Timing of birth and mode of delivery after spontaneous labour onset

Background High caesarean rates are a matter of concern and the influence of hospital functioning in such rates has been debated. Objectives To understand the hourly pattern of deliveries after spontaneous labour onset, and to compare the trajectories of the mean number of deliveries on working days with those on weekends, according to the mode of delivery. Methods Between April/2005 and August/2006, women (n = 8,495) delivering a live birth in public hospitals (level III) were consecutively recruited during the procedure of assembling a birth-cohort in the North of Portugal. For this purpose we selected those with a singleton pregnancy and admitted with spontaneous labour (n = 4,087). For eligible subjects, date, hour and mode of delivery were collected from medical records. Statistical modeling used unidimensional cubic splines with hour, ranging from 0 to 23, as predictor and the mean number of vaginal and caesarean deliveries as dependent variables. These splines are functions defined on the domain of the independent variable and that consist of a finite junction of local cubic polynomial regressions. Fitting and selection of the models was assessed through the percentage of explained deviance, generalized cross-validation score and residuals inspection. Comparison of curves was based on the 95 % pointwise prediction confidence intervals. Results The percentage of explained deviance in the models was 85 and 68 % for vaginal and 90 and 51 % for caesarean deliveries, on working days and weekends, respectively. On working days, a pattern of relative deficit of nocturnal deliveries with a minimum between 5 a.m. and 6 a.m. was observed for both vaginal and caesarean deliveries. The hourly variation in diurnal births was different according to the mode of delivery. Vaginal deliveries steadily increased from 6 a.m. onwards, reaching a peak at 6 p.m. Caesarean deliveries presented two diurnal peaks; the first at 11 a.m. and the second in the late afternoon, between 6 p.m. and 7 p.m. The hourly pattern of vaginal deliveries on weekends and working days was quite similar, but a significant difference in the mean number of vaginal deliveries was observed between 10 p.m. and 11 p.m., which is lower on weekends. The peaks for caesarean deliveries on working days disappeared on weekends, the only pattern similarity being the decrease observed overnight. Conclusion There was no evidence of the weekend effect in the mean number of caesarean deliveries. Nonetheless, the differences observed in the hourly pattern of vaginal and caesarean deliveries on working days suggest the influence of hospital functioning in the time when caesarean is performed