Senate Budget Committee Issues Testimony From Brookings Institution Senior Fellow Fiedler
* * *
Chairman
My testimony will examine the administrative costs generated by interactions between health care providers and payers(including both public programs and private insurers), as well as how changes in public policy might help to reduce those costs. I will make five main points: 1. A reasonable estimate is that the administrative costs generated by provider-payer interactions amount to around half a trillion dollars per year or around 11% of annual health care spending. Most of these costs are incurred by providers while negotiating contracts, collecting information about patients' coverage, obtaining prior authorization for care, submitting claims for payment, and reporting on quality performance. Payers incur additional costs to perform their part in these interactions.
2. Administrative costs are ultimately borne by consumers and taxpayers. Costs incurred by providers are reflected in the prices that providers negotiate with private insurers; similarly, these costs require public programs like Medicare and Medicaid to pay higher prices to elicit adequate provider participation. Higher prices directly increase patient costsharing in many instances and, together with the administrative costs incurred directly by payers, increase premiums and program spending (as applicable).
3. Many administrative processes serve valuable purposes, so efforts to reduce administrative costs can involve tradeoffs and should proceed thoughtfully. For example, it is essential to have some set of procedures for compensating providers.
Similarly, payers' prior authorization requirements can prevent delivery of inappropriate services, and audit processes can be effective tools for identifying and deterring fraud.
1 The views expressed in this testimony are my own and should not be attributed to the staff, officers, or trustees of the
* * *
4. Certain targeted reforms could reduce administrative costs with few substantive downsides. One is eliminating Medicare's Merit-Based Incentive Payment System, which places large reporting burdens on clinicians with few benefits. Another is replacing the cumbersome arbitration process that determines payment rates for certain out-of-network services under the No Surprises Act with a simpler "benchmark" payment regime. A third is reforming Medicare Advantage's risk adjustment system to reduce plans' ability to increase payments by documenting additional diagnoses. More generally, policymakers should be attentive to how policy choices across many domains affect administrative costs.
5. Standardizing billing, coverage, and quality reporting processes across payers could generate larger savings but could also present tradeoffs. The wide variation in rules across the menagerie of public and private payers that operate in
One worthwhile goal is standardizing how providers and payers share claims information. One strategy policymakers could consider is creating a central clearinghouse that would accept claims from providers in a standardized format and route them to payers, an approach used in some other industries and in some other countries' health care systems. This approach would likely more fully standardize claims transactions than past federal efforts that have established transaction standards to govern decentralized provider-payer interactions. However, it would be essential that a clearinghouse be well-run, as a poorly run clearinghouse would likely generate few benefits or even do harm.
Policymakers could also consider standardizing the substance of some payer rules related to billing, coverage, and quality reporting. This approach might generate larger administrative savings but would also limit payers' ability to tailor rules to their circumstances or experiment with novel approaches. Setting rules through a centralized process might also produce rules that are systematically better or worse than existing rules. These factors could cause standardization to produce offsetting costs, like greater use of inappropriate services, that would need to be weighed against administrative savings. Thus, the desirability of this type of standardization is likely to be highly case-specific. Quality reporting may be a domain where standardization would be particularly attractive, as it could likely both reduce administrative costs and increase the utility of the resulting quality data by increasing provider-level sample sizes and easing cross-payer comparisons.
The remainder of my testimony examines these points in greater detail.
Background on the Administrative Costs of Provider-Payer Interactions
Health care providers devote substantial effort to interacting with payers; activities include negotiating contracts, collecting information about patients' coverage, seeking prior authorization for care, submitting claims for payment, and reporting on quality performance. Estimating these costs is challenging because it requires detailed information on provider costs and because, even when these data are available, it is not always obvious what costs are incurred for which purposes.
However, one synthesis of survey estimates concluded that provider costs associated with these activities total 13.0% of revenue for physician practices, 8.5% for hospitals, and 10.0% for other providers, as shown in Figure 1.2 Other studies using more recent data (but that are narrower in scope or categorize costs somewhat differently) obtain estimates of a broadly similar magnitude.3 Under current projections of health care spending, the estimates in Figure 1 imply that providers in
* * *
[View chart in the link at bottom.]
