How UofL Health and Norton Healthcare Use Data to Improve Medicaid Quality

For years, the promise of data-driven healthcare often collided with the reality of fragmented, untrusted information. At the recent NCQA Health Innovation Summit, executives from UofL Health and Norton Healthcare revealed how they bridged this gap, moving from a culture of skepticism to a collaborative model that has significantly improved Medicaid quality metrics across Louisville, Kentucky.

The program, which operates as a state-directed payment initiative, began with 10 quality performance metrics before expanding to 21. These measures—covering everything from pediatric screenings and behavioral health to chronic condition management—are mandated by the state’s quality strategy for Medicaid populations. As reported by Healthcare Innovation, the two health systems have divided their focus to maximize impact: UofL Health concentrates on the adult population, while Norton Healthcare tackles the pediatric side.

Building Trust Through Custom Analytics

The transformation was not immediate. When Robert Hayes, UofL Health’s value-based purchasing manager, joined the program in 2023, he found that the existing data infrastructure lacked the confidence of the staff. While electronic health records like Epic offer built-in dashboard tools, they often failed to capture the nuances required for these specific Medicaid programs.

The solution was a custom build using Power BI, a move that shifted the internal conversation from debating the validity of the data to addressing actual clinical gaps. By creating a database that pulls in specific requirements for procedures like mammograms or colonoscopies, the team provided primary care clinics with a clear, actionable view of their patient needs.

When you have actionable data, you can actually start to build processes that allow clinics to offer better care.

This approach turned data into a tool for operational change. At the Mary & Elizabeth family medicine clinic in Louisville, the team utilized this new visibility to integrate mammogram screenings directly into primary care visits. Rather than requiring patients to return for a separate appointment, the clinic now coordinates with its on-site mammography suite to provide care during the initial visit, a process made possible only because the clinic finally had accurate, real-time patient lists.

Shared Infrastructure and Social Determinants

While UofL Health and Norton Healthcare maintain separate systems, they have aligned their efforts through a monthly workgroup facilitated by the consultancy Sellers Dorsey. Amanda Elhart, associate director of clinical performance and impact at Sellers Dorsey, describes the firm’s role as an intermediary that aggregates these efforts and shares best practices, both between the two Kentucky systems and from national trends.

This collaboration has extended beyond clinical metrics into the social determinants of health (SDOH). Norton Healthcare, for instance, identified food security as a persistent barrier for their pediatric patients. By gathering data directly from parents, they were able to justify the placement of food banks within their pediatric primary care practices.

UofL Health has similarly scaled up its SDOH screening efforts. While these assessments previously existed only in isolated pockets, they are now being integrated system-wide. This allows staff to identify community needs and connect patients with necessary resources like food banks at the clinic level.

The Looming Fiscal Reality

Despite these successes, the path forward faces significant economic headwinds. The integration of community health workers and support staff is becoming increasingly difficult due to the current fiscal climate. As Hayes noted, there is a looming concern regarding a potential “cliff” in Medicaid reimbursement, compounded by the uncertainties surrounding federal policy changes like the impact of H.R. 1.

Norton Healthcare faces data breach lawsuit from former employee, current patient

These financial pressures underscore why the “right people at the table” matter. According to the team, the lack of coordination between quality improvement and clinical operations had previously stalled progress for years. Once the correct team was assembled, they managed to define and implement 21 measures in just five months.

For other health systems looking to replicate this model, the lesson is clear: proactive data is the baseline. As Elhart pointed out, many organizations still rely on lagging data—often three to six months old—which renders it nearly useless for real-time clinical intervention. By investing in SQL development and analysts who understand the complexities of EHR systems, health systems can move from reacting to the past to managing the health of their communities in the present.