Recent trends in cesarean section reduction in extreme south of Brazil: a reality only in the public sector?

This study aimed to describe changes in cesarean section (C-section) prevalence from 2007 to 2019, in public and private sectors, according to maternal characteristics. We included all puerperal women who gave birth in Rio Grande, Rio Grande do Sul, Brazil, at years 2007, 2010, 2013, 2016 and 2019. A questionnaire was applied up to 48 hours after delivery. We assessed C-section rates over time and described the prevalence according to independent variables. Poisson regression was used. A total of 12.415 puerperal women were included. The prevalence of C-section increased between 2007-2013 (from 51.2% to 61.2%) and decreased between 2013-2019 (48.9% in 2019). This decrease was observed only in the public sector and was higher among the youngest (-10.0 percentual points) and high educated women (-10.3 percentual points). While in the private sector C-section occurrence increased even more (95.7% in 2019). In the public sector, women that were older, with a partner, primiparous, who performed prenatal care in the private system and with adequate prenatal assistance presented higher prevalence of C-section. In the private sector the prevalence was high independently of the maternal characteristics. In order to reduce C-section rates, efficient delivery care policies mainly focused on the private sector are necessary.


introduction
A cesarean section can prevent maternal and perinatal mortality and morbidity when medically justified. However, according to the World Health Organization, there is no evidence that an occurrence higher than 15% could benefit both mothers and newborns 1,2 . In addition to exceeding this percentage, C-sections have steadily increased over the past three decades 3,4 .
Considerable variations in cesarean rates are found worldwide, the lowest in Africa (7.3%) and the highest in Latin America and the Caribbean (40.5%) 5 . Brazil is one of the countries with the highest prevalence of C-sections globally. In the 1970s, the country exceeded the 15% published by WHO, reaching 55.0% in 2016; it has become about four times higher than the justifiable limit 6 .
Social inequalities are observed in C-section rates. Higher rates are more frequent in women from higher socioeconomic status, who had more prenatal consultations, primiparous, had a previous cesarean section, delivered the child with a doctor after a prenatal period, and had low gestational risk [7][8][9] . However, the main inequality seems to be the funding of childbirth care. According to data from the Brazilian National Supplementary Health Agency (ANS), 85% of births in the private sector were performed by C-section in 2018, while in the public sector (SUS -Unified Health System), this rate was 43% in 2018 10 .
Frequent monitoring of this rate is essential to understand the effectiveness of implemented policies and guide new ones. Nevertheless, population-based studies assessing birth trends over time with a short periodicity between evaluations are scarce. This study aims to describe changes in C-section prevalence from 2007 to 2019, in public and private sectors, according to maternal characteristics in Rio Grande, Brazil, to learn the trends of C-sections over 12 years in the south of Brazil and the relative contribution of each sector (public and private). These studies included all puerperal women living in urban or rural areas of the municipality whose newborns weighed at least 500 g or reached 20 weeks of gestational age. These women were interviewed within 48 hours following delivery after signing the informed consent form authorizing their participation in the study. We excluded multiple births from the analysis.
The survey questionnaire was applied by trained interviewers who looked for puerperal women daily in the maternities. Each survey had at least three interviewers to apply the questionnaire to the puerperal women. This questionnaire investigated demographic, occupational and socioeconomic characteristics, reproductive history, habits of life and behavior, pattern of morbidity, and use of health services during pregnancy. All information contained in the pregnant women's card used in prenatal consultations was also collected, as well as the newborn's medical records.
The outcome of interest in this study was C-section occurrence. The analyzed independent variables were: i) sociodemographic: maternal age (11-19, 20-24, 25-29, 30-37 years), self-reported skin color (white, mixed, black), living with a partner (yes or no), completed years of schooling (0-8 years, 9-11 years, 12 years or more), monthly family income in minimum wages (in quartiles); ii) past obstetric history: parity (primiparous, multiparous); iii) maternal morbidity during pregnancy: hypertension (yes or no), diabetes (yes or no), depression (yes or no) and anemia (yes or no); iv) current obstetric history: prenatal care (public or private), adequacy of prenatal (adequate or inadequate). Prenatal care was considered adequate when pregnant women attended six or more appointments starting before the fifth month of pregnancy, performed two or more qualitative tests of urine and two or more diagnostic tests of HIV and syphilis.
After consistency errors, data were analyzed using Stata 16.1 software 11 . First, we presented the total births and C-section (CS) occurrence according to maternal characteristics and perinatal surveys. Secondly, we graphically presented the total number of births and the prevalence of cesarean section, according to childbirth care financing -private or public healthcare (SUS -Unified Health System). Due to expressive differences between public and private C-section prevalence, the following analyses were stratified according to childbirth care financing. Thirdly, we presented the adjusted prevalence of cesarean and the respective 95% confidence interval (95%CI), according to maternal characteristics stratified by childbirth care financing. We then performed a multivariate analysis using Poisson regression with robust confidence intervals, followed by the "margins" post estimation command. Three hierarchical levels were considered according to a conceptual framework. At the first level, family income and maternal sociodemographic variables (age, skin color, schooling, and living with a partner) were included. At the second level, parity, adequacy of prenatal care, and public or private prenatal care were included. At the third level, we included hypertension, diabetes, depression, and anemia during pregnancy. All these variables were inserted into the model using backward selection, each level at a time, excluding those with p<0.20.
Lastly, we present a summary of the main changes in the occurrence of C-sections over the study years using the linear trend test. Percentage points (pp) were calculated for the public and private sectors considering the difference between the cesarean rate at the end (2019) and the beginning (2007) of the study period. A p-value lower than 0.05 for a two-tail test was specified for these analyses.

