INTEROBSERVER AGREEMENT IN EVALUATION OF DIFFERENT PARAMETERS OF CTG IN LABOUR

Abstract

The aim. To assess the interobserver agreement of intrapartum cardiotocography of different parameters using the 2015 FIGO CTG classification. Methods. A retrospective cohort study in two large birth centers of Lithuania. Medical records were randomly selected from a birth database with inclusion criteria: term singleton pregnancy in cephalic presentation with meconium-stained amniotic fluid. A total of 84 CTG tracings were extracted and analyzed by obstetricians having different clinical experiences. Interobserver agreement was assessed using the Kappa and Intraclass Correlation Coefficient. Results. Good to moderate interobserver agreement was found for most CTG parameters (variability, presence of decelerations, bradycardia, reduced variability with tachycardia, and the sum of the area within all decelerations). Assessment of accelerations had moderate interobserver agreement (kappa coefficient 0.29–0.47). The assessment of the type and form of decelerations had the lowest agreement (kappa coefficient 0.12–0.23). Clinicians having more than 10 years’ experience showed higher agreement, whereas the mixed group of doctors with more than 10 years and less than 3 years’ experience showed lower interobserver agreement. Categorization of CTG also had low or only moderate interobserver agreement (Fleis kappa 0.19–0.44). The highest agreement was achieved when evaluating the total width of decelerations in a 30-minute CTG interval (the intraclass correlation coefficient was between 0.48 and 0.80). Conclusions. The main CTG parameters – baseline rate, variability, presence of decelerations, and the sum of the area within decelerations - are CTG signs that can be interpreted objectively and with high interobserver agreement after proper training, even without having clinical experience. Since the overall CTG classification into categories (normal, suspicious, and pathological) has a low or medium interobserver agreement, it would be rational to refuse overall CTG classification, and for clinical decisions to use more reliable and higher agreement parameters (baseline rate, variability, presence of bradycardia, presence of decelerations and the total width of decelerations in a 30-minute CTG interval).