* * *
Private insurers and public programs incur additional costs to play their part in provider-payer interactions. Commercial insurers reported costs for claims processing, cost containment, and...
2
3
4 These and subsequent calculations in this section use the most recent National Health Expenditure projections.
* * *
...quality improvement activities equivalent to 2.9% of claims spending in 2021.5,6 Unfortunately, comparable data are not available for other payer types. It is plausible that the corresponding percentage is higher for private insurers that deliver Medicare and Medicaid coverage since these programs generally pay lower prices for care, so the same per-claim administrative cost would represent a larger share of the cost of the claim.7 These insurers may also make greater use of prior authorization and other utilization controls, which may also raise administrative costs.8 On the other hand, this percentage is clearly smaller in traditional Medicare; total administrative spending for Medicare Part A and Part B (including expenses unrelated to traditional Medicare's interactions with providers) amounted to only 2.2% of traditional Medicare's claims spending in 2021.9
Nevertheless, the estimate for commercial plans can provide a sense of the magnitude of the costs that payers generally incur in connection with their interactions with providers.10 Applying this estimate across all payers implies that payers' costs to interact with providers will total
These administrative costs are ultimately borne, at least in large part, by consumers and taxpayers. In commercial insurance, the prices negotiated between insurers and providers are likely to reflect the administrative costs borne by providers, at least in the long run. Those higher prices, as well as the administrative costs incurred directly by insurers, are then reflected in premiums, cost...
5 This estimate was derived from the medical loss ratio filings of individual, small group, and large group insurers.
6 Some studies report far higher estimates of payer-incurred "billing and insurance-related" costs. These studies often include costs of payer activities that are not directly related to interactions with providers (e.g., marketing, underwriting, and premium collection) and sometimes include insurer-paid taxes and insurer profits. These broader measures can be relevant when considering proposals to change how insurance is provided but are less relevant to assessing proposals that would change how providers and payers interact. See, for example,
7
8
9 This estimate was derived from the 2022 Medicare Trustees report. See Boards of Trustees of the
10 A caveat is that this estimate includes some costs associated with some insurer activities that are not directly related to interactions with providers, like internal analytic activities and enrollee-facing wellness activities. On the other hand, this estimate may not fully reflect insurer spending on employees and non-labor inputs that are used both for interactions with providers and for other purposes.
* * *
...sharing, or both.11 Part of those costs is paid by consumers and part is paid by the federal government (which subsidizes most forms of private coverage via the tax code).
In public programs, increases in the administrative costs borne by providers require those programs to pay providers higher prices to elicit any given level of provider participation.12 The costs are then ultimately financed by taxpayers or through premiums paid by program beneficiaries. The same is true of the administrative costs that programs incur directly.
The complexity of health care providers' interactions with payers appears to vary widely across countries. One recent study collected detailed data on the number of minutes of work that is required to collect payment for inpatient services in six countries.13
* * *
[View chart in the link at bottom.]
* * *
This finding likely reflects, at least in part, the fact that
11 Strictly speaking, this is only likely to be true of administrative burdens that increase the marginal cost of delivering a health care service or covering an additional enrollee, but that is likely generally the case in practice.
12 For empirical evidence on this point, see
13 See
* * *
...insurers operating Medicare Advantage plans, still more private insurers that offer private plans in the commercial market, the state's fee-for-service Medicaid program, and private insurers that operate Medicaid managed care plans, among others. Even within a given insurer and coverage type, rules may vary depending on what specific plan a patient is enrolled in.
Targeted Steps to Reduce Insurance-Related Administrative Costs
Given the size of the administrative costs generated by providers' interactions with payers, it is natural to ask whether these costs can be reduced. In looking for ways to do so, it is important to recognize that administrative spending is not inherently wasteful. Administrative processes serve important purposes: billing processes are needed to compensate providers for delivering care; prior authorization requirements can prevent delivery of inappropriate services;14 and audit processes can help uncover and deter low-value utilization.15 Thus, policy efforts to reduce administrative burdens should be attuned to tradeoffs and proceed thoughtfully.
