Eighteen of the 114 accountable care organizations (ACOs) that joined during the first year of the Medicare Shared Savings Program are part of the PACT Population Health Collaborative, a program established in 2010 by Premier, Inc., to help members of the hospital and health system alliance succeed at value-based contracting. The Premier collaborative focuses extensively on guiding ACOs (and prospective ACOs) in their efforts to restructure care delivery models and invest in technology systems to manage populations of patients and reduce health care spending. The collaborative also helps ACOs interpret the complex regulations underpinning the program and the shifting formulas Medicare uses to determine whether the ACOs have met their quality and cost targets. Quality Matters asked Wes Champion, senior vice president of Premier Performance Partners, Premier's consulting arm, about the successes and challenges of the program thus far.
Quality Matters: In the first year of the Medicare Shared Savings Program (MSSP), roughly half of the ACOs saved money for Medicare, but only 29 percent saved enough to qualify for a bonus. Did the ones that were successful have any common structural or operational characteristics?
Champion: We've seen in both the MSSP and commercial ACO contracts, the ones that tend to succeed have established the foundation for managing population health using, among other things, patient-centered medical homes and care management programs. They also tend to have a high degree of clinical integration and sophistication with health information technology and informatics, which are needed to identify high-risk patients and create workflows that enable providers to efficiently meet patient needs. Having previous experience managing risk, for instance through prior ownership of a provider-sponsored health plan, also helps. In the commercial contracts, we've noticed that successful ACOs often have a strong payer–provider partnership that is built on transparency, shared value propositions, aligned incentives, and shared risk.
Quality Matters: It's interesting that two of the four MSSP sites electing to take upside and downside risk (exposing them to penalties if they failed to meet savings targets in the first year) did better in the first year. Did they take a different or more aggressive approach to cost containment and quality improvement?
Champion: The short answer is yes. If a system is taking downside risk, they are usually much more aggressive in their approach to population health initiatives. But it's more of an issue of pace than a difference in the types of initiatives they've implemented. They seem to be working faster, which for some may be attributable to having previously invested the infrastructure needed for population health management.
Quality Matters: What were the biggest challenges for the ones that didn't earn a bonus?
Champion: Patient turnover was a big problem because of the MSSP's attribution method. Our members tell us that only about 75 percent of their ACO population remained the same from one year to the next—so the ACOs aren't getting credit for the work they've done with some of their patients. One way of addressing this is to allow beneficiaries to actively align with an ACO if they choose to. This, combined with the existing attribution process, would stabilize at least a portion of the population. And that in turn would help stabilize the spending targets.
Another issue may have been the risk adjustment used to calculate spending targets. Currently, CMS does not allow an increase in the acuity of a population to increase an ACO's payment, but it does allow it to reduce payment. It's lopsided and unfair. It should either allow a positive adjustment, or remove the negative adjustment.
Quality Matters: There also seems to be a lot of confusion about how the Centers for Medicare and Medicaid Services (CMS) is calculating the initial benchmarks for spending and savings targets. To what extent were those based on regional or national numbers?
Champion: The formula is very complicated but in general it works this way: CMS uses three years of historical data for the beneficiaries who would have been assigned to the ACO had it been operational at that time. To establish the baseline for the first year, CMS takes that number and adjusts it using the national percent change in spending. Then CMS adds an absolute dollar change in per capita spending, based on national numbers, to predict what spending would be the following year. That's the spending target. One of the challenges is this doesn't happen until after the fact because CMS must finalize the retrospective beneficiary attribution and risk adjust the numbers. So, to answer the question, ACO-specific data are used as the foundation for spending targets, but it's national numbers that are used to project spending rates going forward.
Quality Matters: Does that methodology favor ACOs operating in high-cost or low-cost regions?
Champion: This kind of bias was a major concern in the development of the program. CMS addresses it by using a benchmark defined by regional spending and adjusting it based on trends in the national rate of growth in Medicare costs. As a result, ACOs operating in low-cost regions benefit by CMS adjusting the baseline and spending targets using the national trend, and the high-cost areas benefit by having more cost to be able to pull out of the system. This approach is designed to incent both regions to participate.
