Skip to content

Welcome to Society of Hospital Medicine

Your Browser Is No Longer Supported. Please upgrade your browser from Internet Explorer 10 to Internet Explorer 11 or higher

SHM Offers Comments on MACRA Implementation

November 17, 2015

SHM's Policy Efforts

SHM supports legislation that affects hospital medicine and general healthcare, advocating for hospitalists and the patients they serve.

Learn More

Download Letter


Centers for Medicare & Medicaid Services
Department of Health and Human Services
Attention: CMS-3321-NC
Baltimore, MD 21244-8016

Dear Acting Administrator Slavitt:

The Society of Hospital Medicine (SHM), representing the nation’s nearly 50,000 hospitalists, appreciates the opportunity to provide feedback on the Request for Information on the implementation of the Medicare Access and CHIP Reauthorization Act of 2015 (MACRA). The passage of MACRA set in motion an alignment of the various quality and value-based programs at CMS under the umbrella of the Merit-based Incentive Payment System (MIPS). It also created a pathway out of the MIPS for alternative payment models (APMs).

Hospitalists are front-line providers in America’s hospitals, providing care for millions of hospitalized patients each year, many of whom are Medicare and Medicaid beneficiaries. In their role, hospitalists manage the clinical care of acutely ill, hospitalized patients, while working to enhance the performance of hospitals and healthcare systems. The unique position of hospitalists within the healthcare system affords a distinctive role in facilitating both individual physician-level and systems/hospital-level value based purchasing and approaches to alternative payment models. Their experiences inform our comments on the MIPS and APM pathways.

A. The Merit-based Incentive Payment System (MIPS)

SHM encourages CMS to exercise caution in implementing the MIPS, an expansive program with many untested and untried components. The scope and complexity of the MIPS generates a high potential for unintended and unexpected consequences.

MIPS EP Identifier and Exclusions

CMS contemplates what identifier or identifiers should be used to identify MIPS eligible professionals (EPs). They note that under previous programs, and even within current programs, participants were identified in a variety of different ways: PQRS uses a combination of Tax Identification Number (TIN) and National Provider Identifier (NPI) or just the TIN for group reporting; the value modifier uses TIN; Pioneer ACOs use a unique ACO identifier. There are advantages and disadvantages for using each different identifier type. SHM recommends CMS use a combination TIN-NPI, as long as the principles outlined below can be met:


  1. Consistency of Identifier across MIPS Domains: As CMS indicated, many of the MIPS component elements in their current forms have different identifiers. This lack of consistency has caused a great deal of confusion among provider groups, and prevented real alignment across programs. SHM recommends that CMS institute a single identifier policy across the MIPS domains.
  2. Flexibility for Groups to Self-Select: It is important that groups are able to decide whether they wish to report together as a group, or separately as individuals. It is also important that groups that choose to report as a group are able to create an association that is meaningful for clinical and performance improvement efforts. This may require CMS to develop more streamlined processes around group-formation and a robust set of clear and simple educational materials.
  3. Reduction of Administrative Burdens: SHM recommends CMS choose an identifier that reduces the amount of administrative confusion and burden for provider groups. SHM also recommends CMS consider its ability to administrate whichever identifier it selects to ensure that the identifier is accurate and up-to-date.
  4. Transparency and Accuracy: It is imperative that CMS be fully transparent about what providers (NPIs) are associated with the identifier and what patients are associated with specific measures in the MIPS. NPIs associated with the given identifier must be accurately matched to the identifier in real time or as close to real time as possible. CMS must establish efficient mechanisms for groups to resolve inconsistencies or problems.
  5. Establish Clear Rules around the Group Identifier: One of the most persistent questions that provider groups ask is how and when a single provider is attributed to the group for a penalty or bonus payment. In hospital medicine, it is common for providers to change groups at any time throughout a year, potentially confounding how the MIPS payment adjustment is applied. CMS must develop and disseminate guidelines for how changing groups will affect an individual’s payments.