This section discusses three targeted policy changes that could reduce administrative costs with few substantive downsides: (1) eliminating Medicare's Merit-Based Incentive Payment System; (2) reforming the No Surprises Act's method for determining payment for certain out-of-network services; and (3) making the Medicare Advantage risk adjustment system more resistant to plans' diagnosis coding efforts. The next section considers some wider-ranging reforms.
Eliminating Medicare's Merit-Based Incentive Payment System The Merit-Based Incentive Payment System (MIPS) was created by the Medicare Access and CHIP Reauthorization Act of 2015 (MACRA); it adjusts most clinicians' Medicare payments upward or downward based in their performance in several domains, including cost and quality. While clinicians who participate in certain alternative payment models are exempt from MIPS, only a minority of clinicians participate in those models, so most are subject to MIPS.16
Unfortunately, there is little reason to believe that MIPS is achieving its goal of improving the quality or efficiency of patient care.17 One fundamental problem is that MIPS allows clinicians to...
14
15
16
17 For more discussion of these points, see
* * *
...choose many of the measures that they are evaluated on. In practice, different clinicians choose different measures, and they likely do so at least in part based on which measures they expect to perform best on. This makes it impossible to use MIPS scores to meaningfully compare clinicians and, thus, doubtful that MIPS can motivate better outcomes.
Measuring cost and quality performance at the level of individual clinicians or practices, as MIPS aims to do, is also challenging. Patients' outcomes are shaped by the efforts of many different providers, which makes it difficult to determine who is responsible for what, plus it can be hard to constructstatistically reliable performance estimates at the provider level. This is a recipe for weak, incoherent incentives, and it is likely why a plethora of programs that have adjusted providers' payment rates based on provider-level measures of cost and quality performance (including programs that avoid some of MIPS' distinctive shortcomings) have failed to improve care.18
MIPS does, however, impose large compliance burdens on practices. Much of the information used to compute a practice's MIPS score--notably its performance on quality measures--is reported by the practice itself. Practices are also responsible for deciding which quality measures to report, as well as how they want to be scored in other MIPS domains. These activities are costly. A recent study that interviewed practices about their MIPS compliance costs estimated that practices spent nearly
...From the Merit-Based Incentive Payment System (MIPS): What Will Be Learned and What Next?," JAMA 321, no.12 (
18 See, for example,
19
20 This estimate was obtained using CMS' estimate of the total number of MIPS-eligible clinicians in 2019 and the Medicare Trustees' estimate of total spending under the physician fee schedule in that year.
21 There is some reason to suspect that this estimate overstates practices' actual costs. These estimated costs exceed the difference between the largest positive and largest negative MIPS payment adjustment applied for 2019. (
* * *
In sum, I see little reason to believe that MIPS generates benefits that justify its substantial costs. With colleagues, I have argued for repealing MIPS and replacing it with small, targeted incentives for practices to undertake specific high-value activities: (1) using a certified electronic health record, which can help advance broader federal efforts to ensure that clinical data can flow across providers when needed; and (2) reporting data to a clinical registry, which can help facilitate valuable clinical research.22 In parallel, policymakers should strengthen incentives to participate in advanced alternative payment models and, ideally, streamline quality reporting requirements under those models.23 The
Reforming the No Surprises Act's mechanism for determining payment for out-of-network care The No Surprises Act limits patients' exposure to "surprise bills" when they receive certain outof-network care, including out-of-network emergency services and services delivered by an outof-network physician at an in-network facility. Under the law, insurers must cover these services and apply only in-network cost-sharing, while providers cannot bill patients for more than the innetwork cost-sharing. The payment the provider receives from the insurer is then determined via negotiations between the two parties or, if they cannot agree, via the Independent Dispute Resolution (IDR) process: a "baseball style" arbitration process in which the insurer and provider each make an offer and the arbitrator chooses between the offers based on statutory criteria.
The IDR process has created substantial administrative costs for both providers and insurers. From
...suggests that the costs faced by typical practices may not have been quite this large. Additionally, costs may have declined since 2019 as practices have gained experience and as CMS has tried to simplify the program.
