Disparities in caesarean section prevalence and determinants across sub-Saharan Africa countries

Background Access to safe Cesarean section (C-section) in resource-constrained settings such as sub-Sahara Africa (SSA) region is a foremost approach to reduce maternal mortality. C-section is an obstetric operative procedure used appropriately to improve delivery outcomes. However, errors in the procedure have enormous potential harm that may outweigh the benefits. This study assessed the prevalence and determinants of C-section in several SSA countries. This study examined the prevalence and determinants associated with cesarean delivery in SSA countries. Methods Secondary data of women of reproductive age (15-49 years) from the current Demographic and Health Survey (DHS) in 34 SSA countries was utilized in this study. The mode of delivery among women was the primary outcome variable. Percentage and descriptive statistics were used to conduct univariate analyses. Furthermore, multivariable multilevel logistic regression was used to investigate correlates of C-section among SSA women. Results Results showed disparities in the percentage of C-section among women from 34 SSA countries. C-section at public healthcare settings ranged from 3% in Burkina Faso to 15.6% in Ghana. However, in private healthcare settings, C-section ranged from 0% in Sao Tome and Principe to 64.2% in Rwanda. Overall, C-section was 7.9% from public healthcare and 12.3% from private healthcare facilities respectively. In the adjusted regression model; women aged 35–49 had increase in the odds of C-section, while a unit increase in the number of children ever born had 17 and 20% significant reduction in the odds of C-section in public and private healthcare respectively. Assessing public healthcare settings; women from richer/richest households, male and large size children at birth had increase in the odds of C-section, while those from rich neighbourhood had reduction in the odds of C-section. In private healthcare settings, women with high decision making power and multiple births had increase in the odds of C-section, while those who attended ANC visits had significant reduction in the odds of C-section. Conclusion The findings from this study would help formulate health policies and implement actions that would improve the outcome of C-section care. Monitoring of emergency obstetric care services is necessary to address issues connected to poor C-section outcomes in resource-constrained settings. Also training of medical personnel including midwives and nurses in emergency obstetric care, ensuring accessibility to life-saving drugs and supplies should be encouraged in health care system.