Comments on Use of TIN, NPI, TIN-NPI, and Unique MIPS Identifier

CMS considers multiple identifiers for administering the MIPS. The NPI, the smallest unit of measurement, is associated with the individual provider throughout their practice and would remain unchanged even if a provider switches groups or moves to a different location. The TIN, or the next larger unit, is used to identify groups of providers who have agreed to associate together for the purposes of billing. As groups grow and evolve, there may be multiple TINs in play: some associated with a site of service, some associated with specialties within a group, and some are artifacts of previous group arrangements. The TIN-NPI is a hybrid unit, allowing both identifiers to define more precisely the provider as an individual part of the group. The largest unit would be a unique MIPS identifier, which could be any structure of providers coming together for the purposes of performance in the MIPS. The closest proxy to this hypothetical identifier would be the Pioneer ACO identifier.

The NPI is a consistent identifier associated with a provider for the duration of their practice of medicine. If MIPS payment adjustments are applied to the NPI, it might be more possible to prevent gaming the identifier system by ensuring that regardless of what TIN (or other identifier) the NPI is associated with, it will have a payment adjustment applied to the NPI consistently. At the same time, using the NPI suggests that measures in the MIPS can truly be applied to the individual’s performance, when in reality many of the measures are systems- or team-based metrics, not reflective of the majority or plurality of a provider’s services, and/or may not have enough data to be statistically reliable or relevant. Although this sort of individual-level performance improvement might be a goal of the MIPS and other pay-for-performance programs, it is not a reality within the current programs and measures for hospitalists and many other providers.

If CMS uses the TIN as the unit of measurement for providers, SHM has some concerns that many TINs have not been constructed with a pay-for-performance application in mind, which could lead to potential downstream implications for payments and quality improvement efforts. To mitigate these concerns, CMS could develop a more accessible mechanism for de-TINing and re-TINing into more appropriate group structures. However, SHM acknowledges there are severe limitations and implications for creating a more fluid TIN environment that will be difficult to administer and maintain consistency.

Individual providers can switch from one TIN to another or bill under multiple TINs, creating a number of structural issues in the use of the TIN as the unit of measurement for the MIPS. It would be unwieldy to require providers to participate in the MIPS fully under each TIN, and likely measures under a given TIN may not have enough patients for them to be statistically reliable. CMS will also need to develop policies to ensure that providers billing temporarily under a TIN, such as what occurs with locum tenens, do not inadvertently disadvantage TIN performance.

SHM cautions against developing a new MIPS identifier. We have numerous concerns about the increased administrative burden for groups to maintain another identifier separate from those already in use by Medicare and other payers. Provider groups would need to be vigilant about maintaining the accuracy of their list of providers in this new identifier, which will significantly increase administrative cost. Recent experiences with some of the alternative payment models (bundled payments, ACOs, etc.) also raise serious concerns about the ability of CMS and its systems to keep track of an entirely new identifier with fairly fluid boundaries.

After considering the options, SHM recommends CMS use a combination TIN-NPI, as long as the principles outlined above can be met and this identifier can be fairly applied and administered. This TIN-NPI combination uses widely understood identifiers already established in the healthcare system. In order for this to work, CMS will need to prioritize ensuring that TINs are valid and up-to-date, and establish simple processes for resolving any issues with NPI attribution.

Quality Performance Category

The Quality Performance Category will subsume the entirety of the existing PQRS program and continue to build on that existing structure of measures and reporting mechanisms. Currently, PQRS participants can choose from reporting by claims, registry, qualified clinical data registry (QCDR), Group Practice Reporting Option (GPRO) web-interface, GPRO registry reporting, direct electronic health record (EHR) reporting, and EHR vendor reporting. As detailed below, SHM encourages CMS to establish policies that:

  • Maintain the claims reporting option for the foreseeable future
  • Reduce the number of required measures for reporting
  • Exercise caution when mandating reporting requirements, such as the CAHPS

Reporting Methodologies

Hospitalists typically report by either the claims or registry reporting options in PQRS. Billing data is the most readily available information for hospitalists, and for many of them, the only information available. As they work almost universally on hospital EHR systems, the nature of their relationship with the facility determines their access to more detailed data for registry and other methods of reporting. Given pay-for-performance pressures faced by hospitals, the provider-level reporting requirements for PQRS is understandably not a high priority for hospital IT and quality measurement efforts.

