We often think of an ACO as a way for providers to coordinate care, improve quality and outcomes for patients, and reduce costs for payers. This is the value construct for Medicare ACOs and most commercial ACO-like initiatives. But applying the ACO framework to Medicaid runs head on into a stubborn challenge: the disproportionate impact of socioeconomic factors on health in the Medicaid population.
Poor health outcomes in low-income populations are often exacerbated by unstable employment and housing, transportation difficulties, and lack of access to nutritious food. Add in a high prevalence of mental illness and substance abuse, then layer it all over a mélange of cultural, language, and health literacy barriers. It doesn’t take long to realize that the doorway to good health lies not only at the doctor’s office but through a maze of unmet social needs.
“We don’t have the payment structure in Medicaid in any state yet to adequately address those needs,” says Gregory Paulson, executive director of the Trenton Health Team (THT), which operates a New Jersey Medicaid ACO demonstration program. “Affecting quality, cost, and satisfaction for Medicaid beneficiaries is going to involve creative use of community resources.”
That realization is driving some true innovation at the state level. According to the not-for-profit Center for Health Care Strategies (CHCS), nine states have fledgling Medicaid ACO demonstrations serving a total of 2 million beneficiaries (see map below). While each state defines quality metrics and care-management requirements, Medicaid ACOs have wide latitude to design programs that serve the unique needs of their populations.
Source: Center for Health Care Strategies, March 2016
Recent reports by CHCS and the National Association of Medicaid Directors have identified promising components of success. These include payment models that directly or indirectly encourage coordination with non–health care services; highly targeted data analysis; and shared savings for federally qualified health centers (FQHCs). These attributes, along with a fourth—a belief in the power of collaboration over competition—characterize two Medicaid ACO demonstration projects that provide real-time lessons in serving complex populations.
The Federally Qualified Health Center Urban Health Network, or FUHN—pronounced “fun”—is an alliance of 10 FQHCs in the Twin Cities. In 2013, FUHN became one of the first of six Medicaid ACOs in Minnesota’s Integrated Health Partnerships demonstration. “It was pretty gutsy, and the fact that we didn’t have a large integrated health care delivery system took people by surprise,” says Jaeson Fournier, FUHN’s board treasurer and immediate past chairman. Today, with over 33,000 lives attributed to it, FUHN is one of only two all-FQHC Medicaid ACOs in the country.
FUHN’s leaders strongly believed that an emphasis on primary care—an FQHC’s sweet spot—was the key to success in a value-based system.
“We viewed it as an opportunity for FQHCs to have a seat at the table,” says Fournier. “We kidded around that the choice was to be under the bus, ride on the bus, or drive the bus. We decided we were best positioned to drive the bus.”
Care coordination makes sure that “patients who are moderately or persistently mentally ill are connecting with the medications they need,” says Jaeson Fournier of the Federally Qualified Health Center Urban Health Network (FUHN) in the Twin Cities.
The road to success on which the FUHN bus travels is paved with intensive care coordination and patient activation. Care coordinators are more accessible to patients than ever before and reach out to high emergency department utilizers who haven’t been seen in the clinic for a while. FUHN-attributed patients are drawn from the diverse communities within the Twin Cities, including immigrants—many of whom find navigating the U.S. health care system challenging—so referrals to specialty care can be high-intensity events. At one FQHC, West Side Community Health Services, where Fournier is CEO, a staff member may personally accompany a patient to a specialist’s office if necessary, then follow up with the patient to make sure he is getting needed services.
Many of the patients attributed to FUHN’s ACO have mental health or substance abuse needs, or both. “That’s one of the most significant areas where intense care coordination has been helpful, in terms of making sure that patients who are moderately or persistently mentally ill are connecting with the medications they need and routinely coming to follow-up care,” Fournier says.
New Jersey’s Medicaid ACO demonstration is unique in its provider–community approach. To participate, ACOs must have representatives of primary care and specialty providers, community organizations, and residents on their boards. “This project is a way to engage the community in the process of improving care delivery to Medicaid beneficiaries,” says THT’s Paulson.
THT has been at this for a while. Founded in 2006 in response to the closure of one of Trenton’s hospitals, THT was a way for the city’s health department, its FQHC, and its two remaining hospitals to assess community health needs and align efforts. Nine years later, THT launched one of three ACOs in New Jersey’s demonstration project. In New Jersey, Medicaid ACOs are accountable for the care of all beneficiaries in a particular region. In Trenton, that means that THT is responsible for 47,000 lives—more than half of the people living in a city of 84,000 people.
Servicing Medicaid beneficiaries can be tricky, says Gregory Paulson of the Trenton Health Team. “The health care system wants them to be healthier, but often that’s not a person’s most important need.”