Background
By the end of year 2015, the Sustainable Development Goals (SDGs) emerged with a target to bring a reduction in maternal mortality ratio (MMR) to less than 70 per 100,000 live births worldwide, and to ensure healthy lives for all at all ages by 2030 [1]. Despite the immense global interventions to reduce the problem of mother and child deaths due to complications in pregnancy and delivery, the magnitude of maternal mortality remains unabated specifically in sub-Sahara Africa region [2]. This implies the necessity to provide evidence-based, quality and high-impact maternal healthcare services, particularly; universal access to emergency gynecological and obstetric care be made a priority on the global health agenda. Predominantly, developing countries are known to account for approximately 99% of the estimated 303,000 maternal deaths that occur per year worldwide, where access to antenatal care, family planning, postnatal care and emergency obstetric services have been reported inadequate [3].
In the quest to achieve SDG-3, equity and equality in availability to emergency obstetric care including assisted vaginal delivery together with safe caesarean section (C-section) is exceedingly essential [4]. C-section is a known life-saving procedure for both mother and child. Contemporary obstetrics and gynecological practice for medical, economic and social reasons have observed higher rate of cesarean section universally [5,6]. There is an increasing attention that C-section rates have consistently been on the rise, regardless of race, age, medical condition and gestational age. Global attention over such upsurge have persuaded World Health Organization (WHO) to recommend that C-section prevalence should not surpass 15% [7], with numerous evidence signifying that C-section prevalence beyond 15% were not linked to further reduction in maternal and child morbidity and mortality [8]. Notwithstanding, there are disparities in the prevalence of C-section in developing countries ranging between 2 and 39% [5,9].
C-section has become a prominent indicator of accessing progress in emergency obstetric care, and a method to avert complications during labour and delivery [10]. The role of caesarean section in poor-resource settings is difficult to obviously describe. In many low-and middle-income countries, the tool has been reportedly underutilized especially among the disadvantaged populations, and overemployed by the privileged group, while no consideration has been adopted to ensure that universal access is obtained [11]. C-section has become a priority in agenda setting to curb the menace of maternal death through improved quality and use of services for the management and treatment of complications in pregnancy, labor and delivery [7]. Holistically, a foremost strategy in the reduction of maternal morbidity and mortality includes promoting skilled birth attendance or institutional delivery and utilization of prompt C-section as a remedy to salvage delivery crisis [12]. Evidence-based studies have shown numerous factors connected to consistent rise in the rate of C-section in many communities. As reported in a previous study, women with higher economic class underwent more C-section than women without formal education and those in low wealth index class. Furthermore, women who delivered in private health facilities are known to have more C-section than the women who utilized government owned health facilities [13].
Disparities exist in C-section trends across diverse populations, economic class amongst other factors in sub-Saharan Africa region and the world at large [14]. Elsewhere in sub-Sahara Africa, a study showed huge inequalities in C-section levels across diverse socio-environmental and demographic factors indicating differentials in accessibility to health care services [15]. More so, population based studies involving large number of women who delivered at referral centers in sub-Sahara Africa countries, identified individual woman factors related to C-section, while variations were noted across the rate of intrapartum, emergent and elective C-sections [16]. Considering the various types of C-section, prominent maternal risk factors are history of previous C-section, hypertension, premature rupture of membranes amongst others [16].
As a major abdominal surgery, C-section commonly happens under critical conditions as signified during fetal distress, hemorrhage, cephalopelvic disproportion and eclampsia. Unfortunately, this procedure is sometimes performed by incompetent and poorly trained personnel [11]. Despite it is being performed in overall high-risk women with baseline risk of adverse outcomes, C-section can worsen the outcomes through damage to pelvic organs, increased blood loss and increased risk of thromboembolism, to name but a few [11]. However, C-section can prevent stillbirth, maternal morbidity and complications-related mortality during pregnancy and childbirth. It is in the light of the above that this study examines the prevalence and correlates of caesarean section in sub-Sahara Africa region.

Data source
Secondary data analysis was conducted involving individual woman component of the dataset from the Demographic and Health Survey (DHS cluster design based on a list of enumeration areas (EAs), which are systematically selected units from localities and constitute the Local Government Areas (LGAs). The LGAs are subdivisions of each of the administrative States (including the Federal Capital Territory) and classified under geographical zones/divisions [17].

Variable measurement Outcome variable
The dependent variable was the mode of delivery among women of reproductive age. Data about the method of last pregnancy delivery was collected in dichotomous form as either caesarean section or otherwise.

Independent variables
The explanatory factors include; current age of a respondent (15-19, 20-24, 25-29, 30-34, 35-39, 40-44, 45-49 years), place of residence (urban vs rural), respondent and partner's educational attainment (none, primary, secondary and higher), religion (Christianity, Islam, traditional and other religion), sex of household head (male vs female), frequency of reading newspaper, frequency of listening to radio, frequency of watching television, birth type (singleton vs multiple), sex of child (male vs female), total children ever born, age of respondents at first birth, number of antenatal visits, size of child at birth (very large, larger than average, average, smaller than average, very small) and employment status (employed vs unemployed), women's decision making power; was evaluated with questions on who made decisions about women's own health care, household purchases, visits to family members and husband earnings. All decision components were categorized into a decision made by husband or other person, a decision made together with the woman, or a decision made by the woman only. In addition, on wealth index; the calculation of household socioeconomic level involved the use of items such as possession of TV, radio, bicycle, type of floor, roof, toilet facility, water source and so forth. The principal component analysis (PCA) was used in assigning a score which were then summed and standardized for the households. The standardized scores places the households on a continuous scale based on relative wealth scores. The scores were thus categorized into quintiles to rank the household as poorest/poorer/middle/ richer/richest. The place of delivery was grouped as private vs public health facility. Neighbourhood-level factor was operationalized with a principal component analysis using the proportion of respondents with: no formal education, rural resident, unemployed and living below the poverty level (asset index below 20% poorest quintile). A standardized score with mean 0 and standard deviation 1 was computed from this index; with higher scores indicative of lower socio-economic position (SEP). We grouped the scores into tertiles to allow for nonlinear effects and provide results that were more readily interpretable in the policy arena.
Country-level factor included human development index, a measure of country's intensity of deprivation, which is the average percentage of deprivation experienced by people in multidimensional poverty. The country-level variables were also grouped into tertiles (low, middle and high levels).

