Mediator

A humanized, digital translation system that provides a more personalized communication experience for medical patients with limited English proficiency by addressing not just their linguistic, but also emotional and cultural needs.

SKILLS:

UI/ UX

Human-Centered Design

System Thinking

User + Stakeholder Research

User Journey Mapping

UX Scenario Building

Information Architecture

User Testing

Iterative Ideating

Wireframing

Interactive Prototyping (Figma)

+ my role

TYPE:

Team (of 3)

TIMELINE:

8 weeks

Anchor 1

BACKGROUND

US federal law requires all healthcare organizations to provide services in a language that patients with Limited English Proficiency (LEP) understand. Despite this, healthcare systems have not adapted to best serve this population.

Language and cultural barriers contribute to extensive health disparities with higher rates of medical errors and worse clinical outcomes amongst non-English speaking communities.

What systems are currently in place?

While professional medical interpreters do exist, they are difficult to access everywhere and for everyone. Consequently, most LEP patients rely on digital translators or Ad Hoc translators (bilingual family, friends, or staff), but many challenges come with this.

We decided to investigate these challenges LEP patients face within current healthcare systems with further user and stakeholder research.

USER RESEARCH + INSIGHTS

Anchor 2

Nine in-depth user interviews were conducted with users of various levels of English proficiency.

Our interview questions focused on understanding LEP patients' emotional, cultural, and linguistic pain points with the quality of their medical visits.

Two detailed stakeholder interviews with healthcare providers from smaller and larger-scale settings helped us understand their activity flow and challenges in providing care for LEP patients.

Anchor 3

IDENTIFYING DESIGN REQUIREMENTS

Responding to the main problems identified above, we defined some requirements to guide the direction of our design:

1

A digital translation system, but humanized.

Both parties agree this provides quicker and more objective translation. Now, add the human touch it lacks.

2

Account for varying levels of English proficiency.

This influences the level of help they require: from standard, spoken English to complex, medical jargon.

3

Account for the doctors side of the conversation.

Bridge awkward communication gaps by making translation transparent to the doctor, and fit doctor's current activity flow.

4

Address emotional needs.

Decrease the nervousness, embarassment, and discomfort LEP patients feel in the tense medical setting.

5

Address cultural differences.

Some LEP patients are immigrants. Contextualize culture-specific concepts.

6

Provide post-visit support.

Provide a translated visit summary to address feeling lost in translation, and enable translation of doctor-provided documents.

Therefore, our project goal became:

How might we provide a more effective and personalized communication experience for LEP patients during medical visits, while addressing not just their linguistic, but also emotional and cultural needs?

THE DESIGN SOLUTION

Anchor 4

Here is a short video demonstrating and explaining our final design within a user scenario:

Here is a shorter breakdown of the final design presented above:

1. ONBOARDING + USER PERSONALIZATION

The help you need, exactly the way you want it.

The user can tailor their translator to specific preferences by choosing:

A translation level that adapts translation functions and UI to their fluency level (click here for feature details.)

An avatar that represents their cultural identity.

The translator's pitch and pace, adjusted to personal comfort.

2. STARTING THE CONVERSATION

One step towards building a doctor-patient connection.

LEP patients often miss chances for routine conversation to build doctor-patient rapport while caught up in translation.

This feature helps LEP patients start the conversation on the right foot.

(doctor's side of the screen flipped for English readability)

3. TRANSLATION FEATURES

Humanizing technology.

*Natural Language Processing (NLP) technology plays a critical role in our design.

NLP utilizes AI to respond to text or voice data in the same way humans do, taking irregularities in human language into account.

It also allows for 'Named Entity Recognition', to categorically identify words as useful entities. This enables the following functions:

a) Decoding Difficult Medical Jargon

NLP technology detects medical terms to generate infographics that supplement the doctor's explanations.

b) Cultural Diets

Addresses cross-cultural barriers in making diet recommendations.

c) Prescriptions

De-mistifies medicines to help the user feel at ease with unfamiliar names, that may only be popular regionally.

