An article that appeared on Tuesday in the Health Affairs Blog was both fascinating and bracing. Authored by Nicolas Stine, Dave Chokshi, Janine Knudsen, Megan Cunningham, and Ross Wilson, “How America’s Largest Safety-Net Health System Built A High-Performing Medicare ACO” really is must-reading for anyone tracking the ongoing evolution of accountable care organization (ACO) development work specifically, and the shift towards value-based healthcare delivery and payment more generally.
As the article’s authors, senior leaders at the 11-hospital New York City Health & Hospitals Corporation (NYC Health + Hospitals), write, “The strategic implications of ACOs for safety-net providers and health systems are less clear. Some have questioned whether the safety-net health systems that serve a large portion of the country’s Medicare, Medicaid, and uninsured population are capable of the financial and structural investments necessary to succeed as ACOs. Others have identified unique strategies to help safety-net ACOs do well but predicted only modest and incremental success. Ultimately, it is not yet clear how the shift from volume to value will impact the financial viability of safety-net providers,” they write.
“Meanwhile,” they say, he need for a sustainable business model for safety-net systems could not be more acute. Along with rapid payment transformation and regulatory uncertainty affecting all health systems, safety-net organizations face particularly difficult financial circumstances. The simultaneous squeeze of decreasing subsidies and a payer mix tilted toward Medicaid and uninsured patients makes fee-for-service financing uniquely challenging. The future of safety-net systems, therefore, may be particularly dependent on their ability to adapt for success in the new value-based payment architecture. In this environment, the largest safety-net health system in the country, NYC Health + Hospitals, formed an ACO in 2012 to participate in the Medicare Shared Savings Program (SSP). Exhibit 1 offers an overview of the ACO at a glance,” they report.
The results of this initiative, to date? Impressive. As the authors write, “In its first four performance years, the NYC Health + Hospitals ACO (HHC ACO) has reduced costs by 4–12 percent annually compared to benchmark, while continually improving quality; it is the only ACO in New York State to achieve shared savings in all four Medicare SSP performance years. Overall costs to Medicare have been reduced by more than $31 million, generating shared savings incentive payments of nearly $14 million.” Honestly, those results would be impressive for any hospital-based patient care organization in participating in the MSSP. For a public hospital system? One could legitimately call them spectacular.
What’s more, data and analytics have been enormous drivers of NYC Health + Hospitals’ success. As the authors note, “Key insights from our initial analyses of ACO claims data were major drivers of how we set strategic priorities and drove performance. Among the most valuable tools provided to an ACO are the feeds of comprehensive claims data received monthly from Medicare. While we historically have had access to claims data for certain managed care subpopulations, the Medicare fee-for-service data are unique in that prices are regionally uniform and patients do not have network limitations. The ACO claims data therefore provide for more unfiltered analysis of how our patients interact with health care providers across settings and how our population’s use compares to local and national benchmarks.”
Now, the authors do concede that they actually started out with a few advantages, from an operational standpoint. “First, compared to our ACO cohort,” they concede, “we had low rates of certain types of high-cost elective outpatient use at baseline. For example, our expenditures per patient on magnetic resonance imaging (MRIs) are the lowest of all ACOs nationally. We hypothesize that this is driven by our role as the primary provider of many costly elective services (for example, advanced imaging and specialty care) for New York City’s large uninsured population. We have a limited supply of these resources and provide them according to clinical need and regardless of ability to pay.”
As a result, the NYC Health + Hospitals leaders note, “Our clinicians therefore become careful stewards of imaging and specialty care, and seek to promote access through avoiding unnecessary tests or referrals. Consequently,” they write, “we were able to devote less time and energy attempting to curb excessive or wasteful elective use compared to most other ACOs. Indeed, many health systems have historically developed fine-tuned financial machinery embedded throughout their operations to wring revenue out of the fee-for-service market. It was clear from our utilization data that we were already lean in this regard and could focus more on reducing costs in other areas.”
