StigmaGuard
FHIR-native clinical documentation review using CDS Hooks to detect potentially stigmatizing language at note sign-off and provide explainable, context-aware recommendations while keeping clinicians in control.
Core trick
A CDS Hooks service runs at note sign-off, flags stigmatizing terms with explanations, and proposes alternatives the clinician can accept or reject.
Tags
Description
FHIR-native clinical documentation review using CDS Hooks to detect potentially stigmatizing language at note sign-off and provide explainable, context-aware recommendations while keeping clinicians in control.
README excerpt
# StigmaGuard — Real-Time Clinical Documentation Review StigmaGuard is a FHIR-native clinical documentation review service that evaluates clinical notes before signing, explains potentially stigmatizing language, and suggests patient-centered alternatives while keeping the clinician in control. Built at the Anthropic / Abridge / Lightspeed Healthcare Hackathon, July 2026. > **Status:** This repository contains a hackathon prototype demonstrating a standards-based approach to real-time clinical documentation review. The focus is on workflow integration, explainable AI, and FHIR/CDS Hooks interoperability. --- # The Problem Language in the medical record influences future clinical decision making as well as future patient engagement with the system. Terms such as *"addict," "drug-seeking," "non-compliant,"* and *"dirty urine"* have been shown to contribute to clinician bias and can persist throughout a patient's longitudinal record. Research has documented this problem since at least 2018 (Goddu et al.), and multiple studies have suggested real-time documentation review as a future direction. More recently, Zhou et al. (UC Irvine, arXiv 2606.00019) demonstrated that stigmatizing language is often introduced during clinician editing of ambient AI drafts rather than by the ambient AI system itself, making note signing a particularly valuable intervention point. --- # What Was Built - **FHIR-backed EMR simulation** using HAPI FHIR with patient charts and documentation