Today, we submitted to CMS our comments on the proposed changes to the Medicare Shared Savings Program. This year represents the second major revamp to the program and comes with a new name, Pathways to Success. On the whole we believe the ACO program is going to be stronger than ever before. Below are our full comments on the proposed regulation. Fair warning these are technical comments on a 600+page regulation so there are a few words more than a normal blog. For more background on what CMS is proposing we encourage you to take a look at our overview of the proposed changes and our deeper dive into the benchmark changes.

Aledade (www.aledade.com) partners with 339 primary care physician practices, FQHCs and RHCs in value-based health care. Organized into 20 accountable care organizations across 24 states, these primary care physicians are accountable for over 320,000 Medicare beneficiaries. More than half of our primary care providers are in practices with fewer than 10 clinicians. We are committed to outcome-based approaches to determine the value of health care. 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.

We appreciate CMS’s effort to continue to align MSSP’s financial incentives with value creation in Medicare. Better alignment means that savings accrue from unnecessary hospitalizations that were avoided, wasteful testing that was eliminated, uncontrolled diabetes that become managed, and better quality of life for Medicare beneficiaries. Poor alignment can leave savings up to chance, or worse, contingent upon stinting and cost shifting or other factors that do not positively affect the lives of Medicare beneficiaries.  We offer our comments on increasing financial alignment and increasing participation in the Medicare Shared Savings Program (MSSP) so that more Medicare beneficiaries can receive the benefits of lower costs and higher quality that physician-led ACOs offer.

By every measure, the MSSP is saving Medicare money and improving quality. Today it is simply the most cost effective way to provide Medicare benefits to the American people. Yet performance among ACOs is uneven, and with the right policy environment, the potential for much greater cost savings and even better quality is within our grasp.

In the proposed rule, CMS provides a comprehensive and in-depth assessment of the MSSP. CMS found that Medicare ACO efforts reduced total FFS Medicare Parts A and B spending in 2016 by about 1.2%, or $4.2 billion (after accounting for shared savings payments but before accounting for additional savings from the potential impact on MA plan payments). If FFS spillover effects are included,  the savings rise to 1.7 percent, or $5.95 billion. CMS should include the full FFS spillover effects in the savings calculations of the final rule. CMS also found that the savings varied markedly between low revenue ACOs and high revenue ACOs with low revenue ACOs accounting for nearly all of the savings to Medicare. This finding is supported by recent research from Harvard that found savings concentrated in ACOs composed solely of physicians, rather than those that include hospitals. We also compared MSSP performance to the Next Generation ACO program and found that low revenue ACOs even outperformed Next Generation ACOs.

It is this performance that should drive Pathways to Success. First, CMS should encourage more low revenue ACO participation. Second, CMS should increase financial alignment with value, particularly for high revenue ACOs. Our comments all emanate from these two objectives, which will ultimately promote cost savings in ways that are good for patients, good for health care providers and good for society.

As CMS charts the evolution of MSSP into the Pathways to Success program, we encourage CMS to value a dollar of cost savings generated through ACO performance more highly than a dollar saved through reduced shared savings payments to ACOs. The current impact analysis of the proposed rule has Pathways to Success saving $330 million in “net federal impact”  through 2024 compared to the current state, but that was accomplished through $390 million in reduced payments to 169 fewer ACOs. The loss of participation- especially by low revenue ACOs- causes $60 million in INCREASED claims costs, and comes at a real cost to beneficiaries. CMS should prefer reduced costs that represent fewer unnecessary hospitalizations, less waste in the system and healthier people. We encourage CMS to set a goal for the final regulation to generate more savings from reduced costs than from reduced shared savings payments.

The most direct way to further reduce medical costs is to increase low revenue ACO participation. To achieve this, we suggest the following:

  • Increase the proposed range of the shared savings rate in Basic (A-E) from 25 – 50% to 40 – 60% to provide an adequate return on investment and to encourage more risk taking
  • Reduce the minimum savings rate for low revenue ACOs to decrease uncertainty and provide early returns on investment to support the transition to risk
  • Eliminate or, at minimum, increase the proposed cap on the regional benchmark to create long term stability and strong incentives for ACOs to achieve and, just as important, maintain a high level of regional efficiency
  • Finalize the proposal to allow ACOs to choose retrospective or prospective assignment so that ACOs can match their interventions to their preferred assignment methodology

We offer greater detail on each suggestion later in this letter.

