CMS Proposes Changes to the MSSP Benchmarking Methodology

Transcription

1 Policy Brief February 3, 2016 CMS Proposes Changes to the MSSP Benchmarking Methodology On January 28 th CMS released the proposed rule updating the benchmarking methodology for renewing ACOs in the Medicare Shared Savings Program (MSSP). In the final rule update to the MSSP in June of 2015, CMS signaled their intent to more thoroughly evaluate the various options for creating a more accurate and equitable financial benchmarking methodology in a future proposed rule, the program revision most frequently requested by stakeholders. The major aspects of this proposed rule include 1) inclusion of regional expenditures as a major factor in establishing an ACO s benchmark in their second and subsequent performance periods, 2) an additional option for ACOs to transition more quickly to riskbearing tracks and 3) a number of administrative modifications that illustrate growing sophistication within CMS to act as a value-based purchaser. This document provides an overview of these major aspects of this rule, including some implications of the proposed changes. For a detailed summary of the rule, see the Appendix on page three. 1. Regional Trends as a Major Factor in Determining Benchmark CMS clearly believes that, despite some potential for unintended consequences, incorporating regional factors in assessing how ACOs perform financially relative to their benchmark will more accurately reward improvements in efficiency and benefit both CMS and ACO participants. The proposed changes include incorporating regional expenditures in benchmark rebasing in two different ways: CMS would establish and update the rebased benchmark using the regional growth rate, rather than the national, to trend forward an ACO s historic spending. Under the proposed approach, CMS would calculate the risk-adjusted county FFS expenditures for the ACO s regional service area, weighted by the proportion of the ACO s assigned beneficiaries in each county. Using regional trend factors would incorporate regional FFS spending and population dynamics specific to the ACO in the rebased benchmark. CMS would adjust the benchmark by a percentage of the difference between the ACO s average expenditures and the average expenditures in the ACO s service area. Although this is the more aggressive approach to incorporating regional expenditures, CMS is proposing a blended approach of historical and regional benchmarking that would gradually favor the regional component over time. The first agreement period would continue to be a solely historical benchmark, but in the second agreement period, the rebased benchmark would be 65% historical and 35% regional, and 30% historical and 70% regional in the third agreement period. Implications: CMS anticipates that these changes will keep ACOs enrolled in the program when downside risk is mandatory because ACOs will have more certainty that prior efficiencies will continue to be beneficial. Using regional expenditures will minimize the impact of localized exogenous factors such as a bad flu season or unusually large wage index adjustment. Additionally, replacing the national trend with the regional trend may help to ensure that ACOs do not receive undeserved savings or losses. For example, cases that result in lower rebased benchmarks will prevent ACOs from earning savings unrelated to efficiency improvements, and cases that result in higher rebased benchmarks will prevent ACOs from abandoning the program after factors beyond their control prevented them from receiving savings. On the other hand, ACOs with low benchmarks

2 relative to their region could receive savings unrelated to improved efficiency, and ACOs with high benchmarks relative to their region may choose not to participate, foregoing the opportunity to impact populations with the greatest opportunity for improvement. There is potential that regional growth rates could be reduced by a prevalence of high performing ACOs in the region, but CMS believes this would not prevent an overall improvement in the incentive for ACOs to improve efficiency. CMS expects the net effect of the changes to be an increase in program savings. 2. Additional Option for Transitioning to Risk-Bearing Tracks Although officials acquiesced to calls that ACOs be allowed to continue a second agreement period under the upside only Track 1 in the June 2015 Final Rule Update, CMS feels there are ACOs who would benefit from an intermediate option. CMS proposed adding a new participation option that would allow eligible Track 1 ACOs to defer their entrance into a performance-based risk model (Tracks 2 or 3) by one year, extending their first agreement period under Track 1 for a fourth performance year. This option would be available to ACOs that are eligible to renew for their second agreement period in Implications: This additional year will allow ACOs to further develop infrastructure and prepare for downside risk, but will also force ACOs to make a tradeoff between deferring risk and establishing a new benchmark with the updated regional methodology. 3. Growing Sophistication within CMS to Administer ACO Programs CMS proposed a number of technical changes intended to refine and improve the administration of the program: Incorporating regional HCC risk scores in the risk adjustment of the rebased benchmark to account for differences between the ACO s beneficiaries and those in the region, and to protect against varying coding intensities. Adopting a less operationally burdensome approach for adjusting the benchmarks to account for changes in ACO provider lists that relies on a single reference year. Making county-level data available to enable ACOs to better predict their benchmarks. Clarifying timeframes and guidelines for reopening past payment determinations to give ACOs more financial certainty. Projected Financial Impact of the Proposed Changes To predict the expected financial impact of the proposed changes, CMS conducted a Monte Carlo simulation that consisted of 1,000 random trials. The median expected impact of the proposed rule would be a $370 million reduction in claims and a $250 million dollar increase in shared savings payments, resulting in a $120 million net decrease in federal costs. A high estimate (90 th percentile) resulted in $490 million net decrease in federal costs, while the low estimate (10 th percentile) resulted in a $250 million net increase in federal costs. CMS believes that the overall financial benefits of these changes will come as a result of more accurate shared savings and loss determinations, increased predictability and greater incentives for ACOs to focus on cost reductions. The full text of the proposed rule can be found here. CMS will accept comment on the proposed rule until March 28th.

