Inter-hospital comparison for caesarean section in nulliparous, term, singleton, vertex caesarean deliveries.

Background In order to improve comparisons among hospital performances to promote a reduction of unnecessary caesarean sections (CS), we analyzed data on CS from Birth Units in Emilia-Romagna (Italy) using Robson’s Ten Group Classification System. We focused in particular on the first two groups which include nulliparous, term, singleton, vertex (NTSV) caesarean deliveries, with a low “a priori” risk of performing a CS. The aim of our study was to determine whether the adjustment for clinical and socio-demographic variables of the mother and the fetus is necessary for inter-hospital comparisons of CS rates in NTSV group. It’s important to consider NTSV CS rates rather than overall CS rates, as reducing CS in this group may be effective in reducing overall CS rates. Methods All women who delivered in Emilia-Romagna from January 1, 2007 to June 31, 2009 were identified by linking Hospital Discharge Abstracts and Birth Certificate databases. Socio-demographic variables and maternal and neonatal clinical factors were selected with a stepwise procedure in a logistic regression model and included into the risk adjustment model. We estimated adjusted RRs of CS in NTSV group by hospital in order to perform a risk adjusted comparison among the 29 Birth Units. Results The NTSV group included 46650 records and had a CS rate of 24.10%. Rates among the Birth Units varied substantially from 12.76% to 43.04%. Only one Birth Unit changed its rank of more than four positions after adjusting for clinical and socio-demographic variables. Conclusions Our data show that risk adjustment seems to be unnecessary to compare hospital CS rates in NTSV group, confirming that this group has an intrinsic adjustment. The variability in NTSV CS rates among structures could be related to non clinical determinants.