Ethical approval
We conducted the analyses using publicly available data from demographic health surveys. Prior to each interview, participants gave informed consent to participate in the survey. DHS Program is consistent with the standards for ensuring the protection of respondents' privacy. ICF International ensures that the survey complies with the U.S. Department of Health and Human Services regulations for the respect of human subjects. No further approval was required for this study since the data is secondary and is available in the public domain. More details about data and ethical standards are available at: http:// goo.gl/ny8T6X.

Data management plan
The baseline socio-economic, demographic and other characteristics of respondents were computed using summary statistics together with percentages. The complex survey module (svyset) was used to account for sample weight. In addition, the association between explanatory variables and C-section among women was investigated using multivariable multilevel fixed-and-random effect logistic regression was used to obtain stratified models for public and private health facilities. The results of fixed effects (measures of association) were reported as odds ratios (ORs) with their 95% credible intervals (CrIs). Significance level was set at 5%. Data analyses was done using STATA version 14.0 (Statacorp, College Station, Texas, United States of America).

Sample characteristics
We analysed information on 159,327 respondents from 32 countries in sub-Saharan Africa (Table 1). Table 1 shows the countries, year of data collection, and the surveys characteristics. The median number of neighbourhoods sampled was 434, ranging from 104 in Sao Tome & Principe to 1504 in Kenya. The median number of respondents was 3960 (range: 1119 to 12,478). As shown in Fig. 1, there was a wide variation in the percentage of C-Section. The characteristics of the pooled sample is shown in Table 2.

C-section at public healthcare settings
The prevalence of c-section ranged from 3% in Burkina Faso to 15.6% in Ghana (Fig. 1). The results of multilevel model is shown in Table 3. In the fully adjusted model controlling for the effects of individual, neighbourhood and country level factors, maternal age, wealth, sex of child, number of children and size of the baby at birth were significantly associated with odds of c-section. Women aged 35 to 49 years old were more likely to have had c-Section to those aged 15 to 24 years old (OR = 2.75, 95% CrI 2.33to 3.22). Women from the richest households were almost as twice as likely to have had c-Section than those from poorest households (OR = 1.97, 95% CrI 1.54 to 2.55). Male children were more likely to have been delivered via c-section than female children (OR = 1.21, 95% CrI 1.10 to 1.33. The odds of c-section decreased with increasing number of children ever born (OR = 0.83, 95% CrI 0.81 to 0.86). Children large size at birth were more likely to have been delivered via c-section (OR = 1.15, 95% CrI 1.05 to 1.26).

C-section at private healthcare settings
The results of multilevel model is shown in Table 3. In the fully adjusted model controlling for the effects of individual, neighbourhood and country level factors, maternal age, decision making power, multiple birth and number of children were significantly associated with odds of c-section. Women aged 35 to 49 years old were more likely to have had c-Section to those aged 15 to 24 years old (OR = 4.31, 95% CrI 2.83 to 6.67). Women with high decision making power were more likely to have had c-section (OR = 1.52, 95% CrI 1.14 to 1.97). The odds of c-section decreased with increasing number of children ever born (OR = 0.80, 95% CrI 0.75 to 0.86).  and Ethiopia reported 33.3 and 30.1% respectively from private settings. Overall, public healthcare settings reported lower C-section in several SSA countries including Sao Tome & Principe (0%), Congo (2.3%), Burkina-Faso (3%) amongst others (see fig. 1 for details).

