blog icon
10.16.2020

Aledade Comments on the Physician Fee Schedule and Other Changes to Part B Payment Policies

by Travis Broome, SVP for Policy and Economics

On August, 3 2020, the Centers for Medicare & Medicaid Services (CMS) issued a proposed rule that includes their proposed changes to the Physician Fee Schedule (PFS) for 2021. 

Here at Aledade, we know these policy decisions are crucial not just for the hundreds of independent primary care practices in Aledade ACOs, but for physicians across the country. We provided full comments to CMS on a variety of topics in the proposed rule. The big proposal for primary care is to significantly update the physician or advanced practice provider time spent on evaluation and management (E&M) codes for the first time in over 20 years. 

Below are our full comments on the proposed regulation. Fair warning these are technical comments on a very long regulation so there are a few words more than a normal blog.

Aledade (www.aledade.com) partners with over 600 primary care physician practices, FQHCs and RHCs in value-based health care. Organized into 46 accountable care organizations across 26 states, these 5,659 clinicians are accountable for over 400,000 Medicare beneficiaries. More than half of our primary care providers are in practices with fewer than 10 clinicians. We are committed to outcome-based payment models to improve the value of health care delivered to Medicare beneficiaries and other Americans. We are committed to using technology, data, practice-transformation expertise and, most important, the relationship between a person and their primary care physician to improve the value of health care.

Our primary recommendations are:

  • The proposed Medicare Shared Savings Program (MSSP) quality measures are the right measures; however, the percentile scoring system creates too narrow a range and heightens disparities in health care - an alternative is needed;
  • 2021 is not the right time to change MSSP reporting methods as the new method would require implementation in January 2021, not February 2022 when the report is due; and
  • The revision to the valuation of E&M codes is long overdue and has been consistently foreshadowed and should be finalized for 2021.

Our full comments are below. Thank you for the opportunity to provide input into the proposed CY 2021 Medicare Physician Fee Schedule.
 

 

MSSP Quality
Measure Set and Benchmark Methodology

We support the proposal to move from ten clinical quality measures down to these three specific measures. Successfully treating high blood pressure is the best way to save the most lives.
Uncontrolled diabetes inevitably leads to comorbidities. Recognizing and treating depression is the most direct overlap between primary care and behavioral health. By removing the other, mostly topped out, screening measures CMS will reduce reporting burden.

We do have concerns about the Unplanned Readmission Rate for MIPS and the All-Cause Unplanned Admissions for Multiple Chronic Conditions for ACOs. These measures are both particularly sensitive to risk adjustment and have very narrow ranges. In particular the current Risk-Standardized, All Condition Readmission measure has a range of just 15.75 to 14.56 per thousand. This means in a 5,000 member ACO there are only 5 readmissions between the 30th and 90th percentile. In no case does moving from one percentile to the next represent even a single readmission. The range is slightly wider in MCC admissions going from 65.99 to 41.39 in the percentile changes. This still is a small number of people with only 125 admissions from terrible to great in a 5,000 person ACO. MCC has the further concern that it is right on the edge of the reliability threshold. On both of these risk-adjusted measures our community health center ACOs do poorly compared to private practice ACOs. Yet the community health center ACOs generate savings more reliably than private practice ACOs. We believe that the risk adjustment in these measures is inadequate, especially in light of the narrow range.

We suggest that CMS apply the same eligibility category definitions utilized in cost calculations to account for failings in risk adjustment to create more accurate quality score peer-based benchmarks. As these are claims-based measures, CMS knows which beneficiaries are in each eligibility category. This would allow for more precise risk adjustment of these measures.

We find it difficult to comment on the proposal to consolidate the CAHPS survey into a single measure without a more detailed proposal from CMS. The current measures used in CAHPS have either a very tight range or a capped out wide open range. This concentrates the performance on CAHPS into performance on Health Promotion and Education, Shared Decision Making and Stewardship of Patient Resources. As CMS develops the composite CAHPS measure, we encourage CMS to account for these effects to create a composite score that reflects performance in all areas of CAHPS, not just a subset. The narrow ranges of both the readmission measure and the CAHPS measure show the need for an alternative to a strict percentile approach to benchmarking.

