Documentation of Stigmatizing Language in Electronic Health Records and Birth Outcomes

Area of Science:

• Perinatal Health
• Medical Informatics
• Health Equity

Background:

• Stigmatizing language in healthcare settings reflects underlying biases.
• Understanding the impact of such language is essential for enhancing perinatal health outcomes and promoting health equity.

Purpose of the Study:

• To investigate the association between stigmatizing language used in electronic health records (EHR) and various birth outcomes.
• To identify specific categories of stigmatizing language and their correlation with adverse perinatal events.

Main Methods:

• Analysis of EHR data from 18,897 patients admitted for childbirth between 2017 and 2019.
• Utilized natural language processing to identify stigmatizing language (e.g., marginalized identities, difficult patient, authoritarian decisions).
• Examined associations with low-risk cesarean birth, postpartum hemorrhage, and chorioamnionitis.

Main Results:

• Any stigmatizing language in EHRs was associated with increased odds of low-risk cesarean birth (aOR=1.36), postpartum hemorrhage (aOR=1.68), and chorioamnionitis (aOR=1.23).
• Specific categories like 'difficult patient' and 'unilateral/authoritarian decisions' showed significant associations with adverse outcomes, including higher odds of cesarean birth and postpartum hemorrhage.

Conclusions:

• Stigmatizing language documented in clinical notes is significantly associated with negative birth outcomes.
• Findings underscore the necessity of addressing biased language in EHRs and implementing policies to mitigate bias for improved perinatal health.