How ACOs Use Data Analytics to Improve Outcomes, Reduce Financial Risk

April 26, 2023 - Accountable care organizations (ACOs) have significantly influenced the shift in the healthcare payment process from fee-for-service models to value-based care since their introduction in 2010 through the Affordable Care Act (ACA).

However, ACOs must effectively gather data and track patient journeys across the care continuum to provide value-based care. The technology, data, and clinical capabilities of an ACO are a key part of its strategy to successfully improve patient outcomes and address financial risk. But how do ACOs effectively use these capabilities together to drive these goals?

Shawn Bassett, vice president for the North Region of Collaborative Health Systems, and Sheila Magoon, MD, executive director of Buena Vida y Salud ACO and South Texas Physician Alliance, shared with HealthITAnalytics in Zoom interviews how their ACOs are using data analytics to support high-quality care delivery and prudent healthcare dollar spending.

GETTING ACTIONABLE DATA AND ESTABLISHING INFRASTRUCTURE

Since tracking patient journeys and spending is critical for ACOs, gathering and accessing data is one of the first steps toward using that information to minimize risk and improve outcomes, Bassett and Magoon agreed.

However, ensuring that the data are timely and actionable can present a host of challenges.

“When we look at the IT challenges of our business nature, it's making sure to get the real-time data at the right place at the right time because that can make the clinical aspect really frustrating,” Bassett explained. “We see it quite often with some of our providers that if we're just operating off of claims data, that's a little bit delayed. They could already be addressing a patient's needs or concerns, and when we give them data, they start to question it.”

Giving providers the right data at the right time requires multiple sources of data and the appropriate infrastructure to support the management and communication of that data, he continued.

Keeping a focus on the ACO’s goals, which include helping patients remain healthy at home, can help organizations sift through the wealth of data generated to pull what is actionable and omit irrelevant information or interesting but non-actionable ‘curiosities,’ Magoon noted.

“It's really about what's actionable,” she said. “How much can you consume? There's so much data, [and] you could spend your whole life studying just the data. But then, how do you help your patients do better?”

“When everything goes around that test… then it really narrows down what data you really need and look at,” Magoon explained. “It's no longer a research curiosity; it's really about clinical care. So I think our goal is that as long as we keep it all focused on clinical care and keeping our patients healthy at home, then that helps us identify what we need to look at from those data standpoints.”

On top of ensuring the data make it into an ACO’s system, interpreting it, and efficiently delivering it to the appropriate providers, Bassett noted that ACOs must also contend with educating providers on how to utilize the data on the clinical side.

Magoon echoed this sentiment, stating that her ACO applies different approaches to help clinicians effectively use data.

After identifying actionable areas of clinical improvement, she explained that some providers will leverage the ACO’s analytics platform themselves and pull the appropriate data, while others prefer to sit down as a team with ACO leadership regularly to go over the data together.

These providers may lack the time or extra staff numbers needed to pull the data themselves, so Magoon and her team evaluate the capacity and capability of each provider office within the ACO to help guide them through the analytics process and choose an appropriate approach to benefit their specific needs. In doing so, those teams can gain insights into their patient populations that they didn’t have before.

Magoon further stated that this flexibility can be applied more broadly, indicating that while ACOs have different challenges based on factors such as geography or patient population makeup, many of the potential hurdles and benefits of using analytics are the same.

“Each [ACO has its] own risks and challenges… but for us, and in my personal experience, we've had a whole variety of different challenges, ranging from in the beginning being able to get the necessary data, to then learning what is actionable, and then to engaging other healthcare providers and facilities to meet those individual goals,” she said. “So what our philosophy is, then, is that if we can just provide really good quality care, [and] meet our patients' needs as best as possible so we can keep them healthy at home, that we think that those financial goals will follow.”

LAYERING DATA SOURCES AND ANALYTICS TOOLS

Addressing the challenges associated with improving care quality and reaching financial goals rely on more than just having actionable data and educating providers.

When Bassett’s ACO works with providers, it doesn’t just provide simple tools or data reports. Rather, the organization has community health workers, practice transformation coaches, and other teams to lay the foundation for transitioning that provider to value-based care.

