Interpretations
Lab results people can read — the interpretation layer practitioners picked Fullscript for. Labs series, part 2.
- Role
- Product Designer — Patient Experience
- Timeline
- April 2024 – February 2025
- Team
- 1 Product designer (me) — patient views and the practitioner builder · 1 Staff product designer — co-designed the practitioner builder · 1 Content designer — language system · 1 Product manager · Clinical advisors (chief medical officer, dietitian, clinical education)
- Scope
- Patient-facing result interpretation · Practitioner interpretation builder · Range visualization · Biomarker trend graphs · Plain-language and AI-drafted summaries · Web + mobile
My role
- Designed the patient-side interpretation views: range bar rendering, result cards, and biomarker trend graphs.
- Co-designed the practitioner interpretation builder — the authoring surface where practitioners write, edit, and send interpretations — with the staff designer.
- Presented the trend-graph concept to the clinical advisory group and ran the biomarker-trends design sessions.
- Extended the Labs biomarker color system into the interpretation surfaces.
- Partnered with the content designer who set the plain-language rules for every summary.
TL;DR
In short
Problem
A lab result without context is a number that either means nothing or scares someone — and every diagnostics platform hands patients exactly that: values, ranges, and no translation.
Approach
Worked both sides of the interpretation layer — the patient-facing result views and, with a staff designer, the practitioner authoring builder — while a content designer set the rules for what the words could claim.
Outcome
Interpretations shared with patients grew 5× in 2 months after launch, and a practitioner told customer success the feature alone sold him on Fullscript as his labs partner.
Context and problem
The situation we walked into
Business context
Fullscript Labs shipped result delivery in 2024, but delivery isn’t understanding. The bet behind interpretations: if practitioners can attach context to every biomarker — where a value sits, what it means, what to watch — the result becomes a conversation instead of a document, and the platform becomes the place that conversation happens.
User context
Two users share one artifact. Practitioners need to annotate results quickly enough to do it for every patient. Patients need to read those annotations without clinical training — and without being told a number is “good” or “bad” when the truth is more conditional. The design had to serve the writer and the reader at once.
Constraints and complexity
Clinical range definitions belonged to the clinical team, and the language rules were strict: fact-based wording, no emotional labels, no diagnosis. AI-drafted summaries raised the stakes further — drafts had to be reviewable and editable by the practitioner before a patient ever saw them. Design owned how all of this read, not what it claimed.
Goals and success metrics
What success looked like
A patient who opens an interpreted result understands where each value sits and what their practitioner wants them to know — and practitioners find interpretations fast enough to write that they attach them routinely.
Success criteria
- Interpretation views ship for both patient and practitioner in 2024.
- Coverage expands from the initial panels to the full Quest catalog.
- Practitioners share interpretations with patients in growing numbers month over month.
Approach
How we got there
- 01
Designed the range bar as the anchor
Every biomarker renders on a range bar — low, suboptimal low, optimal, suboptimal high, high — using the Labs biomarker palette. The zones came from the clinical team’s marker review; my work was making the same bar read correctly at card size, in dark mode, and for values that sit far outside any zone, where proportional rendering would break the scale.
- 02
Brought trends to the advisory group
Presented the concept for showing previous results to the clinical advisors, jamming with the content designer and staff designer on how set-value and binary biomarkers should differ. The sessions settled how longitudinal context appears: a trend graph once history exists, gated so a single result never fakes a direction.
- 03
Held the language to fact-based rules
The content designer set the editorial system for summaries: no “positive” or “negative,” no emotional framing, facts a patient can act on. When early interpretation content missed the clinical quality bar — our chief medical officer said so plainly — the summaries were rebuilt against those rules rather than softened around them.
- 04
Scaled from panels to the full catalog
Interpretations entered production in August 2024, AI-drafted summaries followed in September, and coverage expanded to every Quest test in October — the same month Fullscript acquired Rupa Health. Auto-sending arrived in early 2025, closing the loop from draft to patient inbox.
Key decisions
Forks in the road
Gate the trend graph behind a second result
Showing a graph from the first result implies a direction that doesn’t exist; hiding it entirely risks patients assuming something is broken. One data point can’t trend, and a fabricated slope on a health metric is worse than an empty state.
- Decision
- The trend graph appears only once a biomarker has 2 or more results; a single result gets a deliberately written empty state instead.
- Tradeoff
- First-time patients see no graph at all — accepted, because an honest blank beats a misleading line.
Fact-based language vs reassuring language
Reassuring copy (“this looks good!”) is friendlier and tests well in isolation, but a biomarker’s meaning depends on context only the practitioner has — and a cheerful label on a conditionally fine value is a claim the platform can’t stand behind.
- Decision
- Adopted the content designer’s editorial rules: no good/bad labels, no emotional framing, facts the practitioner can own.
- Tradeoff
- The copy reads cooler than consumer health apps — accepted, because trust here is clinical, not conversational.
AI drafts inside an editorial cage
AI-generated summaries made interpretation-per-patient feasible at scale, but an unreviewed generated sentence about someone’s blood work is a liability, not a feature. The alternative — fully manual notes — kept quality high and volume near zero.
- Decision
- AI drafts the summary against the editorial rules; the practitioner reviews, edits, and owns every interpretation before it reaches a patient.
- Tradeoff
- Review adds a step to the practitioner’s flow — accepted, because the practitioner’s name is on the note.
Solution
What we shipped
Outcomes and impact
What it moved
Shared interpretations
500
5× in 2 months
Practitioners had shared 100 interpretations by December 2024 and 500 by February 2025.
Coverage
All Quest tests
Expanded from the initial panel set to the full Quest catalog in October 2024.
Practitioner choice
Deciding factor
A practitioner told customer success the interpretation feature sold him on choosing Fullscript as his labs partner.
Business impact
Interpretations became a stated reason practitioners chose Fullscript for labs — customer success relayed a practitioner who picked the platform because of the feature, and another who registered the same night after seeing interpretations and trending. The layer this project built became the foundation the Journeys product later stood on.
User impact
Patients open results that come with their practitioner’s context attached: where each value sits, how it’s moving, and what it means in plain terms. Practitioner-relayed feedback was consistently strong; direct patient-behavior data is the open question the next measurement pass answers.
In their words
What people said
“Interpretations and trending is wonderful… I’m registering tonight!”
Practitioner — comparing lab platforms, November 2024 “The launch of interpretations for all Quest results and Lab templates was a hit — practitioners were genuinely excited.”
Neil Shah — Product Manager, Fullscript Labs — GA announcement
Learnings and reflections
What I’d take with me
The quality bar came from a clinician, not a dashboard
The hardest feedback in the project was our chief medical officer flagging early interpretation content as not ready. That single review reset the standard: the summaries were rebuilt against the editorial rules, and the bar never dropped again.
A feature can be loved and invisible at once
Months after launch, a practitioner who gave Labs a glowing testimonial turned out to be managing results in a spreadsheet — she didn’t know interpretations existed. Adoption isn’t a quality problem; it’s a discoverability problem, and it needs its own design work.
Design the reading, not the ranges
The clinical team owned what the zones were; I owned how they read. Keeping that boundary explicit made advisor reviews faster and kept design debates about legibility instead of medicine — the version of the collaboration where everyone argues their own expertise.