PLoS ONE (Jan 2022)

Prediction of odds for emergency cesarean section: A secondary analysis of the CHILD term birth cohort study.

  • Mon H Tun,
  • Radha Chari,
  • Padma Kaul,
  • Fabiana V Mamede,
  • Mike Paulden,
  • Diana L Lefebvre,
  • Stuart E Turvey,
  • Theo J Moraes,
  • Malcolm R Sears,
  • Padmaja Subbarao,
  • Piush J Mandhane

DOI
https://doi.org/10.1371/journal.pone.0268229
Journal volume & issue
Vol. 17, no. 10
p. e0268229

Abstract

Read online

IntroductionPreviously developed cesarean section (CS) and emergency CS prediction tools use antenatal and intrapartum risk factors. We aimed to develop a predictive model for the risk of emergency CS before the onset of labour utilizing antenatal obstetric and non-obstetric factors.MethodsWe completed a secondary analysis of data collected from the CHILD Cohort Study. The analysis was limited to term (≥37 weeks), singleton pregnant women with cephalic presentation. The sample was divided into a training and validation dataset. The emergency CS prediction model was developed in the training dataset and the performance accuracy was assessed by the area under the receiver operating characteristic curve(AUC) of the receiver operating characteristic analysis (ROC). Our final model was subsequently evaluated in the validation dataset.ResultsThe participant sample consisted of 2,836 pregnant women. Mean age of participants was 32 years, mean BMI of 25.4 kg/m2 and 39% were nulliparous. 14% had emergency CS delivery. Each year of increasing maternal age increased the odds of emergency CS by 6% (adjusted Odds Ratio (aOR 1.06,1.02-1.08). Likewise, there was a 4% increase odds of emergency CS for each unit increase in BMI (aOR 1.04,1.02-1.06). In contrast, increase in maternal height has a negative association with emergency CS. The final emergency CS delivery predictive model included six variables (hypertensive disorders of pregnancy, antenatal depression, previous vaginal delivery, age, height, BMI). The AUC for our final prediction model was 0.74 (0.72-0.77) in the training set with a similar AUC in the validation dataset (0.77; 0.71-0.82).ConclusionThe developed and validated emergency CS delivery prediction model can be used in counselling prospective parents around their CS risk and healthcare resource planning. Further validation of the tool is suggested.