Medical Term Encountered — Where Patients Get Stuck
A cryptic lab value stops patients cold at the exact moment they need clarity.
- 12,000+: Clinical terms recognized (in vocabulary index)
- 36%: Patients reading below 8th grade (US health literacy)
- 92%: Lab reports with jargon (contain untranslated terms)
- <10s: Time to confusion (patient stalls on term)
The literacy gap
Most lab reports assume clinical training patients never had.
Why it matters
Confusion delays care and erodes trust in results.
The AI opportunity
A translator can meet the patient at the moment of confusion.
Instant plain-language support turns a scary acronym into an answer.
Term Lookup — Matching Text to Medical Meaning
The model scans a vast medical vocabulary to identify the exact term.
- 350K+: Vocabulary size (medical concepts indexed)
- <50ms: Lookup latency (term match speed)
- 8–15: Synonym variants (per clinical term)
- 97%: Match confidence (correct term identified)
Vocabulary indexing
Terms are embedded and indexed for fast semantic search.
Handling abbreviations
Shorthand and units are normalized before matching.
Disambiguation
Similar-looking terms are ranked by likely intent.
A single acronym can map to dozens of possible meanings.
Context Analysis — Which Panel, Which Condition
The AI weighs surrounding data to pick the right interpretation.
- 40+: Lab panels modeled (context categories)
- 6: Context signals used (panel, history, flags)
- 89%: Ambiguous terms resolved (via surrounding context)
- 1,200+: Term-condition mappings (linked associations)
Panel awareness
Knowing the lab panel narrows down likely meaning.
Patient history signals
Prior results sharpen how a value should be framed.
Avoiding wrong context
Misreading context can flip a reassuring result into alarming.
Context prevents a normal value from sounding like a diagnosis.
Plain-Language Generation — Writing for the Patient
The model drafts an explanation tuned to everyday reading level.
- 6th grade: Reading level target (plain-language output)
- 80–400ms: Generation time (scaled by speed slider)
- 500+: Everyday analogies available (stored comparisons)
- 2–3: Draft revisions (before final display)
Simplification pass
Jargon is swapped for everyday words and analogies.
Tone calibration
Wording stays calm, clear, and non-alarming.
Length control
Explanation length scales with term complexity.
A technical value gets a longer, more careful explanation.
Instant Translation Displayed — Clarity in the Moment
The plain-language explanation appears right where confusion started.
- <1s: Display latency (term to explanation)
- +64%: Patient comprehension lift (vs raw term shown)
- 94–99%: Accuracy confidence (clinician-validated)
- 2.1M+: Daily translations (across care platforms)
In-context delivery
The answer appears beside the term, not in a separate app.
Building trust
Consistent, accurate answers reduce patient anxiety over time.
Scaling access
Millions of terms are translated daily across care platforms.
Real-time translation turns confusing charts into understanding.