Rate of caesarean sections according to the Robson classification: Analysis in a French perinatal network – Interest and limitations of the French medico-administrative data (PMSI)

Abstract Introduction The objective of our study was to determine, in accordance with WHO recommendations, the rates of Caesarean sections in a French perinatal network according to the Robson classification and determine the benefit of the medico-administrative data (PMSI) to collect this indicator. This study aimed to identify the main groups contributing to local variations in the rates of Caesarean sections. Material and methods A descriptive multicentric study was conducted in 13 maternity units of a French perinatal network. The rates of Caesarean sections and the contribution of each group of the Robson classification were calculated for all Caesarean sections performed in 2014. The agreement of the classification of Caesarean sections according to Robson using medico-administrative data and data collected in the patient records was measured by the Kappa index. We also analysed a 6 groups simplified Robson classification only using data from PMSI, which do not inform about parity and onset of labour. Results The rate of Caesarean sections was 19% (14.5–33.2) in 2014 (2924 out of 15413 deliveries). The most important contributors to the total rates were groups 1, 2 and 5, representing respectively 14.3%, 16.7% and 32.1% of the Caesarean sections. The rates were significantly different in level 1, 2b and 3 maternity units in groups 1 to 4, level 2a maternity units in group 5, and level 3 maternity units in groups 6 and 7. The agreement between the simplified Robson classification produced using the medical records and the medico-administrative data was excellent, with a Kappa index of 0.985 (0.980–0.990). Conclusion To reduce the rates of Caesarean sections, audits should be conducted on groups 1, 2 and 5 and local protocols developed. Simply by collecting the parity data, the excellent metrological quality of the medico-administrative data would allow systematisation of the Robson classification for each hospital.