Greater financial alignment with value creation can also generate more cost savings. CMS’s proposals greatly increase financial alignment. To increase it even further, we suggest the following:

  • Remove the ACO assigned beneficiaries from the regional benchmark and regional trend because their inclusion disadvantages rural ACOs, and the “small numbers problem” can be easily addressed as other CMS programs demonstrate
  • Do not finalize the blended regional/national trend proposal; it does not address the problems created by the inclusion of the ACO’s assigned beneficiaries in regional trends and it creates uncertainty for the ACOs
  • Increase the cap on changes in risk score from +/- 3% to +/- 5% to focus only on outliers and to remove any incentive for ACOs to avoid risky beneficiaries
  • Increase the final regional benchmark percentage to 70% from the proposed 50% to create long-term stability in the program and move to a structure more similar to Medicare Advantage
  • Finalize the proposed glide paths to two-sided risk
  • Finalize the ability to transition to higher levels of risk as soon as the ACO is ready
  • Shorten the time frames for surety bonds to reduce the costs of going to risk without compromising the financial guarantees to CMS
  • Do not increase the administrative burden on ACOs with either disproven or unproven beneficiary notification methods

We offer greater detail on each suggestion later in the letter.

We greatly appreciate the considerable thought and analysis that CMS has devoted to the proposed Pathways to Success program. Our comments focus on refining the proposals to create even greater participation and alignment; only rarely are we in opposition to the proposals. We look forward to working with CMS to increase the savings to Medicare through Pathways to Success, and ultimately deliver better Medicare experience for beneficiaries.

Increasing ACO Participation

CMS enjoys a remarkable return on its investment in MSSP; Track 1 low revenue ACOs generated savings to Medicare of $73 per beneficiary against benchmark in 2016 and those savings appear to be growing each year. Each additional low revenue ACO that joins the MSSP represents future savings and better care for Medicare beneficiaries. But for an ACO to join Pathways to Success, a reasonable return on investment to the ACO must also be attainable. In order to determine a rate of return to the ACO, assumptions on the investment must be made.

To calculate the type of return an ACO needs to achieve, we look at CMS’ own estimates of what it costs to run an ACO from the AIM and CPC+ models.  For a 10,000-person ACO, AIM makes a $1,570,000 investment in the first year and a $960,000 investment in the second year, totaling $2,530,000. For a 10,000-person practice, CPC+ Track 2 makes a recurring annual investment of $3,360,000 – considerably more than the AIM. The average benchmark for an MSSP ACO in 2017 was $10,554 making the total spend for this same 10,000 person ACO $105,540,000. For AIM, this represents an investment of 1.5% of the total cost of care in the first year. For CPC+, the investment is a whopping 3.2% of total cost of care every year. It is against these investments that CMS should consider appropriate shared savings rates and other policies in order for Pathways of Success to succeed.

Increase the Proposed Basic Shared Savings Rate

We cannot recommend that CMS finalize its proposal to reduce the shared savings rate from 50% to 25% for one-sided risk. While we support the principle of rising share rates with rising risk,  a 25% share rate is simply inadequate to support the investments an ACO makes. We recommend that CMS use 40%–60% as the range for Basic (A-E) because this will justify the requisite ACO investments and encourage ACOs to take on more risk.

At a share rate of 50%, ACOs must reduce costs by 3% to break even on AIM-level investments and 6.4% to break even on investments at the CPC+level. Only 16% of first-year ACOs achieved 3% savings and only 6% achieved 6.4% savings. So the vast majority of ACOs will begin at with an  investment deficit, and it can accumulate year over year. If CMS were to lower that rate to 25%, only 7% of first-year ACOs would break even at AIM levels of investment, and not a single ACO would have repaid CPC+ levels of investment. By raising the share rate to 40%–60%, CMS maintains the possibility of a return on investment for the best performers in the early years of ACO participation, while maintaining its proposed glidepath to risk in the Basic track.

Basic Level A Level B Level C Level D Level E
NPRM 25% 25% 30% 40% 50%
Aledade Recommendation 40% 40% 50% 55% 60%

 

By encouraging greater participation in Pathways to Success, CMS would be making its own investment in future savings that will more than offset the greater shared savings payments under our recommendation.