Given hospitalists’ reliance on claims-based reporting, we recommend maintaining the claims reporting option for the foreseeable future. This would provide a viable pathway for groups who do not have the ability to report on measures through any of the other reporting mechanisms.

Number of Measures and MAV Process

When CMS increased the number of measures required for successful PQRS reporting in FY 2014 Physician Fee Schedule final rule, hospitalists’ ability to successfully participate in PQRS was severely diminished. Thus, maintaining the PQRS requirement of nine (9) measures for successful participation in the Quality Performance Category would be a mistake. There are numerous specialties whose providers have only a handful of applicable measures. Hospitalists, for example, are generally able to report on only four (4) measures in the current PQRS program, and those measures have only loose applicability to hospitalist practice. The other measures that have inpatient codes in the denominator are simply not relevant to hospitalists, or are not within their scope of practice.

As an arbiter for specialties unable to meet the nine (9) measures requirement, CMS allowed groups to report on fewer than nine (9) measures and be subject to the Measure Applicability Verification (MAV) process. The MAV process uses clusters of measures to determine if a provider could have reported on other, related measures. Instead of setting arbitrary thresholds and policing through the MAV, CMS should work with groups up front to help them identify measures, regardless of the number, that are relevant and meaningful to those providers.

If a threshold must be used, we recommend CMS reduce the number of required measures, preferably back to three (3). Groups should report only on the measures that fit their practices, not on an arbitrary number of measures. Even if providers are reporting on fewer measures, if they are reporting on actionable and meaningful measures, they would be meeting the goals of the quality performance assessment portion of the MIPS.

Requirement of CAHPS Reporting

As part of the reporting requirements for PQRS, CMS indicated its intention to require the reporting of the CAHPS for PQRS (i.e., CG-CAHPS) for successful participation. In the FY 2015 Physician Fee Schedule Proposed Rule, CMS clarified that CAHPS for PQRS does not include hospitalists as “focal providers”. In effect, because of the relationship between hospitalists and their patients, the CAHPS for PQRS surveys could not be effectuated for hospitalists. We urge CMS to keep this principle in mind as they implement requirements for CAHPS reporting and other measures reporting: there are differences in the practice types and relationships with patients that complicates, necessarily, the ability to institute universal reporting mandates. At the same time, SHM is deeply concerned about the proliferation of patient surveys as a consequence of CMS’ implementation of PQRS and the physician value-based payment modifier. Input from patients is an important addition to assessing provider performance; however, this must be balanced by acknowledging the limitations of surveys and survey fatigue, particularly as patients are confronting acute illnesses.

Resource Use Performance Category

The Resource Use Performance MIPS Category would subsume the cost and efficiency measures from the physician value-based payment modifier. These measures include: Total Per Capita Costs for All Beneficiaries, Total Per Capita Costs for Specific Conditions, and Medicare Spending per Beneficiary (MSPB). Given fundamental issues with current Resource Use measures, SHM recommends against expanding into other cost measures at this time. CMS needs to refine the current set of measures before adding new measures to the mix.

Peer Comparison Pools

SHM has consistently raised concerns with the peer comparison groups for cost measures. It is critical that CMS develop fair comparison pools for providers in order to ensure that intraspecialty differences do not disadvantage one type of provider over another. For example, hospitalists certified as internal medicine physicians look categorically more expensive than their outpatient peers. This is a structural feature of the normal differences in practice between hospitalists and outpatient primary care physicians, not a reflection of the efficiency of an individual provider.

One simple way for CMS to rectify this problem is to fast-track approval of Medicare specialty billing codes, which would allow specialties to define with greater granularity differences in the physician practices. SHM applied for a specialty billing code for hospitalists in May 2014, but has not heard any response from the relevant CMS office. If CMS intends to use physician specialty as a source of peer comparison pools, the agency must prioritize approval of additional specialties and sub-specialties to ensure comparisons are fair.

CMS should also consider additional layers of comparison built around variations in case-mix index and regional cost differences to create more equitable pools. To increase the validity and accuracy of the measures, providers ought to be compared only to others with similar practice patterns and patient populations.

