Beyond the Numbers: An Approach to Patient Panel Management at Duke Primary Care

How many patients is too many?

It’s a deceptively simple question that looms large for leaders trying to balance supply and demand in today’s primary care landscape. In Episode 30 of the All Access Pass podcast, we sat down with Dr. Kevin Shah, Vice President of Primary Care and Chief Medical Officer at Duke Primary Care, to explore the sophisticated approach his team has taken to make panel management both more objective and more actionable.

The Problem: Demand That Outpaces Supply

Duke Primary Care serves a rapidly growing population in Central North Carolina, where the influx of new residents far exceeds the pace at which the health system can recruit and onboard primary care providers. “We cannot grow primary care fast enough to meet the demand,” says Dr. Shah. But simply hiring more providers isn’t a sustainable answer. Instead, the Duke team is asking: How do we maximize the capacity we already have while ensuring high-quality, equitable care?

The Foundation: Define the Panel

Step one was getting clear on what a “panel” actually is. Dr. Shah and his team began by distinguishing between a total panel (patients for whom a provider is listed as the PCP in the EMR) and an active panel (those who have had an ambulatory visit at Duke in the past two years).

That two-year window was intentional. “We know that some patients may go a year or more without needing to be seen, especially if they’re relatively healthy,” explains Dr. Shah. The team found that narrowing the window further didn’t significantly improve accuracy, and the two-year cut-off struck a balance between inclusivity and clinical relevance.

To further refine their understanding of panel composition, Duke incorporated variables such as age, gender, insurance type (especially Medicare), and pediatric status. They also added continuity metrics, capturing whether patients were consistently seeing the same provider, the same practice, or the Duke system more broadly.

The Model: Normalizing Visit Rates

From there, the team tackled one of the thorniest issues in panel management: visit rates. If one provider sees their patients twice per year and another sees theirs once, their panels can’t be directly compared. And historical data alone can be misleading, particularly if providers have been constrained by access issues.

Enter multivariate regression.

Duke used clinical complexity markers—like COPD, diabetes, depression, and heart failure—to develop an index that estimates expected visit rates. This allows the model to account for panel acuity when calculating how many patients a provider can reasonably manage.

The resulting formula is simple in structure but powerful in impact:

Ideal Panel Size = Adjusted Visit Capacity ÷ Expected Visit Rate

This output helps identify whether a provider is over-paneled, under-paneled, or appropriately matched.

The Mindset: Data-Informed, Not Data-Driven

Crucially, Dr. Shah emphasizes that the tool is not prescriptive. “This is not the source of truth,” he says. “It’s meant to be an objective analysis to help inform clinical and operational decision-making.”

Local leaders still hold the authority—and responsibility—to open or close panels, taking into account on-the-ground realities like provider availability, urgent care utilization, and patient preference. In some cases, panel reassignments occur manually if patients have migrated their care organically. The tool simply offers a shared language and consistent framework for having those conversations.

The Future: Reimagining the Care Model

For health systems looking to expand panel capacity without burning out providers, visit rates must be addressed—but not by rushing patients through appointments. Instead, Dr. Shah advocates for team-based care models and work redesign that redirect appropriate follow-ups to other members of the care team or alternative channels.

“There’s a confluence of goals here—quality, access, panel size, patient experience,” says Dr. Shah. “And they all point to the need for different ways of delivering care.”

Final Thoughts

Patient panel management is not just about counting names in a database. It’s about aligning clinical demand with operational supply in a way that supports sustainable growth, equitable access, and high-quality care. The approach at Duke Primary Care offers a thoughtful, structured, and adaptable model for how to get there.

For organizations facing rising demand, constrained capacity, and a mandate to do more with less, Dr. Shah’s insights provide a roadmap for what’s possible when data and local leadership work hand in hand.

Want more insights like this? Listen to Episode 30 of the All Access Pass or explore our resource library at www.patientaccesscollaborative.net.