Lower the Minimum Shared Savings Rate for Low Revenue ACOs

We support encouraging low revenue ACO participation by offering a lower MSR. As discussed in the previous section, even at higher levels of sharing rates, the first few years of population health are unlikely to generate savings against a historical benchmark that are sufficient for an early return. As ACOs begin to generate savings and shift the cost variation curve to the right it becomes more likely that they will cross the MSR, but not guaranteed until later years. This makes it difficult for low revenue ACOs to achieve sufficient returns to bear risk with adequate reserves. We recommend that CMS lower the MSR for low revenue ACOs in Basic Level A and Level B by 2% from the current MSR. For example, a 10,000 person ACO has a current MSR of 3%; under this recommendation it would be 1%.

This proposal would make it much more likely for low revenue ACOs to see early returns on their investments, which would support earlier transitions to risk and prevent ACOs from dropping out of the program. Although we understand  CMS is concerned about ACOs that do not make investments and still receive savings payments, we believe several forces mitigate this possibility. First, low revenue ACOs generate net savings to Medicare even in Track 1. Second, the availability of the lower MSR is time limited to two years under the proposed Basic track. Third, as shown in the analysis that accompanied the NPRM, savings against benchmark under represents the total savings to Medicare. CMS still retains 100% of the savings not reflected in benchmark performance regardless of MSR. According to the analysis in the regulation, a low revenue ACO with 10,000 Medicare beneficiaries would have saved Medicare $730,000 in 2016. CMS simply can’t afford not to have more low revenue ACOs. We believe that the performance of existing ACOs will continue to improve, but if we ever hope to turn 2016’s $6 billion in savings into $60 billion, we will need considerably more ACOs. Creating a more predictable, reasonable return on investment will spur the private investment needed to make this happen.

As part of this proposal, CMS should be vigilant about potential gaming of the low revenue designation. We define gaming as a scenario in which owners of ACO participants are not official  participants in the ACO, thereby allowing the ACO to qualify for the low revenue designation. We suggest two possible solutions to address this concern. CMS could simply lower the low revenue designation from 25% to 15%. We note that the actuaries’ impact analysis used the even lower 10% to draw the distinction. Alternatively, CMS could continue the Track 1+ policy of requiring ownership attestations from ACO participants.

Apply Regional Adjustment Earlier but Remove the Proposed Cap

We support CMS’s proposal to apply a regional adjustment in the first contract period, but we do not recommend that CMS finalize its proposal to cap the regional benchmarking adjustment. At a modest 25% or 35%, this adjustment provides a duly-earned incentive for efficient practices in their first contract period, without discouraging less efficient providers to enter an ACO and improve. It also provides predictability and simplicity for ACOs as they seek to understand the nuances of the regulatory environment.  Finally, it accelerates the process of making MSSP more like Medicare Advantage, which is a CMS goal that we support. However, we are concerned about the 5% cap on the regional adjustment, and we propose eliminating it while CMS gains more experience with Pathways to Success. If CMS does not eliminate the cap, we recommend increasing it to 8% and applying it at the aggregate level not the eligibility category level, which is an efficiency return similar to what Medicare Advantage plans can receive net of their administrative costs for administering the plan (~7%).

We understand CMS’s concern about allowing large windfalls for regionally efficient practices, but such a cap should capture only true outliers, which would suggest a cap greater than 5%. In considering a cap, it is imperative to consider the inherent policy tradeoff, and we believe that CMS is overly concerned with biased ACO selection, in which ACO participants skew heavily towards already efficient providers. There are several reasons why this should be less of a concern, and why increasing the cap to 8% at the aggregate level strikes a better policy balance. First, even if ACOs were able to identify such practices using historical data and to sign up a portion of those practices, such efficiency is likely to revert to the mean, including during the gap between when data is available, and when the performance period starts. Furthermore, by capping the regional adjustment of an ACO at 5%, CMS is reducing the incentive for the practices within an ACO to continue to improve. Such a policy would create essentially a 100% tax rate that sends the wrong behavioral message to ACOs, to rest on their laurels. This is especially true if CMS applies the cap at the individual eligibility category level. If an ACO were to cap out on ESRD patients or dually eligible patients, the ACO would be incentivized to no longer improve on those patients even though much more may still be done for them. It is extremely difficult to achieve regional efficiency, and even more difficult to sustain it. Medicare should not implement policies that penalize physicians for further improving on high-quality, cost-efficient care.  Finally, just as in Medicare Advantage, a dynamic free market, profits will allow providers that have achieved and sustained ever higher levels of patient care and efficiency to expand their practices and attract greater number of patients, which is beneficial to both patients and taxpayers.