22 Fiedler et al., "Congress Should Replace Medicare's Merit-Based Incentive Payment System"; Fiedler, Medicare physician payment reform after two years: Examining MACRA implementation and the road ahead.
23 For a recent review of the evidence on this point, see
24
25
26
27
* * *
...is in addition to the expenses that they will incur to conduct negotiations prior to entering IDR or that they will incur during the IDR process (e.g., to respond to arbitrators' inquiries).
It is plausible that these costs will wane somewhat over time. Higher fees may help to reduce IDR volume. Additionally, IDR volume may decline as the parties gain experience with the process. This is because going to IDR only makes sense if the two parties have divergent beliefs about what price the arbitrator will ultimately select; otherwise, they would both be better off reaching an agreement at a price close to the price that they expect the arbitrator to pick and avoiding the costs associated with
These administrative costs are avoidable. During the debate that led to the No Surprises Act, policymakers considered approaches under which payment for an out-of-network service subject to the law's protections would equal a statutorily specified "benchmark" price. For example, one bill specified that an insurer would be required to pay the median contracted rate it had paid for the service before enactment of the No Surprises Act.29 (The insurer's historical median contracted rate is currently a factor that arbitrators are supposed to consider in IDR.) Another approach would have been to set the benchmark price equal to a multiple of the price Medicare pays for the service.30 These approaches could be revived in light of the dismal experience with IDR.
Some may worry that the "benchmark" approach would result in providers being paid less appropriate prices than under IDR. But this concern is likely ill-founded. Notably, policymakers could set the benchmark so that the overall level of payments to providers is at whatever level they deemed appropriate; for example, they could set a benchmark that would ensure that providers are paid the same amount, on average, as under IDR.
Moreover, there is no reason to believe that the IDR process will do a good job of tailoring prices to particular cases. Arbitrators have no clear economic incentive to want to arrive at the "right" price (even if it were clear what that price is). Rather, arbitrators' main incentives are to: (1) minimize their costs of deciding cases; and (2) maximize their future volume.
The first incentive will tend to encourage arbitrators to reach decisions by applying simple rules rather than by carefully considering the facts of any particular case; the guidance arbitrators have received is compatible with this approach, as they have broad latitude to decide how to weigh the statutory factors. The second incentive will tend to reinforce the first incentive since, under the law, arbitrators are generally selected by mutual agreement of the two parties. Thus, an arbitrator is likely to wish to decide cases however it expects other arbitrators to decide cases. Otherwise, it...
28 For more discussion of this point, see
29
30 See, for example,
* * *
...is likely to be perceived as more favorable to either providers or insurers than the "typical" arbitrator and will run the risk of being vetoed by the disfavored party in future cases.
Even if arbitrators do carefully consider the circumstances of a particular case, it is far from clear that this will lead to the "right" prices. Notably, apart from the insurer's historical median contracted rate, the most concrete factor that arbitrators are supposed to consider is the provider's recent contracted rates. These rates are often highest for the providers that were most aggressive about using their ability to surprise bill patients as leverage in contract negotiations with insurers.31 There is little reason to want to favor these providers over others.
Making the Medicare Advantage risk adjustment system more resistant to plan "coding" efforts Under the Medicare Advantage (MA) program, the federal government establishes a payment rate for each participating plan based on a bid submitted by the plan and a "benchmark" based on traditional Medicare spending in the plan's county. That payment rate reflects what the plan would be paid to cover enrollees with the same risk profile as traditional Medicare enrollees. Actual payments are then "risk adjusted" to ensure that payments to the plan are commensurate with the cost of serving the beneficiaries who enroll in the plan. To facilitate risk adjustment calculations, MA plans submit information to CMS on what medical diagnoses their enrollees have, which CMS uses to calculate average "risk scores" that are used to adjust payments.