These teams work with providers to help improve clinical workflows before actionable ACO data is introduced, which enables these providers to ascertain “what they need to know, when they need to know it, [and] how they can act on it,” Bassett explained.

Magoon and her team take a similar approach, developing clinical workflows by identifying opportunities in care utilization and quality and focusing on narrow targets related to those opportunities.

“[Doing so] helps us develop care coordination and outreach efforts, whether it's patient engagement, provider engagement, or engagement with other healthcare providers within the community to address those specific challenges,” she noted.

Using this strategy, the ACO tackled high emergency room (ER) utilization among its patient population. Magoon and others identified patients who more frequently used the ER, made a list of these patients, and gave that list to providers and care coordinators for use in outreach.

On reaching out to those patients, care coordinators were able to provide patient education about when to call a healthcare provider before making an ER visit and help identify and address patient needs that were tied to ER utilization.

In addition to optimizing clinical workflows, forging data analytics partnerships is critical to reaching ACO care quality and financial risk goals.

Bassett and Magoon’s ACOs both have a relationship with the Health Data Analytics Institute (HDAI), an analytics company that specializes in providing quantified health risks, personalized care profiles, and risk modeling methodology to improve health outcomes.

The partnership with HDAI is a major part of his ACO’s data analytics strategy because the collaboration allows broader access to data, resources, and healthcare analytics expertise, Bassett explained.

For instance, working with HDAI has offered Bassett more insights into providers he is considering working with or bringing into the ACO.

“When I go into a new market, I no longer have to recruit a doctor and hope for the best once I get data; I'm already able to have some insight as to what their historical performance is,” he said, further explaining that using historical performance data allows him to see more than just high or low physician performance.

“What do those trends over the years look like? Because a doctor was a high performer two years ago, [it] doesn't necessarily mean they are a high performer today,” Bassett said. “Or, on the flip side, I've seen several doctors [for whom] we've watched the data, and we've seen their numbers just going high, which makes you nervous, but is there a story behind that? Did they start picking up a different practice?”

For example, data trends can indicate that a provider may have become a skilled nursing facility (SNF)-based provider, which would have an overall negative impact on costs for the ACO but may not result in a dip in care quality. Seeing the bigger picture of what’s going on with a provider in this way allows Bassett and other leaders to make more effective decisions about which physicians to work with and how to work with them effectively.

Bassett also noted that HDAI’s work in artificial intelligence and predictive analytics further supports the ACO’s mission by helping it to prioritize when and how to work with patients.

“I don't have enough resources to take care of 180,000 lives,” he stated. “I need to know which ones I need to impact at which time to be able to relay that to the provider.”

Using historical trends in the patient’s health data, predictions about their future needs can be generated, Basset explained. Then, these data can be compared to that of other patients in the same clinical scenario to determine which patients take clinical priority to help boost health outcomes.

With these insights, which are generated from a combination of current and historical claims-, Centers for Medicare and Medicaid Services (CMS) Virtual Research Data Center (VRDC)-, EMR-, prescription-, social determinants of health (SDOH)-, and admit, discharge, transfer (ADT)-based data, alongside ER, inpatient, and SNF utilization metrics, ACOs can bring down costs and reduce waste, as these areas typically create a majority of healthcare costs.

According to Magoon, by layering these data and leveraging different analytics tools, ACOs can also bolster interoperability.

“We work with a health information exchange, so we pull data from a variety of sources,” she explained. “Someday, I would love to have data flow like water and breathe like air. But in the meantime, that's just not happening yet.”

While interoperability has significantly improved, ensuring seamless data flow is still complicated.

“So, what we've decided to do until [data] can really flow smoothly is we are looking at what I call a ‘Lego model,’ Magoon stated. “We layer together a variety of our different data sources, and now we're layering together two different analytic platforms, so that way we can build a more comprehensive view of our patients. That way, we have a better idea of what's going on with our population.”

By creating this more holistic view of the patient population, Magoon and her team aim to create more opportunities to drive improvements that allow their patients to remain 'healthy at home,' among other key performance indicators (KPIs).

LEVERAGING ANALYTICS PARTNERSHIPS TO DRIVE KPIs

In terms of KPIs, ACOs aim to reach the quality performance benchmarks outlined by CMS. For the 2020-2021 performance years, the benchmarks are encompassed within four domains: Patient/Caregiver Experience, Care Coordination/Patient Safety, Preventive Health, and At-Risk Population.