The proposal to cap the regional adjustment is also not in keeping with CMS’s desire to promote physician-led ACOs, which are organizationally and empirically in a better position to succeed in shared savings arrangements. In a changing landscape of value-based care, these physicians face immense pressure to join hospital systems, which can offer them higher salaries and technological and regulatory support. However, we know that independent physicians would prefer to remain autonomous and hope to resist the financial temptation of the hospitals. We also know that such hospital-physician vertical consolidation is harmful to Medicare and our health care system more broadly. Hospital acquisition of physician practices reduces patient choice, promotes more concentrated markets, increase prices and reduces incentives to deliver high-quality care. Rewarding physician practices for their efficiency compared to large hospital systems may very well be the best way to keep them independent. At a minimum, CMS ought not to implement policies that reduce the financial proposition of remaining as an independent.

Allow Choice of Assignment Methodology

We support the proposal to allow all ACOs to choose between retrospective and prospective assignment. We appreciate CMS’s efforts to include waivers in both methodologies. The inclusion of the waivers in both methodologies is key to making this policy a true choice for the ACO. It is not obvious that one approach is better than the other for population health. It is also very likely that ACO characteristics mean the better approach may vary by ACO. For example in a recent National Association of ACOs report, prospective assignment only assigned 79% as many beneficiaries that retrospective assignment did. Smaller ACOs might not cross the 5,000 threshold with prospective assignment. Other ACOs might prefer the “tune-up” period for new patients that prospective assignment creates where they are a new patient in say June of 2018, but won’t be assigned until 2019. Given the variability, we believe CMS makes the right proposal to leave the decision in the hands of the ACO.

Aligning Financial Incentives with Value

Remove ACO Beneficiaries from Regional Comparisons

We strongly recommend that CMS remove an ACO’s assigned beneficiaries from regional benchmarks and regional trends. We continue to believe that the adverse effects of this policy are underappreciated. In 2016, CMS introduced the regional benchmarking methodology for ACOs that included the ACO’s assigned population in the comparison group. This has two effects which increase as the ACO’s share of the county population grows. First, in order to demonstrate savings, ACOs must fight against the very progress they are creating. A rural ACO with 50% market share would have to generate 6% in real savings to be recognized for a 3% savings against regional trend. Second, the ill effects of this policy are exacerbated by imposing a cap an ACO’s risk score, yet still incorporating uncapped risk increases of an ACO’s population in the region’s risk score. Every ACO loses calculated savings compared to regional trends whether their risk score increases or decreases. This discourages ACOs from moving to two-sided risk and reduces  the Medicare savings ACOs create. Most egregiously, this policy systematically disadvantages rural ACOs compared to urban ones. By adopting the policy of removing ACO beneficiaries from regional comparisons, CMS will increase the participation of rural ACOs in Pathways to Success.

The table below demonstrates the effects of this policy on ACOs. In our example, we have an ACO with a historical benchmark of $9,000 in a region where all the beneficiaries outside the ACO have average annual expenses of $10,000. For simplicity, risk is held constant; hence, beneficiaries in the ACO cost $1,000 less per person, risk adjusted, than the beneficiaries outside of the ACO. We also assume that the ACO reduces costs by 5% in the performance year, from $9,000 to $8,550, generating $450 in historical savings. Assuming a regional bonus of 35% of the regional benchmark difference, the combined regional savings and historical savings of this ACO is $800 per person ($1,000 x 35% + $450). Yet, as the table illustrates, if the ACO’s beneficiaries are included in the regional comparisons CMS starts to capture 100% of those savings at a rate that increases in direct proportion to market share.

Clearly, an ACO that represents 5% of its market and captures 95% of the generated savings is more likely to move to risk than an ACO that represents 35% of the market and only captures 65% of its generated savings. It is also true that the ACO with 5% market share is almost certainly urban, while the ACO with 35% market share is likely rural. We include the market share of Aledade ACOs in 2016 as an example. The primary driver of the percent market share is location, not size nor composition, as all the ACOs are all made up of primary care physician offices.

