Design and launch an MVP app to carry out clinical trials in the UK to obtain medical device approval.
Five months start to finish working closely with founders and dev team — remote project spread across four different time zones.
• UX research
• Interaction design
• Prototyping and usability testing
• Design library and developer handoff
Preliminary results indicate AutoDVT to be 30x faster than the existing clinical pathway without sacrificing accuracy.
People are affected each year
People die each year
cost to treat DVT each year
A non-specialist will use a small handheld ultrasound machine and the AutoDVT app to conduct compression ultrasound scans assisted by AI guidance for a provisional diagnosis.
The provisional diagnosis is sent to the cloud dashboard for review, where a radiologist can log in remotely and make a final diagnosis.
Medical professionals without formal ultrasound training need a way to diagnose DVT. Currently, a radiologist is required, but it’s expensive, time-consuming, and not always possible.
AutoDVT will provide an ultrasound software solution for non-specialists to detect DVT by harnessing AI-based guidance and diagnostics. We’ll know we’re right if our solution takes less than 15 minutes, has less than 1% false negatives, and 10% false positives.
Conducted 2 stakeholder interviews with founders and reviewed 10+ sources about DVT, diagnosis, and treatments.
Collectively created a problem statement, use cases, executive summary, and design principles to create a shared vision and understanding.
Designed a high-level flow chart with user actions, user decisions, and main screens.
Mapped out mid-fidelity wireflows with all the necessary screens to complete an ultrasound exam.
Created a hi-fi interactive prototype to conduct remote usability testing with 5 users to test satisfaction and performance.
I created a design library to keep the design consistent between the Android mobile app and the dashboard web app and ensure a seamless design handoff.
The design library was built in Sketch and was handed off to developers using Zeplin.