Statistical analyses
Data were analysed using SAS version 9.4. Proportions were expressed as percentages. Non-normally distributed continuous data were expressed as medians and interquartile ranges. We described differences in raw caesarean delivery rates between Group A and Group B by Robson group. Robust Poisson regression was used to generate relative risks for the outcome of caesarean delivery, using the following explanatory variables: time-period (Group A or Group B), maternal age, maternal body mass index, parity group (nulliparous, parous, or parous with ≥ 1 previous caesarean deliveries). Due to interactions between plurality, malpresentation, and preterm birth with time-period, the data were divided into four strata: (1) term cephalic singleton births; (2) preterm singleton births; (3) multiple gestations; (4) term non-cephalic singleton births. Robust Poisson regression was performed only in the first stratum (term, cephalic, singleton births) because maternal characteristics were not associated with the increase in caesarean deliveries within the other three strata.
Linearity of maternal age and BMI was assessed by categorizing these variables in to four groups and plotting the β-coefficients in the regression equation against the midpoints of these groups. If there was a non-linear relationship, the categorized variables were used. Missing data were handled using multiple imputation with 20 iterations.
To estimate the contribution of maternal factors to changes in the caesarean delivery rate within the first stratum ([1] term cephalic singleton births), we calculated the increase in number of caesarean deliveries attributable to maternal factors using the following formula:
\begin{equation} {Increase\ in\ number\ of\ caesarean\ deliveries\ attributable\ to\ maternal\ factors=\backslash n}{Increase\ in\ number\ of\ caesarean\ deliveries\ among\ women\ \times\ Proportion\ of\ the\ increase\ in\ the\ caesarean\ delivery\ rate\ attributed\ to\ maternal\ factors\backslash n}\nonumber \\ \end{equation}
Where,
\begin{equation} Increase\ in\ number\ of\ caesarean\ deliveries\ among\ women=\ \left(P_{B}-P_{A}\right)\times N_{B}\nonumber \\ \end{equation}
where Pi = proportion of all women with a term cephalic singleton birth in Group i who had a caesarean delivery, and