*ACO 15 includes Philadelphia

*ACO 14 includes New York City

*ACO 13 includes Orlando

*ACO 12 includes Philadelphia

CMS has raised one concern with the removal of an ACO’s beneficiaries from the regional benchmark. This concern is that without the ACO’s beneficiaries, some rural counties will not be large enough to calculate a meaningful regional benchmark and/or regional trend. First, while this poses a challenge, simply sacrificing those most harmed by the policy cannot qualify as an tenable solution. Second, CMS has historically dealt with this small numbers issue in a much fairer way. Medicare Advantage, whose members are not included in the calculation of their rates, encounters the small numbers problems due to high MA penetration. CMS’s solution is not to put Medicare Advantage members back into the rate calculation, but simply to expand the geographic area under consideration from the county to the state. Given the variety of size amongst states, we recommend a more narrow geographic expansion. We recommend that for counties with small populations, the geographic area be expanded from the county of beneficiary residence to all the counties that are in the same hospital referral region (HRR). We considered recommending that all contiguous counties be included; however, for some suburban counties, this could result in the addition of one more urban county in the direction of the tertiary care, and several more rural counties further from tertiary care, ultimately distorting the comparison. By including all the counties in the same HRR, CMS would create a region that is based on common health care service utilization. HRRs are defined by zip codes, but for simplicity we recommend that CMS include all of the counties that are in the same HRR, even if the entire county is not in the HRR. Failure to address this easily solvable problem unjustifiably disadvantages rural ACOs and thereby discourages their participation.

We recognize that our proposal, which would necessitate county benchmarks unique to every ACO, may be burdensome for CMS if the underlying infrastructure is not flexible. While not as precise, CMS could apply a mathematical adjustment to the existing regional calculations to simplify the process. The regional efficiency (regional benchmark compared to ACO benchmark) and regional trend could be multiplied by 1 divided by (1-ACO’s market share).

True Regional Trend = (Regional Trend including ACO beneficiaries) x (1/(1-ACO Market Share))

True Regional Efficiency = (Regional Risk Adjusted Costs including ACO beneficiaries – ACO Risk Adjusted Costs) x (1/(1-ACO Market Share))

Because this adjustment would be a less accurate measure of the regional trend and regional benchmarks, we would encourage CMS to consider using assignable, rather than assigned, beneficiaries in calculating the adjustment, in order to capture spillover effects. We note that this does not directly address the small number problems and that expansion to HRR may still be warranted in some cases. Overall, we encourage CMS to calculate CMS regional benchmarks and regional trends without the ACO’s beneficiaries; however, if the burden is too high, we believe that this adjustment methodology is better policy than both the current inclusion of the ACO beneficiaries and CMS’s proposed blending of the regional and national trends.

CMS claims in the proposed rule that by blending regional trends with national trends at the percent of market share alleviates the problems created by including ACO beneficiaries in regional comparisons. First, it does not address at all the reduction in the regional benchmark. CMS’s own analysis shows that nearly half of ACOs already disadvantaged by this policy would see additional cuts as a result of the blended national rate. Referring back to our table, an ACO at 35% market share would only receive 68% of the regional bonus that an ACO at 5% market share would receive, even if it has an identical $1000 difference between its beneficiaries and the others in their regional. The blended trend does nothing to effect this disparity. Second, we do not believe the blended trend addresses the problem with including beneficiaries in regional trend. For example, consider a hypothetical ACO that has 20% market penetration and reduces costs by 5% in a region where trend increases by 5% in the 80% of beneficiaries not assigned to the ACO. Under current methodology, the regional trend would be 4% for the region (5% x 80% + 0% x 20%). Under the proposed rule, the regional trend would still be 4%, but it would only account for 80% of the trend (3.2%). National trend would make up the other 20%. If national trend is 3%, then the hypothetical ACO would be at a lower trend than the current policy would dictate (4% x 80% + 3% x 20% = 3.8%). This example ACO would have been been better off under the current policy. Even in a favorable scenario, where national trend is higher than regional trend would be under current policy, this is only adequate if national trend is higher than regional trend would have been without ACO beneficiaries included. This effectively shifts savings from high-trend to low-trend areas. We address this policy goal later. We believe we have demonstrated that the blended trend does not solve the problems created by inclusion of the ACO’s beneficiaries in regional trend, which moves savings out of rural areas and into urban areas.

Use Regional Trends as the Most Accurate Measure of Value

We cannot recommend that CMS finalize its proposal to blend regional and national trends based on ACO market share. We support the use of regional trends in all years. CMS gave two reasons for the blended trend rate. First, it was billed as a partial solution to the ill-advised inclusion of ACO beneficiaries in regional benchmark and regional trend. As discussed previously, it is a rare case in which the blend accomplishes its goal and, in those cases, it is merely by chance. This simply does not work and should not be the basis of a finalized CMS policy. Second, by including national inflation, CMS has yet again decided to reward low-trend regions and to penalize those areas with high trends. The original statute enacting MSSP, as well as the first regulation for MSSP, also sought to use national trends to similarly sacrifice direct measures of value creation, in favor of other policy ends. The flaw has always been that it measures ACOs based on uncontrollable factors: an individual ACO is not responsible for whether its area is high cost or low cost, and it is often not responsible for cost changes in the region.

