KidneyMate

AI-driven meal tracking and personalized support for kidney and diabetes management.

My Role

Research

Interviews

Wireframe

Design System

Prototype

Usability Test

Team

2 Product Designers

Tools

Figjam & Figma

Gemini 2.5

Chat GPT 5

Google Sheets

Procreate

Duration

Aug - Oct 2025

(12 Weeks)

Overview

Problem

Patients with irreversible kidney disease face a heavy burden managing complex dietary restrictions, especially when diabetes is also involved. While knowing how safe a meal is for their condition would be highly valuable, no existing app clearly provides this guidance.

How might we make chronic disease management easier and more consistent?

Main Feature

Solution

Building Confidence in Self-Managed Kidney Care with AI-Powered Nutrition Guidance.

Flow 1

Auto-logging dinner with before & after Food Scanner

Centered on an AI Food Scanner accessed from the Home CTA, Kidney Mate analyzes before & after-meal photos to estimate intake, calculate key nutrients, and deliver real-time feedback.

Flow 2

Reviewing past data and stored alerts

All food logs, AI insights, and key nutrients are automatically saved and displayed by date in a color-coded calendar, while medication reminders, dialysis schedules, and alerts are centralized on the Home screen.

Flow 3

Manually correcting AI-generated entries in Intake Info

To ensure accuracy, users can edit or add meals when needed. The Home meal log supports multiple items over time with editable food details, combining AI auto-logging with manual input.

Flow 4

Exploring recommended healthy meals and checking recipes

The Meal Plan provides a personalized weekly diet that resets every Monday and adapts to both AI and manually logged data, offering tailored meals, nutrient insights, and flexible meal options.

Flow 5

Discovering Healthy Dining Options Nearby

To support safe dining out, Explore Restaurants is added to the Meal Plan as a fixed CTA, allowing users to search and filter both kidney and diabetes friendly menus nearby for worry-free choices.

Flow 6

Viewing synced lab results and exploring the community

The Report syncs hospital data to display lab results by date through simple trend views and detailed charts, and summarizes daily nutrient intake. The Community offers peer support among patients.

Design Process

Empathize

What challenges do CKD patients face in daily self-management?

* CKD : Chronic Kidney Disease

Secondary Research

Understanding the Challenges of Kidney Failure Patients

To understand CKD patients’ challenges and unmet needs in existing health apps, we conducted secondary research including AI-driven market, trend, and technology analysis (Delft CKD AI Research, KHS 2025), along with competitive benchmarking of connected health and diet management apps.

Market research

content analysis (reddit, play store, app store)

Trend analysis (delft CKD AI research)

Trend analysis (technology research)

Benchmarking (kidney and diabetes management, diet, and fitness apps)

Emerging AI trends in healthcare at KHS 2025

Hypothesis from Secondary Research

Personalized diets are challenging

Patients have different conditions, making it hard to know how healthy their nutrient intake is for them.

Limited Supply of Kidney Health Apps

Despite strong demand, kidney patients lack specialized care apps compared to the abundant diet and diabetes app market.

1:1 In-Depth User Interviews

Deep dive into patient pain points in diet management and kidney apps

To deepen our understanding of dietary challenges and limitations of kidney apps identified in secondary research, we conducted anonymous phone interviews with 9 participants from a kidney disease community. We extracted 95 insights and synthesized them into 11 categories and five key themes through affinity mapping.

[Participate]

9 participants aged 30 – 60 : Dialysis patients (3), Kidney transplant patient (1), Diabetes patients (2), Caregiver (1), Renal-friendly meals startup founder (1), Dialysis unit medical staff (1)

Interview sessions with a dialysis nurse who consented to being recorded

Color-coded interview script

Categorizing interview insights through affinity mapping

Key Insights from 1:1 In-Depth User Interviews

Complex diets and challenging self-management

The complexity of managing care and dietary rules creates significant cognitive and emotional burden.

Lack of Personalized and Accessible Support

Current solutions lack personalized, user-friendly support, limiting long-term engagement, especially for older users.

Define

Defining the Core Self-Management Challenges for Kidney and Diabetes Patients

User Persona · Journey Map

Defining a Primary Persona for Kidney and Diabetes Care

Grounded in interview insights, we defined David Miller as a persona to prioritize kidney patients while supporting diabetes users, and used a journey map to identify key moments for meaningful app support.

“My diet keeps changing, and I’ll need to keep tracking everything even after transplant. I just want an easier way.”

David Miller (56)

Diagnosis : Chronic diabetes with end-stage renal disease (ESRD) on hemodialysis.

He manages both diabetes and chronic kidney disease, making dietary decisions complex and stressful, while strict restrictions and dialysis fatigue limit the value of single-disease health apps.

GOALS

· Maintain stable health until kidney transplant.

· Follow a diet suited for both diabetes and kidney disease.

FRUSTRATIONS

· Lacks tools that adapt food data for kidney disease.

· Needs a new diet balancing diabetes and kidney restrictions.

Problem statement · HMW

Reframing the Problem into a Design Question

Users managing both diabetes and chronic kidney disease lack clear, reliable, and integrated guidance to confidently manage complex dietary and health decisions long term.

How might we make chronic disease management easier and more consistent?

Ideate

Exploring Ways to Turn Data Logging into Personalized Nutrition Insights

Impact–Effort Matrix

Prioritizing Do-Now MVP Items for Key Kidney Patient Needs

Using team voting and an Impact–Effort matrix, we selected Do-Now items that were most needed by kidney patients and quick to resolve to support an efficient MVP phase.