Discussion
The focus of this study was to investigate the prevalence and determinants of C-section mode of delivery in SSA countries using current nationally representative data from DHS conducted between 2008 and 2016. The findings from our analysis revealed disparities in the prevalence of C-section across various countries, while some reported substantial increase during the study period, some countries had very low percentage of C-section. The prevalences obtained were similar to previous reports from developing countries [18][19][20]. According to WHO, the prevalence of C-section in any population should be within the interval of 5-15%, however, some countries in SSA were found to have below the minimum 5% recommended by public and private healthcare facilities. Studies have shown that the  prominent reasons for the low coverage of C-sections were insufficient provision of equipment and medicines in the available emergency obstetric health units, lack of skilled birth attendants, unavailability of life-saving obstetrics services, long distances and poor landscapes without proper transportation could cause major geographic barrier in access to emergency obstetric care [21,22]. Conversely, despite the reasons for the high prevalence in some countries are not completely known due to multifactorial nature, it could be connected to supply-side and demand-side factors. Maternal factors, such as education and other characteristics driven by health professionals, health care system, awareness, perception and socioeconomic factors could be responsible.
The multilevel logistic regression model used to examine significant predictors of C-section, indicated that age, wealth index, parity (number of children ever born), sex of child, size of child at birth, neighbourhood socioeconomic status, women's decision making power, multiple births (i.e: twins, triplets etc.) and antenatal care visits were significantly associated with having a C-section model of delivery. Consistent with previous studies [23,24], our findings showed that women who belong to the poorest households had lower odds of C-section than those from rich households in SSA countries. The current data have confirmed poverty as a major factor responsible for the low utilization of C-section among women. The cost implication of assessing a crucial life-saving procedure such as C-section is a major factor that could hinder the process of achieving equality health care services utilization as evident in this study. However, women from least disadvantaged neighbourhood had significant reduction in the odds of C-section. Affordability of adequate antenatal care visits; a medium of behavioural change communication through feeding habits,  exercising and warning signs at onset of pregnancy-related complications enlightenment could help reduce the occurrence on emergency C-section. In this study, older women had higher C-section, which could be due to complications resulting from advanced reproductive age or due to life time C-section data used for this study. This is consistent with the findings from previous study [25]. Further, women with large number of children had less of C-section; the experience from several child births could reduce the risk of pregnancy related and intrapartum complications among the multiparous women. Child factors associated with C-section were sex, size of baby at birth and type of birth. Multiple births such as twins, triplets or more and large baby size or weight could cause cephalo pelvic disproportion or malposition. Child sex was also significantly associated with C-section among women in SSA countries. Previous studies have reported similar factors significantly associated with C-section [6,26].

Strength and limitation
This study used large sample size involving multiple nationally representative datasets from several countries in SSA to investigate prevalence and correlates of C-section in sub-Sahara Africa region. Nonetheless, the data lacks information relating to clinical indications for C-sections; as the data did not distinguish between elective and emergency C-sections. Also, the use of this information for decisionmaking and comparison should consider the cross-sectional nature of the data which is inadequate to sufficiently establish causality. Maternal empowerment has been known to affect several health care services. In this study, women's decision making power was significantly associated with increased C-section. In an effort to educate women about their birthing rights and options, guides are provided for women on healthy birth mode [27]. Educating expectant mothers with information on pregnancy, delivery and hospital C-section rates could assist women in making informed choices, leading to increase in C-section.

Conclusion
In practice, healthcare system is grouped into public and private hospitals with specific features according to the country. The differences in C-section prevalence between public and private healthcare settings are either due to the difference in prenatal and delivery care between these two settings that could influence the delivery outcome or the preference of patient mode of delivery. The high prevalence of C-section in private healthcare settings is a broad concept connected with several factors. Therefore, the interventions and programs should be targeted to address both maternal preference and professional attitude towards the mode of delivery. Educational interventions to improve quality of painless labour and vaginal delivery should be introduced in both public and private healthcare settings to lower potential C-section (elective) rate. Women's awareness towards social beliefs as C-section (elective) is safer than normal (vaginal) delivery and information regarding complications of C-section and their outcomes should be enhanced. Active involvement of the policy sector is needed to strengthen equity and universal health coverage in maternal healthcare. The results suggest the need for accurate and timely screening of women during obstetric care and, the choice of performing C-section should be based on clear and well-supported justifications.