Strengths and weaknesses
All data in this study had been collected prospectively from consecutive cases by one observer as part of an in-unit ongoing safety analysis. This approach allowed a large number of in-labour and delivery variables to be collected, including position and station of the fetal head and experience level of the final deliverer, factors that can bear a considerable influence on the outcome of a delivery and which need to be accounted for in any analysis of delivery types. Deliveries were also analysed in an ‘intention to deliver’ fashion, reflecting the fact that an operator makes a judgement on which delivery mode to embark on without knowing if this will lead to success or failure, therefore the risks the patient is exposed to are not only those of the original delivery attempt, but also any subsequent or final mode of delivery. A further strength of the study was the use of specific groupings to allow for all methods of instrumental delivery to be compared with a control group of primary emergency full dilatation Caesarean.
This study was limited by data collection from only one delivery unit where there is a high degree of experience amongst consultant staff in the use of Keilland’s forceps, it was notable that a greater proportion of Keilland’s forceps were performed by consultant staff (Supplementary Figure 1). Although this suggests that such deliveries are safe in experienced hands, the findings of this study will be less generalisable to units where this experience with Keilland’s forceps is not present, or where there is a predominance of alternative instrumental delivery types such as ventouse. The study was designed to look at the frequency of immediate outcomes that may present serious risk to mothers and babies, and due to the relative scarcity of such outcomes a composite score was designated. This carries limitations in two ways: firstly, such outcomes do not capture the longer-term picture, and any further research should endeavour to collect data on long term outcomes for both mothers and babies. Secondly, although producing a composite score of outcomes is pragmatically necessary when such outcomes are scarce, such an approach can be problematic as it assumes that all parts of the composite are of equal value to patients and clinicians which is not likely to be the case when comparing a neonatal death with, for example, an isolated low Apgar score, or a need for transfusion in an otherwise well woman. This underlines the urgent need for the development of core outcome sets in Obstetrics 16, with consideration of standardised outcome ‘weighting’ towards outcomes that are more significant to patient’s real-life considerations when using a composite score 17.