Is the Robson's classification system burdened by obstetric pathologies, maternal characteristics and assistential levels in comparing hospitals cesarean rates? A regional analysis of class 1 and 3

Objective: To assess if maternal risk profile and Hospital assistential levels were able to influence the inter-Hospitals comparison in the class 1 and 3 of the “The Ten Group Classification System” (TGCS). Methods: A population-based analysis using data from Institutional data-base of an Italian Region was carried out. The 11 maternity wards were divided into two categories: second-level hospitals (SLH), and first-level hospitals (FLH). The recorded deliveries were classified according to the TGCS. To analyze if different maternal characteristics and the hospitals assistential level could influence the cesarean section (CS) risk, a multivariate analysis was done considering separately women in the TGCS class 1 and 3. Results: From January 2011 to December 2013 were recorded 19,987 deliveries. Of those 7,693 were in the TGCS class 1 and 4,919 in the class 3. The CS rates were 20.8% and 14.7% in class 1 (p < 0.0001) and 6.9% and 5.3% (p < 0.0230) in class 3, respectively in the FLH and SLH. The multivariate logistic regression showed that the FLH, older maternal age and gestational diabetes were independent risk factors for CS in groups 1 and 3. Obesity and gestational hypertension were also independent risk factors for group 1. Conclusions: TGCS is a useful tool to analyze the incidence of CS in a single center but in comparing different Hospitals, maternal characteristics and different assistential levels should be considered as potential bias.