Why the study?
Neural networks are increasingly used to assess physiological processes and predict crises like hypotension, prompting evaluation of their ability to predict hypotension under spinal anesthesia during Cesarean section.
Does a neural network analyzing non-invasive arterial stiffness variability predict hypotension and vasopressor requirements in obstetric patients undergoing spinal anesthesia for Cesarean section?
Population
Obstetric patients undergoing Cesarean section under spinal anesthesia
Comparison
Single hidden layer neural network of 12 nodes vs discrete-feature discrimination approach
Design
Pilot study
Key result
A neural network analyzing arterial stiffness variability from a non-invasive blood pressure device predicted severe hypotension after spinal anesthesia with an AUC of 0.89.
Authors
Loading...
Neural network approach may aid hypotension prediction in Cesarean delivery; leaves open clinical utility pending larger validation.
Observational (n=49)
No
Does a neural network analyzing non-invasive arterial stiffness variability predict hypotension and vasopressor requirements in obstetric patients undergoing spinal anesthesia for Cesarean section?
Absolute Event Rate: 0.89% vs 0.87%
p-value: p=<0.001
A neural network analyzing non-invasive arterial stiffness variability can effectively predict hypotension and phenylephrine requirements in obstetric patients undergoing spinal anesthesia.
Gratz et al. (2020) conducted an observational in Hypotension during spinal anesthesia for Cesarean section (n=49). Neural network prediction using arterial stiffness variability vs. Discrete-feature discrimination approach was evaluated on Prediction of severe hypotension (Area Under the Curve) (p=<0.001). A neural network analyzing arterial stiffness variability from a non-invasive blood pressure device predicted severe hypotension after spinal anesthesia with an AUC of 0.89.