In 2012, the Centers for Medicare and Medicaid Services (CMS) launched the Medicare Shared Savings Program (MSSP) with the goal of reducing medical costs, enhancing care quality, and improving the health of traditional Medicare beneficiaries. The program allows accountable care organizations (ACOs) to share in generated savings (or losses as applicable). Over the course of the program's roughly 10 years of operation, ACOs have collectively achieved billions in gross savings.1 However, one of the key risk-mitigation features of the program also has collectively eliminated significant savings that some ACOs would have shared: the MSR/(MLR).
MSR/(MLR) background
The MSR/(MLR) that is applicable to an ACO depends on the participation track it has selected. MSSP offers different tracks with a series of escalating levels of risk (i.e., shared losses) and reward (i.e., shared savings), depending on an ACO’s risk appetite. An ACO must generate savings (or losses) above (or below) the MSR/(MLR) to share in gross savings (or losses).
For one-sided tracks (i.e., BASIC Level A or B), the MSR is calculated on a sliding scale based on the number of assigned beneficiaries to the ACO and ranges from 2.0% to 12.2%. For all other MSSP tracks, the symmetrical MSR/(MLR) is selected by the ACO and can be either (a) size-based (as described above for BASIC Levels A and B), or (b) 0.0% to 2.0% (in 0.5% increments).
It is important to note that once an ACO in a two-sided track selects an MSR/(MLR), it is locked in for the remainder of the agreement period (currently five years) and can only be modified once the agreement period ends or if the ACO exits and reenters the program (i.e., rebases). If an upside-only ACO transitions to an MSSP track with downside risk during the agreement period, it has the option at that point to select an MSR/(MLR), which then applies for the remainder of the agreement period.
Figure 1 illustrates how the MSR/(MLR) functions for ACOs in both one-sided (upside-only) and two-sided (upside and downside) tracks.
Figure 1: Illustrative example of MSR/(MLR)
MSR/(MLR) mechanics
To determine whether an ACO meets the MSR/(MLR) and is eligible to share in savings (or losses), the ACO’s gross savings amount is first calculated. The gross savings is the difference between the benchmark expenditures and the performance year (PY) expenditures.
Gross savings = benchmark expenditures – PY expenditures
The gross savings rate is then measured as the ratio of gross savings and benchmark expenditures.
Gross savings % = gross savings / benchmark expenditures
This rate is compared to the applicable MSR/(MLR). If the gross savings rate falls outside of the MSR/(MLR) corridor, then the ACO is eligible to share in any savings or losses generated, not just savings in excess of the MSR/(MLR). If the gross savings rate is within this corridor, however, then the ACO does not share in any savings or losses for the performance year.
For example, an ACO is participating in MSSP track BASIC Level A and has an MSR/(MLR) of 2.5%. If the ACO has $12.5 million in benchmark expenditures and has generated $0.2 million in total gross savings, then this ACO has a gross savings rate of 1.6% ($0.2 million / $12.5 million) which is less than the MSR/(MLR). This hypothetical ACO would not share in any savings in this situation (i.e., shared savings is $0).
As a second example, another ACO is participating in the MSSP ENHANCED track and has selected an MSR/(MLR) of 1.5% for the agreement period. If this second ACO has total benchmark expenditures of $50.0 million and has generated a gross loss of -$1.0 million (-2.0% of benchmark), then because this falls outside of the MSR/(MLR) corridor the ACO will share in the total gross loss. The amount of gross loss that is shared depends on the track parameters. For example, under the ENHANCED track, the loss sharing rate is 40%; assuming the ACO meets the quality threshold, the total shared loss for the performance year will be -$0.4 million.
MSR/(MLR) options
For ACOs in either BASIC Level A or B, or that have selected a size-based MSR/(MLR) for the agreement period, the MSR/(MLR) will be based on the number of beneficiaries assigned to the ACO for the performance year (see the sliding scale in Figure 2). The MSR/(MLR) is calculated by linearly interpolating between the MSR/(MLR) values of each beneficiary range, based on the number of assigned beneficiaries in the ACO.
Figure 2: Size-based MSR/(MLR) calculation
Based on the CMS public use files (PUFs), Figure 3 illustrates the distribution of two-sided ACOs and their MSR/(MLR) selections in PY2022,2 including the proportion of ACOs that generated savings or losses within each MSR/(MLR) selection. These ACOs all had the option to select their MSR/(MLR) and we see that 17% chose a size-based MSR/(MLR). Interestingly, we see that few ACOs selected an MSR/(MLR) at a half-percentage increment (i.e., 0.5% or 1.5%), showing that the choice of MSR/(MLR) does not seem to be quantitatively driven for many ACOs (in which case we would expect to see a more uniform distribution across the different options).
Please note that the shared savings/(loss) values throughout this paper were calculated from the CMS PUFs for two-sided ACOs only, and ignore the impacts of the COVID-19 public health emergency (PHE) loss forgiveness as well as the sequestration removal during the PHE. This was done to ensure that comparisons between years are consistent.
Figure 3: Distribution of selected MSR/(MLR) for two-sided ACOs (PY2022)
Figure 4: Distribution of selected MSR/(MLR) for two-sided ACOs (PY2020-PY2022)
As shown in Figure 4, there are a few interesting observations about MSR/(MLR) selection in the past three performance years:
- The proportion of ACOs selecting the higher MSR/(MLR) of 2.0% has decreased sharply over recent years.