3 Appendix: Detailed Rule Summary I. Executive Summary and Background A. Executive Summary Purpose: o Address concerns regarding financial benchmarking methodology o Establish additional options for ACOs to enter performance-based risk arrangements o Create policies for reopening of payment determinations on shared savings and losses Summary of the Major Provisions: o Incorporate regional expenditures to make cost target more independent of historical expenditures o Modify risk adjustment methodology to account for regional adjustment o Revise benchmark methodology to account for changes in ACO (TIN) participation o Add a renewal option to encourage ACOs to enter risk-based tracks earlier o Create policies for reopening of payment determinations on shared savings and losses Summary of the Costs and Benefits: o Net increase of $120 million in savings o Greater program participation o Greater incentive for ACOs to invest in care management infrastructure o Increase participation in Tracks 2 &3 for certain ACOs o Greater net savings to Medicare B. Background: MSSP Final Rule I in November of 2011; MSSP Final Rule II in June 2015 II. Provisions of the Proposed Regulation (p. 12) A. Integrating Regional Factors When Resetting ACOs Benchmark Alternative Approaches to Reset the ACO s Benchmark: Proposals for Defining the ACO s Regional Service Area o Proposed definition: CMS opted to use the county as the geographic unit of measure to align with other CMS programs, provide greater definitional stability, and allow for increased transparency. Proposals for Establishing the Beneficiary Population Used to Determine Expenditures for an ACO s Regional Service Area o Proposed methodology: Assignable beneficiaries means all those who would be assigned to an ACO by virtue of their primary care utilization. Use all assignable beneficiaries, including ACO-assigned beneficiaries Regional service area is any county with at least one assigned beneficiary Calculation will include all assignable FFS beneficiaries in the relevant counties

4 Weight county-level expenditures by the ACO s proportion of assigned beneficiaries in the county Proposals for Determining Country FFS Expenditures o [Outlines expenditure calculation based on proposed additions and current policy] Proposals for Adjusting the Reset ACO Historical Benchmark to Reflect Regional FFS Expenditures o Trend the historical benchmark using regional FFS expenditures o Calculate a regional FFS adjustment to the historical benchmark using regional average expenditures Proposals for Transitioning to a Higher Weight in Calculating the Adjustment for Regional FFS Expenditures o Proposed: An ACO s historic spending would still be used for the first agreement period; in the second agreement period, CMS would adjust an ACO s benchmark to reflect a percentage of the difference between an ACO s historic costs and the regional FFS expenditures in the ACO s regional service area. First agreement period: 100% historical Second agreement period: 65% historical, 35% regional Third agreement period: 30% historical, 70% regional CMS wants the transition to be gradual but quick to ensure continued participation by ACOs in low cost/low growth areas and to watch for unintended consequences. o Rationale: The transition helps ACOs worried about being penalized for previous reductions in expenditures. o Alternative approach: 50% of regional expenditures starting in the third agreement period. Proposals for Regional Growth Rate as a Benchmark Trending Factor o Proposed: Replace the national trend factors used for trending an ACO s BY1 and BY2 expenditures to BY3 in calculating the rebased historical benchmark with regional trend factor. Proposals for Updating the Reset Benchmark During the Agreement Period o Proposed: For ACOs in their second or subsequent agreement period, annual updates to the rebased benchmark will include a regional growth update factor as opposed to the previously used national growth rate. Proposals for Calculation of Regional FFS Expenditures o Will continue to use completion factors based on national FFS claims. o Will likewise exclude payments for DSH and IME and uncompensated care, while including demonstration or pilot payments as per the 11/2011 rule. o Will continue to truncate expenses at the 99th percentile and risk adjust for severity of health status using the CMS-HCC model. o Will continue to calculate expenditures separately for each beneficiary population (ESRD, disabled, aged/dual eligible, aged/non-dual eligible.) o Will also use beneficiary person years to calculate expenditures for each Medicare enrollment type.

