Racial Disparities and the Use of Artificial Intelligence for Predicting Maternal Mortality: A Literature Review

Area of Science:

• Maternal Health
• Artificial Intelligence
• Health Equity

Background:

• Maternal mortality is a critical global health issue, disproportionately impacting Black and Indigenous women.
• Hypertensive disorders and postpartum hemorrhage are leading causes of maternal death.
• Artificial intelligence (AI) offers potential for predicting these complications, but equity concerns remain.

Purpose of the Study:

• To synthesize recent evidence on AI's role in preventing maternal mortality from hypertension and postpartum hemorrhage.
• To assess the equity and validation of AI tools across racial groups.

Main Methods:

• A rapid review of five databases (PubMed, EMBASE, Web of Science, Scopus, LILACS) was performed.
• Studies from the last five years including racial or ethnic data were selected and analyzed.
• Narrative analysis was used to synthesize findings.

Main Results:

• Ten studies met inclusion criteria, demonstrating high AI predictive accuracy (AUROC >0.95) for severe maternal outcomes.
• Few AI models incorporated racial variables or underwent external validation in diverse populations.
• Unrepresentative datasets risk perpetuating health inequities.

Conclusions:

• AI tools show strong technical performance in predicting maternal complications but lack equitable application.
• Ensuring broader racial representation and external validation is crucial.
• Ethical governance is essential for AI to reduce, not worsen, racial disparities in maternal mortality.