- This seems nearly entirely offset by the increase in ACOs selecting a lower MSR/(MLR)—in particular 0.0% and 1.0%.
- The selection of 0.5%, 1.5%, and size-based MSR/(MLR) is relatively constant over the past few years.
In short, it seems that ACOs are trending toward lower MSR/(MLR) selection already, perhaps implying that a larger proportion of ACOs is electing to forgo limited downside protection in favor of expected generated savings.
MSR/(MLR) impact
An ACO’s MSR/(MLR) can protect against losses as well as eliminate potential savings, as seen in Figure 5. Each dot on this graph represents an ACO’s performance in the PY2020-PY2022 period (i.e., a single ACO could be represented with three dots if participating in all three performance years), with colors indicating whether the ACO generated shared savings above the MSR (black), shared savings within the MSR (gray), shared losses within the MLR (yellow), or shared losses outside the MLR (red). All of these are also broken out by MSR/(MLR) selection. The table in Figure 6 shows the counts of ACOs in each of these categories, and the table in Figure 7 shows the aggregate shared savings/(losses) in each of these categories.
Figure 5: MSR/(MLR) impact on two-sided ACO results (PY2020-PY2022)
Figure 6: Count of two-sided ACOs impacted by selected MSR/(MLR): PY2020-PY2022
| Savings/Loss Category | 0.0% | 0.5% | 1.0% | 1.5% | 2.0% | Size-based | Totals |
|---|---|---|---|---|---|---|---|
| Shared Savings | 123 | 29 | 153 | 21 | 132 | 76 | 534 |
| In MSR Corridor | 0 | 2 | 7 | 10 | 25 | 30 | 74 |
| In MLR Corridor | 0 | 1 | 6 | 2 | 15 | 6 | 30 |
| Shared Losses | 10 | 5 | 9 | 3 | 2 | 2 | 31 |
Figure 7: Two-sided ACO shared savings impact of MSR/(MLR) selection ($M): PY2020-PY2022
| Savings/Loss Category | 0.0% | 0.5% | 1.0% | 1.5% | 2.0% | Size-based | Totals |
|---|---|---|---|---|---|---|---|
| Shared Savings | $1,084 | $536 | $1,279 | $109 | $859 | $2,856 | $6,722 |
| In MSR Corridor | $0 | $1 | $7 | $5 | $90 | $263 | $365 |
| In MLR Corridor | $0 | ($0) | ($5) | ($1) | ($30) | ($113) | ($149) |
| Shared Losses | ($4) | ($11) | ($26) | ($2) | ($2) | ($5) | ($49) |
Figures 6 and 7 show more interesting observations relating to the impact of the MSR/(MLR):
- There are more ACOs with savings eliminated by the MSR ($365 million) than ACOs with losses forgiven by the MLR ($150 million).
- This is especially true for the higher MSR/(MLR) selections (2.0% and size-based).
- As a result, there are more shared savings left on the table with the MSR than shared losses protected by the MLR.
- However, the higher MLRs (e.g., 2.0%) did protect a number of ACOs from sharing in losses.
- Additionally, the average loss per ACO for those protected by the MLR was larger than for those outside the MLR.
MSR/(MLR) choice
An ACO’s choice of MSR/(MLR) is very important, especially because this assumption is fixed throughout the entire agreement period. With the 2023 Medicare Physician Fee Schedule (MPFS) final rule changes, low-revenue ACOs in the BASIC tracks are now eligible to receive half of the shared savings forgone in the event that the MSR is not met (but are still not at risk for any losses below the MLR).3 This is an additional consideration for ACOs that qualify as low-revenue, and may make a nonzero MSR/(MLR) a more attractive choice.
When choosing an MSR/(MLR), an ACO should consider:
- Its risk appetite and financial goals
- Its size, i.e., what would the MSR/(MLR) be if the size-based option is chosen
- Its historical performance (and its participant providers)
- Known changes to ACO dynamics that could impact performance (e.g., large participant list shift, new geographic service area)
Based on PY2020-PY2022 results, if a 0.0% MSR/(MLR) had been selected universally, more savings would have been shared than lost. However, a 0.0% MSR/(MLR) also increases an ACO’s exposure to downside risk. The choice of MSR/(MLR) can be supplemented by actuarial analyses and modeling (e.g., with Milliman’s ACO Builder®4) to understand the likelihood of savings or loss in future performance years.
Conclusion
Within MSSP, the MSR/(MLR) acts as a risk-mitigation mechanism, and for two-sided ACOs this assumption is a choice. Choosing a higher MSR/(MLR), e.g., 2.0%, has been shown to historically eliminate more savings than protect against losses. However, choosing a higher MSR/(MLR), e.g., 2.0%, has been shown, historically, to eliminate more savings than it has protected against losses. There are several factors ACOs should consider when choosing their MSR/(MLR), including robust financial modeling given their specific participant providers and risk appetite.
1 See https://www.naacos.com/medicare-aco-program-results--2022-edition.
2 CMS (2022). Performance Year Financial and Quality Results. Retrieved January 30, 2024, from https://data.cms.gov/medicare-shared-savings-program/performance-year-financial-and-quality-results.
3 For more information see: https://www.milliman.com/en/insight/impactful-mssp-rule-changes-2023-mpfs.
4 See https://www.milliman.com/en/products/milliman-aco-builder.