Issues with Attribution of Patients

The current cost measures in the physician value-based payment modifier are not defined around the costs attributable to the provider, but system costs associated with the patient. In most cases, the assigned provider cannot control all or most elements of costs associated with the patient, rendering the cost measures ineffective for individual quality improvement. While the provider may have control over a portion of the costs associated with the patient, they are not the only caregiver throughout the measure episode and may not, in fact, be the primary source of costs for the patient.

It is difficult to attribute patients to individual hospitalists due to the nature of their practice. The attending, or billing, physician on a patient’s record may not be the hospital providing the most care to the patient. Because hospitalists operate on a shift-based model, handoffs between providers are common and are not accurately captured in the attribution methodology. This reality raises stark questions about the feasibility of ever attributing patients equitably to a given hospitalist. Cost measures aggregated at the TIN level may mitigate some of this dynamic, but does not address larger concerns about the inability for these measures to evaluate the costs for which a group or individual are responsible.

Most hospitalists will have patients attributed to them in the MSBP measure, as it is associated with the inpatient stay. However, hospitalists are increasingly practicing in other facilities, such as skilled nursing facilities (SNFs), and are having patients attributed to them in the two-step primary care attribution process for other measures. Because measures with this attribution process are oriented around primary care, the costs for hospitalists with patients attributed in these measures are disproportionately higher, contributing to a dynamic of failure on cost measures. In this way, the patient attribution process also reinforces the importance of establishing fair comparison pools, at a minimum, to improve the fairness of the measures.

Clinical Performance Improvement Category

SHM supports the concept of a domain that allows providers to receive credit for participating in an array of activities aimed at improving quality, expanding access, and/or increasing efficiency. CMS should develop policies that ensure credit is awarded for existing and future quality improvement efforts in both facilities and practices, preferably through an attestation system that has a minimum of administrative burden for reporting.

The overarching goal for the Clinical Performance Improvement Category should be to allow credit for the important work physicians are doing, in many cases already in progress, to improve their practice and care overall. This category should not add undue administrative burdens to physician groups and should not require the performative or measurement elements already encapsulated in the other categories of the MIPS.

Hospitalists are front-line leaders on systems, process, and quality improvement in their hospitals—this is a central feature of the specialty. Unfortunately, because of the way CMS historically has structured quality improvement and pay for performance programs, this important facility-level work has not counted towards hospitalist performance assessment. SHM believes the Clinical Performance Improvement Activities Category is an opportunity to correct this unfortunate oversight. CMS must ensure hospitalists doing continuous quality improvement activities in their hospitals receive credit for this work.

Some of the activities contemplated in the RFI and articulated in the law are unambiguous, such as expanding office hours or accepting new patients. The category should be flexible enough to account for these differences and should not prioritize one type of activity over another. However, we note that not all activities require the same amount of resources or time and many quality improvement activities, particularly those around culture change, are significant multi-year efforts.

SHM recommends CMS consider an attestation system, similar to what is used for claiming Continuing Medical Education credits, for providers to indicate their involvement in a Clinical Performance Improvement Activity. An attestation system would allow providers the choice of the full range of activities available to them, while minimizing the reporting burden. It would also allow for the proliferation of types of clinical performance improvement activities as providers strive to meet the requirements of this category in good faith. Certain activities, such as completion of Maintenance of Certification (MOC), have a validation element built into the activity, while others do not. While SHM supports using MOC and other certification-based pathways, CMS should ensure this category contains a broad spectrum of activities and should not restrict performance in this category only to those activities that are designed to be part of a larger certification or reporting process.

Meaningful Use of Certified EHR Technology Performance Category

The Meaningful Use Category essentially incorporates the current Meaningful Use program into the MIPS. While SHM agrees that EHR technology is important, we caution that CMS should not view all providers as identical when contemplating this category. Hospital-based providers were given a blanket exemption under the HITECH Act, which was subsequently narrowed in the Continuing Extension Act of 2010. As a result, many hospitalists are being inappropriately considered eligible professionals (EPs) under Meaningful Use, and are subject to penalties. Over the past two years, CMS has issued an automatic exception for hospitalists that reincorporates hospital observation services, a short-term fix for hospitalists caught by the hospital-based threshold. We encourage CMS to continue this automatic exception and pursue all other available options under the MIPS to ensure that providers who should not be subject to Meaningful Use are not penalized.

