Rudy Douven, J. Michael McWilliams, Thomas G. McGuire
R. Douven, T. G. McGuire, and J. M. McWilliams, “Avoiding Unintended Incentives in ACO Payment Models,” Health Affairs, Jan. 2015 34(1):143–49.
One goal of the Medicare Shared Savings Program is to encourage health care providers to reduce spending on services through better coordination of the care delivered to beneficiaries across care settings. The Centers for Medicare and Medicaid Services (CMS), which administers the program, will share savings with accountable care organizations (ACOs) if they keep spending below a benchmark level over the three years of their contract. The benchmark is a weighted average of spending for the ACO’s attributed beneficiaries during the three years preceding the start of the contract.
The method CMS uses to calculate ACO spending benchmarks can potentially create perverse incentives. Spending for the three years before ACOs enter into or renew their contracts is weighted unequally: greater weight (0.6) is placed on the year immediately before the start of the contract, while lower weights are placed on the two previous years (0.1 and 0.3 in years 1 and 2).
Providers therefore have an incentive to increase spending in the year prior to entering their contract—by inflating their benchmark, they make it easier to earn shared savings. Under current rules, an ACO can expect to earn between $1.48 and $1.90 in shared savings for every dollar increase in spending in the year before the contract starts.
As CMS plans the next phase of the ACO program, it should consider improvements to ACOs’ incentives for achieving savings, the authors suggest. Such changes might include modifications to the benchmark weighting system, or new benchmarks that combine an ACO’s spending history with that of other Medicare providers.