loading page

A Risk Calculator to Predict the Need for Maternal or Neonatal Hospital-Based Peripartum Intervention: Modelling National Surveillance Data
  • +2
  • George Zhang,
  • Frances Wang,
  • Ha Vi Nguyen,
  • Jessica Bienstock,
  • Marielle Gross
George Zhang
Johns Hopkins Medicine School of Medicine

Corresponding Author:[email protected]

Author Profile
Frances Wang
Johns Hopkins University School of Medicine
Author Profile
Ha Vi Nguyen
Johns Hopkins University School of Medicine
Author Profile
Jessica Bienstock
Johns Hopkins University School of Medicine
Author Profile
Marielle Gross
University of Pittsburgh School of Medicine
Author Profile

Abstract

Objective: Given growing interest in alternatives to hospital birth, particularly given the COVID-19 pandemic, we developed a peripartum intervention risk calculator (PIRC) to estimate maternal and neonatal risk of requiring hospital-based peripartum intervention. Design: National cohort study. Setting: United States. Sample: Hospital births captured by the Pregnancy Risk Assessment Monitoring System from 2009-2018. Methods: The cohort was stratified by receipt of hospital-based interventions, defined as: 1) operative vaginal delivery (forceps or vacuum), 2) cesarean delivery, or 3) requiring neonatal intensive care unit admission. Gravidas with prior cesarean delivery or fetal malformation were excluded. Main Outcome Measures: Risk of requiring hospital-based intervention. Results: A total of 63,234 births were evaluated (72.6% full-term, 48.5% nulliparous) including 37.9% who received one or more hospital–based interventions. Gestational age was the most predictive factor of requiring hospital-based intervention, with lowest odds at 400/7-406/7 weeks. Previous live births (Ref: none; 1, OR 0.41; 2, OR 0.35; ≥3, OR 0.29; p<0.05 for all) were protective. Other predictors included advanced maternal age, high pre-pregnancy body mass index, maternal diabetes, maternal hypertension, and not exercising during pregnancy. The resulting seven-factor model demonstrated strong discrimination (optimism corrected C-statistic=0.776) and calibration (mean absolute error=0.009). Conclusions: We developed and validated the PIRC for predicting individualized risk for hospital-based intervention among gravidas based on seven readily accessible prenatal factors. This calculator can support personalized counseling regarding planned birth setting, helping to close a critical gap in current clinical guidance and providing an evidence-based risk assessment for those contemplating alternatives to hospital birth.