A Patient-Centered Data-Driven Analysis of Epidural Anesthesia

Abstract

This study investigates safety and efficacy of a large-dose, needle-based epidural technique where the anesthetic dose is administered through an epidural needle prior to insertion of the epidural catheter. Using a data-driven machine learning (ML) approach, the findings show that the needle-based approach is faster and more dose-effective in achieving sensory level than the catheter-based approach. The authors also find that injecting large doses in the epidural space through the needle is safe. And a needle dose of at most 18 ml offers lower hypotension complication. ML predicts hypotension with 85% accuracy and shows that total dose, injection duration, weight, and physician experience are top features impacting sensory level. The findings facilitate pain relief improvement and establish new clinical practice guideline for training and dissemination of safe administration. The successful prediction of hypotension allows for early intervention. Although almost 50% of drug combinations used involve fentanyl, the findings show that fentanyl has little effect on outcome and should be avoided.