Intra‐ and inter‐observer variability in the assessment of intrapartum cardiotocograms

TLDR

The diagnostic value of a test depends on its variability and accuracy, and reducing this variability is essential before the true predictive value and cost/benefit of electronic fetal monitoring can be estimated. This study aimed to estimate observer variability and accuracy when experienced obstetricians assessed intrapartum cardiotocograms. Fifty cardiotocograms were evaluated twice by four obstetricians, who identified compromised‑infant cases using provided criteria and a one‑third incidence assumption. Only 22 % of cardiotocograms were assessed identically by all obstetricians in both evaluations, with inter‑obstetrician accuracy ranging from 50 % to 66 %, indicating that observer variability severely impairs the clinical value of electronic fetal monitoring.

Abstract

The diagnostic value of a test depends on the variability of the test results and the accuracy of the test. The object of this investigation was to estimate the observer variability and the accuracy, when intrapartum CTGs were assessed by experienced obstetricians. Fifty CTGs were evaluated twice by four obstetricians. They were asked to identify the CTGs belonging to the compromised infants. They were told the criteria for a compromised infant and the incidence (one‐ third). Eleven (22%) of the CTGs were assessed in the same way of all the obstetricians in both evaluations. Between the obstetricians, the accuracy ranged from 50 to 66%. We conclude that the considerable observer variability found in this and other investigations must severely impair the clinical value of electronic fetal monitoring (EFM). The variability must be reduced before the “true” predictive value and the cost/benefit of EFM can be estimated.