This system gives MA plans a strong incentive to devote effort to reporting as many diagnoses as possible for their enrollees. Unsurprisingly, MA plans report far more diagnoses for their enrollees than those enrollees would accrue if enrolled in traditional Medicare.32 In many cases, the additional diagnoses reflect conditions that beneficiaries do have, but that tend to go unrecorded in traditional Medicare. In other cases, the additional diagnoses are not supported by beneficiaries' medical records.33 MedPAC estimates that MA plans' diagnosis coding efforts increase the risk scores of MA enrollees by 10.8% above what they would be if they were enrolled in traditional Medicare. CMS does apply a "coding intensity adjustment" to the risk scores of MA enrollees that is intended to offset plans' coding efforts, but it is currently just 5.9% (the statutory minimum).34
While the most important effect of MA plans' coding efforts is to increase how much CMS pays MA plans, these activities also increase administrative costs. Much of these additional costs is incurred by plans, but some is incurred by health care providers because MA plans use a variety of strategies to enlist providers in the search for additional beneficiary diagnoses. For example, MA plans often offer bonus payments to providers who report additional diagnoses.35
31 Fiedler, Adler, and Ippolito, "Recommendations for Implementing the No Surprises Act."
32 For an up-to-date review of this evidence, see
33
34
35
* * *
For this reason, some reforms that would reduce the susceptibility of the MA risk adjustment system to plans' diagnosis coding efforts could also reduce administrative costs. One longstanding recommendation from MedPAC is to begin using two years of data on beneficiary diagnoses for risk adjustment purposes, rather than one year as is done at present.36 The logic of this proposal is that using two years of data will increase the likelihood that beneficiary diagnoses are captured even without the special efforts undertaken by MA plans. That may reduce the return to MA plans' efforts to identify additional diagnoses and cause them to reduce the intensity of those efforts. (Using two years of data is also likely to increase the number of diagnoses captured in traditional Medicare and, thus, reduce the coding advantage held by MA plans.)
Another approach is to exclude diagnoses that are particularly susceptible to plans' coding efforts from use in risk adjustment. CMS recently took a step in this direction when it updated its risk adjustment methods for the 2024 benefit year, but it would be worth looking for other opportunities in this vein.37 It is important to recognize that excluding diagnoses from risk adjustment does involve tradeoffs. While it reduces how susceptible the risk adjustment system is to plans' coding efforts, it may also reduce how effective the system is in adjusting for true differences in health status across populations.38 This may create opportunities for MA plans to profit by selectively enrolling healthier beneficiaries, something that already appears to be a substantial problem in Medicare Advantage.39 Thus, this policy approach should be used judiciously.
More ambitious steps: standardizing processes across payers
The three targeted steps described above would achieve meaningful administrative savings while presenting few tradeoffs. There are likely other targeted changes that are worth considering. Indeed, the discussion above illustrates that many different policy choices affect administrative costs and that being attentive to those effects can pay dividends. But achieving large savings would require reforms that target more than just narrow slices of provider-payer interactions.
One approach to broader reform would be to standardize some billing, coverage, or quality reporting processes across the menagerie of public and private payers that operate in
36
37
38
39 Vilsa Curto et al., "Health Care Spending and Utilization in Public and Private Medicare,"
40 Richman et al., "Billing And Insurance-Related Administrative Costs."
* * *
Standardizing information exchange through a claims clearinghouse
One challenge is that different payers may require providers to transmit information in different formats or using different procedures. Standardizing these formats and procedures could make conducting some types of transactions less complex and thereby reduce costs.
The HIPAA standards are valuable, but they have not fully standardized how different providers and payers exchange information. One problem is that the standards do not address all relevant types of information. Notably, there is currently no HIPAA standard governing claims attachments, which are used to transmit additional information that a payer needs to adjudicate a claims (e.g., medical records), despite the fact that HIPAA directed HHS to establish such a standard; this is likely one reason that transmission of claims attachments remain largely manual.41 Although HHS is now in the midst of rulemaking aimed at establishing a standard, this experience illustrates how long-lasting, consequential gaps have sometimes arisen under the HIPAA standardsetting regime. Another problem is that different entities may implement the same standard in slightly different ways, partially undermining the benefits of having a standard.
