Main Findings
In this study we found an overall CS rate of 32.7% at EH from 2014 to 2019, this is close to the national CS rate of 35% 6. Group 5 had a relative contribution of 35.2%, making it the main contributor, followed by group 2 with a relative contribution of 26.6% and then group 1 with a relative contribution of 10.3%. These results are similar to other studies11, 14, 15 and also similar to previous data from Australia 5. The CS rate at EH increased 1.5% per year, from 30.7% in 2015 to 36.0% in 2019, mainly due to increase at BHH. This is the same scenario experienced by other institutions in Australia and in several other developed countries11, 13. This finding allows us to identify these as target groups for interventions aiming to reduce CS in our institutions.
In both hospitals, nulliparous women (groups 1 and 2 together), were the main population contributing to the overall CS rate accounting for just over a third of all CS cases. This finding is very important as woman’s first birth experience has a profound influence on the method likely to be used in her next pregnancy 16.
Interventions aiming to reduce the CS rate in group 5 include increasing the availability of vaginal birth after CS (VBAC) but also reducing the CS rates in both groups 1 and 2. Women’s decisions on mode of delivery after a CS are influenced by their previous birth experience and their current expectations. Antenatal counselling and education might be beneficial for this group to increase their awareness and perceptions about VBAC. Additionally, hospital policies and clinician perceptions towards VBAC are also key aspects to address 14.
The size of groups 1 and 2, (nulliparous women) are within the expected range of 35-42% as per the Robson guidelines 9 and previous papers 17, however, the size of groups 3 and 4, multiparous women, are slightly higher than the 30% recommended in the guideline, suggesting that we could be serving a population with higher fertility rates. The size of group 5 is relatively high and could indicate that we serve many women with previous CS, perhaps many of whom were in groups 1 or 2 in past years at our own institutions, although our current data does not allow us to explore this. Group 8 and 10 are both at the higher end of the expected size, which gives an indication that EH may serve a higher risk population. However, the details of the population variation and level of complexity of our population was not possible to capture by looking at the Robson classification alone.
Although the nulliparous distribution is within the expected range, we found a less than 2:1 ratio between the sizes of group 1 and 2, and groups 3 and 4. This suggests a high incidence of induction of labour and pre-labour CS in both nulliparous and multiparous women. Similar finding was reported by Robson in 2015 17 and efforts to understand indications for induction of labour and pre-labour CS were suggested as a strategy to reduce CS rates.
Women who had Induction of labour, both nulliparous and multiparous, (groups 2a and 4a) had a high CS rate, which was due mostly to increased need for emergency CS. This finding is similar to what has been found in Queensland 16 and in the Nordic countries11. Improvement of surveillance and overall management of induction of labour, including the use of oxytocin and ripening agents are potential interventions to consider. Interestingly for group 4, although induction of labour was still common, the CS rate was much lower than for group 2 in our institutions, similar to results found in Queensland 16. This suggests than perhaps inducing labour in multiparous women does not carry the same risk of CS as when inducing nulliparous women. However, we found differences in our CS rates in groups 2a, 3 and 10 between BHH and AH, suggesting that perhaps we need to provide slightly different strategies in each of our hospitals to reduce our overall CS rate at EH.
Other independent risk factors for CS have been described18 and obstetric pathology like hypertensive disorders, diabetes, obesity and advance maternal age are also independent variables capable of influencing the CS rate. Other studies1 have shown how ethnicity, specially being Asian-born, along with other maternal characteristics, were also associated with increased risk of CS. Therefore, variables such as BMI, maternal age, previous maternal comorbidities and pathologies and foetal compromise should be included when analysing this type of data. In addition, the type of population that EH serves seems to be more complex that captured by the Robson classification and gathering further data regarding their epidemiological and demographic characteristics could help to better understand the CS rates and identify strategies to reduce it 11.