5 Proposals for Modifying the Calculation of National FFS Expenditures, Completion Factors, and Truncation Thresholds Based on Assignable Beneficiaries o Will use the same assignable beneficiaries population for calculation of any factors based on national FFS expenditures. Proposed Timing of Applicability of Revised Rebasing and Updating Methodology o Sets out a schedule for when the new revised rebasing methodology will apply to existing cohorts. Risk Adjustment and Coding Intensity Adjustment: (p.91) Proposals for Adjusting in Determining the Regional Adjustment to the ACO s Rebased Historical Benchmark and Seeking Comment on Approaches for Risk Adjusting Rebased Benchmarks o Context: In the June 2015 final rule, CMS indicated it would address the following issues related to risk adjustment Accounting for differences in risk scores between an ACO s beneficiaries and beneficiaries in their region Protecting CMS against excessive payments as ACOs improve documenting and coding beneficiary conditions o Proposed Change: In determining the regional adjustment to the ACO s rebased historical benchmark, CMS would adjust for differences in health status (using HCC scores) between an ACO and its regional service area in a given year. Would take effect for those rebasing on or after Jan 1, 2017 If finalized, CMS plans to rigorously monitor for the impact of coding initiatives on ACO benchmarks o Rationale: New approach would serve as a partial coding intensity adjustment but may not fully adjust for different coding intensity used by an ACO relative to its region. CMS is considering several alternatives that might be used in the future. B. Adjusting Benchmarks for Changes in ACO Participant (TIN) Composition Proposed Revisions Context: The current approach for adjusting benchmarks for changes in ACOs participant lists is a monumental operational burden for CMS. Additionally, ACOs submitting changes in their participant lists between performance years is highly common. o CMS had to adjust benchmarks to reflect participant changes for 74% of ACOs for PY 2014, and 78% of ACOs for PY Proposed Change: CMS would use a new approach that adjusts an ACO s historical benchmark based on the ACO s participant list from the most recent prior performance year and makes adjustments using expenditures from a single reference year (BY 3). o CMS would apply this streamlined approach program-wide even for ACOs in their first agreement period.

6 Rationale: This new approach would allow CMS to adjust the benchmark using only one BY, rather than three. Initial modeling suggests that this alternative methodology is highly accurate. C. Facilitating Transition to Performance-Based Risk Proposed Revisions Context: In response to the Dec 2014 proposed rule, CMS received comments suggesting ACOs should be able to extend 1 additional year under Track 1. CMS chose not to adopt these suggestions in the June rule. However, based on the recent renewal rate in Track 1, CMS chose to reconsider these issues. Proposed Change: CMS would add a new participation option that would allow eligible Track 1 ACOs to defer their entrance into a performance-based risk model (Track 2 or 3) by one year, extending their first agreement period under Track 1 for a fourth performance year. o For ACOs that are eligible to renew for a second agreement period under Track 1 but instead choose to move to risk two years earlier. o If finalized, would be available to Track 1 ACOs with 2014 start dates. D. Administrative Finality: Reopening Determinations of ACO Savings or Losses to Correct Financial Reconciliation Calculations, and a Conforming Change Circumstances for Reopening Initial Determinations and Final Agency Determinations of ACO Shared Savings or Shared Losses to Correct Financial Reconciliation Context: CMS is concerned that the current uncertainty regarding reopening a payment determination could introduce problems for ACOs, including discouraging ACOs from moving more quickly to two-sided models. Proposed Changes: CMS would have discretion to reopen a payment determination: o At any time in the case of fraud or similar fault o Within 4 years after the date of notifying the ACO of the initial determination of shared savings or shared losses if there is good cause Rationale: Limiting to 4 years would allow CMS enough time to evaluate errors while providing ACOs with more financial certainty. III. Collection of Information Requirements IV. Regulatory Impact Analysis (p. 138) A. Statement of Need B. Overall Impact C. Anticipated Effects (p. 140) Effects on the Medicare Program Quality and financial results from the first two performance years are within the range originally projected in the 2011 final rule.

