1st percentile and the 99th percentile. IDR = independent dispute resolution. QPA = qualifying payment amount. For this figure, all values exceeding 1,000% were top-coded at 1,000%.

The total amount of revenue that an OON provider received for care would be less than the prevailing offer, as the provider would have had to pay an administrative fee to the federal government to participate in the IDR process, which was $115 per dispute for most of 2024,⁴² and a fee to the IDR entity if its offer was not selected.⁴³ For high-level context, the median prevailing offer among all analyzed emergency service disputes was $627 (and the median QPA among all analyzed emergency service disputes was $226).⁴⁴

The median QPA was slightly higher than the median in-network rate (112% of median in-network rate). Although more times than not, the QPA was also higher than the median in-network rate of providers participating in the IDR process, the QPA was frequently below the median in-network rate (roughly 44% of analyzed disputes). Similar to the median QPA, the median insurer offer was slightly above the median in-network rate (131% of median in-network rate), which aligns with CMS's observation that insurers frequently benchmark their offers to the QPA.

State-Level Analysis

Conducting the same analysis—comparing selected first half 2024 IDR emergency service dispute QPAs, insurer and provider offers, and prevailing offers relative to the 2025 median in-network rates of IDR-participating providers—at the state level in 14 states shows distributions that varied by state (Figure 2).⁴⁵ For example, in Texas, the state with the largest number of analyzed disputes (46% of analyzed disputes), the median QPA and the median insurer offer relative to the median in-network rate of providers in the state (119% and 131%, respectively) were similar to the corresponding aggregated medians. At the same time, the median provider offer and the median prevailing offer relative to the median in-network rate of providers in the state (390% and 386%, respectively) were higher than the corresponding aggregated medians. In contrast, and relative to the state's median in-network rate, Nevada had a median QPA (35% of the state's median in-network rate), median plan offer (36% of the state's median in-network rate), median provider offer (118% of the state's median in-network rate), and median prevailing offer (118% of the state's median in-network rate) lower than the corresponding aggregated ratios.

For all 14 states in this analysis, the median prevailing offer was above the median in-network rate of IDR-participating providers, and in nine states it was over three times the median in-network rate. The degree to which the median prevailing offer exceeded the median in-network rate, however, varied.⁴⁶ In Nevada and California, the median prevailing offer was slightly higher than the median in-network rate (118% and 129% of the state's median in-network rate, respectively), whereas Colorado saw a median prevailing offer that was over six times the median in-network rate (619% of the state's median in-network rate).

There also was variance among states in how their median QPA compared with their median in-network rates. Six states (Arizona, California, Louisiana, Nevada, Utah, and Washington) had a median QPA that was less than the median in-network rate of the IDR-participating providers. In the remaining eight states, the median QPA was above the median in-network rate of the IDR-participating providers. This means that in these eight states, an OON payment that equaled the median QPA would represent a higher payment than over a majority of the IDR-participating in-network rates for the service.

Figure 2. Median IDR Emergency Service QPA, Offers, and Outcomes Relative to Median In-Network Rates, by State

Source: CRS analysis of Payerset Transparency Data for May 2025 and Departments of Health and Human Services, Labor, and the Treasury, Federal Independent Dispute Resolution (IDR) Public Use File (PUF) for 2024, Q1, Federal IDR PUF for 2024, Q2. Notes: After the "All Disputes" row, states are organized in ascending order of median prevailing offer value. Disputes resolved in Q1 and Q2 of 2024 can pertain to services provided prior to 2024. Median in-network rate is the 2025 median rate for a given service among all providers in the state that participated in an emergency service dispute that was resolved in the first half of 2024 and has negotiated rates included in Payerset Transparency Data. It is based on each unique combination of billing code, state, and whether the provider was an individual provider or a facility. The median in-network rate of IDR-participating providers is depicted by the dotted vertical line at 100%. IDR = independent dispute resolution. QPA = qualifying payment amount.

The median insurer offer relative to the median in-network rate also varied by state. In five states (Arizona, California, Nevada, Utah, and Washington), the median insurer offer was below the median in-network rate.⁴⁷ Of the remaining eight states, four states (Alabama, Michigan, Louisiana, Colorado) had a median insurer offer that was over two times the median in-network rate (282%, 247%, 209%, and 208% of the state's median in-network rate, respectively). Provider win rates were slightly lower in three of these four states; providers won in 85% of Alabama disputes, 79% of Michigan disputes, 80% of disputes in Louisiana, and 84% of Colorado disputes, relative to 85% of disputes across all states in this analysis (Table 1). Overall, there was little/slightly negative correlation between a state's median insurer offer and provider win rate (-0.21).

