Comparison of a computer system evaluation of intrapartum cardiotocographic events and a consensus of clinicians

To compare between computer analysis of intrapartum cardiotocography (CTG) features by the Omniview-SisPorto 3.5 and a consensus of clinicians.

Agreement study using 50 consecutively acquired tracings (206 h of signals) with >60 min duration, <10% signal loss and recorded in labor at term by internal fetal heart rate (FHR) monitoring. Tracings were divided into 10-min segments and independently analyzed by three experienced clinicians, in order to estimate the FHR baseline and identify periodic events. A consensus was reached using a three round Delphi procedure. Results were compared with the analysis provided by the Omniview-SisPorto 3.5 system.

For baseline estimation, agreement between the computer and the consensus was high [intraclass correlation coefficient (ICC)=0.85; 95% confidence interval (CI) 0.46-0.93], with a mean difference of 3.7 bpm (limits of agreement -4.4-11.9 bpm), and 99% of differences under 15 bpm. A concordant identification was observed in 71% of accelerations (95% CI: 69%-73%), 68% of decelerations (95% CI: 66%-70%), and 87% of uterine contractions (95% CI: 85%-89%).

A high agreement was observed between the Omniview-SisPorto 3.5 and a consensus of clinicians in evaluation of intrapartum CTG baseline, accelerations, decelerations and uterine contractions.