We suggest that CMS develop an alternative to the binary choice of strict percentiles or strict numerical deciles for measures. In some measures this creates a very narrow range without
 
distinctions that resonate with health care providers. Going from the worst to the best due to a few admissions or due to a few answers on a survey does not serve to motivate health care providers. Conversely, strict numerical deciles do not motivate behavior change because everyone ends up in the top two tiers. We suggest that CMS create expanded percentiles. We suggest identifying as having too tight a range any measure (topped out or not) where the difference between the 40th percentile and the 90th percentile represents less than a 20 percent difference. The range for the identified measures would then be expanded to force a 20 percent difference. Example below:
 

Method

40th

50th

60th

70th

80th

90th

Difference

Observed

82

84

86

88

90

92

(92-82)/82= 12.2%

Expanded

75.6

78.88

82.16

85.44

88.72

92

(92-75.6)/82 = 20%

Strict Decile

40

50

60

70

80

90

(90-40)/82 = 63.4%


As you can see, the expanded percentiles stay grounded in the observed experience (90th percentile is the 90th percentile in both) while not exploding the range that way the strict decile does. While the “expanded” will result in concentration of performers to the higher percentiles, it does so to a much less extent than “strict” while creating meaningful differences between the percentiles.
 

Quality Effects on Shared Savings

Given the reduction in measures, we support CMS’s proposal to remove the domain concept from the measures. We are concerned about the proposal to change from the 30th percentile to the 40th percentile for the score needed to qualify for shared savings. First, two measures do not have established benchmarks in MIPS so it is impossible to know the difference between the 30th and 40th percentile. Furthermore, in the measures with very tight ranges the difference between the 30th and 40th percentile may not be a meaningful difference. For example, for the readmission measure less than a single readmission could represent the difference between the 30th and 40th percentile and, therefore, an ACO’s performance could be entirely dependent on risk adjustment. We note that the only Aledade ACOs that have ever scored below the 30th percentile on any measure since 2016 are those that serve disadvantaged populations and the only measures in those ACOs that have dipped below the 30th percentile are the hospitalization measures that are subject to risk adjustment.

It is important for CMS to address the failure in the risk adjustment of these measures whether by using eligibility categories as we suggest or through other methods prior to raising the benchmark threshold for eligibility for savings.
 
We support CMS’s proposal to create a gate effect on quality in which an ACO would be eligible for the maximum shared savings rate if the average percentile score of the six measures were greater than the 40th percentile or 30th percentile (if CMS does not finalize the increase to 30th percentile). This approach is utilized in many commercial ACOs.
 

Quality Reporting Mechanism

The proposed change from the Web Interface to EHR API or registry has two major implications with different consequences to consider.

 

From Manual Samples to Automated Population Measurement
 

There is no rationale for considering the movement from a sample-based, manual web interface to a population and automated submission method, a burden reduction for ACOs. Currently, ACOs are given a sample of assigned Medicare beneficiaries for each measure. ACOs can then mix EHR automation and chart review and a person can manually go into the web interface to fill out the information for each patient. This could even be done completely manually. The entire manual option and/or augmentation with chart review becomes impractical when moved to an all-patient population. If CMS were proposing to continue using a sample population, the automated reporting would be less burdensome than manual submission However, any burden reduction from the automated submission is more than offset by the increased burden of reporting on an entire patient population. Some ACOs have not invested in EHR interfaces. Other ACOs have only partial coverage. Nearly every ACO uses multiple EHRs. EHR interfaces that support quality are not cheap nor are they universal. EHR interfaces routinely cost $10,000 a year or more. Most ACOs will need several EHR interfaces and then another tool to aggregate the information from various EHRs. Today, if one EHR is especially troublesome or expensive it can be bypassed manually. Under the proposal, the only EHR that matters is the most troublesome, expensive EHR interface. Burden is now driven by the worst case scenario and is no longer driven by the average scenario.

We do believe that quality measurement should move to all-payer, all-patient and to electronic reporting; however, the environment does not support the move in 2021. First, moving to a population measurement that requires automation means the implementation date is not by February 2022 as it could be considered in a sample methodology, but rather by January 2021 as there is no realistic way to chart review all patients in the new denominators. If, in January 2021, one EHR in the ACO is not set up to capture a depression screen in exactly the right place for measurement then whether the depression screening occured is lost for measurement purposes even if the screenings are actually performed. Given the pandemic and lack of previous foreshadowing of this change, the move to automated population measurement should be postponed. We also believe that its burden should be appropriately recognized. An estimate of $100,000 per every ACO that uses multiple EHRs would be a very conservative estimate. Large ACOs with double digit EHR vendors could easily spend a million dollars a year more than they do today. We believe this move should happen eventually, but proposed timing of the proposal ignores the complexity, difficulty and expense required for compliance. CMS should give more time and learn more about the burden associated with the transition.

