Intrapartum surveillance: Computer cardiotocography

Several factors seem to limit the effectiveness of the intrapartum cardiotocographic (CTG) monitoring, and the most important factor is the high interobscrver and intraobserver variation. Computerized systems allow consistent and reproducible analysis and interpretation of CTG tracings, but few data are available for intrapartum monitoring. We studied 177 post term pregnancies during induction or augmentation of labour, comparing the CTG data with neonatal outcome. The CTG evaluation by a single perinatology expert was compared to a computerized system, elaborating classical indices, power spectral density indices and approximate entropy, interpreting all results with a neural network to provide a normal-abnormal prediction. Computerized system shows a better overall accuracy, but the expert has a reduced number of false negative diagnoses. In the clinical practice, the role of computerized analysis seems to be the confirmation of a positive diagnosis made with visual analysis.