We read with interest the useful comments of Ayres-de-Campos et al.1 on our paper quantifying inter- and intra-observer agreement of intrapartum ST-analysis of the fetal electrocardiogram (ECG, STAN®).2
In their letter, they first correctly comment upon the different categories for classification of the cardiotocogram (CTG) according to FIGO3 and STAN®. The advantage of the four-category STAN® guidelines, compared to the three-category FIGO guidelines, is the possibility of separately classifying a (pre)terminal CTG. Strict use of the STAN® clinical guidelines allows clear recognition of these worse-case-scenario CTG traces, which should be followed by immediate intervention. In our paper, we explained that the STAN® guidelines are based on the FIGO guidelines, thus not representing them.
In our study, we found satisfactory levels of agreement for the normal and (pre)terminal CTG trace. As expected, agreement for the intermediary and abnormal CTG classes was poor, but we disagree with the argument of Ayres-de-Campos et al. to merge these categories. Of course, merging these categories would make life simpler. However, we doubt whether this simplified ‘traffic-light-approach (red, orange and green)’ also makes it easier, since there is always some doubt about how to act when the light becomes orange. When should we intervene when the CTG traces ‘intermediate’ and ‘abnormal’ will be merged in this ‘orange’ category? Using (existing) cut-off points for intervention (based on ST information) of the abnormal CTG trace will lead to more false positive interventions, whereas cut-off points of the intermediary trace may create a higher false negative rate. In our opinion, the three-category STAN® guidelines, as used in the United States, will increase the number of unnecessary caesarean sections (CS) in countries with low CS rates, as in The Netherlands with a CS rate of 15%.
Indeed, we found a higher agreement on clinical decisions, which is in agreement with the previous paper by Ayres-de-Campos et al.4 We do not think that our findings were because of the dichotomous nature of the options (intervention versus no intervention), since intervention may imply many options, such as lowering oxytocin augmentation, amnion infusion, fetal blood sampling, asking a more senior colleague for help, ventouse/forceps delivery or CS. Therefore, the decision not to intervene, requires a great deal of thinking leading to the conclusion that the obtained information was sufficient to refrain from any of the intervention options.
In general, we agree with Ayres-de-Campos et al.1 that classification of the CTG and the subsequent clinical decision making process should not be more complex than actually needed and that there is still a great need for improvement. However, we feel that, rather than simplifying the STAN® guidelines regarding CTG classification categories, the solution to the large inter-observer variability may be found in computerised backing of CTG classification. In this way objective additional information (e.g. on poor fetal heart rate variability) is provided by computer quantification, a topic very familiar to Ayres-de-Campos et al.5