Computerized predictive models in hospital settings for detecting severe maternal complications: a systematic review...

Area of Science:

• Public Health
• Medical Informatics
• Maternal Health

Background:

• Maternal mortality is a significant public health issue, particularly in low- and middle-income countries.
• Key preventable causes include hypertension, gestational diabetes, pre-eclampsia, postpartum hemorrhage, and infections.
• Computerized predictive models offer potential for early risk identification and prevention of severe maternal complications in hospital settings.

Purpose of the Study:

• To systematically review the evidence on the effectiveness of computerized predictive models in intra-hospital environments.
• To assess the impact of these models on adverse outcomes related to maternal mortality.

Main Methods:

• Systematic review following PRISMA-P guidelines.
• Inclusion of randomized and non-randomized studies from nine major databases.
• Searches conducted in Portuguese, English, and Spanish, with no publication date restrictions.

Main Results:

• Data extraction and quality assessment performed by independent researchers using GRADE and ROBINS-I tools.
• Narrative synthesis guided by SWiM guidelines, with potential for meta-analysis.
• PROSPERO registration number: CRD42024573613.

Conclusions:

• Findings aim to enhance understanding of predictive technologies in maternal healthcare.
• Results will support clinical decision-making, educational initiatives, and health service planning.
• Evidence generated will contribute to reducing preventable maternal deaths in hospital settings.