Blockchain PHR App Research
Measuring consumer trust in a personal health records management app via remote 1:1 usability tests for peer-reviewed publication
UX Researcher and Presentation Designer
Identified critical trust barriers and usability issues that informed design recommendations for peer-reviewed published study on consumer adoption of decentralized health data systems. Delivered heuristics evaluation and usability report, including SUS score and 6 prioritize recommendations with severity ratings.
-
Client: Dell Medical
-
Timeline: 2 months
-
Team: 4 UX Researchers
-
Impact: Informed design decisions to improve usability, peer-reviewed publication (Frontiers in Big Data, 2023)
THE SITUATION
Will users understand and trust our app enough to adopt it and benefit from it?
Conflict
Medilinker — a blockchain-based PHR app enabling patients to control and share their own health records — was preparing for development with no usability validation. The product's security architecture depended on users understanding and trusting blockchain ownership mechanics, but no evidence existed that target users could navigate core flows or grasp what made the system secure.
Cost of the conflict
The app's differentiation — patient-controlled, blockchain-verified health records — had no adoption value if users couldn't perceive the security benefit or complete key tasks like connecting to a clinic, uploading records, or sharing data with a provider.
THE STRUGGLE
Some users didn't trust the technology; We had limited access to participants; My original team was dissolved
Biggest challenge
Blockchain's security model was invisible in the UI — only 1 in 3 participants understood how it worked, making generic "this is secure" messaging insufficient. Trust also varied by data type: participants accepted sharing insurance information but resisted sharing government ID, meaning trust assumptions could not be applied uniformly across flows.
As a student-run study, the team was also constrained to convenience sampling — participants were primarily close connections, which limits generalizability and increases bias. This is a known tradeoff in academic research and was acknowledged in the study scope.
A failed attempt
The initial team assignment was not a functional match. The 2 teammates were new to UX and not pursuing it as a career path, creating an unbalanced workload that fell disproportionately on me. That team was dissolved, and I was reassigned to a different group.
THE SOLUTION
We conducted remote moderated usability testing with 12 participants, de-risking development by catching trust and navigation failures pre-build.
Breakthrough
Conducted remote moderated usability testing with 12 participants across 7 task flows, supplemented by SUS scoring and post-test interviews. Tasks were rated on a 5-point difficulty scale; the two hardest — disconnect from clinic (3.58/5) and send requested information (3.92/5) — mapped directly to the highest-friction UX patterns and drove 6 prioritized design recommendations.
IMPACT
Findings de-risked pre-build decisions enabling the product team to fix trust and navigation failures that would have driven user abandonment post-launch.
The usability insights we delivered means:
-
The development team received a prioritized list of 6 design recommendations tied to specific task failure points, giving them a clear action list rather than open-ended feedback to interpret.
-
The product team had evidence to justify UI changes to stakeholders based on observed participant behavior across 12 sessions.
-
The research team (UT Austin) had findings substantial enough to contribute to a peer-reviewed publication in Frontiers in Big Data (2023).
-
Future users were protected from a product that would have asked them to trust a system they couldn't understand — blockchain's security model was invisible in the UI and would have driven abandonment post-launch.
Why this matters
Usability failures in a trust-dependent product like Medilinker aren't recoverable post-launch — users who don't understand what they're consenting to don't come back. Catching these pre-build is where research has its highest leverage.
WHAT I DID
Designed the screener and script, recruited, moderated, took notes, analyzed the data, designed the read-out deck
My Tasks
-
Design the screener logic and help recruit
-
Help write the overall research plan and script
-
Prepare the screens for the study (designed provide by UX designers working with the product team)
-
Conduct 1:1 sessions, alternating between moderator and notetaker
-
Clean, organize and analyze the data
-
Convert the team's rough slide deck into a polished presentation
Delivered usability insights that:
-
Gave the Medilinker team a validated, evidence-based roadmap for UI/UX improvements before coding
-
Identified the 2 highest-friction flows so development effort could be prioritized
-
Established a usability baseline (SUS 78.95, task difficulty scores across 7 flows) that made follow-on studies measurable against a known starting point
REFLECTION
Team composition is a research variable, not just a logistical one. It directly affects what you're able to produce and at what depth.
Learnings
-
Lead with action items. Stakeholders need the "so what" upfront--consolidated recommendations at the top of reports, not buried after methodology.
-
Trust is a design problem, not a technology problem. Blockchain's security architecture meant nothing if users couldn't perceive it in the UI.
-
Task difficulty rankings reveal priority. The hardest tasks (disconnect, send info) pointed directly to where the experience needed the most work.
What I'd do differently
-
Expand recruitment beyond convenience sampling. Our participants skewed toward people in researchers' networks—limiting demographic diversity and potentially introducing bias.
-
Include target users earlier. Ideally test with actual patients managing chronic conditions across multiple providers, not just general consumers.
-
Add think-aloud protocol more consistently. Some sessions captured richer qualitative data than others based on moderation style.
