For each year, funnel plots have been used in order to investigate the degree of variation in the three different caesarean rates amongst the NHS hospitals. These plots indicate whether the crude caesarean section rates differ significantly from the respective national averages (the horizontal red lines in each graph) and indicate the existence of outliers.
Regarding the elective c-section rates, their dispersion seems to increase after 2003. Moreover, in every year there is a number of trusts with very high elective c-section rates (15% or more). For the total c-section rate, the number of hospitals exceeding the threshold of 30% is increasing in every year, driving the national average upwards. As noted before, the non-elective c-section rate follows a quite modest trend, however, it seems that the variation amongst NHS trusts is also increasing every year. Therefore, there is some first descriptive evidence according to which there is a substantial degree of variation amongst hospitals regarding the incidence of caesarean sections especially during a time when guidelines regarding the promotion of normal birth are being issued and variation in clinical practice is often cited as undesirable. Also, reducing the number of unnecessary caesarean sections will bring about benefits for both mothers (not only they will be able to return to their homes quicker, but the probability of complications will be reduced) and the NHS. Caesarean sections are far more costly from vaginal deliveries and, according to the Parliamentary Office of Science and Technology, a 1% rise in c-sections rates costs the NHS approximately 5 million pounds per year. Moreover, apart from the nation-wide variation in c-section rates, the behaviour of individual hospitals with respect to the c-section rates seems to vary greatly over time as well. In the following “caterpillar” plots, trusts have been ranked according to their overall performance. The mean c-section rates (the yellow diamond symbols) of many hospitals are well above (or below) the national average (the red vertical lines) for the total period, while the plotted interquantile ranges (the horizontal boxes) and the 95% confidence intervals (the horizontal capped spikes) indicate that the performance of each hospital has not remained stable during the period 2000-2013. Yet, as noted above, there are some hospitals which seem to have persistently high (or low) rates of c-sections.