Scored issues and a 2×2 priority matrix

Navigation Concept

Shaping Navigation Strategy Grounded in Patient Experience

Based on the prioritized MVP problems, we translated David’s dietary challenges into a usable experience and designed navigation around his core behavior flow.

Navigation planning with key features and reference screens

Color indicates priority : primary and secondary strategies

MVP Information Architecture

Optimizing AI Camera Navigation Through Information Architecture

To translate the strategy into user experience, we restructured Kidney Mate’s information architecture. By prioritizing the AI Camera as the primary CTA, we reduced navigation complexity and improved access to core MVP features.

Structuring the information architecture

Task Flow

Designing a Flow to Validate Automatic Meal Nutrition Input

To user-test personalized nutrition insights, we designed an MVP task flow centered on log → understand → adjust → apply, connecting AI meal feedback with log review, editing, and meal checks for everyday use.

Task flow 1 - Home (ai food camera → past meal log → manually edit)

Task flow 2 - Meal plan (browse weekly meal plan → get location-based dining)

Sketch

Designing a Flow to Validate Automatic Meal Nutrition Input

We sketched the home screen to validate core features, visualizing key nutrients and fluid tracking with familiar, mental-model-aligned interactions.

Home screen sketches

Prototype

Defining the Core Self-Management Challenges for Kidney and Diabetes Patients

Wireframe

Testing Wireframe Layouts to Reduce Cognitive Load

During sketch-based wireframing, we tested layouts to improve readability and reduce cognitive load by prioritizing key data and using selective interactions.

Wireframe (Home - green, Report - blue)

Color System

Exploring Color Combinations to Improve Information Recognition

To reduce patient burden, we refined the clinical green to a warmer lime, improved contrast for clarity and accessibility, and defined color tokens.

Adjusting color contrast and system tokens for accessibility

Typography

Selecting Inter font to Ensure Superior Data Legibility

To ensure accuracy and trust in a data-driven medical app, we selected Inter for its strong numeric legibility, consistent rendering across iOS and Android, and multilingual support over alternatives like Roboto and SF Pro fonts.

Typography System

Design Token & Component System

Building a Scalable Design System

We systemized colors and typography with tokens, modularized components into a reusable library, and ensured scalable UI consistency and efficient collaboration.

Design tokens for light, dark mode and shared components

Prototype

Completing Prototypes for MVP Validation

After defining the design guidelines for the Kidney Mate MVP, we created a prototype to validate the effective use of the AI Food Camera and health and nutrition data.

Prototype session

Test

Refining Ideas Through Iteration to Improve Clarity, Readability, and Accessibility

Usability test

UT & SUS Results Validating the MVP’s Value for Kidney Patients

To evaluate Kidney Mate’s long-term value, we conducted task-based usability testing with both kidney patients and kidney-care–aware non-patients. Due to ethical and regulatory constraints, non-patient participants complemented patient testing. An average SUS score of 80 showed tasks were generally completable, while confusion around the after-meal flow and manual input pointed to areas for further iteration.

[Participate]

6 participants including chronic patients and general users.

[Session Length]

60 minutes per participant

[Task]

Home (Auto-logging dinner with AI Food Scanner)

Logging meals to assess potassium/phosphorus intake and daily dietary alignment.

Home (Reviewing past data and stored alerts)

Reviewing diet data and feedback from the past two days, and checking accumulated alerts.

Home (Manually correcting in Intake Info)

Identifying an incorrect breakfast log, shortening the title, and adding a missing ingredient.

Meal Plan (Exploring recommended 28 meals)

Reviewing this week’s kidney-friendly meal plan and viewing a recipe in the meal details.

Meal Plan (Explore nearby healthy restaurants)

Searching for kidney-friendly fish dishes nearby and opening a saved restaurant.

Report, Community, More

Viewing lab results in list and range-bar views, then exploring the Community and More.

[Post-Task Questionnaire]

· How did you feel when viewing the real-time nutrient dashboard?

· Did using the AI Food Camera before meals feel easy and intuitive?

· After scanning your meal, was it easy to understand its impact on your health at a glance?

· Does logging meals before and after increase your confidence in the AI’s data accuracy?

· As a kidney patient, how likely are you to continue using this app regularly?

Raw task completion data for 6 user (2025-10-14)

Task-based error analysis synthesizing key usability issues

SUS result for 6 user

Iteration

Prioritizing and Iterating on MVP Issues Using a 2×2 Matrix

To improve MVP usability, we prioritized key user-testing issues using a 2×2 impact–effort matrix, iterated on home alerts, dashboard visualization, and the food scanner flow, and confirmed the need for ongoing validation.

Scoring Key Pain Points by Desirability & Feasibility (5-Point Scale)

Iteration Priority Matrix (Yellow - Immediate, Green - Secondary)

Home dashboard & alarm

The dashboard was simplified to show key kidney nutrients using green for normal and red for risk, with improved alerts.

Intake info (when eating multiple foods over time)

The nutrient intake page was improved for clarity by reducing header space and enhancing button affordance.

After eat scanner camera mode

Combining food and barcode scanning in the camera caused cognitive overload, so we simplified it to a single core function.

After eat scanner before confirm

The post-meal screen shows before-and-after photos together, removing the unnecessary retake button to reduce confusion.

Screens updated through post-usability-test iteration

Reflection

Clarity Over Information Density

Usability testing showed that too much information causes cognitive overload, while prioritizing essential content creates a clearer, more usable experience.

Design Beyond Assumptions

Observing real user behavior revealed interaction issues beyond our assumptions, reinforcing that testing and iteration are essential for usability, especially for older users.