UnityPoint Health's foray into CMS' NextGen ACO model is helping the health system transition its entire book of business to a value-based model.
UnityPoint Health's participation in the initial rollout of CMS' Next Generation ACO Model helped them succeed beyond their expectations in their transition to reimbursement risk.
In performance year 1 (2016), UnityPoint's 85,000-member Iowa Health ACO earned more than $10.5 million in shared savings.
Like most health systems, UnityPoint still gets more revenue from fee-for-service than what Vice President and Chief Analytics Officer Betsy McVay calls fee-for-value, but she says the NextGen ACO model offered the health system an opportunity to transition to a value-based payment model.
"We know we need to be successful in risk models and, ultimately, full capitation, to provide the best quality care," she says. "It’s absolutely necessary."
Though most top leaders concede that not only is value-based care necessary, it's also ethically better to make that transition, many executive leaders at hospitals and health systems struggle with how quickly and fully to embrace the commitment, given that the two reimbursement systems are diametrically opposed, and how thin healthcare margins can be.
UnityPoint leaders say they believe navigating those two payment models can be done effectively and without absorbing huge losses in the process. Here's what they did to achieve savings.
Step 1: Target unnecessary costs by leveraging IT
Aric Sharp, the vice president of UnityPoint Accountable Care, which has responsibility over the NextGen ACO as well as several other commercial value-based contracts for the company, says its 15-person ACO division is critical in targeting interventions that can save costs through better patient management and elimination of variation.
The ACO team manages what Sharp calls a "broad-matrixed team effort" to move the organization toward value. UnityPoint's network consists of about 7,500 providers (approximately 2,000 are employed by UnityPoint) and 40-plus hospitals.
The organization's value-based contracts cover about 250,000 patients, a little more than a quarter of the patients in the entire health system, including about 85,000 in the NextGen ACO.
Sharp says UnityPoint wanted to move to the NextGen ACO partly to increase the patient population under risk contracting, but more broadly to force cultural change around patient care. Also, with its characteristic prospective as opposed to retrospective patient attribution, NextGen's architecture gets closer to where analytics can make a difference on cost and quality.
"You don't get the churn that you experience in the retroactive models," says Sharp, whose background is in multispecialty physician group management.
"You have to be data-driven in this work," he says. "One of the key steps is having the ability to mesh your claims data with your EHR data, which gives you a more robust picture of what’s going on in the population you’re accountable for."
That said, getting the data to mesh is difficult, but once you can make inferences from the combined data sources, you can see the opportunities for improvement, he says, in certain chronic diseases, procedures, and in identifying and interacting with rising risk and high-risk patients.
"You start to realize what you need to work on because you can’t do it all," he says.
In the initial years, UnityPoint's ACO focused more on quality metrics and identifying gaps in preventive care. As they got better at that, the focus morphed to gaps in chronic disease care.
"Now we’re focusing more on the postacute space, where we’re really lagging in performance, and using data to quantify the opportunities to use hard-wiring and decision support tools to drive quality improvement as well as decreases in costs," he says.
For example, in 2015, UnityPoint discovered it was two standard deviations higher than the mean in SNF average length of stay than the rest of the NextGen cohort. So, it put a lot of effort into building a network of preferred skilled nursing facilities, instituted EHR-based decision support for discharge nurses, and implemented programs that hard-wired follow-up with patients and caregivers.
SNF average length of stay has dropped from 26 to 21 days over the first two years of UnityPoint's efforts. Over this period the SNF network has also grown as UnityPoint Accountable Care has expanded its NextGen ACO into other regions. New regions joining the effort have experienced similar drops in SNF average length of stay as well as favorable movement in SNF hospital readmission rates. For example, in 2016 when a series of new SNFs were added the SNF hospital readmission rate moved from 16.28% to 11.31%.
"Not only length of stay, but also readmission rates have gone down, which is exactly what you would want," says Sharp. "Not only are we proud of the work—and it's taken hundreds of people to pull that off—but look what we can do when data shows us where we can improve."
Step 2: Use data analytics to improve care quality, clinical effectiveness, and patient experience
Indeed, data analytics can play an increasing role, says McVay.
Now the emphasis is on predictive modeling—using data to predict which patients are likeliest to be admitted to inpatient care within six months, for example.
It’s exciting to be at the front end of this change in industry norms, of the shift to value, Sharp says, but he says leaders must keep in mind that the models are evolutionary and returns diminish over time, so shared risk models are not the end game.