The median insurer offer relative to the median QPA also varied by state, indicating that insurers operating in different states appear to take different approaches with respect to IDR offers. In all 14 states, the median insurer offer was at or above the median QPA, with five states (Arizona, Connecticut, Nevada, Utah, and Washington) having no (or minimal) difference between the median insurer offer and the median QPA. Alabama, Louisiana, and Michigan had the most separation between the median insurer offer and the median QPA. Overall, there was little/slightly negative correlation between the size of the difference between median insurer offers and median QPA in a state and the state's provider win rate (-0.15).

Some of the state-by-state variation could be due to differences in the types of emergency service disputes going through the IDR process in a given state. As an example, in most states, the most common disputed emergency service was for a moderate-level emergency room visit (CPT Code 99284); in Connecticut, Nevada, and Utah, by contrast, the most common disputed billing code was for a high-level emergency room visit (CPT Code 99285).⁴⁸ In another example, in 11 states, facilities accounted for 1% or less of the state's emergency service disputes, but facilities accounted for a much higher percentage of state disputes in Texas and Colorado (26% and 38% of the state's disputes, respectively).⁴⁹

Summary and Policy Implications

This analysis compares median QPAs, insurer and provider offers, and prevailing offers to the median in-network emergency service rates of IDR-participating providers in selected states. The findings indicate several broad patterns. First, over 95% of disputes had a prevailing offer that was above the median in-network rate among IDR-participating providers. Second, the median prevailing offer in the IDR process for emergency service was slightly more than three times the median in-network rate among IDR-participating providers. Third, the median QPA was slightly higher than the median in-network rate.

An overall median prevailing offer in the IDR process for emergency service that is over three times the median in-network rate among IDR-participating providers suggests the IDR process could be used by IDR-participating providers in negotiations with insurers to put upward pressure on in-network emergency service rates.⁵⁰ Depending on the extent to which these negotiations result in increased in-network rates, this could increase health care costs for emergency services.⁵¹

Across all states examined, the median prevailing offer was above the median in-network rate of IDR-participating providers. However, there was substantial state level variation in the median prevailing offers relative to the median in-network emergency service rates of IDR-participating providers in the state. For example, Nevada and California had median prevailing offers that were only modestly higher than their respective state median in-network rates, at 118% and 129% of the state's median in-network rate, respectively. By contrast, Colorado had a median prevailing offer that was more than six times the median in-network rate, at 619% of the median in-network rate. Overall, 9 of the 14 states analyzed had median prevailing offers that exceeded three times the state's median in-network rate.

This state level variation implies that the extent to which providers may be able to use IDR outcomes in contract negotiations with insurers to exert upward pressure on in-network emergency service rates may differ across markets. In states where prevailing offers closely track median in-network rates, the negotiation leverage associated with IDR outcomes may be more limited. In states where prevailing offers diverge substantially from in-network rates, however, IDR outcomes may be able to play a more pronounced role in shaping contracting behavior and, potentially, health care spending for emergency services.

Variation also was observed in QPAs relative to median in-network rates. In six states, the median QPA was less than the median in-network rate of IDR-participating providers; in the other eight states, the median QPA exceeded the median in-network rate. This variation suggests the QPA may be perceived differently by OON IDR-participating providers across states. In states where the QPA falls below the median contracted rate, some OON providers, particularly those with relatively high negotiated rates with other insurers, may view the QPA as an inadequate benchmark. Depending on initial OON negotiations with a disputing insurer, providers with high negotiated rates with non-disputing insurers may be more inclined to pursue the IDR process, which also may influence which negotiated rates are represented in the analytical sample.

More research on the IDR process, including what percentage and types of OON disputes are going through the process, could provide indications about the extent to which the IDR process could be affecting health care costs, including in-network rates; network participation; and, by extension, the federal budget. In addition, research that looks at OON disputes and IDR participation by different criteria (e.g., specific geographies, non-emergency OON services, provider/insurer characteristics) may provide additional, more nuanced insights into how the IDR process could interact with underlying market dynamics.