Other Measures

Hospitalists do not fit into the “one-size-fits-all” approach of the current Quality and Resource Use Performance categories. These programs were designed with outpatient and/or procedure-focused specialties in mind, where there are clearer boundaries for measuring care and more ready access to data for quality measures. SHM reminds CMS of the structural issues faced by hospitalists in physician pay-for-performance programs and strongly encourages the agency to prioritize developing pathways for adoption of facility and other metrics.

It is critical that physicians are able to report on measures that reflect their unique practice patterns and characteristics. Hospitalists are, by the nature of their practices, team-based and systems-oriented providers; many of the measures in PQRS, and therefore the MIPS, are not designed for inpatient provider teams. SHM has been steadfast in advocating for the development of an option for hospitalists to align their performance in physician pay-for-performance programs with that of their institutions. In prior rulemaking, CMS contemplated two different pathways for facility alignment: adopting facility performance and retooling measures for provider-level reporting. Section 1848(q)(2)(C)(ii) of the Social Security Act makes retooling IQR performance metrics to the physician level unnecessary. SHM is much more supportive of CMS developing a simple methodology to adopt facility performance.

This option will allow providers who voluntarily elect it, to align their performance on selected measures with their hospitals. This level of voluntary alignment, while it may not be appropriate for every hospitalist and hospitalist group, embodies the drive towards coordinated, team-based care in the hospital. However, we stress that this should remain an option, as this level of alignment would be inappropriate for many hospitalist groups. We recognize the option to use facility metrics may not be a priority for CMS at this time, but with upwards of 50,000 hospitalists being inappropriately measured and penalized, we encourage CMS to implement this option without undue delay. As CMS develops options for providers to report on measures from other Medicare payment systems, we encourage CMS to keep in mind the following principles:


  1. Maintain as an Option. CMS should not move towards requiring any providers to align their performance with facilities. It may be an appropriate option for some facility-based provider groups to pursue, however there are many who would not elect this option.
  2. Choice in Measures. Providers should have the flexibility to choose facility measures they feel are adequately under their scope of influence and reflective of the care they are providing. This would also enable providers to pick clinical areas for focused quality improvement efforts.
  3. Flexibility in Attributing Facilities. Hospitalists often practice in multiple locations, making it difficult to elect a single facility for performance alignment. CMS should consider a methodology that enables proportional attribution.
  4. Reduction of Reporting Burden. It does not make sense to have hospitalists report separately on quality measures for patients already reported on by facility-level quality metrics for similar clinical concepts. CMS should take into account how hospitalists are incentivized and/or actively working on improving facility performance on quality metrics.

Much of the systems- and quality-improvement work undertaken by hospitalists is meant to improve facility performance and outcomes on facility-level metrics. Indeed, many hospitalists are incentivized to help their facilities meet HVBP and Meaningful Use performance goals. It makes sense to capitalize on this alignment by counting the work hospitalists are doing to improve for their facilities as their performance for the MIPS. We look forward to continuing to work with CMS on developing options for providers to align performance with their facilities.

Flexibility in Weighting Categories

In statutory authority granted by MACRA, the Secretary has the ability adjust the weighting of categories if a provider does not have enough measures or activities applicable for a particular category. SHM expects hospitalists will ultimately be excluded from Meaningful Use, effectively requiring the weighting for the Meaningful Use category to be redistributed to other categories. This raises some serious concerns, given some of the structural limitations that disadvantage hospitalists in the other categories. On balance, if CMS must transfer of weight from one category to another, SHM encourages preferential transfer towards the Clinical Performance Improvement Activities Category.

