Overview

• Researchers analyzed 18,646 labor and delivery clinical notes from 2017 to 2019 using natural language processing to uncover racial and ethnic disparities in documentation.
• Black patients were disproportionately subjected to stigmatizing language, with 54.9% of their notes containing such descriptors compared to 49.3% overall.
• Hispanic and Asian/Pacific Islander patients experienced different patterns of bias, with less frequent positive language and underrepresentation in certain categories.
• The study highlights how biased language in clinical notes can perpetuate healthcare inequities, impacting patient trust, care quality, and outcomes.
• Authors advocate for interventions including culturally sensitive guidelines, provider training, and AI tools to address and monitor biased language in medical documentation.