Results
A total of 12.645 puerperal women were interviewed, with a response rate of around 98,0% in all study years. After multiple births exclusions (n=230), 12,415 women were included in the analysis (Figure 1). Table 1 shows the sample description and the occurrence of C-sections according to the year. The prevalence of C-section increased from 51.2% in 2007 to 61.2% in 2013, while from 2016 to 2019 this occurrence decreased (from 54.0% to 48.9%, respectively). We observed an increase in C-sections from 2007 to 2013 among all maternal characteristics. However, we found a decrease from 2013 to 2019, except in the private sector, which increased overall. Although the total number of deliveries decreased in the private sector in 2016 and 2019, the prevalence of C-section increased, reaching 95.7% in 2019. On the other hand, this prevalence decreased from 2013 to 2019 in the public sector (from 46.5% to 39.8%, respectively) ( Figure 2). Table 2 presents the adjusted prevalence of C-section according to independent variables and study year in the public and private sectors. Among those who gave birth in the public sector, older women, those who have a partner, primiparous, performed prenatal care in the private sector, with adequate prenatal assistance, and those women who have hypertension or diabetes presented higher C-section occurrence. While among women who gave birth in hospitals from the private sector, the occurrence of C-section presented little variation according to maternal characteristics and, in general, it was higher than 80%. Table 3 summarizes of the changes in C-section prevalence in public and private sectors from 2007 to 2019 according to the characteristics presenting more variation. We observed a decrease in global C-section prevalence due to a decrease in this procedure in the public sector. In this sector, the decline was higher among women with higher education (-10.3pp) and the youngest (-10.0pp). The increase in C-section rates in the public sector were among older women (3.0pp), with hypertension (3.6pp) and diabetes (0.7pp). Despite this reduction in global and public C-section rates, there was an increase in almost all variables in the private sector. This increase in C-sections was more expressive among women with adequate prenatal care (15.0pp) and diabetes (10.6pp).