The combination of these two policies has created a great deal of uncertainty in ACOs. For example, in the first performance year of one of Aledade’s ACOs, a new rehabilitation hospital opened early in the year. This increased rehab costs by ~1% of the total cost of care, creating a permanent headwind for the ACO. In another rural Aledade ACO, patients began receiving adequate access to certain Part B drugs, which increased Part B drug expenditures by 27%. This regional development was excellent for Medicare beneficiaries, but made it virtually impossible for the ACO to generate savings. Unobservable on national trends, both examples had profound effects in the region. Further, non-medical examples abound such as when hurricanes or other natural disasters impact any area, the regional trend justifiably deviates from national cost growth. CMS should move away from national trends because they confound the measurement of results that are controlled by and attributable to the providers in an ACO.

Improving Utilization of Risk Adjustment

We support CMS’s proposal to eliminate the the distinction between newly and continuously assigned patients and apply the HCC scores of the assigned population to adjust the benchmark. This simplifies the risk adjustment methodology, aligns it with value generation for greater predictability, and makes the MSSP operate more like Medicare Advantage.

To reduce gaming, the current model allows risk scores of continuously assigned patients to decrease, but not increase. Therefore, practices are inherently disadvantaged by changes in the acuity of their patients: if their population becomes sicker over the course of the contract period, the benchmark is not increased; if their patients become healthier, the benchmark is decreased to reflect the lower predicted spending. Considering the sensitivity of the benchmark to risk scores, these small increases in population acuity over the course of the contract period can meaningfully harm ACO performance. The disadvantage of asymmetric risk adjustment for continuously assigned patients is mitigated, in part, by allowing the newly assigned population to receive symmetric changes in risk scores. But the effect is minimal, and the bifurcation of the ACO population (newly and continuously assigned) introduces complexity and unpredictability. As such, we support the elimination of this distinction, coupled with a symmetric risk adjustment model.

We support CMS’s proposal to adjust benchmarks based on risk, in order to decrease actuarial risk, particularly for smaller physician-only ACOs. We are also sensitive to CMS’s concerns about gaming, and we thus support a risk adjustment cap. However, such a cap ought to be set as an outlier policy that prevents excessive upcoding. If CMS retains a cap, not only should it be raised, but CMS should also apply the cap at the aggregate level of the ACO, not the individual eligibility categories (ESRD, Disabled, Dual, Aged/Non-Dual), and CMS should consider overall ACO size when choosing an appropriate cap.

There are other reasons besides focusing on outliers to increase the size of the cap. If the cap is set too low, CMS is incentivizing “cherry picking” and “lemon dropping,” wherein providers face strong incentives to select healthy patients and avoid the sickest patients, once the contract period begins. We believe that the proposed 3% cap should be raised to 5%, in order to avoid these perverse incentives. Indeed, in the proposed rule, CMS reviewed trends in HCC risk scores in the MSSP and states that “a 3% cap…would limit positive risk adjustment for less than 30% of ACOs, even when there is a 5-year lapse between BY3 and the performance year.” Thirty percent is far too great of a proportion to harm in this way, and it will unjustifiably penalize ACOs who care for a population whose risk increases over the course of the contract period. We encourage CMS to re-run its analysis to determine a cap that would hinder 10% or less of ACOs. We hypothesize that a +/- 5% cap at the aggregate level of the ACO would be sufficient. As alluded to above, it is critical that the cap is applied at the aggregate level because some categories have much smaller populations and therefore will vary naturally more than the aggregate ACO population; this methodology will ensure that the risk adjustment cap is an outlier policy that discourages lemon dropping. Further, we encourage CMS to consider whether the cap should vary by ACO size, considering that, for example, a 5,000 assigned beneficiary ACO will have more variation than a 50,000 assigned beneficiary ACO.