Magoon explained that most ACOs use these national, established benchmarks to guide KPIs based on their populations’ needs, targeting metrics like ER admissions, hospital readmissions, and fall prevention. From there, ACOs need to build programs addressing these KPIs. Analytics partnerships can play a major role in developing the programs.

But having an analytics partner that understands healthcare is crucial, Bassett noted.

“We wanted to know that they understand the population they're working with, that they have the real-time deliverables of the data, and that they can perform and understand what it is we're dealing with,” he explained.

“Now, we have great analysts, and they do an amazing job, but sometimes analysts are just analysts. You can't take someone from Wall Street and just expect them to understand the nature of the healthcare industry because they simply say, ‘Here's a report that I ran with the numbers,’” he continued.

Instead, analysts need to know the story behind the data or understand enough about the healthcare industry to dig deeper and find the story hidden in that data.

To illustrate his point, he highlighted an incident where an analyst drew up a report showing that the ACO’s physicians were generating double the national average in revenue. Based on his experience in the industry, Bassett knew that such a phenomenon was unlikely and concluded that there was an error in the analytics used to generate that report.

Having a partner who has that kind of knowledge and can help explore the story that the data are telling to find such analytics errors is significant, he stated. Such a partner can play a vital role in helping to identify the ‘why’ behind care quality and cost drivers. Once the ‘why’ is identified, ACOs can take action to lower costs and improve care.

Bassett explained that most KPIs revolve around SNF, ER, and inpatient admissions. His ACO also looks at the overutilization of specialists in the market, such as those in orthopedics and ophthalmology, because these are high cost-drivers.

“So when we're looking at the SNFs, we're looking at decreasing length of stay or understanding the quality behind that SNF,” he said. “Is it risk-adjusted correctly? I'm going to take a look at — is it the length of stay? Are they dealing with a higher-risk patient? There are some of those SNFs [where] the reason their costs are higher is because their acuity of the risk is higher. I've got to be able to see that and know that it can be adjusted.”

Inpatient and ER admissions are similarly explored to help see where adjustments can be made to meet national ACO benchmarks.

In addition to using the data to gain insights into historical trends, Magoon indicated that working with HDAI has also allowed her ACO to leverage predictive analytics. The two organizations recently signed a contract to use these analytics to identify patients at greater risk of adverse events and higher resource utilization to help prevent avoidable encounters and improve care quality.

Some of these adverse event reports are used to predict unplanned hospital admissions, she explained. Using analyses of historical data, the ACO has already been able to prevent some hospitalizations. But with predictive analytics, Magoon and her team aim to flag patients before they begin to decompensate and help address the factors that would have put them in the hospital.

“That report's now gone out to all of our physicians,” Magoon stated. “So they have their patient list, and we're asking them, ‘We'd like you to bring these patients in once a month and see them.’ [These patients have] a number of different chronic diseases, so there's medical necessity that's built into this. And because they are considered high risk for a hospital admission, then let's see if we just bring them in more often, can we get these patients to stay healthier, identify any issues prior to a significant event happening?”

This program, in collaboration with HDAI, is set to run for at least 12 to 18 months to determine if it significantly improves the ACO’s care quality and costs.

“At this point, I think that we need to look at our data and what's going to be clinically actionable in helping our patients, and the other component that we think is going to be really important for us to continue to drive those improved outcomes is [predictive analytics,]” Magoon said. “Because that, I think, is the only way, as an ACO, that we're going to be able to stay in the game long-term… we're going to have to start figuring out what's about ready to happen before it happens, so we can mitigate it.”

While analytics and predictive modeling may be key to an ACO's success, Bassett noted that everything comes back to “doing healthcare justice” and keeping people at the center of these efforts.

“There's the key value of the data itself, but we are healthcare, no if, and, or buts,” he stated. “We are dealing with people, human beings… the patient, the doctors, [and the] staff, all those factors have to be considered.”

“If I created a solution, or a strategy, or a plan that focuses solely on the data, then I can say right now that we won't be as successful as we could in the industry we're in. We've got to take into account human beings,” he continued.