SHM is concerned about the implication of redistributing weight from the one category to the others, in that the net effect is higher risk in a different category. For example, if ten percent were redistributed to the Quality Measures Category, the weight of an individual measure would rise proportionally. So, for hospitalists, more weighting would be added to the few applicable measures they can report on, effectively tying more of their reimbursement to largely irrelevant metrics. A similar set of issues arise if weighting is distributed to the Resource Use Category. Without better measures, improved beneficiary attribution, and fair comparison pools, any additional weighting put towards Resource Use would severely disadvantage hospitalist providers in the MIPS overall.

SHM believes preferential transfer towards the Clinical Performance Improvement Activities Category will encourage quality improvement while reducing administrative burdens and structural disadvantages. It is the least static category with more opportunities for providers to engage in activity that is relevant to their practice. It is also the category most closely aligned with and responsive to clinical realities. Providers can analyze data and execute local-level performance improvement efforts much more readily than attempting to engage in the tedious and costly process of developing and shepherding new quality or resource use measures.

Feedback Reports

The current experience with Quality and Resource Use Reports (QRURs) is that practicing clinicians are often unable to access their reports, rendering any data contained in them ineffective tools for quality and process improvement. Hospitalists report nearly insurmountable barriers to accessing their QRURs in the current CMS Enterprise portal. It is difficult to determine who within a group has the authority to access reports and the log-in process is cumbersome and unnecessarily complicated. In addition, the group member who has authority to access the reports is frequently not a clinician or front-line provider, the intended audience of the reports. SHM believes a significant portion of why groups are not retrieving their QRURs is because of the difficulty involved in accessing the reports.

If the goal is to provide meaningful and actionable information to front-line providers for improving the quality and efficiency of care, CMS should make every effort to streamline the process for retrieving these feedback reports to ensure they are widely and readily available. The reports must also be as transparent as possible, containing specific information about the NPIs associated with the data and clear explanations on how to resolve inconsistencies or errors with provider attribution, patient attribution and calculation of measures.

B. Alternative Payment Models

Payment Incentive for APM Participation

Hospitalists face several barriers towards participation in an APM. As facility-based providers, most hospitalists have little to no control over whether or not their facility participates in a Medicare EAPM. Hospitalists are also not in a position to choose their patients and generally have no control over whether a given patient, or services provided in caring for a patient, would or would not count towards EAPM payment thresholds. Additionally, most hospitalists are either salaried employees of a hospital or contracted to provide services for the hospital. These employment structures will make it difficult, if not impossible, to trace percentages of APM revenue directly to individual or groups of hospitalists.

With this in mind, SHM has serious concerns about the ability of providers to meet the 25% payment threshold in 2019, which is scheduled to increase to 75% by 2023. We are also concerned that an impending rush to establish specialty focused APMs and subsequent counting of payment or patient thresholds will likely result in APMs competing for patients, selecting patients based on APM applicability, erosion of cooperation among providers, and entrenchment of provider siloes within the healthcare system; all to the detriment of patients and the payment system.

To address these concerns, SHM recommends that CMS take a flexible approach in defining “services furnished under this part through an EAPM entity” – a single definition may not be realistic for every provider type or APM approach. An APM that involves revenues for physician and professional services only will need to use a different method for determining revenue thresholds for the participating physicians than would an APM that involves revenues for hospital and post-acute care services. Thus, these methodologies should be left to the discretion of the APM Entities, who should be asked to describe the method they will use when they submit an APM proposal.

Patient Approach

Eligible physicians should not be required, prospectively, to choose the patient or payments approach. They should retain the option to use whatever approach comes closest to satisfying the threshold. CMS should develop policies that allow groups the option to pursue either approach, and maintain the option to switch approaches once they know where they stand at the end of a performance year. We note that the patient approach may be the more realistic and straightforward option for most providers.

We encourage CMS to count a patient in any APM towards meeting the threshold for all providers taking care of that patient regardless of whether they are in the same APM. If a patient is receiving services through an EAPM and is subsequently seen by a provider within a different APM, this should count toward services furnished through the second provider’s EAPM as well. The justification for this approach is that each patient goes into the denominator for every provider who cares for that patient. Therefore, it is logical that any patient enrolled in any APM be included in the numerator of every provider who cares for that patient during the risk period defined in that APM. This approach would recognize that care delivered within an APM does not occur in isolation, and would help encourage working relationships and cooperation among differing APM models and their participating providers. These providers will still follow the protocols, processes and interact regularly with the program staff of an APM in which they are technically not participating. Their interests are still aligned with the unrelated APM since failure to perform would risk losing the patients to another provider group.