Finally, as more ACOs bear risk and become efficient, high-quality alternatives to Medicare Advantage, CMS should consider policies that equalize current actuarial disparities that result from risk adjustment. Medicare Advantage allows for annual and uncapped changes in risk scores, creating strong incentives to code intensely. In 2016, risk scores were 8% higher in MA than FFS for similar patients, according to the Medicare Payment Advisory Commission (MedPAC). And accounting for the coding intensity adjustment, which applies an across-the-board coding reduction for MA patients, risk scores remained 2 to 3% higher in MA, compared to analogous FFS patients. By setting the MSSP risk adjustment cap as an outlier policy, and thereby allowing accurate and comprehensive diagnostic communication of risk for ACOs, CMS will begin to create more equivalence between MA and FFS.

Increase Regional Benchmark Percentage to 70%

As we noted in our comments addressing the regional adjustment cap, we support CMS’s proposal to introduce a regional component of the benchmark in the first contract period. We agree that the initial regional benchmark percentage should be set at 25%/35%, but this percentage should be increased to 70% in subsequent contract periods, rather than capped at the proposed 50%. It is worth emphasizing that CMS has proposed not only to reduce the regional benchmark percentage to 50%, but also to impose a cap of 5% on the regional adjustment. We understand CMS’s impetus for these proposals, but it is a substantial over-correction – and it creates a new set of ill-advised incentives that reduce the long-term viability of cost-efficient ACOs. We recommended eliminating the cap. Alternatively, we recommended an 8% cap tied to regional efficiency gains that can be expected in Medicare Advantage. If CMS does finalize a cap, we believe it is redundant to also lower the percentage from 70% to 50%.

ACOs are Medicare’s most efficient delivery system; the MSSP saves money relative to FFS and MA, and it performs better on quality metrics. Given that MA benchmarks are effectively calculated with a 100% regional component (no historial factor), CMS’s proposal of 50% regional adjustment for ACOs is a regression from the goal of MA-MSSP harmonization; it would codify a substantial divergence from MA and stymie cost-effective ACOs that wish to become risk-bearing ACOs. A 70% regional factor for second-contract ACOs justifiably aligns these risk-bearing providers with Medicare Advantage.

Finally, we anticipate that CMS may be deciding between finalizing the risk adjustment cap of 5% or reducing the regional factor to 50%. In this scenario, we would encourage CMS to choose to eliminate or raise the cap.

Create a Glide Path to Two-Sided Risk

We support the proposal to create a glide path to two-sided risk in Basic based on the use of revenue-based risk. We believe revenue-based risk properly motivates positive change without threatening the financial viability of physician practices. There has been considerable focus in the ACO community on the right time to require ACOs to transition to risk. We do not take a position on the optimal year and instead focus on the progression to risk. We support CMS starting with low levels of risk and progressing towards risk adequate to justify the AAPM bonus. We also support the proposal for ACOs to stay at revenue based risk for another full contract in Basic Level E. Finally, in light of all these proposed changes, we would encourage CMS to allow ACO participants to switch ACOs on the July 2019 start date, even if the ACO participant is in an ACO with an existing ACO agreement that runs past July 2019.

Allow ACOs to Transition to Risk Faster

We support the proposal for five year contracts, but only if ACOs can more quickly move to risk within Basic and can move to Enhanced in any performance year. CMS should not prevent ACOs from moving to risk as soon as they are ready to do so.  While a clear minimum glidepath advances the goals of risk bearing without deterring participation , the ability to assume risk more quickly is equally crucial for ACOs that are ready. If CMS wants to maximize ACO risk taking, this policy is key.

Shorten the Time Frame for Maturity of Surety Bonds

We cannot recommend that CMS lengthen the time frame of the repayment mechanism from 5 years to 7 years. We suggest that CMS institute a 3 year repayment mechanism that is renewed annually. These recommendation are based on our experience securing surety bonds for risk taking ACOs.

In 2018, three Aledade ACOs entered into two-sided risk contracts under the current Medicare Shared Savings Program.  Of these three ACOs, two participate in the “higher reward – higher risk” Track 3 contract (representing 50% of all 2018 entrants into Track 3). We secured surety bonds to meet the 1% requirement for the Track 3 ACOs. We look forward to bringing many more of our ACOs into two-sided risk; however, the cost is already substantial.

Under the proposed rule, all ACOs will be required to take two-sided risk within three years. As a prerequisite to take on two-sided risk, ACOs must demonstrate their ability to repay shared losses by setting up repayment mechanisms.  The repayment mechanisms can be (1) cash in an escrow account, (2) surety bond, or (3) letter of credit from a bank.

