Interobserver reliability to interpret intrapartum electronic fetal heart rate monitoring: does a standardized algorithm improve agreement among clinicians?

Our aim was to investigate the accuracy in predicting intrapartum fetal acidaemia and the interobserver reproducibility of a mathematical algorithm for the interpretation of electronic fetal heart rate (FHR) monitoring throughout labour. Eight physicians (blinded to the clinical outcomes of the deliveries) evaluated four randomly selected intrapartum FHR tracings by common visual interpretation, trying to predict umbilical artery base excess at birth. They subsequently were asked to re-evaluate the tracings using a mathematical algorithm for FHR tracing interpretation. Common visual interpretation allowed a correct estimation of the umbilical artery base excess in 34.4% of cases, with a poor interobserver reproducibility (Kappa correlation coefficient = 0.24). After implementation of the algorithm, the proportion of correct estimates significantly increased to 90.6% (p < 0.001), with excellent inter-clinician agreement (κ: 0.85). To conclude, incorporation of a standardised algorithm reduces the interobserver variability and allows a better estimation of fetal acidaemia at birth.