Risk calculator for hypertension related postpartum readmission


      To develop a calculator that estimates the risk of hypertension related postpartum readmission using our validated composite predictive model.

      Study Design

      We created a user-friendly calculator to aide providers in predicting hypertension related postpartum readmission. We validated our predictive model using data from a retrospective cohort study from a single Midwestern academic center of all women who delivered between 2009-2018. The primary outcome of interest was hypertension-related postpartum readmission within 42-days postpartum. We used a cost-sensitive random forest method to determine variable predictors of postpartum readmission. The receiver operating characteristic curve, sensitivity and specificity was calculated for the model.


      32,645 women who delivered from 2009-2018; 170 were readmitted in the postpartum period due to a hypertension-related diagnosis. Random forest method achieved a sensitivity of 85%, specificity of 79%, and balanced accuracy of 82% for predicting readmission. The most important variables for predicting readmission included systolic blood pressure at 48 hours postpartum and systolic blood pressure in labor. Overall, Fifteen variables were included in the final model and resulting user-friendly online clinical calculator.


      This validated model predicts the risk of hypertension related postpartum readmission. In contrast to prior known predictors of readmission this model uses distinct clinical variables that are easily identifiable in the medical chart. This user-friendly online clinical calculator may be used as an individualized objective tool during discharge planning, outpatient management, and perhaps those who may most benefit from daily remote patient monitoring or closer outpatient surveillance to prevent readmission.
      Figure thumbnail fx1
      Figure thumbnail fx2