Insulet
Omnipod 5 Insulin Pump App Design

Overview

Insulet – a healthcare company specializing in tubeless insulin pump systems

Insulet is a healthcare company that develops tubeless insulin pump systems for people with diabetes. Its Omnipod product line replaces traditional insulin delivery methods with a wearable, waterproof device designed for everyday use.

Omnipod 5 is supported by a companion mobile application (iOS and Android) that allows users to manage insulin delivery and monitor key health data.

As the product expanded globally, the app needed to support multiple languages, accessibility requirements, and safety-critical interactions without introducing ambiguity or usability risks.

This project focused on improving UX structure, localization logic, and accessibility consistency in the mobile experience.

Research

Understanding accessibility needs in insulin pump users

To better understand usability barriers, we reviewed internal research and conducted usability sessions with participants representing a wide range of visual abilities.

Research highlighted a significant portion of insulin pump users living with visual impairments or vision-related conditions, directly affecting how they interact with medical interfaces in daily life.

To validate usability challenges, we conducted qualitative testing with:

• 11 participants

• 9 women, 2 men

• Ages 25–74

• All participants were either blind or had low vision

Key Insights

Accessibility was not a secondary scenario, but a core usage condition for a meaningful portion of users.

Key patterns included:

• Difficulty interpreting small or low-contrast UI elements

• Heavy reliance on screen readers or voice feedback for critical actions

• Challenges with abbreviated labels and unclear hierarchy

• Increased cognitive load when medical values lacked clear structure

• Strong dependency on predictable interaction patterns

The findings reinforced that accessibility limitations directly affect safety, not just usability.

Problem Statement

The mobile app was not designed for linguistic and accessibility variability at a global scale

As Omnipod 5 expanded into new markets, the interface revealed systemic UX limitations. These were not isolated translation issues, but structural problems in how content was designed, structured, and rendered across languages.

Rather than a single failure point, multiple recurring patterns exposed the same underlying issue: the system relied on English-centric assumptions and did not account for linguistic variability, accessibility needs, or layout flexibility.

Example 1: Variable-driven grammar breaks

The interface displayed “1 items selected”, which is grammatically incorrect in English and becomes even more problematic in languages with complex plural and declension rules.

Because the number was embedded within a fixed sentence structure, the phrase could not adapt correctly across languages. As a result, the entire expression became grammatically incorrect, reducing clarity and increasing cognitive load.

This was especially problematic for users relying on screen readers, where incorrect grammar directly impacts comprehension and trust in the interface.

Example 2: Unit break disrupting readability

Another recurring issue was unit abbreviations breaking into a second line (e.g. “mmol/L”), separating them from their associated values.

While the layout worked correctly in English (e.g. “mg/dL” remaining on a single line), longer or less flexible abbreviations in other languages caused unintended line breaks.

This disrupted visual hierarchy and made values harder to scan and interpret. Because units are essential for understanding medical data, even small inconsistencies introduced friction and increased the risk of misinterpretation.

Example 3: Text overlap with UI elements

Another issue observed was text overlapping with UI components. In this case, a longer German label extended into the space of the “i” (information) icon, visually colliding with it and reducing clarity.

While the layout functioned in English, longer words in other languages required more space than the interface allowed. Without support for text expansion, labels began to overlap, truncate, or spill into adjacent elements.

This created a cluttered and visually confusing experience, making it harder to distinguish between content and interactive elements—particularly critical in a medical context.

Synthesis

Across all examples, the root cause was consistent:

• Content was tightly coupled to fixed sentence structures

• UI layouts were not designed to adapt to text expansion

• Language-specific grammar rules were not supported

• Accessibility needs were not considered at a structural level

As a result, users were required to spend additional effort interpreting information, increasing cognitive load and introducing risk in a safety-critical environment.

The problem was not translation quality, but a UX system that did not support linguistic and accessibility variability at scale.

Bridge to Solution

Addressing these issues at the surface level would not scale. Fixing individual strings or screens would not resolve the underlying problem.

A system-level redesign was required—one that separates content from structure, supports linguistic variability, and ensures accessibility and layout scalability across all user scenarios.

Solution: System Redesign Approach

Instead of addressing each issue as a one-off fix, the redesign focused on creating a scalable system that could support multiple languages, accessibility needs, and safety-critical content.

The examples identified in the problem statement pointed to the same underlying issue: UI content, grammar logic, accessibility, and layout behavior were too tightly connected. To make the experience more reliable across markets, the system was redesigned across four interconnected layers:

• content structure

• linguistic logic

• accessibility

• layout scalability

Figma audit and redesign mapping

To support the redesign, I reviewed key screens and organized recurring issues into pattern groups inside Figma. This helped connect individual UI problems to broader system-level improvements.

The audit focused on:

• screens where content broke across languages

• repeated patterns with fixed layouts or fragile spacing

• labels, values, and units that required clearer structure

• flows where accessibility and reading order were especially important

• reusable components that needed to support longer localized text

This process helped move the work from isolated screen fixes toward a more consistent design system approach.

1. Content structure simplification

UI text was redesigned to separate meaning from fixed sentence structure.

One example was the selected item count. The original structure displayed the value inside a fixed phrase:

“1 items selected”

This created a grammar issue even in English and made the string difficult to adapt across languages with more complex plural and declension rules.

The improved structure separated the label from the variable:

Selected: 1 item

This made the message clearer, easier to scan, and more reliable across different word orders, plural forms, and assistive technologies.

2. Linguistic flexibility (i18n logic)

Language-specific content was redesigned to support locale-aware output instead of relying on static strings.

This included adapting:

• labels

• time formats

• date formats

• measurement units

• medical terminology

For example, the same event detail screen needed to work differently across locales. In English, values used formats such as 6:10 PM – 6:40 PM and mg/dL, while the German version required 18:10 – 18:40 and mmol/L.

These differences affected more than translation; they required structured i18n logic so medical information could remain clear, consistent, and understandable across supported regions.

3. Accessibility-first UX structure

The interface was redesigned for clarity in assistive contexts.

Key improvements included:

• replacing abbreviations with full semantic labels

• improving reading order for screen readers

• prioritizing critical medical information in hierarchy

• ensuring clarity under zoom and reflow conditions

This improved usability for visually impaired users and reduced interpretation errors.

4. Layout and scalability system

UI components were redesigned to support flexible content expansion.

Improvements included:

• support for longer translated strings

• consistent spacing and hierarchy across layouts

• reduced dependency on fixed-width structures

This ensured the interface remained stable across all supported languages.

Summary

Outcome & Impact

The redesigned UX system improved clarity, accessibility, and global scalability across the Omnipod 5 mobile app.

Improvements observed during validation and design review included:

• Reduced localization-related UI inconsistencies across languages and flows

• Improved task comprehension in accessibility testing sessions

• Fewer reported screen reader interpretation issues in usability feedback

• Increased structural consistency across global product interactions

The system enabled a more predictable and safe interaction model for a safety-critical medical product, particularly for users relying on assistive technologies.

What I Learned

UX structure is more important than translation quality in global products

• Language rules should be treated as system logic, not content fixes

• Accessibility must be designed into the system, not added later

• Designing for edge cases improves the entire product experience

• Medical UX requires precision at both language and interaction levels