Value Based Payment

Analysis: 2018 Medicare ACO results: Promising but not sufficient

November 5, 2019 2:02 pm
  • CMS recently released the results from the 2018 Medicare Shared Savings Program.
  • Overall, the program’s 548 ACOs reduced total costs relative to the benchmark by $1.7 billion, according to the Health Affairs blog.
  • CMS kept $739.4 million of the $1.7 billion in savings, according to Health Affairs.

CMS released the results from the 2018 Medicare Shared Savings Program (MSSP) on Sept. 30. The MSSP model directly targets the waste areas of failures of care delivery, failures of care coordination, overtreatment and to a certain extent, pricing failures as discussed in my Oct. 15 blog that cites a recent JAMA study.

On the surface, there’s cause for optimism. Overall, the program’s 548 ACOs reduced total costs relative to the benchmark by $1.7 billion, according to the Health Affairs blog ($169 PMPY1) in 2018. Of that amount, CMS kept $739.4 million, Health Affairs reported ($73 PMPY ,HFMA analysis) and “sharing” $966 million ($96 PMPY1) with the MSSP participating ACOs. (Note: The Health Affairs blog referenced in my blog today focuses solely on savings to CMS, which explains the difference in some of the numbers cited.)

MSSP ACOs and Medicare FFS spending

MSSP ACOs are “responsible” for the spending of approximately a quarter of all Medicare fee-for-service (FFS) beneficiaries (10 million out of 40 million). As a hypothetical, to give you a sense of the MSSP’s ability to reduce federal healthcare spending, if you extrapolate CMS’s portion of the savings ($73 PMPY) to all 40 million FFS beneficiaries, the reduction in outflows from the treasury (Part B, outpatient services) and Trust Fund (Part A, inpatient and skilled nursing facilities) would be approximately $3 billion. By any standard, that’s a lot of money. But in the context of Medicare FFS’s projected $500 billion2 in 2018 or $804 billion dollar federal deficits, it looks less impressive. It accounts for approximately .59% of 2018 projected Medicare FFS spending and .36% of the federal deficit.  

Better performance of both risk-bearing and physician-led ACOs?

Much has been made of the better performance of both risk-bearing and physician-led ACOs. In 2018, the 95 risk-bearing MSSPs on average generated $214 PMPY of savings overall ($96 PMPY of which goes to CMS), the 235 physician-led MSSPs (defined as “low revenue”) saved $353 overall ($180 PMPY going to CMS). On the surface, both seem to perform much better than the “average” ACO.

Assuming the performance is replicable across the country, if every Medicare FFS beneficiary was managed by a physician-led MSSP, it would yield $7.2 billion in savings (1.4% of projected 2018 Medicare spending 2  to the Treasury and Trust Fund.

However, that assumption may be flawed. Average benchmarks for both physician-led and risk-bearing MSSPs are higher (between $200 and $300 more) than the average MSSP, implying these MSSPs are in areas with greater utilization. And that difference in benchmark accounts for the discrepancy in performance (savings) on average. Said in plain English, groups of providers, regardless of who organizes the ACO, that are on the hook to pay CMS back are probably not participating in a risk-based model if their odds of success aren’t greatly improved by a market that has greater than average utilization (and therefore savings opportunities). Same thing goes for the MSSP in general for groups of independent physicians. Whereas health- system-based MSSPs are more likely looking at this more as an experiment that will inform future care transformation efforts.

ACO administrative costs

Speaking of expenses . . . MedPAC estimates that ACOs’ administrative costs, the expense of setting up and managing the ACO, including data analysis and reporting quality measures, approaches $200 PMPY. Some 343 MSSPs (representing 6.1 million beneficiaries) didn’t generate shared savings4. Assuming MedPAC’s estimate is correct, these ACOs’ infrastructure cost approached $1.2 billion dollars. Beyond taking a loss on infrastructure, there were no dollars to share with participating physicians for their efforts. The 205 MSSPs that generated “sharable savings” were able to keep on average $250 PMPY of the savings they generated. These ACOs had 3.9 million beneficiariesattributed to them. Net of infrastructure costs, they’ll have on average approximately $950k each to share with their participating physicians.  


From the government’s perspective, unless something significantly changes, the MSSP and other similar programs aren’t going to produce enough savings to right-size federal healthcare spending. That’s not to say that we should stop experimenting with value-based models like the MSSP if providers are willing to participate. Any reduction in spending coupled with an increase in quality improves value and is therefore worth the effort. But as a society, we also need to start thinking about additional ways to close the deficit between our means to pay for federal healthcare programs and what these programs cost.

From the MSSP participant side, assuming administrative cost estimates are in in the right ballpark, these costs are too high. One must wonder how much longer those MSSPs that are not generating sharable savings will continue to participate. And even for those who continue participating, after a year or two of not receiving a shared savings payment, how engaged are the community physicians in these models?

For those MSSPs generating shared savings, $950k net of administrative costs doesn’t go very far when the average number of primary care physicians in an ACO is 250. Some of these ACO administrative costs are relatively fixed in nature. Ideally, as more lives (e.g. individuals covered under ERISA plans, attributed individuals from MA plans) are moved into these models, economies of scale might take over. However, I’m somewhat skeptical that will occur in the near term without an unforeseen catalyst. The other option might be shared services across ACOs to spread the fixed costs out over more lives which is more or less what some of the analytics platform providers are doing today.  


1. HFMA analysis of 2018 MSSP PUF.

2. HFMA Analysis of 2018 CMS Office of the Actuary National Health Expenditure Projections, estimated Medicare Advantage spending removed) spending, (includes Part D).

3. Methodological note: I realize the issues with comparing the savings to the benchmark and a counterfactual using difference in difference analysis. When MedPAC has looked at MSSP spend using a counter-factual population, the total savings (including CMS’s cut) was between 1% and 2% , which for purposes of directionality isn’t radically different than my back of the envelop (or pivot table).

4. Eleven  repaid losses, 59 were upside only and generated a loss, and 273 reduced spending but did not exceed the MSR and therefore did not share savings, analysis of 2018 MSSP PUF file.  


googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text1' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text2' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text3' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text4' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text5' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text6' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-text7' ); } );
googletag.cmd.push( function () { googletag.display( 'hfma-gpt-leaderboard' ); } );