Discovery
Operating within a lean startup environment with limited recruitment resources, I used convenience sampling to kickstart our discovery phase with 5 users. While I recognized the limitations of this sample, it allowed us to identify low-hanging fruit usability issues and high-level mental models without stalling project momentum.
Finding 1
Users told me that seeing a single lab result doesn't tell them if they are getting better or worse. They need to see the trend over time to understand the "why."
Finding 2
People hate logging into multiple hospital websites to find one piece of information. Most give up because it’s too much effort.
Finding 3
Users worry that their medical history is "gone" when they move to different cities or change doctors.
Collaborating with the team, we prioritized four foundational ideas to balance technical constraints with user needs. We focused on two primary challenges: Effortless syncing and Clarifying results.
Users wanted all their medical history in a single, secure place. I focused on building Visual Trust and managing the technical complexity of syncing data from multiple hospital portals.
The single health connector
A secure, one-time "handshake" that links multiple hospital portals. By authenticating once, users bridge the gap between fragmented systems and eliminate the need for repetitive logins.
Privacy by choice
A control center that gives users total control over their records. Instead of a permanent "all-or-nothing" sync, users can toggle specific data types (e.g., "Lab Results only") and revoke access to any provider at any time.
Users struggled to find meaning in decades of scattered records. I focused on designing a dashboard that turned raw data into Clear, Readable Trends so users could finally see the "Big Picture" of their health.
Pinned favorites
A customizable summary at the top of the dashboard. This allows users to prioritize their most critical metrics—like blood pressure—for an instant status check without searching.
Comprehensive health trend
Color-coded feedback and reference ranges that explain what the numbers actually mean. This allows users to understand their results in context by seeing how they track across their entire health trend.
Initial design
Based on our initial ideation, we mocked up the Adding Connections flow and a Manual Dashboard. Our goal was to streamline the data-linking process while giving users total agency over their most critical health trends.
This initial flow focused on a high-trust, linear path to bridge 20 years of medical history. We prioritized transparency and granular control to ensure users felt secure before initiating a deep-data sync.
To manage a high-density clinical library without cognitive overload, I designed a system that allows users to pin any metric to a 'Hero' position, enabling them to view critical health trends at a glance.
These small cards provide a quick overviews of the health data. I focused on making the numbers easy to read and the categories easy to find, so users don't have to dig through complicated menus to see their latest results.
Challenges & Iteration
Design is rarely a straight line. After our initial V1 concepts, we received critical feedback from usability testing and latest technical specs from our engineering team. These three insights forced us to move past "ideal" designs and solve for real-world constraints.
The Problem: We designed the app for a quick 10-second sync, but the hospital servers actually took 5 minutes to send the data. Asking a user to stare at a loading screen for five minutes was unrealistic and would have caused most people to give up.
The Fix: I moved to a "fire and forget" model. Now, users trigger the update and can immediately close the app. I implemented a push notification to alert them the moment their data is ready, turning a technical lag into a respectful, non-intrusive experience.
Message informing users this is gonna take 5 mins
When the sync is completed, they will get a notification.
The Problem: V1 used large square cards and prominent sparklines to create a "breathable" aesthetic. However, usability testing revealed that users found the layout "too spaced out", leading to interaction fatigue during extended browsing sessions.
The Fix: I transitioned to Condensed Horizontal Tiles to prioritize comparative scannability. By removing the read more links at the dashboard level, I increased information density by 2x. This allows users to view significantly more data above the fold.
The V1 square card layout suffered from poor vertical space efficiency
The new horizontal cards is much more efficient when browsing
The Problem: We originally designed for a standard profile of 50 biomarkers, but testing with power users managing multiple chronic conditions revealed libraries of 200+ metrics. Our "show-all" layout quickly collapsed under this volume, turning a streamlined dashboard into a source of major cognitive overload.
The Fix: I defined a new logic to implement progressive disclosure. Instead of a full list, we now show the 20 most recently updated cards. For deeper analysis, I introduced a "View All" entry point paired with search and filters to ensure the dashboard remains clean while providing power users with seamless access to their entire health history.
Have a 'view all' button after the initial 20 cards
Search and filters make it much easier to find a specific card
The Problem: Users found the carousel confusing and tiring. Most didn't even realize there were more cards to see. Because they couldn't tell what was coming next, they usually just stopped at the first one.
The Fix: I replaced the carosel with a simple vertical list and added an AI option to fill it in automatically. By letting users just scroll down naturally and moving the extra date settings to a deeper page, I made the experience much more intuitive where the most important health info is never hidden.
Empty state favorites
Vertical orientation of favorites for improved scannability
Final results
After iterating, I ran a final validation round with 4 participants to test the new changes. Here are the findings:
Faster Browsing: By replacing inefficient square cards with a condensed horizontal layout, I doubled the information visible "above the fold." Combined with a continuous vertical feed, this eliminated the need to swipe left and right, which significantly speeds up task completion.
Solving Data Overload: I moved the exhaustive list of biomarkers to a dedicated "All Results" page with search and filters. This cleaned the primary dashboard of visual noise, allowing users to focus on key health trends without feeling overwhelmed.
Decreased Drop Rate: Introducing a "Permission to Leave" screen successfully neutralized the anxiety of the "active wait." By removing the pressure to stay on a loading screen, we saw a dramatic decrease in abandoned sync sessions as users felt more in control.
The Impact
Final design
Add new connections
Add new connections screen
Connections onboarding
Set sharing permissions
Syncing info screen
Sync complete notification
Dashboard favorites
Empty dashboard favorites
Manually add to favorites
Favorites added, vertically oriented for rapid browsing.
All results
All results section with space-efficient horizontal cards for streamlined result browsing.
Search and filters make it much easier to find a specific card
Individual result detail page
Detailed result analysis featuring a date range picker and summary grid for historical context.
Hover state for a datapoint on the chart