Discussion
Our findings show that C-section prevalence increased from 2007 to 2013 and decreased from 2013 to 2019. Moreover, this recent decrease occurred unevenly between public and private health services. A real drop in the cesarean section rate was seen only in the public sector, which concentrates more births, while in the private sector, the prevalence increased even more during this period. In addition, C-section rates were higher among older women in the public sector, with a partner, primiparous, who performed prenatal care in the private system, and with adequate prenatal assistance. However, C-section prevalence was high independently of maternal characteristics in the private sector.  Monthly household income in minimum wages; b Considered adequate when pregnant women attended six or more appointments began prenatal care before the fifth month of pregnancy and did two or more qualitative tests of urine, two or more tests diagnosis of HIV and syphilis.  Caesarean section can save women's and infants' lives and should be universally accessible 12 . Nevertheless, its overuse has been a serious problem worldwide. Data from 169 countries estimated that between 2010 and 2015, the number of babies born by C-section almost doubled (from 12% to 21% of all births) 13 . This occurrence was even higher in Latin America and the Caribbean in 2015 (about 44% of all births). Brazil was the second country with the highest number of C-sections; rates increased considerably between 2000 and 2011 (from 37.7% to 53.9%) 7 , reaching 57.0% in 2014 and presenting a slight decrease in 2016 (55.0%) 14 . This reduction in the C-section rates has been the first one since the 1970s and has possibly demonstrated the beginning of a change in childbirth care practice 15 .
In the present study, we observed a decrease of 2.3pp in C-section rates over the years of study. This decrease was 12.3pp when we considered the period from 2013 to 2019. Possible explanations include improvements in the Unified Health System (SUS) and implementation of protection measures for vaginal delivery, such as the presence of a companion in the delivery room, inclusion of an obstetric nurse, and the need for parturient authorization for procedures such as C-section and episiotomy. These efforts may have encouraged women to have a vaginal delivery and could have inhibited the medical team from performing unnecessary cesarean sections, especially in the public sector. Moreover, an obstetric regulation focused on reducing not clinically recommended cesareans was implemented by ANS (National Agency of Supplemental Health). However, given the increase in cesareans trends in the private sector, this norm does not seem to be efficient 16 .
C-sections in the private sector presented the highest increase between 2016 and 2019 (about 6.7pp), reaching 95.7% of all births, and this finding agrees with the literature 9,17 . This increase is indeed not explained by clinical necessity alone. Pregnant women assisted in the private sector have greater power to decide on the type of delivery. The private sector allows the scheduling of C-sections on request based on the professional's choice, the team, and even the health establishment 18 . In a study conducted in the state of Parana (Brazil), the high occurrence of C-sections in the private sector (93.8%) was mainly determined by the women's desire for this type of delivery in early pregnancy (OR=18.3) 17 . Another argument explaining high C-section rates in the     private sector is that providers generally receive a higher payment for C-sections than for vaginal births 19 . Physician practice styles and familiarity with high-tech procedures may be more prominent in private hospitals, influencing recurrent C-section practice 20 . Despite this increase in C-section rates in the private sector, it is important to note that the total proportion of births in this sector decreased by 4.6pp over time. Furthermore, deliveries were even more concentrated in the public sector (about 80.0%). These differences in the proportion of births would explain why the global C-section rate in the municipality has decreased, despite the increase observed in the private sector. In addition, in our study, C-section rates were high even in cases that do not depend on maternal clinical characteristics and which commonly characterize groups with lower obstetric risk, such as women who performed adequate prenatal assistance, those with prenatal care in the private sector, and who lived with a partner. This scenario was seen in both private and public sectors, but more markedly in the latter. Likewise, a study performed in Brazil concluded that C-sections were more common among women at low risk of maternal or fetal death, suggesting, once again, that the option for this type of delivery was mostly elective 7 . A previous publication also using data from the "Estudos perinatais" observed that C-section occurrence by request of pregnant women doubled, from 11% in 2007 to 22% in 2016 21 . Only these percentages of caesarean sections alone already exceed the justifiable limit published by WHO.
Another important result observed in this study was the higher occurrence of C-section in primiparous women in public and private sectors. This may be related to medical convenience because the progression of labor tends to take longer among primiparous women 22 . In addition, fear of childbirth and concerns related to vaginal trauma can influence primiparous women to request a C-section even before labor 23 . This is alarming, once a previous cesarean section could increase the risk of a repeat surgical delivery 24 . A more careful evaluation of indications for primary cesarean is essential.
The literature has shown the presence of social inequalities in C-section occurrence 7,25,26 . In Brazil, family income and maternal schooling are the main determinants of this occurrence. Studies performed in the south of the country in 1982,1993,2004, and 2015 observed that the richest and the most educated women presented the highest cesarean section occurrence 9 . Our data agree with this, showing the same pattern in the public and private sectors. This finding could be explained by the higher purchasing power of these families, the empowerment of women with respect to the right to choose, the freedom to arbitrate about their health, and the access to new technologies, in addition, for some women, a cesarean section implies better health care [26][27][28][29] .
When interpreting the results, we need to consider that most of the information comes from the pregnant women's reports obtained through a single approach and may have been affected by the recall bias. However, we sought to minimize this bias through the interview in the first 48 hours after delivery. Besides, the clinical indications for C-sections were not verified, which could better explain our results. As a strength, it is noteworthy that this study represents a census with a large sample size, a low percentage of losses (1.9%), carried out periodically since 2007, enabling the temporal evaluation of delivery indicators.

Conclusions
Our results showed a decrease in cesarean section rates between 2013 and 2019, which could result from public policies implemented in the Unified Health System (SUS). Nevertheless, in the private system, this occurrence is almost universal. It is essential to point out that this rate remained extremely high, even in the public sector, and improvements are still needed. Future interventions focused mainly on the private sector may bring more effective results in reducing the overall occurrence of cesarean deliveries.

Collaborations
LP Marmitt: methodology, writing -original draft preparation, formal analysis. AKF Machado: writing -original draft preparation and formal analysis. JA Cesar: conceptualization, methodology and supervision.