Yan Hu contributed to this case study.

Key takeaways

• Outcomes accountability in patient support helps leaders use real-world evidence to decide which programs should grow, change or stop.
• Data integration with outcomes can show whether patient support improves medication fulfillment and persistence, not just enrollment or service utilization.
• Pharma patient support analytics can reveal which patient segments benefit most, helping teams focus resources where support changes behavior and improves outcomes.

Across pharma, patient support now requires significant resources. Yet many leaders still lack evidence to decide which programs should grow, which need work and which should end. Enrollment, coaching touch points, satisfaction scores and website usage still dominate reporting, but these measures show only what happened inside the program. They don’t account for whether support changed patient behavior or answer the core question leaders are accountable for: Is patient support changing behavior in ways that improve patient outcomes?

Novo Nordisk faced that decision with WeGoTogether (WGT), its flagship patient support program for people using GLP-1 therapies. Working with ZS, the team moved beyond engagement metrics to measure whether support changed patient behavior. By linking program data to pharmacy claims, Novo Nordisk shifted the question from how many patients engage to what engaged patients do differently.

“This is a truly data-driven and holistic analysis,” said Yan Hu, director of data science and advanced analytics at Novo Nordisk. “Because we connect it to our patient journey work, we can look at every touch point within WGT and measure how each one influences patient access, initiation and long-term adherence to therapy. It’s not surface-level engagement metrics. It’s actual behavioral change and patient outcomes.”

Use linked claims data to prove what patient support changes

Before the analysis, the prevailing internal view was that WGT primarily helped patients access and initiate therapy. The team linked program data to claims data and used advanced causal modeling to isolate what parts of the program were truly driving change. This approach also accounts for the fact that the strongest data is often from enrolled patients, therefore standard confounding is not enough to isolate impact.

The outcomes-driven impact assessment informed program refinement. Participation in WGT showed a 3%-6% improvement in 14-day prescription fulfillment across different patient segments, a more precise measure than website registration rates alone because it shows whether patients follow through in the critical early window to actual fulfillment. But the larger impact appeared in persistency. WGT extended time on therapy by approximately six to 22 days among engaged patients compared with nonparticipants, reflecting a scenario where all eligible patients are reached or activated in the program. In a category where adherence drops sharply in the early months of treatment, that extension gave Novo Nordisk a more precise and defensible case for the program, with clinical and program implications that enrollment metrics alone could not show.

This insight clarified the role WGT plays in the patient journey and how the program should be evaluated. Its value lies in helping patients stay on therapy after initiation, with program value demonstrated by time on treatment rather than enrollment activity. That focus strengthens how success is defined, guides resource allocation and reinforces the case for strengthening patient support.

BMI outcomes provided clinical validation of that effect. Registered patients achieved 6%-12% greater BMI reduction than comparable nonparticipants, translating longer time on therapy into a measurable health outcome and demonstrating impact beyond behavioral metrics alone.

That effect was not evenly distributed. Persistence gains concentrated in specific patient segments, which became the second major learning.

“The initial hypothesis included access and affordability needs, such as the e-voucher or coupon, but when you look at the drivers, you see a much clearer differentiation,” Hu said. “Affordability matters for some patients, while others are motivated by health management needs and long-term outcomes. Some have severe comorbidities and care deeply about managing their conditions. For those patients, the coaching and support services provide real value alongside the treatment itself.”

Turn patient segmentation for outcomes into precision support decisions

Once it became clear that value concentrated in specific patient types, the resource-allocation logic shifted.

A broad, undifferentiated support strategy no longer made sense. The evidence showed that the program’s mechanism of impact ran through coaching and support for patients managing serious health conditions, while patients with immediate affordability or access needs may need a different support path. Treating those groups the same diluted effectiveness.

The design logic changed immediately. Content and outreach could now align with the segments where support changed behavior, rather than treating patients with different needs as if they were navigating the same experience. For Novo Nordisk, that meant aligning coaching resources with patients managing serious comorbidities and simplifying access-focused support for patients whose immediate needs center on affordability.

“Once you understand the patient segments, you can design targeted interventions much more precisely,” Hu said. “Some patients, middle-aged males for instance, are more driven by health outcomes. Others are more appearance-driven. So, the content you deliver will differ to emphasize cardiovascular benefits for one group and lifestyle benefits for another.”

Segment-level clarity turned a large program into a set of targeted decisions about where to focus resources, what to emphasize and which experiences to deepen or simplify. Patients with multiple cardiovascular conditions receive more intensive coaching and follow-up, while patients with immediate affordability needs move through simpler access-focused support. Resources shift to where support changes behavior.

Combine pharma patient support analytics into decision-grade evidence

Connected data-driven impact measurement established whether behavior changed, driver analysis showed which aspects of enrollment and engagement drove that change, and patient segmentation clarified who the program was truly serving and how. ZS brought these components together into a single analytical framework, producing decision grade evidence leaders could use to guide patient support decisions.

“Together, they help us unlock very clear priorities about where and what to invest in next,” Hu said.

The value of this clarity was organizational as much as analytical. Conversations about WGT shifted from belief to evidence. The program could be positioned accurately as a persistence engine for defined patient segments, with a quantified effect and a defensible program rationale. That shift reduced noise and gave leaders a clearer basis for action.

Make outcomes accountability in patient support repeatable across programs

The lasting value for Novo Nordisk is not a better read on one program but a repeatable way to guide patient support decisions. While the approach has not yet been rolled out across therapy areas, the goal is to adapt the framework as a standard way to measure impact across programs.

“You build the analytical foundation for one patient support service, and with therapy-specific adjustments, you can adapt it to other programs,” Hu said. “That’s a significant asset for the organization.”

Move from activity tracking to patient support program outcomes

Patient support is now too important to be evaluated by activity metrics alone. Leaders need to know what behaviors support changes, which patients benefit and where resources should move next. As data foundations mature, the barrier is less technical than managerial. The harder question is whether teams are ready to use outcomes evidence to decide which experiences to expand, redesign or retire.

That matters because support services are not neutral at scale. They either reflect what patients are navigating or reinforce assumptions about what patients need. Outcomes evidence gives leaders a way to tell the difference before resources follow the wrong story.

“Medication alone may not be sufficient. The supporting ecosystem matters just as much to the overall patient experience. Before starting therapy, patients face very different challenges, including clinical hesitancy, behavioral barriers and social drivers of health,” Hu said. “All of those are captured in our analysis, and together they give us a holistic picture of what patients are actually up against.”