Indications and appropriateness of caesarean sections performed in a tertiary referral centre in Uganda: a retrospective descriptive study

Introduction The WHO has identified an ideal caesarean section rate for a nation of 10-15%, but much higher rates are seen in tertiary referral centres in resource-poor countries. Interventions by the author to improve care and reduced unnecessary caesareans were undertaken including staff education and production of clinical guidelines. This study aimed to identify indications for caesareans and whether the decision to perform caesareans was appropriate in order to improve care, and whether the above interventions had an impact on this process. Methods Two groups of 100 consecutive cases from October 2014 and 100 from February 2015 were retrospectively selected that resulted in caesarean. These case notes were analysed for demographic data, caesarean indication and appropriateness. Results In 46% of cases the decision for caesarean was considered appropriate. No significant difference (p>0.05) was found between the two groups in terms of patient demographics or appropriateness of caesarean (43% in Oct-14 compared to 48% in Feb-15). The most common group of indications for caesarean was dystocia (43.5%) with 28% appropriate; followed by fetal distress (18.5%) with 30% appropriate; previous scar (17%) with 85% appropriate; malpresentation (10.5%) with 48% appropriate; and maternal compromise (10%) with 80% appropriate. Conclusion The high number of unnecessary caesareans appeared to be related to lack of knowledge and inexperience of staff. Despite attempts to address this through teaching the scope of the problem is so large it needs a fundamental change in the healthcare system in terms of resources, education, continuing professional development and clinical governance.


Introduction
The World Health Organization (WHO) has identified an ideal caesarean section (CS) rate for a nation of around 10-15% [1,2]. This is based on studies that show improving maternal and neonatal morbidity and mortality as rates rise up to this level, but minimal improvements or even negative health outcomes as rates increase past 10% [3,4]. However, it is very difficult for an individual unit or hospital to use this information as an audit tool or a comparison due to the number of complex issues that impact on CS rates. In the UK national bodies have set a target caesarean rate of <23% [5], but in this resource-rich national health-care setting there are less factors that might cause large variations in rates between different hospitals. In resource-poor countries like Uganda, there is often a model of care where each region will have a tertiary centre where complex cases and severely obstructed labours might be referred, small units that perform normal deliveries, and large rural regions with traditional birth attendants. To know what an ideal CS rate for a unit like this is difficult. A number of studies from tertiary referral centres in Nigeria report CS rates of 11.3 -35.5% [6][7][8]. The rate in our unit where this study was based was 32% in 2014 [9]. Whilst these rates are generally well above the WHO recommendation, there are limited studies exploring whether these higher rates are acceptable in the context of the caseload. The objectives of this study were to look at the indications for CS, the appropriateness of this decision for CS and what alternative management might have been offered in order to explore why the CS rate was at this level.
In addition, educational interventions were instigated that might be considered a normal part of clinical governance and quality improvement, to see if these might improve the appropriateness of the decision for CS.  using Graph Pad Scientific Software. Depending on the data set this was an unpaired t-test, or a two-way Fisher's exact test. The level set for statistical significance was p < 0.05. No power calculations were performed as the original data was collected in the context of an audit rather than interventional study. Permission for the original audit work was given by the hospital director and head of department and it was presented locally. Advice was sought from a national ethics board about the need for further approval to use the data in the context of this analysis and it was not felt to be necessary. Data was anonymised appropriately.

Results
Of the 200 cases the mean age was 23.9 years with a range of 15-43 years. There were 78 primiparous women (39%). 129 women had no previous caesarean sections (64.5%), 38 one previous caesarean section (19%) and 33 two or more previous caesarean sections (16.5%). The median gestation for delivery was 39 weeks, Page number not for citation purposes 3 with a range of 28-43 weeks gestation. There was no difference in demographic data between the 2 groups from October 2014 and February 2015 (p>0.05) ( Table 1). The indications recorded for CS were put into 5 main groups consistent with other literature in the field [6,10]. The most common indication was dystocia in 44% (n=88), followed by presumed fetal distress in 18.5% (n=37), high risk of uterine rupture in 17% (n=34), malpresentation in 10.5% (n=21) and maternal/fetal compromise 10% (n=20). Table 2 shows some of the indications broken down further to give an idea of some more specific indications for CS cited. Table 3 Table 4 shows an alternative management that could have been performed safely that might have avoided a caesarean section, or where a caesarean section would have been done at a more appropriate time. There were 109 cases included that were considered not appropriate in the above analysis.

Discussion
In the context of a unit with a known high CS rate, the principle findings of the study were of a disproportionately high number being performed for dystocia (44%). Alongside this was the findings that the decision for CS was appropriate in less than half of the cases (45.5%), and that this was even higher for cases performed for dystocia, where only 28.4% were appropriate. A cohort of interventions over the 4 months of the study period to try and reduce the number of unnecessary CS was unsuccessful with no difference in appropriateness found (p=0.57). The fact that only 4% of CS decisions were made by senior doctors is also an important finding. Finally, despite more complex interventions being available is likely that the number of cases was not sufficient to show a significant impact from the intervention between the two groups.
The study was a single-centre study, so how applicable it is to other units or settings is questionable. Finally, the decision as to whether a CS was appropriate was subject, performed by one individual, making reproduction of the study difficult. The most common CS indication group by far was dystocia. This is in keeping with other studies from sub-Saharan Africa [5][6][7] and resource-rich settings [9].
However, all of these studies saw much lower proportions being performed for dystocia, 28-36% compared to the 44% we found. In