May 22nd, 2012
As Integrated Delivery Networks (IDNs) assess the financial implications of accountable care, six key variables deserve special attention. These variables are unique because they will shape both the accuracy of future financial projections and begin to set the broader strategy for the ACO.
Variable 1: The Halo Effect. As IDNs shift utilization patterns of their covered lives, experience has shown that these efforts tend to spill over and affect non-ACO enrollees as well. This “halo effect” is especially likely among patients within the same payor class (e.g. a Medicare ACO will likely affect utilization of all Medicare beneficiaries). The issue is that reduced utilization or less intense services benefits the ACO in the form of shared savings. However, reduced utilization among non-participants just lowers revenues.
Understanding the halo effect is critical to evaluations of accountable care’s effect on the bottom line. For most systems, the default assumption is 1:1, with every change in ACO utilization matched by the non-ACO population. IDNs may attempt to minimize the halo effect by segmenting care so that only fully participating physicians treat ACO patients, or by increasing the number of enrollees, thereby increasing the potential for shared savings.
Variable 2: Number of Enrollees. There are fixed costs in an ACO, and larger enrollment allows the IDN to spread per-member costs. The size of the enrollment pool also has an effect on the incentive structure. If the number of enrollees is not of sufficient size, or if enrollees are spread among a large number of primary care physicians, it is difficult to create incentives needed to alter provider behavior. For proof, one need look no further than the Physician Group Practice Demonstration (PGP), where participants with smaller enrollments (fewer than 19,000) achieved shared savings less frequently than those with larger populations. At the same time, a larger pool is also an advantage in Medicare, since having more enrollees lowers the Minimum Savings Rate (MSR) the ACO must achieve before it becomes eligible for shared savings payments. This is why most IDNs have projected a size of at least 30,000 members.
However, the number of potential enrollees is limited. An IDN should not expect a larger enrollment than its current market share unless it has explicit strategies for growth. Furthermore, as PGP shows, one can expect that about 25 percent of members will churn out of the ACO each year. Using PGP as guidance, it is safest to assume the highest number of ACO enrollees would be 75 percent of the current market share for a given payer.
Variable 3: The Cost to Build and Operate the ACO. The Government Accountability Office (GAO) and the American Hospital Association (AHA) released conflicting estimates for the cost of building and operating an ACO. A possible source of the discrepancy is simply a difference in point of view. The GAO used PGP to estimate costs of approximately $1.7 million for start up and the first year of operation.
However, many PGP participants already had much of the needed infrastructure in place before joining the program. In contrast, AHA represents a broader range of hospitals, including those that have little infrastructure in place, and their estimate (up to $12 million in development costs) is much higher.
For most hospitals, the cost appears to be somewhere in between. For the majority of ACO models seen to date, cost projections equal 45 – 70 percent of AHA’s estimates.
Variable 4: The Ability to “Bend” the Cost Curve. Based on lessons learned from PGP, ability to generate savings is highly correlated with an ability to manage care. In fact, of all the PGP participants that generated payments, all but one had experience owning an HMO. Although Medicare does not permit ACOs to steer beneficiaries, commercial or Medicaid plans may allow for greater provider management of care setting, and IDNs should examine opportunities to align beneficiaries’ interests with the ACO’s when negotiating with payors.
Despite this, savings are not a guarantee, and they do take time to be realized. In the PGP (which used a different benchmarking methodology than current ACOs), four of the 10 participants generated shared savings payments in three out of five years, including two that generated payments every year. However, the average improvement rate was just 1.2 percent, and others have suggested that the annual decline rate possible for inpatient utilization is around a maximum of 6 percent. Still other participants in PGP increased their utilization relative to the benchmarks.
Variable 5: Structural Considerations. Because ACOs are not expected to achieve break-even until years after forming, and the IRS has stated that physician owners cannot have losses subsidized by the IDN, most IDNs opt for super majority, if not complete, ownership of the ACO. If the IDN has physician owners, most are collecting funding at the outset, versus annual capital calls.
Once break-even is achieved, IDNs need a strategy for funding expenses and provider bonuses. Projections estimate the ACO needs to retain 40 percent for expenses, leaving 60 percent for bonuses and incentives. Of the money available for bonuses, most IDNs elect a 50/ 50 division between hospitals and physicians, although some have proposed 33 percent to the hospital, 33 percent to specialists and 33 percent to primary care providers.
Variable 6: The System’s Ability to Adapt. As previously noted, ACOs in many cases result in near-term losses. For this reason, it is important to estimate the IDN’s ability to adapt to its lower financial performance and develop a plan to restore margins.
For those that financially benefit in the ACO model, most overcome losses by filling beds emptied by Medicare (or Medicaid) patients with better paying commercial patients. Moving commercial market share is one of the most efficient ways to “win” with an ACO, and three to six points of share is often enough to overcome the impact of the ACO. Other options to overcome losses can include increasing managed care rates by about 2 – 5 percent, or reducing overhead expenses by 1 – 2 percent. However, neither of these strategies are easy to implement, and the reality is most systems will need to attempt a range of strategies to achieve even small wins.
Even with the best financial modeling, ACOs present a financial risk. For instance, one IDN with about $500 million in net revenue projected an annual decline in profitability of $2 million to $5 million with a small Medicare ACO. In this case, the IDN did not have confidence in their their ability to lower ACO infrastructure costs, control the halo effect, reduce operating expenses, increase commercial payment rates, or grow existing market share. These cases are not rare. To overcome them, IDNs need to honestly estimate and account for the six critical variables. Not only will they help in developing a more accurate projection of the financial implications, they will help focus the institution on the challenges ahead.Email This Post Print This Post
Don't miss the insightful policy recommendations and thought-provoking research findings published in Health Affairs.
to the #1 source of health policy research.