Inside the wound-native scribe: how structured fields replace prose at capture
product
A deep dive into how a wound-native scribe turns spoken findings into structured fields — measurements, tissue percentages, staging, and codes — at the point of capture.
Most AI scribes produce prose. A wound-native scribe produces structure. The difference isn't cosmetic — it's what determines whether a chart is codeable, trendable, and audit-ready the moment it's signed. Here's what's happening under the hood.
The capture layer: vocabulary before grammar
A general scribe listens for sentences. A wound-native scribe listens for wound anatomy. When a clinician says "70% granulation, 30% slough, moderate serosanguinous drainage, undermining at 3 to 6 o'clock," the system doesn't just transcribe those words — it maps them to fields:
- Tissue composition: granulation 70%, slough 30%
- Exudate: volume = moderate, character = serosanguinous
- Undermining: present, location = 3–6 o'clock
That mapping happens in real time, not as a post-hoc parse of finished prose.
Measurement without a ruler
Spoken measurements are captured, but so are visual ones. Phone-based planimetry, described in the Point-n-Capture deep dive, turns a photo into length, width, and area — populating the same measurement fields the scribe is listening for. The two channels reconcile, so what's in the chart matches what's on the camera roll.
Structure that survives across visits
Structured fields aren't just a documentation nicety — they're what makes healing analytics possible. When every visit produces the same fields, the healing analytics dashboard can render area over time, tissue composition over time, and predicted trajectory. Prose can't do that, no matter how well written.
Codes generated from findings, not from the note
Once fields exist, coding becomes deterministic rather than interpretive. Debridement depth drives the CPT selection. Ulcer etiology and stage drive the ICD-10. E/M level is supported by the documented complexity of decision-making. The code isn't inferred from prose — it's derived from the same structured findings the clinician spoke.
Payer rules checked against structure, not text
Because the chart is structured, payer rules can be evaluated as logic rather than keyword search. Modifier 25 requirements, LCD medical necessity language, and frequency limits are checked against fields at sign-off. The compliance gate flags gaps before the note is signed, not after billing gets a denial.
Why this tier matters
A specialty-aware scribe can produce a better-sounding wound sentence. Only a wound-native scribe produces a wound record — one that codes itself, trends itself, and defends itself in an audit. That's the tier the full wound care platform is built on.