Appendix A. Methods

Payerset Transparency Data

The Congressional Research Service (CRS) used the public-use file (PUF) Details File to identify the 3,800 providers that were part of a dispute involving emergency services (i.e., involved service codes 99281, 99282, 99283, 99284, 99285)⁵² that were provided in an emergency room of a hospital (i.e., had a place of service code 23); that was not a default determination; and that did not involve a bundled payment.⁵³ Based on this list, Payerset generated a dataset of all in-network rates (within the Payerset database) for emergency services codes for the identified set of providers, which totaled 3.1 million in-network rates.⁵⁴ The resulting dataset is referred to as Transparency Data in this report.

CRS trimmed and modified the Transparency Data to better align the data with the types of disputes that went through the independent dispute resolution (IDR) process in the first half of 2024 and to account for certain challenges associated with the underlying Transparency in Coverage data requirements.⁵⁵

First, CRS limited the dataset to include only in-network rates expressed as a dollar amount, a derived amount, or part of a fee schedule (in instances where the fee schedule had no corresponding in-network rate or derived amount for the unique combination of payer, National Provider Identifier [NPI], CPT code, place of service code, and reporting entity name). Rates that were expressed as a percentage of billed charges, were expressed as per diem rates, or had other designations were excluded from the analysis.⁵⁶ Of the initial 3.1 million rates, 55.07% were negotiated dollar amounts, 31.92% were fee schedule rates (of which, less than 1% had a corresponding negotiated rate), 2.07% were derived amounts, and 10.94% were expressed as a percentage of billed charges, per diem rates, and other amounts.

Transparency Data also were trimmed to include only rates associated with services provided in an emergency room of a hospital (i.e., rates associated with service code 23), as 99.73% of IDR disputes involving emergency service codes occurred in an emergency room of a hospital. Because facility rates in the Transparency Data are not required to include service code values, rates for facilities were not trimmed if no service code was listed and the provider was a hospital.⁵⁷ Of the initial 3.1 million rates, 48.42% of rates were for services provided outside of a hospital emergency room.

In addition, Transparency Data were limited to in-network rates that were not part of a bundled payment or capitated payment amount. Bundled payments are single payments for a combined set of items and services. Under capitation, a provider is paid a set amount based on the number of people cared for, regardless of the number of services provided. Of the initial 3.1 million rates, 0.05% of rates were bundled payments and 0.11% were capitated rates.

Also excluded from the Transparency Data were any rates associated with billing modifiers. Billing modifiers are codes that signify an item or service is provided in altered circumstances (e.g., only the technical component of a service was provided, or the service was provided via telehealth) and may cause pricing changes. IDR dispute data do not include information on billing code modifiers. Of the initial 3.1 million rates, 12.69% of rates were associated with a billing modifier.

CRS also excluded from the Transparency Data any non-CPT billing code types (e.g., Healthcare Common Procedural Coding System [HCPCS]), as IDR dispute data associated with service codes 99281, 99282, 99283, 99284, and 99285 were all CPT codes. Of the initial 3.1 million rates, 2.83% of rates were indicated as HCPCS.

CRS modified United Healthcare custom emergency service rates within the data. United Healthcare has custom rates for emergency services that are provided in an institutional setting. These rates are associated with the "billing_code" values EMR, EMR1, EMR2, EMR3, EMR4, and EMR5.⁵⁸ EMR1 through EMR5 were reassigned CPT codes of 99281-99285, respectively, and EMR was not used. All of United's EMR codes accounted for 1.72% of the initial 3.1 million rates.

Duplicate values were identified based on rates that matched according to the following variables: payer, NPI, Taxpayer Identification Number (TIN), billing class, negotiated rate, negotiated type, reporting entity name, and reporting entity type. Duplicates were subsequently removed.

After all trimming and modifications, the Transparency Data universe for analysis contained roughly 740,000 rates relating to over 3,600 providers.

PUF Details File

The PUF Details File was used to match providers (and their in-network rates) to relevant states in which they provided OON emergency services that were subsequently part of an IDR dispute. If a provider was identified as having participated in an IDR dispute in the PUF Details File that pertained to this analysis, the provider was designated as operating within the state in which each disputed service was provided. In some instances, providers had multiple disputes and different disputes involved services provided in different states. For example, of the 3,800 unique, applicable providers in the PUF Details File, 117 participated in disputes spread across two states, 24 participated in disputes spread across three states, 12 participated in disputes spread across four states, 4 participated in disputes spread across five states, and 1 participated in disputes spread across six states. In these instances, providers were designated as operating in every state in which they had at least one relevant dispute.