The repayment mechanism amount required is significant and can range from $100,000 for an ACO with 5,000 assigned beneficiaries under the BASIC track, to $5,000,000 for an ACO with 50,000 assigned beneficiaries under the ENHANCED track. The proposed 7-year term required for repayment mechanisms raises the cost much more than simply adding two years would suggest.

Surety bonds or letters of credit can reduce the liquidity burden of the repayment mechanism, as they allow for lower cash collateral.  However, the requirement to maintain a 7-year term severely limits the availability and attractiveness of surety bonds as an alternative to posting cash in escrow which limits the ability of an ACO to invest in savings.

Typically, surety bonds are nearly always issued for a maximum of 5 years, due to reinsurance and regulatory complications surfacing beyond this time frame.  Most notably, reinsurance treaty prohibits insurers from writing bonds with terms exceeding 5 years.

Aledade has brought CMS’s proposal to the attention of One Beacon Surety Group, Philadelphia Insurance, Liberty Mutual and Swiss Re. After discussion and analysis both organizations came to the conclusion that if the proposed  7-year term is finalized, all surety bonds would require the full 100% collateral. This would impose a significant liquidity and capital burden, limiting a ACO’s ability to invest in innovations that deliver higher quality care at lower cost.  This would be especially problematic for physician-based and small, rural ACOs, neither of which have access to low-cost capital.

As noted in the proposed rule, the 7-year term would also introduce challenges regarding repayment mechanism amount estimations at the start of each performance year, as well as changes to the estimated amount upon contract renewal with term extensions.  While the proposed rule attempts to allow ACOs to re-use existing repayment mechanisms, in practice these repayment mechanisms (other than escrowed cash) will be re-underwritten and reissued on an annual basis.

Therefore, Aledade strongly urges CMS to set the Pathways to Success repayment mechanism duration to 3 years, with a required annual renewal with the appropriate updated repayment mechanism amount. We also recommend that CMS only require the value of the repayment mechanism to change if it increases by 10 percent or more. CMS’s proposal to use the lesser of 10 percent or $100,000 would require nearly all ACOs with a total cost of care of $200 million or more to reset every year, which would greatly increase the burden on CMS and the ACO. These recommendations would allow CMS to (1) continue to protect the financial integrity of the program by ensuring that all continuing and renewing ACOs will remain capable of repaying losses, (2) streamline to one consistent repayment mechanism, and (3) preserve the viability of surety bonds and letters of credit so physician-led and small, rural ACOs access capital and liquidity.

Reduce, Do Not Increase, the Administrative Burden on ACOs

We cannot recommend that CMS finalize its beneficiary notification of voluntary alignment as proposed. We believe in voluntary alignment as an important expression of beneficiary choice. However, the proposed division of labor – wherein CMS crafts the message and the ACO delivers it – does not work. When ACOs were required to send out mailers regarding data sharing, the most frequent response was that the beneficiary did not want the government to have their data. The message was misunderstood by the beneficiaries and ACOs and physicians were left with the fallout from a message they did not craft. We believe CMS should take ownership of the process. It should both craft the message and disseminate it to assigned beneficiaries. If CMS does finalize a requirement for ACOs to educate beneficiaries about voluntary alignment, then CMS should delegate the entire process – both crafting and delivering the message – to ACOs.

Avoid Beneficiary Opt-In Requirements for Shared Savings Programs

We believe that requiring opt-in of all ACOs in the Pathways to Success program would end participation in the Pathways to Success program. The shared savings economic model simply does not support the type of investments that Medicare Advantage makes in an opt-in model. CMS would undoubtedly lose the cost savings demonstrated by ACOs in 2016 and 2017 and certainly would have no chance of greatly increasing those savings in future year. As hypothesized by CMS in the proposed rule, enrollment in Pathways to Success would fall dramatically and CMS would have to implement a much higher shared savings rate in order to support the large ACO investments that enrollment would require.

CMS also discusses the possibility of ACOs opting in to an enrollment model. We believe that anytime a beneficiary is subject to a restriction in providers or a less generous benefit than in traditional Medicare, the beneficiary should choose to enroll in a model. CMS is not currently proposing to offer ACOs either ability in the Pathways to Success program. If CMS were in the future to offer these options in Pathways to Success or in another model, we would encourage CMS to revisit enrollment as an option for ACOs wishing to include more advanced benefit design in their ACO work.

We look forward to continuing to work with CMS to incentivize more value creation in health care. Please contact me or Travis Broome (travis@aledade.com) if you have any questions about our submission and/or if we can be helpful to you and your staff as you consider the finalization of this regulation.