Inside the AI Referral Suggestion Engine: How WoundScribe Picks the Right Specialist
product
A deep dive into how WoundScribe weighs wound status, perfusion, infection, and comorbidities to suggest podiatry, vascular, or ID referrals with traceable rationale.
Naming the right specialist is the hardest part of a referral. WoundScribe's suggestion engine turns that judgment into a transparent, auditable process.
The inputs the engine weighs
Every referral suggestion draws from structured facts already in the chart:
- Wound status — location, depth, tissue composition, duration
- Perfusion — ABI, toe pressures, pulses
- Infection signals — erythema, purulence, culture results, systemic symptoms
- Metabolic control — A1c, glucose trends
- Active diagnoses — PAD, diabetes, osteomyelitis, venous insufficiency
- Care setting — clinic, bedside, mobile
Those inputs come from imaging, scribing, and coding agents working together across the AI-powered EMR for wound care.
How suggestions are ranked
Each candidate specialty is scored against the clinical evidence and payer coverage:
- Covered + high evidence → surfaced as a primary suggestion (e.g., podiatry for a heel DFU)
- Covered + strong signal → surfaced with rationale (e.g., ID for infection signals)
- Requires prior auth → surfaced with a prior-auth flag and the driving evidence (e.g., vascular for ABI 0.6)
Traceability
The right-hand panel shows exactly which facts drove each suggestion. Nothing is a black box — every recommendation is traceable back to a documented finding, the same audit thread that runs through wound healing software.
What clinicians control
Suggestions are prompts, not automation. Clinicians accept, edit, or dismiss each one, and dismissals are logged with a reason so the engine learns your practice patterns without overriding judgment.