In addition, they note, “[O]ur costs were heavily concentrated in hospitalizations for patients with exacerbations of chronic medical and behavioral health conditions. Furthermore, among our most frequent users of acute care services, there was higher use of outside health systems in our dense urban setting. While most ACOs naturally focus on their high-utilizing population, it appeared that our performance would be particularly driven by our ability to better manage this fragmented care for a complex and chronically ill population.”
So, let’s be clear: for an ACO created and operated by a public hospital system, to be able to reduce costs by 4-12percent against its own benchmark, and also to be the only ACO in New York state to achieve shared savings in all four performance years so far? That’s huge. So, first of all, the NYC Health + Hospitals/HHC ACO leaders deserve a mountain of praise, just for achieving those metrics. What’s more, as we all know, public hospital management has historically not been especially innovative or enterprising. Certainly, that is changing, but the NYC Health + Hospitals leaders are obviously way ahead of their fellow public-hospital colleagues in many ways.
And, as quoted above, the NYC Health + Hospitals leaders have been deriving “key insights from our initial analyses of ACO claims data,” which have been able to “provide for more unfiltered analysis of how our patients interact with health care providers across settings and how our population’s use compares to local and national benchmarks.”
As the articles author’s note, “Our costs were heavily concentrated in hospitalizations for patients with exacerbations of chronic medical and behavioral health conditions. Furthermore, among our most frequent users of acute care services, there was higher use of outside health systems in our dense urban setting. While most ACOs naturally focus on their high-utilizing population,” they note, “it appeared that our performance would be particularly driven by our ability to better manage this fragmented care for a complex and chronically ill population.
These first two strategic insights are noteworthy in the context of recent debate around whether ACOs seeking to reduce costs should focus more on reducing low-value services for all patients as opposed to targeting those likely to be particularly high cost. In our case, the data suggested there was less opportunity for reducing systemic waste but a higher concentration of excess use among high-risk patients than other ACOs. We suspect this may be a more common dynamic for other safety-net ACOs. Indeed, the bulk of our savings have been achieved through reduced hospitalizations for patients with chronic conditions.”
In other words, for safety-net hospitals, there is great opportunity in zeroing in on targeting high-cost patients. The NYC Health + Hospitals leaders also came to realize that, as they put it in their article, “[W]e would need to focus on identifying high-risk patients instead of just high utilizers. We reviewed multiple vendor products for risk stratification but had concerns about their predictive validity in a mostly dual-eligible safety-net population. Ultimately, we developed an in-house risk scoring system that could be internally validated using our historical claims data.”
Meanwhile, much innovation in the ACO world still relies on mission-driven personal leadership, as the authors note in their article. “Our ACO infrastructure, with site-specific physician champions and administrative leads,” they write, “attempts to encourage local creativity and engagement. Local medical directors appoint an ACO clinical lead at each of our 20 primary care centers. Each ACO lead engages with multidisciplinary teams embedded in each site’s patient-centered medical home (PCMH), each of which has 2014 NCQA level III accreditation. We found that each local leadership team functioned best under a different mixture of central direction and flexibility, and prioritized finding the right balance with each.”
There are so many lessons here, around the agile and strategic use of data, some potential secrets to effective ACO leadership, and the effective deployment of resources of all types. But the bottom line is this: if the New York City Health & Hospitals Corporation’s leaders can do as well as they have in the MSSP program, that means that their achievement is clearly possible for many types of hospital-based organizations, including those working with resource scarcity and other limitations.
And that means that the “Debbie Downers” in U.S. healthcare who keep pointing fingers every single time that CMS (Centers for Medicare and Medicaid Services) senior officials release new data on the various ACO programs, and we see that some ACOs are doing less well than others, need to sit back a bit and consider how early we are in the forward evolution of ACO development—and how much potential there really is, going forward. Let’s all keep our eyes open and celebrate breakthroughs like this one, and be aware that, in the accountable care and value-based healthcare phenomenon’s “journey of a thousand miles,” we’re still just a few miles from the start of the journey. And in that context, this is a hope-inducing report, indeed.