Once the state(s) of the provider were assigned, this information was joined with the Transparency Data using the NPI variable.

Within the Transparency Data, medians of in-network rates were determined by finding the median value among the rates for each unique combination of billing code, state, and whether the rate applied to a provider or a facility. Providers operating in multiple states had their in-network rates counted in each state that they operated in. Of the roughly 740,000 in-network rates, roughly 445,000 rates applied to over 2,300 providers located in the 14 states included in the analysis.

PUF Outcomes File

The PUF Outcomes File contains information on 578,212 IDR dispute items resolved in the first half of 2024. This file was trimmed to include only emergency service disputes, which represented 40.40% of dispute items (or 233,575 disputes). These disputes were further trimmed to focus on emergency services provided in an emergency room of a hospital, which represented 99.73% of IDR disputes involving emergency services.

Bundled disputes (1.97% of emergency service IDR disputes) also were removed from the universe, because the offer and payment amounts for these items could represent the payment amount for a bundle of services.

Furthermore, 22.67% of emergency service disputes had selected data elements suppressed due to privacy concerns, did not have all data values entered, or had negative qualifying payment amounts (QPAs).⁵⁹ In such instances, the disputes were removed from the dataset. Default determinations also were removed, as these represented disputes where one party did not submit an offer or pay applicable fees. Of the emergency service disputes, 19.10% were default determinations (or had no response for this variable).⁶⁰

The initiating party variable indicates whether a dispute was initiated by an insurer, an individual provider, or facility. The variable was used as a proxy to determine whether individual provider rates or facility rates should apply to a given dispute. Disputes initiated by insurers were excluded from this analysis. Given that individual providers and facilities have been heavy utilizers of the IDR process and have initiated a majority of disputes, trimming insurer-initiated disputes reduced 0.008% of emergency service dispute items.

After accounting for these adjustments, the PUF Outcomes File had 167,147 dispute items.

Finally, the geographic location in the PUF Outcomes File was modified. The PUF Outcomes File lists geographies at the metropolitan statistical area (MSA) level, which can span state lines. Each dispute in the PUF Outcomes File was assigned a state of location that represented either the state in which the entire MSA was located or a state that contained a portion of the MSA.

Sample Used for Analysis

Only disputes in selected states were used for this analysis in an attempt to reduce potential distortions of data resulting from different geographic variables between the PUF Details File and the PUF Outcomes File.

The PUF Details File has disputes at the state level, whereas the PUF Outcomes File has disputes at the MSA level. Because MSAs can cross state lines, these two files cannot be easily linked by the geographic variables. As an example, disputes in the PUF Outcomes File that indicate the service was provided in the Washington-Arlington-Alexandria MSA could be shown in the PUF Details File as occurring in the District of Columbia, Virginia, Maryland, or West Virginia. To minimize mis-categorization of states in the PUF Details File (which would create mis-categorized in-network rates in the Transparency Data), the analysis was restricted to 14 states where at least 95% of a state's trimmed emergency service disputes occurred in an MSA that was entirely encompassed within that state. For purposes of this calculation, the total number of a state's trimmed emergency service disputes included any dispute that occurred in an MSA that contained at least a portion of the state.⁶¹

These 14 states are Alabama, Arizona, California, Colorado, Connecticut, Florida, Idaho, Louisiana, Michigan, New Mexico, Nevada, Texas, Utah, and Washington. The disputes in these 14 states represented 82,119 disputes, which was about half of the total emergency room disputes in the trimmed PUF Outcomes File.

Analysis

After the Transparency Data and PUF Outcomes File were prepped, the files were joined so that median in-network rates were assigned to each dispute item in the PUF Outcomes File based on a matching state and service code and whether the service was provided by a provider or a facility.

Each dispute item's QPA, insurer offer, provider offer, and prevailing offer was then presented as a percentage of the applicable median in-network rate for the same service, in the same state, and whether provided by a provider or a facility.

Appendix B. Selected Characteristics of IDR Dispute Sample Used for Analysis

Sylvia Bryan, CRS Research Assistant, provided general research support and produced the "Distributions of IDR Emergency Service QPA, Offers, and Outcomes Relative to Median In-Network Rates" graphic.

Footnotes