From Medicare Beneficiaries to All Patients
 

As stated previously, reporting on behalf of all patients increases the burden on ACOs; there are scoring considerations as well. There are differences between the population served by community health centers and the population served by private practices which are magnified when we move from reporting on Medicare patients to all patients. The three clinical quality measures do nothing to account for these differences. The claims-based measures use HCC risk adjustment, but that is not an option for these all payer clinical measures. CMS does not have the data to categorize all patients in the equivalent of eligibility categories. However, CMS could use the weighting of an ACO’s eligibility categories to create peer groups on benchmarks that can be assumed to carry forward to the entire patient population. CMS should consider whether the move to all patients improves or worsens disparities and seek to improve measurement and benchmarking in a way that incentivizes improvement of disparities.

Revisions to Payment of Evaluation and Management Codes
Time Values for Levels 2 - 5 E/M Visit Codes

We support the updated valuations developed by the AMA RUC and proposed by CMS. The evidence developed by the AMA RUC is compelling. Anecdotally, our primary care physicians agree that the pre- and post-service time is undervalued. We further support the use of the mathematical sum of component time rather than the total time recommended as we agree it is unclear what explains the difference. Ambiguity should not be the foundation on which to change the standing policy in the PFS of the total equals the sum of the components.

Revaluing Analogous Services

We appreciate CMS considering all services that use E/M visits as a building block in their valuations. Our experience and expertise centers around primary care and population health so we focused on codes related to those areas in our comments.

  • TCM Services: We support the new valuation as the TCM service description clearly ties not only to E/M, but specific E/M codes.
  • Annual Wellness Visits: We support the new valuations for AWV; however, we do note that the Initial Annual Wellness Visit (G0438) is a rare E/M based code that will actually pay less in 2021 than in 2020.
  • Behavioral Health Care Services: We support the continued relativity between E/M and behavioral health care services.
  • GPC1X: We support the continued inclusion of GPC1X as this reflects time not specifically addressed in the AMA RUC survey for the time values for levels 2 - 5. Regarding clarification, the very broad criteria of established patients seen once in the last three years does not match well to the common understanding of a longitudinal relationship. CMS should clarify whether they assume that a longitudinal relationship exists for all established patients based on the collaborative care offered to all patients of the provider or whether CMS prefers providers to evaluate on a patient by patient basis whether a longitudinal relationship exists based on utilization of collaborative care services.
  • 99XXX: We support the continued inclusion of this code and support the counting of time at the end of the level 5 E/M code time range.
Effects on Aledade Partner Practices

Aledade partners with primary care practices across the country. As CMS’s analysis by speciality shows, on average these revisions would result in an increase in revenue for primary care as primary care is weighted more to E/M than to procedures.

total-revenue-change-bar-graph


Our data indicate that actual revenue changes vary by practice; not every primary care practice will see increases, but on average Aledade practices would experience a 4.2% increase in total Medicare revenue. This analysis does not include use of GPC1X. Full analysis is included in the Appendix.

Interaction with Budget Neutrality

We support the finalization of the revisions despite the interaction with the budget neutrality requirement. Both the CMS and the AMA have long foreshadowed the revisions to the payment of evaluation and management codes. The intersection of the revisions with the long standing Congressional requirement of budget neutrality have also been known for years. By statute it is CMS’s responsibility to accurately value the services in the physician fee schedule and Congress’s job to modify budget neutrality if it believes it is warranted. Absent the pandemic, we believe that would be the last word on the subject and the responsibility to consider the impacts of budget neutrality rests solely with Congress.

However, Congress has granted HHS extraordinary flexibility and extraordinary funding to deal with the pandemic. We propose that CMS use both the flexibility and funding provided by Congress to implement a one year suspension of budget neutrality effects caused by the revisions to payment and evaluation codes. According to the 2018 Medicare Utilization and Payment Program Statistics, CMS made physician/supplier payments totaling $109 billion dollars. Even assuming a five percent increase in that number the cost of setting a neutral conversion factor would be roughly $11.5 billion dollars. HHS has significantly more COVID-19 funding than that available from Congress. We lack the legal expertise to evaluate the legality of this policy option; however, we believe it is good economic policy during the pandemic. We strongly encourage CMS to fully explore its pandemic related flexibilities and funding to offset the budget neutrality effects for 2021 and then return the responsibility to Congress where it belongs for 2022.

Finally, we do end with a clear statement that if CMS is unable to offset the effects of budget neutrality, we support the finalization of the revisions and resulting annual decrease in the conversion factor.
 

Telehealth

We fully support the proposal to extend the expanded use of telehealth for the full calendar year before the end of the public health emergency. Telehealth has been a life saving tool for many during the pandemic with many beneficiaries having used the tool for the first time. The number of providers and Medicare beneficiaries utilizing this tool has exploded. We fully appreciate the constraints on CMS under the current statute and appreciate CMS’s efforts to give both Congress and the provider community as much time as possible to address those constraints.