One potential reform is to go beyond standard setting and establish a single, centralized clearinghouse that would accept claims from providers, route those claims to payers, and route payers' responses back to providers.42 The clearinghouse would ensure that information exchanged in each direction was formatted in accordance with relevant standards. Similar systems are used in other economic sectors, like banking, and in the health care sectors of some other countries that have multi-payer health care systems, including
Creating a clearinghouse has several potential benefits. First, because all claims would travel through the single clearinghouse, idiosyncratic variation in how different entities implement existing standards would be eliminated. Second, the rules and procedures adopted by the...
41 CAQH, "2022 CAQH Index,"
42
43 Cutler, "Reducing Administrative Costs in
* * *
...clearinghouse would, in effect, provide new de facto standards where standards do not exist. Third, providers and payers would no longer need to establish bilateral conduits to communicate.
To achieve these benefits, the clearinghouse would need to be well-governed, as it would need to establish well-designed processes, execute those processes efficiently and effectively, and react nimbly to changing circumstances. Indeed, a clearinghouse that performed poorly at these tasks might not only fail to improve on the status quo but actually make claims processing more burdensome than it is today. Any benefits of a clearinghouse would also need to be weighed against the cost of operating it, which would presumably need to be financed through fees on the providers and payers using the clearinghouse, although these costs might be offset by reductions in the costs of private-sector clearinghouses that currently perform similar functions.
Standardizing substantive features of payer payment, coverage, or quality reporting policies
While variation in approaches to information exchange is one source of excess administrative costs, another factor is that different payers have substantively different payment, coverage, and quality reporting rules. For example, some hospital-payer contracts use diagnosis related groups, while others are based on a hospital's chargemaster.44 Similarly, different payers apply prior authorization to different services or require different standards to be met for a service to be covered.45 And different payers require providers to report on different quality measures.46
These differing rules may require providers to submit different information to different payers for the same service, which may be burdensome even if the process for submitting any given piece of information is fully standardized. Moreover, even where differences in different payers' rules and standards do not require providers to submit different information (or otherwise behave differently) with different payers, providers may still have strong incentives to do so. Customizing reporting (or underlying care delivery) in this way likely adds administrative costs.
For these reasons, standardizing payment, coverage, or quality reporting processes across payers might generate meaningful administrative savings.47 However, this type of standardization could also present tradeoffs. Under such a system, payers would be less able to tailor rules to their particular circumstances or experiment with new approaches, which could create offsetting costs, such as increases in the use of inappropriate services. Standardized rules might also differ in systematic ways from existing rules, which could be good or bad depending on the quality of the...
44
45
46
47 Cutler, "Reducing Administrative Costs in
* * *
...process used to set the new standardized rules and the degree to which the private incentives underlying the existing rules are aligned with society's interests.
In sum, the desirability of this type of standardization is likely to be highly case-specific and dependent on the particulars of the proposal under consideration. I do, however, want to touch on one domain where greater standardization may be particularly likely to be beneficial.
That domain is quality reporting. While I previously discussed the burdens created by MIPS, Medicare's quality reporting rules are not the only ones that providers must contend with; private insurers have similar programs, and these programs also generate large administrative costs.48
Policymakers could reduce these costs by establishing a standardized set of quality measures for different categories of providers, require providers to report on those measures to a centralized database, and require all payers to rely on those measures rather than collecting their own bespoke quality measures.49 Standardizing quality reporting may have few downsides since it is unclear whether the current quality reporting regime is creating substantial benefits (consistent with my skepticism about the benefits of MIPS and similar programs, discussed above). Indeed, it is plausible that centralization would make quality reporting more effective by increasing the number of patients observed for each provider and easing cross-payer comparisons.
Conclusion
Interactions between health care providers and payers generate hundreds of billions of dollars in administrative costs annually in
48
49 Cutler, "Reducing Administrative Costs in
* * *
Original text here: https://www.budget.senate.gov/imo/media/doc/101823_drfiedlertestimony.pdf
Senate Finance Committee Issues Testimony From Security Health Plan CEO Hoglund
Aetna CVS Health to enter the Affordable Care Act (ACA) individual insurance exchange marketplace in Maryland for January 1, 2024
Advisor News
Annuity News
Health/Employee Benefits News
Life Insurance News