Introduction This study aimed to evaluate the impact of planned caesarean section (PCS) and investigate variation in its use in 11 hospital maternity services in the Umbria region (Central Italy).

Methods A population-based study using routine maternity discharge data (CEDAP) from 2007 to 2008. The study included 15 898 women (nulliparous, single cephalic, ≥37 weeks, physiological pregnancy and without fetal development defects). The outcome of the study was the “planned caesarean” section rate and excluded women with “emergency caesarean”. Using a multi-level model we evaluated non-observed heterogeneity; the second level variable was represented by different birthing centres and two context variables were included: the presence or absence of neonatal intensive care unit and the capacity of the service (1000 birth per year).

Results Of the 6811 births, 5934 (87.12%) were born by natural labour and 877 (12.88%) by PCS. The logistic model with random effects including some control variables at the individual level indicated a significant value in the residual variance of the second level (ρ 0.086). Introducing the two context variables resulted in a substantial reduction in unexplained variance.

Conclusions Variation in PCS rates in the 11 birthing centres is explained by the presence or absence of a neonatal intensive care unit and by birth centre capacity. These findings can be used by policy-makers to increase the efficiency and availability of appropriate technology in hospital maternity services.