Visual analysis of antepartum fetal heart rate tracings: inter- and intra-observer agreement and impact of knowledge of neonatal outcome

TLDR

The study aimed to assess inter‑ and intra‑observer agreement in visual fetal heart‑rate tracing analysis and to determine how knowledge of perinatal outcome biases interpretation. Four observers independently evaluated 100 fetal heart‑rate tracings; two observers repeated the analysis for intra‑observer reliability, while two others re‑analysed tracings labeled with fictitious perinatal outcomes, and agreement was quantified using proportion agreement for qualitative parameters and intraclass correlation coefficients for quantitative data. Inter‑ and intra‑observer agreement was generally low, with only baseline and normal variability categories showing acceptable concordance; quantitative measures such as variability, decelerations, and accelerations had poor agreement, and knowledge of perinatal outcome significantly biased observers, reducing reliability.

Abstract

To evaluate the inter- and intra-observer agreement of visual analysis of fetal heart rate tracing and to evaluate the bias introduced by knowledge of perinatal outcome in this interpretation.One hundred tracings were independently analyzed by four observers. In a second study period, two observers re-analysed the 100 tracings in order to evaluate intra-observer agreement. The other two observers re-analyzed the tracings, which were labelled with fictitious perinatal outcome to evaluate the impact of this information on reliability. Agreement was analyzed by means of the proportion of agreement for qualitative parameters and the inter- and intra-class correlation coefficient for quantitative data.Poor agreement was found for quantitative variability, low variability category and number of decelerations. Moderate agreement was observed for baseline, normal variability category and number of accelerations. Fetal heart rate variability and number of accelerations and decelerations were found to be significantly influenced by clinical information of perinatal outcome. Biased observers showed lower reliability than unbiased ones.Visual assessment of fetal heart rate tracings is unreliable due to low rates of agreement between and within observers. Only qualitative classification such as normal baseline and normal variability showed good agreement. Knowledge of clinical information introduces subjectivity to the visual analysis, leading to a negative impact on reliability.