7 Adjusting the benchmark by a percentage of the difference between the ACO s expenditures and the average expenditures for the service area strengthens the incentive for ACOs to invest in infrastructure and care redesign because it reduces the amount that positive performance leads to lower a lower benchmark in a subsequent period. Replacing the national trend with the regional trend will minimize the impact of localized exogenous factors such as a bad flu season or unusually large wage index adjustment. The expected net effect is an increase in program savings. Cases that result in lower benchmarks will prevent ACOs from receiving savings unrelated to efficiency improvements. Cases that result in higher programs will prevent ACOs from abandoning the program because factors beyond their control prevented them from receiving savings. Applying the regional trend as a percentage rather than a flat dollar amount is expected to improve the accuracy of updated benchmarks and removes the risk that the flat dollar adjustments could over inflate the benchmark, specifically for ACOs with low historical expenditures. There is potential that regional growth rates could be reduced by a prevalence of high performing ACOs in the region, but CMS believes this would not prevent an overall improvement in the incentive of ACOs to improve efficiency. Removing the savings earned in the previous period from the benchmark rebasing methodology would result in lower benchmarks, but would be at least partially offset by including a portion of the difference between the ACO s expenditures and service area expenditures. A regional benchmark adjustment based on a percentage of the difference between regional expenditures and historical expenditures is expected to have the effect of keeping ACOs enrolled in the program when downside risk is mandatory, because ACOs will have more certainty that prior savings will continue to benefit them. The proposed rule could lead to increased program costs through: o ACOs with low benchmarks relative to their region receiving savings unrelated to improved efficiency o ACOs with high benchmarks relative to their region may choose not to participate foregoing the opportunity to impact populations with the greatest opportunity for improvement o ACOs could make changes to their participant list that would shift their service area to increase assignment of healthy populations and reduce assignment of high cost beneficiaries Monte Carol Simulation CMS used a Monte Carlo simulation to predict the impact of proposed rule. The simulation consisted of 1,000 random trials. Based on the simulation the median expected impact of the proposed rule between 2017 and 2019 would be: o A $370 million reduction in claims o A $250 million increase in net shared savings pay o A $120 million decrease to net federal costs The 10 th percentile expected impact yields a $230 million increase to net federal costs The 90 th percentile expected impact yields a $490 million decrease to net federal costs

8 Effects of the Proposed Rule in Subsequent Agreement Periods (p. 153) Increasing the percentage of the adjustment to the benchmark based on regional FFS expenditures from 35% to 70% in the third period, is expected to prevent ACOs from dropping out of the program because it rewards ACOs for positive past performance rather than letting it make their benchmark more difficult to achieve. ACOs that are below their regional expenditures tend to stay below and those that are above tend to stay above, which creates some potential for bias. CMS has accounted for costs associated with such program selection bias in their calculations. The incentives offered by MACRA may reduce the risk of the incentive bias described above by offering more general incentives for participation in an ACO program. Further Considerations (pg. 158) The proposed rule would include prospective HCC risk adjustment that would account for differences in an ACO s population and the regional service area. The current method of updating risk adjustment relies on demographic risk ratios to prevent against changes to the benchmark related to shifts in coding intensity. In the proposed rule modifications to the HCC risk adjustment factor would be based in part on the difference between the ACO s HCC scores and the regional HCC scores. This is expected to mitigate the risk of coding intensity efforts because the benchmark years for subsequent performance periods would correspond to performance years from a prior period when coding intensity efforts were limited by relying on demographic risk ratios for continuously assigned beneficiaries. Effects on Beneficiaries (pg. 160) Beneficiaries will benefit from ACOs stay in the program and improving program participation Effects on Providers and Suppliers (pg. 161) The proposed rule is expected to provide additional incentives for providers to make investments to improve care coordination by more accurately rewarding improvements in efficiency. Effect on Small Entities (pg.162) Most ACOs would be considered small entities either because of their size or non-profit status. The proposed rule encourages participation by small entities in the following ways: o Making it possible for smaller entities that are risk averse may elect to defer their transition to a risk-based track by 1 year o Streamlining the process for adjusting the benchmark for changes in ACO participant composition o Providing an expected $250 million additional in shared savings payments between 2017 and 2019 Effect on Small Rural Hospitals (pg. 164) How rural hospitals will benefit will depend on their specific ACO.

9 D. Alternatives considered (pg. 165) CMS also considered a methodology that would create the new rebased benchmark by taking the old benchmark and updating it using the regional growth trend. A drawback of this approach was that it would not adjust for changes in provider composition in any way. E. Compliance with Requirements of Section 1899(i)(3)(B) of the Act (pg. 167) The projected impact of the proposed rule does not result in additional program expenditures and thus satisfies the requirement under section 1899(i)(3)(B) F. Accounting Statement and Table (pg. 170) Provides an accounting statement showing the impact of the proposed rule on federal net costs as a yearly amount for each of the 3 years between 2017 and o Primary estimate million o Minimum estimate 73.5 million o Maximum estimate million G. Publicly Available Data to Facilitate Modeling of Proposed Changes (pg. 173) CMS has released additional data along with this rule including a file that contains the average FFS expenditures for each county and the number of assigned beneficiaries for each ACO for each county. This data will improve the ability of ACOs to conduct analysis related to their benchmark. H. Conclusion (pg.176) A median savings of $120 million o As much as $490 million in savings o Or costs of $230 million