Healthcare Analytics, Population Health Management, Healthcare Big Data

Population Health News

How Data-Driven Care Management Improves Population Health in NC

Big data and a coordinated approach to care management is improving population health and lowering costs for Medicaid patients in North Carolina.

By Jennifer Bresnick

- Comprehensive population health management can be a struggle for providers who lack the data, infrastructure, and community collaboration to support patients across the care continuum.  But many healthcare organizations in North Carolina are fortunate enough to have all these resources – and more – to help serve their Medicaid populations.

Population health and care management in North Carolina

Thanks to the hard work of Community Care North Carolina (CCNC), a state-wide care management initiative dedicated to controlling Medicaid costs and improving the quality of care, providers across the state are leveraging analytics insights and real-time patient alerts to target high-risk patients most in need of a helping hand.

Widespread clinical buy-in, an eye for socioeconomic vulnerabilities, and a willingness to invest in continuous improvement are the keys to slashing spending, emergency department use, and hospital admissions for Medicaid patients, according to Annette DuBard, MD, MPH, Director of Informatics, Quality, and Evaluation at CCNC.

“Our community physicians feel they can do a better job caring for Medicaid patients by organizing locally to make sure these beneficiaries have a primary care practice to go to,” she said to HealthITAnalytics.com

“Today, there are more than 1800 participating primary care practices in 14 regional networks that cover around 1.6 million Medicaid recipients in North Carolina.  That’s out of roughly 1.9 million in the state, so we have been tremendously successful in getting Medicaid patients under this umbrella.”

READ MORE: Epic Systems EHR Brings Care Coordination to Walgreens Retail Clinics

CCNC’s quality improvement and cost-cutting efforts have also been tremendously successful.  Between the December of 2014 and the end of 2015, total Medicaid costs dropped five percent.  Inpatient admissions were down by 26 percent, ED visits declined by 7 percent, and preventable readmissions were slashed in half.

Despite increases in clinical complexity among CCNC’s Medicaid population, actual spending per patient per month dropped almost ten dollars below anticipated benchmarks.

CCNC has been working for nearly 20 years to develop the infrastructure and best practices that have supported these outcomes, DuBard said, and the project got off the ground long before big data was a commonplace buzzword.

“This effort started before data was as ubiquitous as it is today – years before electronic health records and most health IT tools,” she explained.  “We’ve always tried to be a data-driven organization, but much of the analytics at the beginning came from manual chart reviews to track quality and progress.”

“We used to say that we had a human health information exchange long before an electronic health information exchange.”

READ MORE: CT Medical Society Announces New Health Information Exchange

The past eight years have seen a “very concerted effort” to develop a statewide health information exchange and analytics infrastructure that could support highly responsive and individualized care management, she added.

“Today, we are using statewide Medicaid claims data complemented by data from other sources, including real-time hospital admission, discharge, and transfer (ADT) data from about 80 percent of the hospitals across the state,” said DuBard.

“We are also using laboratory data and state vital statistics data to build a shared utility for the whole state to use in ways that target patients more intelligently for specific care management interventions and speed up the cycle of quality improvement.  We have full visibility into Medicaid claims, including hospital utilization and pharmacy utilization, thanks to how willing the state has been to partner with us.”

Creating the data-rich medical neighborhood

The principles behind the patient-centered medical home (PCMH) play a large part in helping providers and other community service organizations partner to leverage this data for population health management.

READ MORE: CVS Retail Clinics, VA Partner for Care Coordination, Access

“We are grounded in the primary care medical home, and we have always been focused on building the medical neighborhood,” DuBard explained.  “One of the benefits of having a localized infrastructure is the ability to bring everyone to the table, including the hospitals, the Department of Social Services, the health departments, and other community organizations that provide services for vulnerable populations.”


Read: Identifying Care Disparities for Population Health Management


This organized, collaborative approach is “really about creating a community, and addressing those interstitial spaces that aren’t exactly within the four walls of any given provider system,” she continued. 

“It’s about following the patient and the family wherever they go in these interconnected systems that don’t always have the ability to communicate as freely as they would like.”

Generating, collecting, and sharing data across disparate systems

Data is the lifeblood of this community, and CCNC has successfully connected disparate organizations that can all contribute pieces of the care management puzzle.

“We don’t operate an HIE, in the sense that most people use that term, but we do receive data from many participating partners and use that data to drive care coordination,” DuBard said.  “So we are able to take information, like the real-time ADT alerts, and route it to the appropriate care managers and enhance that information with our predictive modeling and prioritization strategies that are built off claims data and other big data sources.”

