Case Study · Spectral MD, Inc.
DeepView AI® predicts wound healing outcomes — but complex workflows and uninterpretable results were blocking adoption. I led end-to-end UX redesign to make life-critical AI fast, clear, and trusted in clinical environments.
Overview
Problem
Approach
+
Hospitals in clinical studies across UK & US
Participants in formative usability study
Real users in summative validation — passed
Product lines unified in one design system
Core challenges
Field research across 6+ hospitals revealed the same friction points repeating across every clinical team we interviewed.
Workflow & Speed
Complex clinical workflows resulted in long patient wait times. The system was not adaptable for mass casualty events or emergency rooms.
Design response
Quick Scan mode + EHR auto-pull eliminates registration steps in emergencies, enabling scan-and-go in under 60 seconds.
AI Trust & Clarity
AI-generated outputs were complex and difficult to interpret. Without understanding the results, physicians were hesitant to act on predictions.
Design response
Enhanced AI report with visual overlays, TBSA data, fluid resuscitation guidance, and plain-language explanations.
Brand & Consistency
No unified design system caused visual inconsistencies across product lines, impacting brand cohesion and slowing design-to-development handoff.
Design response
Built a complete Figma design system — tokens, typography, components — deployed across all 4 product lines.
Research
Understanding complex clinical workflows across diverse hospital systems is no easy task. We visited 6+ hospitals nationwide, shadowing staff for full-day observations and conducting interviews with physicians, nurses, and IT teams. We also attended major conferences — AAEM, ABA, SRBC, EMS — to understand the broader market landscape.
+
Hospitals visited
Conferences attended
Usability participants
Workflow
"Time is precious in hospitals. Repeated data entry is the most frustrating part of this whole process."
AI Trust
"I don't understand what the AI is telling me — so I can't trust it, even if it's right."
Emergency
"In mass casualty events, I need to grab the device and scan immediately. Registration can wait."
Define
The product serves both government-funded mass casualty triage and commercial burn centers — requiring one interface that scales from battlefield ER to routine hospital rounds.
Government
Commercial
Design
Every major change traces directly to a specific insight from field research. Here are the three decisions that defined the redesign.
01 — Emergency Access
"In mass casualty situations, I need to grab the device and scan immediately. Registration can wait."
We mapped the entire burn management workflow and identified that data entry was the biggest blocker during emergencies. The solution: a "Quick Scan" button that bypasses registration entirely, letting clinicians capture wound images first and fill in patient details after — a scan-and-go model.
We also integrated EHR auto-pull: the system matches patient profiles and merges data automatically, eliminating duplicate entry for non-emergency cases.
Reduced time-to-scan in emergency scenarios
02 — AI Trust
"Understanding AI-generated outputs was a challenge — users found the information complex and difficult to interpret."
Physicians need to understand why the AI reached a prediction before they'll act on it. We redesigned the DeepView report to include visual color-coded overlays distinguishing non-healing from burned areas, TBSA percentages, fluid resuscitation suggestions, and plain-language explanations alongside each AI output.
We also added real-time scan instructions for new staff, addressing high staff turnover in clinical environments where training time is minimal.
Reduced cognitive load; improved clinician confidence in AI results
03 — Design System
"The absence of a unified design system led to inconsistencies in visual elements, impacting brand cohesion."
Working across 4 product lines without shared foundations meant every new screen required re-inventing decisions already made elsewhere. I built a complete Figma design system from the ground up — color tokens, typography scale, spacing system, iconography, and a reusable component library.
This became the single source of truth for all 4 product lines, enabling faster, more consistent design-to-dev handoff and embedding WCAG 2.1 accessibility standards at the component level.
Consistent UI across all 4 product lines; faster dev handoff
"The device helps us focus on critical cases faster while simplifying communication with families — a game changer in high-pressure environments."
Validate
Every design decision was tested with real clinicians — first iteratively in formative sessions, then formally in a summative study that fed directly into our FDA De Novo submission.
Formative Study
An interactive Figma prototype let users explore software in realistic scenarios. Sessions uncovered three major improvement areas: EHR data challenges, real-time instruction needs, and AI interpretation gaps.
Iterative improvements applied
Summative Validation
Carried out with 15 real users to demonstrate ease of use. The report was reviewed by a 60601 lab and FDA in our De Novo submission. As the designer I was not involved in testing per protocol — my co-workers ran it. In one word: we passed.
FDA De Novo submission — passed
Prototype
Part of the prototype is presented here for demo purposes. Due to the complexity and confidentiality of this project, only selected flows are shown.
Prototype coming soon
Only select flows are shared due to project confidentiality