An example best illustrates this methodology: Assume a hospitalist provider is participating in a BPCI stroke bundle. Given the breadth of conditions a hospitalist sees in practice, it is unlikely that the stroke bundle or even multiple BPCI episodes would reach the requisite 25% threshold of payments or patients. However, if an outpatient Patient Centered Medical Home (PCMH) refers a patient to the same hospitalist, this PCMH patient would also count toward the threshold even in the absence of initiating the stroke bundle.

In summary, any provider who takes care of a patient attributed to an APM should get credit for 1 patient in their calculation of their own APM threshold percentages. Allowing a combination of EAPMs for purposes of meeting the patient threshold would not only afford hospitalists and other providers the opportunity to reach the requisite thresholds, but would also encourage providers to work more closely with unrelated EAPMs in providing low cost, high quality care. Such cooperation would be necessary to ensure continual referrals and would also help prevent the potential for unproductive competition and patient selection based on individual APM attribution.

Nominal Financial Risk

SHM understands the intent of this language is to ensure providers have a vested interest in the success of the APM in which they are participating, however, we recommend that financial risk be viewed much more broadly than mere percentages of upside or downside risk and formal downside risk should definitely not be a prerequisite. Nominal financial risk should consider upfront investment, startup costs, and ongoing costs, which would include technology infrastructure, systems reorganization, dedicated administrative staff, maintenance, and other resource-intensive changes that APM participation requires. These are costs that could generally be avoided if providers were to forego APM participation altogether and they are costs that represent a type of built-in downside risk – this investment will not be recouped in the event of APM failure. CMS must develop a methodology to account for startup costs and other investments as part of their analysis of monetary losses under the APM that are in excess of a nominal amount.

For example, the start-up and administration of accepting bundled payment under BPCI is difficult and costly. Hospitalists participating in BPCI Model II have reported significant startup costs arising out of their participation and have reported ongoing BPCI operating expenses averaging $250 per episode/discharge. Further, the average startup costs of for ACOs within the first 12 months of operation have been reported to range between $2.0 million and $4.0 million, with some ACOs reporting even higher costs. In light of the substantial investment of capital, operational startup costs, and ongoing costs involved in implementing any APM model, these costs borne by providers should be counted as a significant element of financial risk. These costs are not recouped should an APM fail and they certainly represent much more than nominal amounts.

Regarding EAPM Entity Requirements

CMS asks what entities should be considered EAPM entities.

SHM recommends that CMS employ a very broad definition of entities that may be considered EAPM entities as a means to promote innovation and recognize the heterogeneity of services and clinicians covered under Medicare. Entities to consider include Managed Care entities. Managed care patients receive care with similar goals to current APMs and many provider practices derive most of their compensation from HMO plans. The small FFS Medicare component of their practice should not prevent them from moving to receive an EAPM designation. Facility-based at risk incentive plans should be counted as EAPM. Patient experience, cost of care, readmissions, mortality rate, goals are all consistent with the APM concepts and hospitalists working within such systems are often not in the position to independently establish their own APM entity. Relatedly, if a hospitalist is employed by an EAPM entity, is held to EAPM measures and is assisting in working toward EAPM goals, the hospitalist should be afforded the ability to “step into the shoes” of the hospital and be deemed an EAPM in their own right.

Additional considerations should include single and multi-specialty practices without limit in size of practice; independent practice associations (IPAs); clinically integrated networks; accountable care organizations (ACOs); health systems; critical access hospitals; rural hospital centers; and federally qualified health center (FQHCs).

Quality Measures

CMS asks what criteria could be considered when determining ‘‘comparability’’ to MIPS of quality measures used within an APM entity.