As Administrator Verma has stated, there is no going back to the way it was. Under the very limited use of telehealth prior to the pandemic, CMS has two payment structures for a service: facility and non facility. Telehealth was moved from facility to non facility for the pandemic.
Neither the facility nor the non-facility amounts are correct for telehealth as neither was priced with telehealth in mind. While working with Congress to update the statute, we suggest that CMS assume long-term, widespread telehealth use and begin the process of creating payment amounts that accurately reflect the provision of telehealth as delivered by providers who also deliver health care services face to face in an office setting. Pricing the future of telehealth will require study and public input and we urge CMS to begin the process as soon as possible.
 

Creation of Category 3

We support the proposed additions to Category 1 and the creation of a pandemic only category 3.

Direct Supervision over Telehealth

We appreciate the flexibility proposed by CMS in direct supervision when telehealth is utilized. We find that physicians are still somewhat uncertain on how to demonstrate they are available during the time of the service when they are not called upon to supervise. We suggest that CMS provide additional guidance that anytime a physician is providing health care services either virtually or in-person regardless of location they are deemed to be available. During times a physician is not delivering services, but is available, we suggest that such times are available in the record. While this does increase the record keeping burden, we believe that is outweighed by the burden being created today by the uncertainty.

Future of Telephone Interactions

We agree that telephone only does not replace a telehealth interaction. The rapid implementation of telehealth has also made telephone an invaluable back up. It is also a lower burden tool that creates efficiencies for both health care providers and patients. We recommend that CMS add additional time frames onto G2012 up to a level 3 E/M visit. Beyond those levels of service, we believe telehealth would be required.

Transitional Care Management

We appreciate CMS’s desire to avoid paying for overlapping services. However, the matrix needed to decide what to bill is a major source of administrative complexity in the program. We support the reduction in overlapping services with TCM. By re-evaluating the cross over, CMS creates flexibility to meet the patient’s needs.

CCM in Community Health Centers

We have reservations about the proposed bundling of G2064 and G2065 in with G0511. This bundling assumes that the mix of physician or advanced practice provider care management and other clinical staff care management is consistent across all community health centers. We do not find that to be the case and see considerable variation. Before finalizing this proposal, we suggest CMS study the variation of code use between different regions of the county and different community health centers. If any significant variation is found, we suggest that CMS delay by one year to collect more data.

Other Changes to MSSP
Dealing with 2020 and the Pandemic

The Medicare Shared Savings Program has been remarkably resilient to the pandemic. Retrospective trend looks like the best decision in value-based payment models. The partial shift to regional trend was a prudent first step but the pandemic has shown the need to shift to fully utilizing regional trends and to removing an ACO’s beneficiaries from the trend to generate the best measure of an ACO’s performance. The decision to remove coronavirus admissions certainly appears to have reduced variation based on the Q2 reports. MSSP has led the value movement from day one and continues to do so to this day.

The pandemic is once in a lifetime event and we support CMS’s continued evaluation. We support the proposal to use 2019 or 2020 quality for ACOs. We support and fully endorse the continued requirement that ACO’s report quality for 2020. While we understand the reasons CMS suspended CAHPS reporting for 2020, Aledade will be conducting the CAHPS survey for 2020 independently. The more information we have on 2020 the better. Yet we have nothing to benchmark that 2020 against so the allowance of using the 2019 results instead in shared savings rate calculations is the right decision.
 

Assignment

We support the additions to the assignment code list. CMS has impressively made changes over the years to keep assignment closely aligned with primary care relationships and we are pleased to see this pattern continue.

Link between AAPM Bonus and TIN

This is indeed a thorny issue. We experienced several situations where physicians received AAPM bonuses based on TINs for which they were no longer actively working. The reassignment of provider enrollment is biased to ensuring that new associations are created timely while there is no incentive to end old arrangements. If the old arrangement has no bills flow through it then effectively it is ended; however, it remains on the books. This calls into question CMS’s proposed approach as the CMS-588 is not a one-to-one relationship with a single QP, but potentially a many-to-one relationship.

We have experienced situations where legal action has occurred based on the employment contracts between QPs and practices. It is impossible for CMS to make the right choice when these private contracts come into play. The best solution is to shorten the timeline for payment dramatically. The MACRA statute intended for this to be a participation bonus. The bonus should be paid within 3 months of being earned. This acceleration of the timeline would end the vast majority of these concerns.

Absent the adoption of a realistic timeline for the AAPM bonus, CMS’s proposal will likely result in fewer disputes around the AAPM bonus when a QP changes practices. However, CMS should be aware that it will not end disputes. The only way to truly solve this problem is to fix the timeline. We will gladly work with Congress on tweaks to the MACRA legislation if CMS believes they lack the authority to make the bonus a participation bonus as Congress intended.