"We ultimately want to be in a place where we’re taking full risk on populations and doing it as good or better than anyone else," he says. "We see the shared savings models as evolutionary and we're happy we’re in them, but how do we make sure we are successful under full risk?"
Ultimately, UnityPoint recognizes that being successful is anchored in efficiency and removing unnecessary utilization. One way to be successful in that endeavor is understanding more about the population, including their health and risks through claims data. Then, more sophisticated operators can apply predictive models to those groups, says McVay, informing them of ways they can intervene.
"Analytics is a big piece of the foundational building blocks of value-based care," she says. "We’ve created an avenue to know who’s been in the hospital in the past 30 days, who’s likely to be readmitted, and patients who are likely to not show up to follow-up appointments."
That allows the analytics department to serve as partners to clinicians or financial leaders rather than taskmasters, says Rhiannon Harms, executive director of analytics at UnityPoint.
"In the [Medicare Shared Savings Program] we didn't earn shared savings, but we saw metrics improve. In NextGen, we’re seeing those savings and are pleased with the results but a lot has come through analytics," says Sharp.
The ability to merge claims and clinical data helps create a holistic picture of patients when they’re seen at UnityPoint, because the EHR data can be overlaid with information from claims.
From a financial perspective, a lot of modeling can be done on per member per month costs, and actuarial methodologies can be used to support metrics that are traditionally found in claims data, says McVay.
She says that UnityPoint's approach to analytics is to collaborate with clinicians on how to best leverage the information and solve problems together.
"It's very tactical. To be successful at value-based care means different partnerships," McVay says.
For example, UnityPoint doesn't employ all the physicians in its network nor does it ever expect to, so they must figure out how to engage independent physicians and postacute partners in a different way—by helping them achieve the patient care and financial goals of the ACO.
Step 3: Transition physicians from a productivity to value-based model
Perhaps the most critical aspect of ensuring that a value-based payment system works is getting physicians aligned with an organization's imperative to reduce unnecessary utilization and readmissions, among other metrics. For those used to productivity-based compensation, it can be difficult for them to make the transition to a compensation plan based on achieving goals in patient experience or quality, for example.
Conversations on transitioning the physician compensation plan began in 2015, says Keith Seashore, executive vice president and chief financial officer at UnityPoint Clinic, the physician organization within UnityPoint.
"From a philosophical perspective, it hasn't been super-challenging," he says, noting that physicians are tired of the "rat race," that is, compensation tied to the [relative value unit] model, which is incompatible with reducing readmissions and increasing patient satisfaction.
"They’re eager to get there," he says.
But it will take time. The unit's compensation committee, which is led by physicians and other practitioners, is leading the transformation. Productivity doesn’t necessarily drive value in value-based contracts, so the challenge is to create metrics that align with them, says Seashore, and find ways to measure compliance and tie compensation to compliance levels.
"They will make the right decision for the organization if you point them in the right direction," he says.
To transition from the RVU model, Seashore says the physicians will gradually transfer to a panel size model and be partly reimbursed on that panel's risk adjustment.
One metric that will factor into the new compensation model is based on thorough documentation and risk coding, including severity of illness. Clinicians and their office staff also spend a lot of effort ensuring that Medicare patients come in for annual appointments, and that they're compliant with medications and screenings. They've also worked hard to coordinate what Seashore calls "whole care services," which encompass transitional postacute care.
"Additionally, we've spent quite a bit of time developing relationships with skilled nursing facilities to ensure alignment with high-cost care," he says. "That's a big component to how you’re going to be successful inside the NextGen products."
It's a slow process. For primary care, the goal was to put 30% of comp at risk within three years from about 10% at the outset. Five pilot sites are testing the model.
"That’s helping us work out the kinks and has developed physician champions who can help move the model forward," Seashore says.
In 2017, they were about 16% at risk, he says, but by 2020, 30% of physician compensation will be at risk, which he calls, "pretty aggressive."
"Many physicians will do very well, and this is budget-neutral, but there will be winners and losers—that's why we have a three-year transition," he says.
In the meantime, UnityPoint will provide coaching, resources, and tools to improve patient experience, quality, or in fixing other problems.
The most challenging issue right now is working through data to ensure accuracy and reliability," he says.
"That’s not new and you’ll hear it from many, but it's not easy to get accurate data on quality and other financial items," he says. "You can’t ignore it because you can’t allow data snafus to undermine the value of this [initiative]. We’re not perfect yet, but we know we have to be good at making sure data is reliable if we’re putting that much comp at risk."
Philip Betbeze is the senior leadership editor at HealthLeaders.