Healthcare Data Mapper
Making healthcare data interoperability scalable—4x faster mapping
The Challenge
Unusable prototype and unclear MVP blocked our ability to scale to more enterprise customers
My Role
Lead Product Designer, 2 years. End-to-end ownership from vision & strategy to implementation. Grew and led UX team of 4.
Impact
Transformed Data Mapper into scalable self-service product: 4x faster workflows, hundreds of millions in contract, double-digit customer growth
Data Mapping was a Blocker to Scale
“I thought it was impossible to meet the complexities of data mapping in a single tool.”
– Internal Data Analyst
Slow Manual Mapping Blocked Scale
Manual SQL methods took 2 years to map a single organization's data. Without acceleration, anything we wanted to do with the data was "dead in the water"
Fragmented & Messy Data
Patient records exist across systems in different formats. Unlocking insights requires transforming data into a consistent standard (FHIR)—called "mapping."
Unusable Engineering Prototype
The existing prototype had no coherent workflow—users couldn't complete end-to-end mapping.
Pivoted Roadmap from Features to Critical User Journeys
The Feature Factory Trap
The team built complete features in isolation (SQL joins, data preview). Users couldn't finish mappings—they needed workflows, not features.
I shifted the team from functionality to user outcomes
Conducted foundational research and recruited User Researcher
Led cross-functional design sprint mapping user workflows
Aligned team on journey-driven roadmap with clear priorities
Data Analysts could finally complete an end-to-end mapping
Designing for Scale: 4x Faster Mapping
Redesigned Information Architecture to Enable 2x Faster Workflows
Restructured mapping organization around user mental models
Replaced rigid FHIR resource organization with flexible user-defined concepts—enabling users to control data grouping for their specific downstream needs.
Built scalable templates that sped up mapping by 2x
Mappings for the same EHR schema were highly similar. Templates enabled "map once, reuse everywhere"—flagged early to build scalability into infrastructure.
“Templates simplifies mapping so you build it once and then you just scale.”
— Customer
Transformed Internal Tool into Customer-Ready Product
Data Preview: Understand data shape instantly
Shows unique values and null data at a glance—eliminating the need to run individual SQL queries.
“The automatic profiling is amazing. Checking distributions and values is something we spend hours manually doing today.”
- Customer Data Analyst
Mapping Tree: Guidance & Mapping
Built-in guidance shows what's required, so users can map and transform data confidently in one place.
Weekly Co-Design
Embedded with data analysts through weekly sessions, constant feedback on in-progress work, and completing full mappings myself to validate the experience
“Your work effectively turned Data Mapper from an internal eng tool to an external facing product”
– Senior Staff Eng Lead
Designing for Quality Trusted Outputs
Speed Was Useless Without Quality
A week-long validation delay forced painful context switching. By the time analysts discovered FHIR errors, they'd moved to other mappings and couldn't remember what they'd done—turning quick fixes into hours of debugging.
Identified Validation Opportunities
Ran 26-person sprint across 2 timezones to brainstorm validation solutions. Identified 6 focus areas and influenced prioritization of 4 validation initiatives on the product roadmap.
Mapping Valid FHIR is Hard
Like language translation, FHIR validation requires grammar (field rules), sentence structure (hierarchies), paragraph flow (dependencies), and accurate meaning (clinical context)
Validation & Mapping are Separate Mindsets
We assumed validation should work like spellcheck—always on. User research revealed mapping and validation are separate mindsets requiring user control: beginners want constant feedback, experts want focus.
Quality Built In: A Real-Time Validation Framework
Dedicated Spaces for Mapping vs Validation Mindsets
Let mappers decide when to switch to validation mindset
Be thoughtful on when and where to surface feedback
Helpful Feedback
Helpfulness = Severity + Actionability
Severity
Impact of error on data quality and downstream dependencies.
Informs when to show
→ High severity = Show immediately
→ Low severity = On demand
Actionability
How easily a user can fix an error.
Informs where to show
→ High actionability = Show near action
→ Low actionability = Surface in validation panel
Invalid Data Transformation
Severity: High—affects data validity. Show immediately.
Actionability: High. Show inline at join action to enable quick fix
Missing Required Field
Severity: Low. Acceptable temporarily. Show in validation panel.
Actionability: Low. May require data transformation/analysis first.
Enabling Self-Servive
Defining Self-Service Strategy
Led PM workshop to define ambiguous "self-service" goals—identified 4 problem themes that led to: dedicated PM hire and customer-focused roadmap
“We walked away with a clear view of our problem space and strategy gaps.”
– Product Management Lead
Created Onboarding Program That Identified Critical Skills Gap; Led to 3 New Hires
Partnered with UXR & Data Analyst to identify critical skills mismatch: partners assigned clinical experts lacking SQL, EHR, and FHIR knowledge. Onboarding workshops with the partner validated the gap—partner hired 3 dedicated employees.
Reduced support burden by simplifying release workflow
Co-led 4 rounds of iteration with 15 engineers and 3 PMs across 2 timezones—shifting from engineering-proposed to user-centered release process.
From Blocker to Accelerator: Data Mapping at Scale
“This is key to streamline workflows at scale & unlock deeper insights from healthcare data”
– TELUS Health
Product & User Outcomes
Enabled novice users to do expert work through templates, validation, and simplified workflows
Built scalable foundation that anticipated infrastructure needs years ahead
Reduced costly production errors through validation framework
Business Results
Enterprise-ready product supporting hundreds of millions in contracts and double-digit customer growth
4x faster workflows unblocked organizational scale
Customer Validation
HCA Healthcare: "Fundamental to our advanced analytics and Responsible AI initiatives"
Highmark Health: Enables us to provide "more personalized and proactive care to our members"