
elenabs/istock via Getty Images
Duke Health to use predictive analytics for population health
Academic medical center Duke Health will use Trilliant Health’s predictive analytics platform to drive medical decisions and improve outreach to underserved patient populations in North Carolina.
Academic medical center Duke Health is turning to predictive analytics tools from Trilliant Health to make its decision-making more precise to meet the needs of patient communities.
In early August, Trilliant Health announced its collaboration with Duke Health, which aims to better understand the challenges of its patient population in North Carolina, according to Stephen Blackwelder, Ph.D., the health system's associate vice president of strategic analytics and consumer insights.
"Duke Health is harnessing the power of analytics and predictive modeling to better understand and address the complex health challenges facing North Carolina's growing communities," Blackwelder said. "By examining factors such as access to care, transportation barriers, food insecurity, housing instability and environmental risks, Duke Health aims to uncover root causes and design targeted interventions that promote health equity."
To achieve health equity, society must remove poverty, discrimination and the factors that result from them, including lack of access to quality education, housing and healthcare, according to the Robert Wood Johnson Foundation.
The collaboration comes as the healthcare industry is shifting toward outpatient and nontraditional settings, such as telehealth.
"Through innovative technologies and data partnerships, Duke Health is forecasting future needs and expanding services -- both in-person and virtually -- to underserved areas across the state," Blackwelder said. “This approach supports strategic growth, enhances community engagement, and ensures that care is accessible, trusted and responsive to the evolving needs of North Carolinians."
Academic medical centers use predictive analytics to forecast demand across multistate referral networks compared with community health systems, which are focused on local market share, service line growth and operational efficiency, according to Hal Andrews, cofounder and CEO of Trilliant Health.
"Our analytics are tailored to support the distinct priorities of both academic medical centers and community providers," Andrews said.
Understanding patient communities using analytics
Organizations such as Duke Health can use data analytics to identify underserved patient populations by analyzing data at a "hyper-local level," Andrews explained.
"These insights reveal where gaps in access exist today and where they are likely to emerge in the future," Andrews said. "By understanding the specific needs of these communities -- from preventative care to acute interventions -- health systems can strategically deploy resources, design targeted outreach programs and expand services to improve access to care."
Predictive analytics allows health systems to move beyond "directionally correct" projections created based on incomplete and outdated data. Instead, Trilliant takes a hyper-local view of supply and demand down to the ZIP-code level, Andrews explained.
"We combine data on healthcare utilization, disease burden, demographic shifts and provider supply to forecast demand across specific care disciplines," he said. "This data-driven understanding of local markets enables health systems to allocate capital effectively, optimize service line growth, avoid overbuilding and improve access to care for the communities our partners serve."
Using psychographic segmentation and network integrity analysis
Duke Health will use Trilliant Health's suite of analytics tools, which includes predictive analytics; psychographic segmentation, a way for researchers to study how communities engage with care across patient life stages; and network integrity analysis, which lets clinicians bolster retention and loyalty to the academic medical center's care ecosystem.
"Psychographic segmentation groups patients by their motivations, preferences and decision-making styles -- the 'why' behind how people engage with care," Andrews said. "We leverage these insights to help health systems understand which psychographic profiles exist in their markets, how these groups differ across life stages and what communication and service strategies will resonate."
The technology could help clinicians understand the profile of some patients in Medicaid populations, who are disengaged or less likely to seek care, according to Andrews.
With this insight, health systems might focus their investments in urgent care facilities and create messaging with an emphasis on convenience, he said.
By using network integrity analysis, health systems like Duke Health can improve their referral retention so patients can experience continuity of care. The analytics can spot "leakage points," which could include out-of-network coverage, Andrews suggested.
"Our analytics pinpoint where and why physicians refer patients out of network, uncovering factors such as limited specialty coverage, lack of convenient locations or capacity constraints," Andrews explained. "By mapping referral patterns and provider relationships, health systems can address these leakage points, align physician networks with patient needs and keep more care in-network, improving both continuity of care and financial performance."
Going forward, predictive analytics and behavioral insights using tools such as Trilliant will allow health systems to forecast future demand and plan capital investments based on how patients look for and use care.
"By combining forecasts of future demand with data on consumer preferences, utilization patterns and access gaps, health systems can identify where services are truly needed and which types of facilities or capabilities will be most effective," Andrews said. "This alignment helps ensure that new clinics, service lines or technology investments are deployed in locations and formats that meet unmet needs, reduce access barriers and deliver the greatest value to both patients and the organization."
Brian T. Horowitz started covering health IT news in 2010 and the tech beat overall in 1996.