Abstract

Louise PARET1, Virginie EHLINGER, Catherine ARNAUD and Christophe VAYSSIERE

To assess interobserver reliability and agreement in the classification of cardiotocograms (CTGs) according to the 2015 revision of the FIGO (International Federation of Gynecology and Obstetrics) CTG classification. Six observers (3 obstetricians and 3 midwives) applied the 2015 FIGO guidelines and independently interpreted 60 intrapartum CTGs, randomly selected after stratification for arterial umbilical pH (pHa) at birth: 1/3 with pHa>7.15, 1/3 with pHa=7.05-7.15, and 1/3 with metabolic acidosis defined by pHa10 mmol/L. Interobserver reliability was assessed by the Fleiss kappa coefficient and interobserver agreement by the proportion of agreement (Pa), calculated according to Grant. The overall interobserver reliability was good (kappa=0.62, 95%CI=0.52-0.73). Interobserver agreement was good for CTGs classified as normal by the observers (Pa=71.5%, 95%CI=67.5-75.2), moderate for those classified as pathological (Pa=57.4%, 95%CI=51.3-63.5) and poor for those classified as suspicious (Pa=36.4%, 95%CI=30.9-41.9). Interobserver reliability was good for baseline, moderate for variability assessment and presence of decelerations, but poor for classifying the decelerations. Results did not differ significantly between obstetricians and midwives. Application of this FIGO classification produced good interobserver reliability and agreement on CTG c