Fetal heart rate tracings: observers versus computer assessment

Area of Science:

• Obstetrics and Gynecology
• Medical Imaging Analysis
• Clinical Decision Support Systems

Background:

• Cardiotocography (CTG) is a standard tool for fetal monitoring during pregnancy.
• The diagnostic accuracy of CTG is limited by substantial intra- and interobserver variability in interpreting fetal heart rate tracings.
• Computerized analysis offers a potential solution to standardize CTG interpretation.

Purpose of the Study:

• To evaluate the reproducibility of Cardiotocography (CTG) readings among different clinicians.
• To compare the consistency of clinician-based CTG readings with those generated by a computerized analysis system (2CTG).

Main Methods:

• Analysis of 63 fetal heart rate tracings by four clinicians (two experienced, two inexperienced) and the 2CTG computerized system.
• Assessment of key variables: baseline fetal heart rate (FHR), long-term variability, accelerations, and decelerations (number and type).
• Interobserver agreement was quantified using the kappa coefficient.

Main Results:

• Interobserver agreement among clinicians ranged from fair to good.
• Agreement between clinicians and the 2CTG system varied significantly across parameters, with lower agreement for FHR baseline and deceleration types.
• The agreement on the type of decelerations between observers and the computer system was notably low (0.01-0.25).

Conclusions:

• Considerable interobserver variability exists in CTG interpretation among clinicians of varying experience levels.
• The 2CTG computerized system shows potential to mitigate the problem of interobserver variability in CTG analysis.
• Further development and validation of computerized systems may enhance the reliability of fetal heart rate monitoring.