Custom Content Builder within AthenaOne EHR
Led the end-to-end design of a configurable clinical assessment builder within the athenaOne EHR, enabling healthcare practices to create custom assessments supporting Value-Based Care, MIPS reporting, and alternative payment models.
Project Overview
Custom Health Assessments (CHAs) enable practices to create custom assessments for care monitoring and patient feedback collection tailored to their specialty within athenaOne. However, CHAs could collect patient responses but couldn’t score those responses or translate them into actionable risk signals to drive value-based care and care improvement strategies.
I led the design of a scoring experience that allowed practices to define scoring rules, interpret overall scores, and associate recommendations with specific risk levels, while keeping the experience simple enough for users without assessment-building expertise.
Role: Senior UX Designer · UX Researcher
Timeline: May–July 2026 · Team: PMs and a UX Designer
The Challenge
The core challenge was balancing flexibility with simplicity.
CHAs can be created by administrators or clinicians across different specialties, each with different assessment and scoring needs. At the same time, scoring involves complex rules that are typically configured through specialized internal tools.
The core question was:
How might we bring this complexity into a practice-facing experience without requiring users to understand the underlying scoring system?
Research & System Understanding
I first studied how athenahealth’s existing licensed screeners, such as the PHQ-9, GAD-7 are built and scored through our internal tooling.
This helped me understand the underlying scoring model, its rules, and the level of complexity involved.
I then contrasted that with the CHA user: a practice administrator or clinician who may have no prior experience creating assessment logic.
This led to a key design principle:
The product should carry the expertise the user doesn’t have.
Rather than exposing the existing system directly, I focused on translating its complexity into a guided configuration experience.
Design Strategy
This created a simple mental model:
Questions → Score → Meaning → Action
The approach also allowed the system to support different specialties without creating separate scoring experiences for each use case.
Key Design Decisions
- Guided workflow instead of one complex builder
- Scoring introduced a different type of cognitive task from writing questions.
- Separating the two helped users focus on one type of decision at a time.
- Adaptive scoring rules
Different question types require different scoring models. Multiple-choice questions use scores per response, while numerical questions use score ranges.
Visible progress and completion
A Not Started → In Progress → Complete status model helped users understand where they were in the process and what still needed attention.
Validation
A comparative usability study validating the scoring wizard's interaction model before development, using two prototype variants built with Claude
Research Objectives
-
Whether participants could complete the wizard unassisted, without disorientation
-
Which of two prototype variants produced a clearer, more efficient task flow
-
Where friction emerged in the interaction pattern — question authoring, score assignment, labeling, or guidance entry
Finding 1 - Structure
A consolidated two-step flow produced stronger task orientation than Prototype A's more segmented, three-step alternative.
Finding 2 - Sequencing
Participants were uncertain whether to complete question authoring before configuring scores — resolved with a single line of sequencing guidance at the entry point.
Finding 3 - Task Confidence
3 of 5 participants hesitated when defining score ranges, unsure how many were required — resolved with an explicit opt-out and a live indicator of what's still missing.
Business & Product Value
Scoring turns CHAs from custom data-collection forms into configurable assessment tools.
It expands the value of CHAs across specialties by allowing practices to:
• Define their own assessment logic.
• Translate responses into meaningful risk levels.
• Attach recommended actions to results.
• Reduce reliance on specialized internal configuration.
• Create assessments that better reflect their own clinical workflows.
The underlying model is reusable across use cases such as pain assessments, preeclampsia risk checks, PT referrals, and pain-interference tracking.
Impact
The feature closes a critical capability gap in CHAs:
collecting information → interpreting it → supporting action.
More importantly, the design establishes a scalable foundation for specialty-specific assessments without requiring a different scoring experience for every use case.
The research also gave us confidence that users could understand and configure the scoring model without needing to learn the underlying system.
Key Learning
My biggest takeaway was that configurability doesn’t have to mean complexity for the user.
The system can remain powerful underneath while the experience progressively reveals only the decisions users need to make.
Early testing was particularly valuable because it exposed gaps in the user’s mental model—not just usability issues. It reinforced the importance of validating the workflow and decision structure early, before investing heavily in the final UI.
What's next
Introduce specialty-specific templates so practices can start from proven assessment structures rather than a blank canvas, while retaining the ability to customize questions and scoring.