Predictive Healing Scores: How AI Flags Non-Healing DFUs Before the Limb Clock Runs Out

Most diabetic foot ulcers that end in amputation were technically "being treated" the whole time. The problem isn't effort; it's that stalled healing hides inside week-to-week measurements that look almost the same. WoundScribe's healing trajectory agent turns those measurements into a predictive score that tells you when a wound is off-track.

What the score is built from

The agent pulls objective signals from every encounter and trends them together:

• Area reduction velocity — is the wound closing at the expected rate for its etiology and week of care?
• Tissue composition shift — granulation vs. slough vs. necrotic percentages over time, drawn from the imaging agent.
• Peri-wound changes — maceration, erythema, callus buildup around a DFU.
• Adherence signal — offloading, dressing changes, and check-ins captured through Pingoo between visits.

The four-week rule, automated

Clinical guidelines say a wound that hasn't reduced by ~50% in four weeks needs a plan change. In practice, that threshold gets missed because no one is doing the arithmetic across scattered notes. The healing dashboard does it automatically and surfaces the flag inside the next encounter — not in a quarterly report.

What clinicians see at the point of care

When a DFU trends toward non-healing, the diabetic foot care view shows a red trajectory band, the specific inflection week, and suggested plan changes: vascular workup, biopsy, advanced therapy candidacy, or referral. The clinician decides; the system just refuses to let the pattern go unnoticed.

Why it matters for limb salvage

Every week a stalled DFU stays on the same failing plan raises infection and amputation risk. Moving the decision-to-escalate forward by even two weeks is meaningful. Combined with the broader wound healing software workflow, the healing agent turns healing analytics from a retrospective report into a real-time safety net.