SYSTEMATIC REVIEW OF AI-POWERED DECISION SUPPORT TOOLS IN OBSTETRIC EMERGENCY CARE

Abstract

Background: Obstetric and gynecological emergencies demand rapid, high-stakes decision-making, where delays or inaccuracies can lead to serious maternal and fetal outcomes. Artificial intelligence (AI)-powered decision support tools are emerging as promising adjuncts to aid clinicians under such critical conditions. Despite increasing interest, the clinical utility, accuracy, and safety of these technologies in emergency women’s health care remain underexplored and unstandardized, warranting a comprehensive synthesis of current evidence.

Objective: This systematic review aims to evaluate the effectiveness, safety, and real-time decision-making accuracy of AI-powered decision support tools in the management of gynecological and obstetric emergencies.

Methods: A systematic review was conducted in accordance with PRISMA guidelines. Comprehensive searches were performed across PubMed, Cochrane Library, Scopus, and Web of Science from database inception to 2024. Inclusion criteria encompassed randomized controlled trials, feasibility studies, qualitative research, and developmental studies evaluating AI applications in obstetric and gynecological emergencies. Data were extracted using a standardized form and assessed for bias using the Cochrane Risk of Bias tool and Newcastle-Ottawa Scale, depending on study design. Due to heterogeneity, a qualitative synthesis was performed.

Results: Eight studies were included, encompassing feasibility, qualitative, and developmental research. AI tools demonstrated high concordance with clinician decisions in simulated obstetric emergencies, improved triage classification, and enhanced workflow efficiency. Clinicians highlighted the importance of transparency, personalization, and ethical considerations in AI adoption. However, most studies were small-scale or simulation-based, limiting generalizability.

Conclusion: AI-based decision support systems show encouraging potential in obstetric emergency care by enhancing diagnostic accuracy and clinical efficiency. Nonetheless, the current evidence base is preliminary. Rigorous, real-world validation and ethical integration are essential for safe and effective implementation in clinical practice.