The ACO’s activities build on THT’s community health assessment, developed from hospital utilization data and 300 one-on-one interviews conducted in the community. The biggest identified need is what Paulson also calls the toughest—behavioral health and substance abuse services. “The needs of patients who suffer from alcohol dependence may not be the same as those with opiate dependence,” he says. “So, we need to be very precise about understanding the problem, then design interventions to meet the needs of those subgroups.”
Those interventions may be inside or outside the health care system. For instance, taking St. Francis Medical Center’s diabetes education program out of the hospital and into churches, where THT delivers an evidence-based curriculum about healthy activity and nutrition. Or dispatching THT’s Clinical Care Coordination team to the streets to listen to patients and guide them to community resources, which can help reduce unnecessary trips to the emergency department. “The health care system wants them to be healthier, but often that’s not a person’s most important need,” notes Paulson. “Maybe they need a place to sleep tonight. Or food. Or to avoid violence.”
The story of one client whom Paulson refers to as “Lady B” speaks volumes about the linkage between overutilization of health care resources and underutilization of community services. Lady B had visited the ER 465 times in a 12-month period, says Paulson. “She was homeless, she had alcohol dependence, and we worked with social services providers and got her housed. Her ER visits dropped to 12 in the following year.”
Real-time data collection and analytics are critical to any ACO, but maybe more so in models where intensive care management is a staple. Both FUHN and THT rely on robust data capabilities for surveillance and to inform case management and outreach efforts.
Take emergency department use. FUHN identifies utilization patterns and leverages staff capacity accordingly, using care coordinators to contact patients while there is a window of opportunity to bring them in for follow-up care, says Fournier, “and, more importantly, to engage them about how they can and should be consuming their health care resources.” It’s part of a larger strategy FUHN developed with its IT partner, Optum, called identification and stratification. The strategy hinges on a scoring system that assesses diagnostic criteria and identifies patients with the greatest opportunity for engagement—and thus, greatest potential savings.
As simple as that may sound, FUHN had to figure out how to make raw data useful. “It’s been our experience that the amount of data coming to a provider organization is analogous to someone directing a fire hose at you. The volume is overwhelming,” says Fournier.
In New Jersey, THT had the infrastructure for effective use of data in place for 18 months before the ACO went live. THT built the Trenton Health Information Exchange (THIE), which connects the vast majority of Trenton-area providers, including labs, behavioral health, and the city’s FQHC, all of whom contribute patient data to the THIE, regardless of payer. As a result, the THIE includes claims and service information on more than 300,000 people. The scope of the THIE’s data repository allows Paulson’s organization to focus on individuals with the greatest need—such as a block of members with diabetes whose HbA1c is not just high but off the charts. Getting to this point is something Medicaid ACOs struggle with nationwide. Paulson says THT has been fortunate to have received a significant amount of private funding, including money from the Nicholson Foundation in Newark.
Fournier acknowledges that when it comes to data-driven coordination and patient activation, significant investment is needed on the front end to yield savings on the back end. “Care coordination is this nebulous thing, and if you’re not very intentional and thoughtful about how you deploy your resources, it can get subsumed in overall patient engagement, so you have to be very purposeful. That was a hard learning for us.”
While FUHN and THT have gained attention for shouldering ACO cost-and-quality burdens in a Medicaid setting, ultimately each will be judged by financial results and their ability to generate long-term savings.
So far, so good for FUHN, which uses a shared savings model with upside-only risk, similar to Medicare Shared Savings Program Track 1 ACOs. Over its first three years, FUHN has saved the state almost $17 million, getting back half of that for meeting cost and quality thresholds. But Fournier worries about being able to sustain this pace.
“One of the things that compromises the long-term viability of an organization like ours is that if you are successful, the Medicaid agency may reset your baseline. So, you end up achieving the goal of improving care and lowering the overall total cost of care, and now that we’ve done that for three years, the state has reset our baseline,” says Fournier.
“What the state fails to recognize is that while we’ve achieved those savings, there has been a lot of work that has allowed that to happen,” he adds, meaning that patients will still need intensive engagement. State Medicaid officials are re-examining this issue, as well as that of risk adjustment for social determinants of health, before releasing the next iteration of its demonstration project.
The Trenton ACO is still less than 18 months old. Looking initially at patient outcomes and satisfaction measures, THT is still learning which small interventions—such as those that lead to health improvements for people with diabetes—work best and can be scaled up. ACO-wide outcomes are a few years off.
Gauging the effectiveness of those interventions also means understanding the concept of regression to the mean. “The concept of trying to isolate the impact of our intervention from things that would have resolved anyway makes this analysis really challenging,” says Paulson. “More and more, it’s part of the national conversation about utilization.”