NYU Langone Health

Overview

I worked as the UX designer for the HiBRID Lab, which sits inside NYU Langone Health as an innovation lab, enhancing the patient and provider experience with AI-driven clinical support tools.


The following projects I worked on is under NDA. To access these projects, write down the code from my resume.

Role

UX Design

Usability Testing

Research

Tools

Figma

Claude Code


Team

Dr. Safiya Richardson

Dr. Adam Berman

April Chien (UX Designer)

Nicole Redfern (Project Coordinator)

Angela Mastrianni

Lynn Xu

Timeline

June 2026 - Present

Projects

I was started on a couple projects that

PROTOTYPE

Clinical Decision Support Tool

Decision support tool for PCPs

SYNTHESIS COMPLETE

Statin-Prescribing Nudges

Epic usability research, five behavioral themes

IN RESEARCH

Cardiology Chatbot

AI Brain, CPCVD patient journey tool

SHIPPED

Content Automation

Airtable, Claude, and Netlify automation

Clinical Support Tool

After running user interviews with cardiologists and primary care providers, we prototyped and ideated with live interactive dashboard.


Context

ASCVD is the leading cause of death in the United States, yet more than half of individuals at elevated risk remain untreated. This gap persists not because the relevant data are absent, but because clinically meaningful risk signals are fragmented across the EHR, rarely synthesized during time-limited encounters, and infrequently linked to actionable next steps.


Problem Statement

How can we consolidate EPIC's data of a patient's history and current vitals into a comprehensive and actionable dashboard to prescribe statins.


Providers struggle to gather and synthesize cardiovascular risk data from fragmented EMR systems — leading to time-consuming workflows and gaps in preventive care.


Second Goal: How can we help providers become more familiar with additional risk-factors in cardiovascular health.
 

Pain points:

  1. Basic data buried under clicks: LDL, calcium scores, and imaging results each reqquire multiple navigation steps to locate

  2. Risk enhancers hidden in scattered patient hisstory: family history, inflammatory markers, and prior findings sit eep in medical history and not surfaced by the EMR

  3. External resutl hard to navigate : Imaging and labs performed outside the system are difficult to locate and fold into the risk picture.




User research was the first step to restructuring the portal. Conducting a multiple-phased test, we landed on a new IA that improved findability significantly.

The Research

Phase 1: Mobile Menu Testing

Designed four distinct menu layouts exploring different hierarchies, groupings, and visual patterns. Each reflected a unique structural hypothesis.

Design C received the highest ease-of-navigation score at 82%.

Option A

Option B

Option C

Option D

Phase 2: Homepage Redesign Validation

Designed and tested three homepage concepts with distinct focuses to determine which layout best supports member’s needs.

The Care-focused homepage won across every measure. 89% found it easy to navigate, 95% said the layout made sense, and 78% felt empowered to manage their care independently.

Benefits Focus

Financial Focus

Care Focus

Design Solutions

Surfacing Items in Multiple Places

Features like 'View Claims' fit multiple categories. Rather than forcing one location, we surfaced them wherever relevant.

Relabeling for Clarity

  • "Coverage & Benefits" clarifies what lives within

  • "Forms & Documents" centralizes downloadable materials

  • "My Account" aligns with common digital conventions

The redesigned IA was officially launched in April 2024

Phase 2 testing was clear: members came to the portal to find care first. So Find Doctors & Care became the first item in the navigation, not buried in the middle.

Impact

“This homepage would make me feel like my health insurance provider genuinely wants to empower me to manage my health care.”

Usability Test Participant

26%

Ease of finding information (35% → 61%)

16%

Self-service capability (44% → 60%)

16%

Overall member satisfaction (54% → 70%)

Second Intervention

Aug 2024 – Dec 2024 · Launched Jan 2025

Closing the Coverage Gap

Every January, hundreds of members with expired plans couldn't reach their documents without calling customer service. The team manually emailed sensitive PHI to each one.


In August, leadership set a December deadline to find a solution. MGBHP's legacy design system added a hard constraint: no rewrite, just what already existed.

Design Solutions

Stripped to what they actually need

Former members need one thing: their tax documents. A status banner surfaced it immediately. Navigation cut from 25 items to 12. Visual design kept identical to the active portal — nothing new to learn.

Mobile-Web First

Rebuilding the native app under a December deadline wasn't realistic. We redirected former members to the mobile-web portal instead – same access, no rewrite, shipped on time.   

Socializing the Change

I wrote quarterly release guides for the CS team – plain-language summaries of what changed and why – so they could answer member questions without escalating every call.   

Impact

The Post-Effective Portal was officially launched in January 2025

“You should have seen the Teams screen today 'light up' with hearts and applaud emojis when they demonstrated the post-effective member experience enhancement. It has made an impact and less activity to CS as a result :)”

Customer Service Manager

45%

Reduction in tax requests (400+ → 222)

ZERO

PHI email transfers (down from 400+)

Reflections

Users don't think in org charts

The old navigation mirrored internal structure. Rebuilding it around how members think about care drove the biggest gain in findability.

Security is a UX problem

Manually emailing PHI was not only a support burden, it was a security risk and a broken experience. Solving both at once made the case for design as a business function.

The data was already in the room

The 2023 survey held answers before the IA project started. The CS team held answers before the former-member project started. Treating existing data and frontline knowledge as research, not background, reshaped discovery.

Small changes, big results

The visual edits for former members were minimal. The impact was not. A focused portal cut call volume and freed the team for higher-value work.