Quality Matters: We've heard some concerns that while benchmarks for spending are being reduced each year, the cost of the infrastructure needed to lower spending (e.g., care coordinators, information technology systems) remains fixed. How do ACOs work around this?
Champion: This is a big issue when you consider that one of the major hurdles to joining the program is the cost of entry, and that includes this kind of infrastructure. ACOs that were not able to recover those capital investments in the first year in the program and/or reduce them will have difficulty taking on downside risk in their second contract. This is an area where CMS could help by not requiring downside risk in the second contract. For instance, CMS could make track two (which requires the ACOs to take upside and downside risk) voluntary, not compulsory for fourth-year ACOs, and allow more than one option for the level of risk and reward.
Quality Matters: What other changes might be needed?
Champion: We've seen CMS make great strides in releasing more data to support ACO operations, but there are still concerns about the data's accuracy, timeliness, and comprehensiveness. We think CMS should invest additional resources in this key area to help providers identify gaps in care and coordination, and improve beneficiary outcomes.
Also, there are six quality benchmarks that require essentially perfect performance for the ACOs to get full shared savings. When you look at comparable benchmarks in HEDIS [Healthcare Effectiveness Data and Information Set], they are much lower. CMS should revisit how they set the benchmarks to make sure ACOs have a reasonable chance of meeting the goals.
Quality Matters: When you look at the MSSP sites you're working with, are you seeing new or interesting approaches to primary care and care management?
Champion: We've seen some of them send high-risk patients to dedicated care centers that take an interdisciplinary approach to outpatient care, which has led to better treatment outcomes and higher satisfaction among patients, staff, and physicians. We're also seeing ACOs increase health coaching and education services, and intensify their focus on mental health and life management issues. And some hospital-based ACOs are investing more heavily in outpatient services.
Quality Matters: How about in the use of information technology and data analysis for population segmentation, risk prediction, and performance improvement?
Champion: Some are using these techniques to proactively develop individualized care plans and automatically generate and distribute prioritized lists about gaps in care. At Premier, we're trying to help providers extend their population health efforts to their local communities. We've partnered with the University of North Carolina (UNC) at Charlotte to help health systems use similar data analytics techniques to target, set, and support their community's health improvement priorities so these communities can appropriately allocate resources. The partnership is leveraging UNC Charlotte's academic expertise and unique community health technology, which can help providers better identify, understand, and predict disparities in care by integrating big data from disparate sources.
Quality Matters: Do you think the ACO model as currently structured is worth scaling and can be scaled to the entire U.S.? If so, what might it take to increase uptake and spread?
Champion: Definitely. The accountable care model is showing great promise in both the private markets and with CMS. Both the MSSP and ACO Pioneer model were able to generate savings for the Medicare Trust Fund and financial rewards for some of the participating ACOs, which is critical for scaling the model. We're also seeing these types of models being utilized in Medicaid plans in numerous states across the country. If they continue to show value, both financially and in terms of improving quality, then the movement toward population health and accountable care will continue. We take it as a promising sign that some providers are building the health information technology infrastructure and core capabilities essential to ACO formation, whether they're in an ACO or not. This implies a new wave of ACO participants are likely to emerge in future years as these partnerships mature.
Quality Matters: In what way do collaboratives help with uptake?
Champion: I think collaboratives help reinforce the importance of taking a data-driven approach to improvement. We encourage the participants in ours to measure their current level of performance with defined, agreed-upon metrics; report those measures transparently; share best and worst practices and execute new strategies collaboratively; and then restart the process.
I think it also helps to identify common problems. From what we've seen, there is a need to remove some of the regulatory barriers and complexity associated with the program. In anticipation of the pending regulation later this year, Premier submitted a comment letter with suggested improvements to the program related to its approach to benchmarking, risk adjustment, and attribution of beneficiaries, among other topics. We hope it helps.
