Translating clinical data with LabViz

LabViz is the participant-facing data visualization layer of a mobile and web app built for a clinical trial in metabolic disease. The trial uses a tightly controlled diet to manage the participant's amino acid (AA) levels as a foundational treatment. From the same app, participants log meals, complete surveys, view lab results, chat with their coach, and work through educational modules, with their AA readings as the central indicator of compliance.

I led the design of how those readings reach the participant. The team was responsible for four applications across the trial (participant mobile, participant web, Dietitian web, and Primary Investigator intake), and this case study covers the most behaviourally sensitive surface in that ecosystem: the chart a participant sees when they open the app.

This is an NDA-compliant retelling of the project. The following has been omitted or visually altered to protect the original work:

• Client name and branding
• The app's design system
• Clinical language and the specific molecules tracked

What you read below is visually abstracted from the source. The story, research, decisions, and outcomes are intact.

Give clinical trial participants an engaging, honest way to read their Amino Acid levels and stay motivated through a difficult treatment.

It's a challenging diet and I put a lot of effort, but I don't really know if I'm doing it right or wrong until my weekly check-ins with you (the Dietitian).

— Participant frustration, surfaced through their Dietitian

The treatment depends on adherence to a strict diet, and participants need timely feedback to know whether their effort is moving the numbers. Without that feedback in the app, motivation breaks down between weekly check-ins, which is exactly the window where adherence matters most.

HIPAA compliance prevented us from interviewing trial participants directly. We ran four interviews across two proxy groups.

• Primary Investigator (1). The clinical lead on the trial. We interviewed them to map the relevant data ranges and identify which metrics deserved space on the participant's screen.
• Registered Dietitians (3). Embedded in the patient journey week to week. They surfaced the pain points their patients voice in check-ins, alongside several of their own.

Four insights set the direction of the work.

1. Too many tracked metrics. The trial's compliance logic ingests an overwhelming number of values. The participant-facing surface needs aggressive simplification.
2. The Dietitian is the only signal. Participants stay in the dark between weekly check-ins, which is where motivation breaks down.
3. Cause and effect are unclear. Participants struggle to correlate recent dietary or behavioural changes with their latest readings.
4. Effort without transparency erodes motivation. Participants put real effort into the diet, and the absence of visible progress flattens their drive to continue.

We translated the insights into three guiding principles for the first round of design.

• One visualization per AA. Cap cognitive load at three discrete charts.
• Direct, clear, honest. The participant must immediately understand whether their value sits within the desired range.
• Contextual elaboration. Pair each reading with copy that explains what it means for compliance and what to do next.

Information architecture

The IA places the AA module on the home page so participants land on their readings every session. A dedicated Trial Data page sits one tap away, expanding each AA with its history and clinical context.

The home page surfaces the three AA levels as a top-level segment tab, with each level rendered as a horizontal range chart.

• Each AA carries a binary verdict: On Target or Out of Target.
• A horizontal bar plots the participant's Baseline, Target, and Current Reading on the same axis.
• A button at the bottom of the module opens the Trial Data page for the longer view.

The first solution shipped, and our weekly proxy feedback loop (Participant → Dietitian → Product team) flagged two issues within the first review cycle.

• The "Out of Target" verdict was the default state. The diet, combined with biological variability, pushed most participants below the threshold most weeks. The binary verdict became a source of demotivation rather than guidance.
• Mixed verdicts blurred compliance. When one AA read On Target and another Out of Target, participants could not tell what their overall standing was for the week.

We reframed the work around two new constraints:

• Compress the three AA readings into a single value the participant can read at a glance.
• Use an affordance the participant already knows how to interpret from outside the clinical context.

We explored a Traffic Light system (Red, Yellow, Green) early. It compressed the data into one indicator, and it carried the wrong emotional weight. Only one of three states felt good, and the other two read as clinical alarm. We ruled it out for that reason.

The 150-point unified score

The winning direction was a composite score derived from the three AA values. Participants already read 0–100 patterns intuitively from school grades, fitness apps, and sports scoreboards. We anchored the new score to that pattern and extended the ceiling to 150.

The 150-point ceiling carried the psychological work. A participant two-thirds of the way to their clinical target lands on a score of 100 instead of 66. The math still reflects clinical reality, and the framing stops punishing participants for the early phase of a long treatment.

Final design

Three changes carry the redesign.

• Gauge chart. The three horizontal range bars consolidate into a single circular gauge. One number, one position on the arc, one read.
• Calmer palette. Clinical reds are gone. The new palette uses softer greens, which lowers the emotional cost of a reading that is still trending toward target.
• Coaching tone. Clinical alerts give way to conversational copy ("Looking great, Jane!" or "Outstanding results!"). The same data now reads as a coach speaking, with the medical context preserved underneath.

The formula is tuned to be lenient. A participant's readings would need to fall severely below target before the composite score reads "OFFTRACK", and a Dietitian intervention typically lands before that result reaches the participant. Three of the four possible score states carry a positive tone, which keeps the emotional weight on the supportive side.

The Trial Data page expands the same logic into three layers.

• Scores. Composite values labelled so higher always reads as better.
• Clinical data. Blood concentration for all three amino acids, where lower is better.
• Contextual support. Dietary, behavioural, and Dietitian-led recommendations tied to the current readings.

Try the prototype

The redacted prototype below is interactive. Tap the CTA on the home screen, navigate the bottom tab bar, switch between the Tips and Chart views, and open the "How it works" sheet from the gauge.

The redesigned module replaced the binary AA charts on the home page and across the Trial Data surface. Participants open the app to a single supportive number and a clear coaching line, and Dietitians report fewer demotivation cues in their weekly check-ins. The product team folded the gauge pattern into other compliance surfaces in the trial as the same problem (multiple metrics, one verdict) appeared elsewhere in the participant journey.

I worked on this product for two years across four applications. The participant-facing surfaces stayed the most engaging part of the job for me. Every decision on those surfaces mapped, however indirectly, to whether someone could keep going through a hard treatment for a hard disease.

UX Design Awards Nominee

The original, unmodified product was nominated for the 2025 UX Design Awards. The project did not take home the award, and the part of that I value most is what the nomination cost: a year of questioning our own assumptions, retiring the first solution after it shipped, and rebuilding the most visible surface of the app from the ground up.

Most of my projects are protected under NDA. If any part of this case study sparks a question, I'm always happy to talk it through over coffee.