From Data to Decisions: Empowering Administrators to Track Healthcare Cost Trends

In the shift toward Value-Based Care (VBC), U.S. healthcare systems are losing millions when patients seek care outside their network—a phenomenon known as Network Leakage. As the Lead Designer and Design Manager, I transformed Innovaccer’s Cost Management Dashboard from a data-heavy reporting tool into a proactive strategy engine that identifies "avoidable" leakage and provides actionable recommendations for administrators.

I wore two hats during this 8-month sprint.

As Design Manager: I led a team of 2 product designers, shielding them from the "death by a thousand stakeholders" environment while ensuring our work aligned with the broader Analytics Charter.

As Module Lead: I got hands-on with the Cost Dashboard. I didn't just review mockups; I facilitated co-ideation workshops with Data Architects and Customer Success leads to ensure our designs were technically feasible within the constraints of our Unified Data Model (UDM).

Across the entire engagement, pain points and opportunities were identified and captured in order to fuel the development of the product backlog and better understand how an all-in-one solution must support and enhance the cost dashboard behavior. 

My challenge wasn't to design a prettier dashboard. It was to build a system that told them where the fire was before the building burned down.

We integrated a "Top Recommendations" engine. Instead of just showing a problem, the dashboard suggested specific actions to reduce Avoidable Leakage Expenditure.

One of the "Senior Truths" I live by is that consistency is a technical problem, not just a visual one. We leveraged the Innovaccer Proprietary Masala Design System (MDS), but because we were building on Power BI, we hit a wall. Power BI's native components are rigid.

I had to manage the "Design-to-Dev" friction by creating custom design overrides. I led the effort to create a hybrid component library that felt like our branded SaaS product but lived within the technical boundaries of a BI engine. This saved the engineering team weeks of rework and ensured scalability. 

This project reinforced a vital truth for my career in 2025: Design leadership is about creating the right environment for clarity. As a Product Design Manager, my success wasn't just measured by the final UI, but by the growth and alignment of the team behind it. Navigating a high-pressure, 8-month delivery cycle required a leadership style that balanced strategic shielding with tactical mentorship.

By establishing a robust design-to-dev handoff using our atomic-design approach, we didn't just save time—we built trust. We moved from a team that "draws screens" to a team that "solves business problems," which ultimately led to the successful adoption of the product by 11 major health systems.

By late 2024 and into 2025, the "AI hype" was everywhere. We integrated AI-driven Propensity Scoring to predict which patients were most likely to seek out-of-network care.

But as a manager, my biggest takeaway was human: Leadership is about alignment, not control. I learned that the most beautiful dashboard is worthless if the Data Analyst doesn't trust the "attribution logic" behind the numbers. My job was to design the trust as much as the interface.

If I were starting this project today with what I know about Agentic Design, I would move away from static alerts. Instead of a "Red/Amber/Green" status, I’d implement a Natural Language Summary. Instead of the user looking at a spike and asking "Why?", the system should have proactively said: "Costs are up 12% because Dr. Smith is using a non-contracted supplier for knee implants. Click here to draft a memo to his department." We were close, but the next evolution is definitely moving from "Analytics" to "Action."