Inconsistency in evaluation of clinical patterns in fetal heart rate waveforms

Acceleration and deceleration patterns in fetal heart rate (FHR) are widely used as an assessment of fetal condition. The basis for both visual and automated quantitative analysis is correctly determined fetal heart rate baseline. There are many methods for automated fetal heart rate baseline estimation. Since there are no definitive criteria or "gold standard" to evaluate algorithms for baseline determination, a clinical expert could be considered as a reference. Taking into account that not the shape of baseline itself but the results of quantitative analysis obtained using this baseline are important, we stated that such evaluation should be done only on the basis of comparison of effects - the recognized accelerations and decelerations. However, the first step for such comparison is to check how well expert meets established FIGO definition of accelerations and decelerations patterns in FHR trace. The paper presents a method for evaluation of expert's interpretation of these definitions. Two indices describing the inconsistency among fetal monitoring system and experts have been defined. The results obtained indicate the essential differences in procedure of accelerations/decelerations recognition by the computerized system and human expert.