In developing regulations for this section, SHM cautions against merely surveying the list of current PQRS measures and transferring them to an APM environment. While some PQRS or PQRS-like measures may be applicable, and even useful to an APM, “comparable” should not be interpreted to mean “the same.” APMs should be left to determine what measures are most useful based on the goals and design of the APM. In making this determination, the primary questions both the APM and CMS should address are:

  1. Does the APM have quality measures?
  2. Do those quality measures align with what the APM hopes to achieve?
  3. Do the measures help ensure that patients are not being put at risk?

As described in sections 1848(q)(2)(C)(ii) & (iii) of MACRA, the Secretary may use measures used for a payment system other than for physicians, such as measures for inpatient hospitals, for purposes of the performance categories. The Secretary may also use global measures, such as global outcome measures, and population-based measures. Through the use of these sections of MACRA, CMS has fairly broad authority to shape the MIPS quality measures category beyond PQRS, and should utilize that authority by thinking outside of traditional routes. For example, bundled payment, ACOs, and certainly capitated payments, represent payment system other than the physician oriented physician fee schedule, yet the given authority could allow their metrics, even when they differ from PQRS metrics, to be included under MIPS. Once the measures are included in MIPS, “comparability” would not be an issue.

This would also help solve “comparability” issues when qualifying a non-Medicare payer, utilizing a different payment system, for the Combination All-Payer and Medicare Payment Threshold. If the non-Medicare payer’s metrics are deemed MIPS measures, comparability criteria become unnecessary.

Use of Electronic Health Record Technology

CMS has assigned priority to the question of how the agency should define “use” of certified EHR technology (CEHRT) by APM participants.

SHM strongly recommends that certified EHR technology for APMs not be tied to Meaningful Use criteria. APM EHR technology should be focused on meeting the needs and goals of the APM rather than meeting predetermined criteria and functionality. Since many new APMs will likely be developed by stakeholders, the traditional concept of CEHRT must be adjusted to allow for the changing landscape and assure the development of specialized health IT modules that support the goals of each APM. The “use” of EHR technology, its requisite functionality, and how the APM plans to meet these needs should be outlined within the APM proposal to CMS – CMS can then work with each APM to certify technology based on whether or not the expressed needs will be met.

Furthermore, “use” of any technology should not imply ownership or control of such technology. Most hospitalists are regular users of CEHRT technology, but are using these systems at the hospital level. This is likely to remain the case for any hospitalist focused APM developed in the future. Similar to Meaningful Use, CEHRT criteria that falls outside of a hospital level EHR without distinct need, would be duplicative, irrelevant or both.

Information Regarding Physician Focused Payment Models

Section 101(e)(1) of MACRA, adds a new subsection 1868(c) to the Act entitled, ‘‘Increasing the Transparency of Physician-Focused Payment Models,’’ or PFPMs. This section establishes an independent ‘‘Physician-focused Payment Model Technical Advisory Committee’’ or PTAC. The PTAC will review and provide comments and recommendations to the Secretary on PFPMs submitted by stakeholders.

The legislative language makes it clear that Congress intended for PFPMs to provide an alternative, transparent avenue for the development of qualified APMs outside of the existing processes. Congress also intended for PFPMs to be implemented – not just recommended, discussed, acted upon slowly, or never acted upon at all. Regulations should establish simple, straightforward criteria, as well as a streamlined pathway for PFPM proposals to be adopted as qualified APMs. CMS should simply ask what costs model participants are likely to incur to participate in the model (nominal risk), what savings the model is likely to achieve for Medicare, what accountability measures should be used to judge whether the model is meeting its targets for costs savings and care quality, and how to hold the APM accountable for measures and cost targets.

The regulations should also make it clear that PFPMs recommended by the PTAC, which also reasonably meet simple criteria, will be accepted by CMS. A more advanced application phase could be used to work out implementation details, but stakeholders should not have to go through multiple, lengthy steps to gain initial approval. HHS needs to organize a reasonable process that will encourage good ideas from specialty societies and other provider organizations to flourish, without overly complicated processes and criteria.


SHM appreciates the opportunity to provide feedback on the Request for Information on the implementation of MACRA. We stand ready to work with CMS to develop policies for the MIPS and APMs that work to improve patient care across the nation. 

Robert Harrington, Jr., MD, SFHM
President, Society of Hospital Medicine