Infographics are supplemented with contextual information about any personal history with that medicine.

d) Conversation Guides + Reminders

Helps LEP patients who forget to ask their doctor critical questions when nervous.

Even post-visit, we've got your back.

Utilizing NLP tasks, a summary of all important information from that visit is generated and stored for post-visit reference.

This clarifies any lost information, alongside the medical jargon explanations excluded from clinic-provided visit reports.

Available in toggleable languages for bilingual patients.

The user can pause summary recording during translation for any sensitive information.

With Neural Machine Translation, any medical documents received during the visit can be translated and stored.

4. TRANSLATED SUMMARY + DOCUMENT UPLOAD

Anchor 5

ADJUSTABLE TRANSLATION LEVEL

To provide more detail on the customizable translation level enabled by this NLP technology, two user journeys are demonstrated here. Users can choose a translation level that responds to their fluency level, and the system UI and functions adapt to their choice.

1 For more English-proficient patients (eg. biligual patients) who only have difficulty with medical jargon:

The screen displays this side-by-side UI to occasionally refer to, and focuses on the translation of only complex medical terms.

The screen displays this face-to-face UI for both the doctor and patient to use, with all conversation details translated.

2 For less English-proficient patients:

The patient can "adjust the translation level" at any point during the conversation if they require more or less detailed translation assistance than anticipated.

Anchor 6

IDEATION

Information Architcture- Language UX-Ver 0.drawio (2).png

Version 1 of user flow

Anchor 7

ITERATING + USER TESTING

Our user research launched a 5-week iterative process of designing a system that catered to these user needs.
This project was developed in three larger phases, with two major rounds of user testing in between them:

1. Feature Testing Cards / Wireframes

We used these cards to test different conceptualized features and identify areas we could better account for patients' emotional and cultural needs.
(Each picture in this section can be clicked into to view details.)

2. Low Fidelity Prototype

From our first round of user testing, we realized more consideration of the context in which users were using our product was required, and streamlined our ideas accordingly.

Our second round of user testing then focused on making this experience more intuitive with a simplified user flow.

3. High Fidelity Prototype / Final Design

Here is an overview of the complete user testing process:

FINAL UX BLUEPRINT

Anchor 8

( click on image to expand )

Anchor 9

THE VISUAL DESIGN SYSTEM

Typeface:

Professional and easy to read.

Visual Hierarchy:

Enabled with text size and weight.

Color Palette:

Calming and aligns with palettes associated with medical contexts to establish user trust in the translation services provided.

Occasional bright pops encourage positivity in the relatively tense medical setting.

Iconography:

Minimalistic line style UI icons.

All medical icons carry a 3D rendered style to add visual simplification to text-heavy screens.

App Branding:

Takes on wordplay.

Anchor 10

FUTURE APPLICATION

While this system has currently been developed for the medical setting, it could be adapted to fit a wide range of contexts that LEP users struggle in, be it with complex jargon or cultural barriers (eg. academic seminars, business meetings, etc.)

At its core, this is a system that emphasizes user empathy.

Anchor 11

REFLECTION

My top 3 learning takeaways from this project:

1

Evaluate the opportunity space.

I understood how important it is to identify and target specific areas within a problem space that require the most attention.
A product that capitalizes on these as opportunities for the greatest impact and innovation will be stronger than a product without focus that tries to address a little bit of everything.

2

Remember the larger picture too.

I learned to consider how my product sits within the larger context it is designed for to create a seamless experience.
It is essential for the product to not only smoothly fall in place with the user's own habits, but also align with the activity flow of other individuals within an environment to avoid unwanted disruptions.

3

Leverage user testing.

I learned how to use user feedback strategically.
Really listening to my users was incredibly important to create a product that with user-driven reprioritization, truly addresses their needs. At the same time, I learned to make smart choices about what feedback to implement. A highly custom requirement may not align with the project scope and mission.