Imagine being a parent in a hospital where you can't understand the language spoken.

The anxiety of not comprehending medical instructions, the fear of miscommunication about your child's care, this is the reality for many non-English speaking families in US hospitals. Healthcare providers, too, face challenges. Relying on video interpreters can delay care, especially during brief interactions like checking vitals or providing post-operative instructions. These delays can lead to frustration and decreased efficiency.

The Vocalis (Low English Proficiency) app aims to bridge this communication gap, facilitating quicker and clearer interactions between nurses and patients/caregivers who don't speak English.

This project was piloted across Emory Healthcare and Children's Healthcare of Atlanta, two of Georgia’s largest hospital systems. The tool supports communication in 9+ languages across inpatient and outpatient settings.

How might we redesign the frontlines of care to support language equity, without overburdening nurses or replacing interpreters?

This project reimagines translation not as a utility, but as a strategic bridge in the patient-provider relationship.

Vocalis:
A Low English Proficiency Application

TLDR
We built a multilingual, context-aware translation tool to support nurses and caregivers in communicating with LEP patients—reducing reliance on interpreters, improving patient trust, and saving costs.

My Role
Strategic Designer leading user research, interface design, and cross-stakeholder alignment with medical staff, translators, and patients.

Motivation behind LEP

Increase communication from hospital staff towards patients
Improve patient’s experience at the hospital
Improve the experience of healthcare providers, which was previously negative with video interpreters and led to delays in care.
Reduce costs and utilization of video/audio interpreter services for short (1 - 2 min) interactions

The app is not meant to replace in-person interpreters. Rather, it is intended to augment conversation experiences where interpreters are unavailable.

Want a more detailed run down of this project?

Part I

Part II

Mapping the Patient Journey & Communication Breakdown

We began by tracing the end-to-end journey of a patient, from check-in to checkout. Each step revealed multiple touchpoints where communication failures could occur, from demographic verification to IV procedures to pain assessments.

We discovered that:

Interpreters are often unavailable during short but critical moments, like administering medication or giving discharge instructions.
Nurses were improvising, using Google Translate (despite non-use rule), gestures, or just skipping non-urgent dialogue.
"Patients are used to the idea of using Google Translate, so they find the usage of the app convenient. Nurses themselves pull out their personal phones sometimes to use G Translate." - Nurse Practitioner, Day surgery, CHOA
LEP patients often relied on children or relatives to interpret, even for sensitive procedures.

This journey map became a systems-level diagnostic tool, helping us identify points where lightweight, contextual translation could reduce harm and improve trust.

Understanding Interpreters, Not Replacing Them

We interviewed professional interpreters and discovered key tension points:

Emotional care often depends on tone and eye contact, hard to deliver over a video call and audio translation tools.
A small mistake in interpretation will trigger a big mistrust response (i.e. asking for, “how is my son?” and the doctor replies “my son is good” instead of your son). She lost faith in the system very fast. - Patient’s Mother/ caregiver
Many interactions last less than 2 minutes but require waiting 15+ minutes for a call to connect.
Gender dynamics, cultural nuance, and literacy levels require more than literal translation.

Strategic Language Curation

We developed and curated a core library of high-utility phrases based on actual nurse-patient conversations. For instance:

Pain assessments (e.g., “Are you in pain now?”)
Procedural clarifications (e.g., “I’m inserting an IV”)
Emotional reassurance (e.g., “I'm getting help”)

Through Pendo analytics, we tracked which phrases were used most and by whom:

“Are you in pain now?” was the most clicked phrase
80% of interactions involved nurses in inpatient care
Frequent use of the “Understand” button signaled uncertainty or discomfort, helping us optimize content

Top Phrases

Designing for Adoption: From Language to Trust

In a hospital setting, every second counts and so does every word. We knew that for this tool to be genuinely useful, it couldn’t just translate words. It had to translate intent, urgency, and empathy, and do so in a way that seamlessly integrated into the nurses’ existing workflows.

That’s why our design focus went beyond static translation. We built features that adapted to real-time contexts, supported nurses under pressure, and gave LEP patients more than a passive role in their care.

1. Language Change Functionality

Hospitals are multilingual environments. With high patient volume and the fact that patients may switch between caregivers, depending on who accompanies them. Our language toggle system made it easy to: switch languages between patients and appointments. We tested how this influenced usage and found that language flexibility reduced nurse hesitation, encouraging more frequent use in mixed-language settings.

Enabling seamless language switching increased nurse confidence and patient comfort, reducing the friction that often stalls adoption of translation tools.

Design iterations: Language change functionality

2. Onboarding for a Clinical Environment

Traditional onboarding didn’t cut it. We designed a clinical-context onboarding flow that:

Aligned with hospital workflows (e.g., shift-change timing, device handover)
Used scenario-based walkthroughs instead of generic tutorials
Trained nurses on edge cases (e.g., what to do when translation fails, how to verify understanding)

Onboarding was framed as a collaboration tool, not another platform to master, paving the way for long-term adoption across staff rotations and new hospital units.

Onboarding screens

3. Homepage Curation

Rather than overwhelming users with hundreds of phrases, we curated two pages. We surfaced top-used phrases based on role and location, turning passive interface space into actionable, adaptive navigation. The homepage could be further curated by the practitioner based on their usage. The second page was designed with task-based clusters, designed through contextual inquiry with frontline staff. For example:

“Vitals & Pain” → checking temperature, asking about pain
“Procedures” → drawing blood, inserting an IV
“Comfort & Reassurance” → calming statements, body position guidance

Curation was not just a UX improvement, it was a trust accelerator, reducing cognitive load and enabling faster task execution during high-pressure moments.

Personalizing homepage phrases

Task based cards

4. Active Translation Feature

Hospital interactions are dynamic. To account for this, we introduced Active Translation, enabling nurses to input custom sentences on the fly and receive voice outputs in the patient’s chosen language.

This was especially valuable during:

Discharge instructions
Medication clarifications
Patient-specific updates

Rather than scripting care, Active Translation allowed the app to follow the rhythm of real conversations, making it a true extension of the nurse’s communication toolkit.

Active Translation feature

Results

Reduced interpreter demand for <2 min interactions, freeing them for more complex needs
Improved nurse experience, reducing frustration and guesswork
Patients signaled understanding more confidently, especially in routine care steps

We're currently working on launching the app on AppStore as we continue research to evaluate long-term impact on readmissions, patient satisfaction, and clinician burnout.

Reflections

Design is not just about what we make, it’s about what we make possible. By designing with frontline workers and marginalized patients, I helped build a bridge across linguistic and emotional divides, without overburdening the healthcare system.