While North Carolina is among the many states that continues to struggle with seamless health information exchange, even basic ADT and claims data can make a huge impact for patients when it is delivered to the right provider at the right time.

North Carolina’s 600-odd care managers can use breakdowns of the patient’s historical utilization patterns and adherence rates to identify beneficiaries who may be in need of a higher level of transitional management. 

“Data about past hospital admissions, ED visits, and adherence to medications or chronic disease management tasks are the metrics that will trigger a notice that the patient is a high priority for more intensive care team management at the time of their hospital discharge,” said DuBard.

“When we receive a notification from a hospital, we can couple that with what we know about that patient from other data sources and send that out as an actionable care alert to the local care manager, who needs to know that there’s a patient going home that day who needs a home visit or a follow-up within 48 hours.  If there are behavioral health needs, social barriers, chronic pain issues, specific chronic conditions or other needs, we can identify those for the care manager who needs to be engaged with the patient.”


Read: Why HIE Data Analytics are Critical for Behavioral Healthcare


CCNC has elected to develop most of its big data analytics and information exchange infrastructure in-house, which has allowed a higher degree of flexibility for its user base across the state, DuBard said. 

“That customization was important for us, because it allows us to be much more user-driven than we could be if we did a big, fancy analytics sweep and implemented an off-the-shelf solution,” she asserted.

Care managers are asked to document their population health management in a statewide care management information system, which was also crafted specifically for CCNC’s needs.

“The care manager isn’t isolated within one practice setting,” DuBard said.  “The role has always been about connecting the dots between all the service providers, including medical and behavioral healthcare providers who may be involved in that patient’s total plan of care.” 

“All of those patient managers are using the same care management information system, which is also something that we’ve developed in-house.  That allows us to route care guidance directly into the workflow of the care manager, and it also allows us to track data in a standardized way across the entire state, including patient assessment information and what interventions have been enacted for them.  It really allows us to analyze what works and what doesn’t over time.”

Developing best practices for continuous improvement

The hardest part of any quality improvement effort is understanding how the patterns in the data can translate to new real-life strategies that will actually produce better outcomes and lower costs. 

Having access to accurate, complete, and timely data is a necessary starting point, but it must be paired with the ability to adapt to what the data is saying.

“In order to be successful, you need to have a culture of quality improvement,” DuBard stressed.  “You need to have a disciplined way of capturing data and measuring impacts, and you must have a willingness to make changes and adaptations based on your data.  Without the willingness to change your processes over time, just collecting the data won’t help you that much.”

“We really value the notion of being a learning health system.  And even though so much of healthcare is local and some of the flavor of exactly what happens to a patient may depend on local resources and local styles, we have really adhered to the principle of learning from each other as much as possible.”

Variability in health IT adoption rates across the primary care system is one major challenge for creating best practices for a state with many different types of communities with unique socioeconomic issues, she continued. 

“Our practices are at many different points on the journey to adopt health IT tools and being able to communicate and exchange data.  And since the care managers are frequently embedded right in the hospital or the primary care practice, there can be some double data entry going on,” she acknowledged. 

“The care managers will be documenting in the statewide system, but they may also be making notes or sending messages using the provider's EHR.  So as we focus on taking care of patients, we also have to be mindful of meeting providers where they are when it comes to health IT.”

In addition to being accommodating to providers, CCNC emphasizes truly personalized care for their patients, and pays close attention to architecting population health management strategies that produce the most bang for the buck.


Read: How to Get Started with a Population Health Management Program


“We have been able to hone in on identifying which patients appear to benefit most from a home visit or from the involvement of a clinical pharmacist, and which patients are more likely to respond to a lighter, low-cost touch like a telephone call,” DuBard explained. 

“As we learn more about those responses, we incorporate them into our care management standards, and we refine the size and scope of our tailored interventions.  That’s how you can achieve return on investment for population health management programs – by tailoring the allocation of your resources to where they will have the biggest impact.”

Providers, patients, and the state Medicaid program all want to see the value of their efforts.  Being able to demonstrate cost savings, as CCNC has done, is “essential” to the sustainability of any population health management programs, stated DuBard.

“We’ve learned just how much the return on investment depends on intelligent targeting of interventions.  Oftentimes we see programs trying to set up care management initiatives or transitional care structures without being as discriminating about targeting as maybe they should be,” she said.

“The patients who are most likely to benefit are not necessarily always going to be the highest cost or highest risk patients.  Some patients are more receptive to attention from a skilled care team.  That may change the trajectory of a disease for certain patients, but not for others."

"So the important thing is being able to use available data to find those patterns that indicate impactability and identify the patients who stand to benefit most from how you choose to deploy your care management resources."

Continue to site...