Assessing Socio-Demographic Factors of Cesarean Delivery in India: Insights From National Family Health Survey-4

Background: The rate of cesarean delivery has considerably increased in the world during the last few decades. This paper aims to investigate the prevalence and socio-demographic correlates of cesarean delivery in India with a focus on Bihar and Tamil Nadu, accounting for the lower and higher proportion of cesarean birth. Methods: The study is based on secondary data, collected from the latest National Family Health Survey in 2015-16 (NFHS-4). The present study is based on 190,898 most recent births during the ve years preceding the survey. Bivariate and multivariate analyses were carried out to identify the factors associated with cesarean delivery. Results: The highest rate of cesarean delivery was observed in Telengana (60%) and followed by Andhra Pradesh (42%), Tamil Nadu (36%). C-section found to be negligible in low-income states, namely Bihar (7%), Madhya Pradesh (10%) and Jharkhand (11%). Multivariate analysis revealed that the prevalence of cesarean delivery was substantially higher among women who married at higher age, with rst birth-order, women in the ‘obese’ category with higher education, those from upper quintile of household wealth. Similarly the incidence of cesarean birth was remarkably higher in private hospitals both in the Indian sample (OR 3.9, 95% CI: 3.77-4.03) and in the selected states, Bihar (12.86, 95% CI: 10.92-15.15) and Tamil Nadu (OR 2.74, 95% CI: 2.40-3.13), compared to the public hospitals. Conclusion: Our study revealed that there are a high proportion of women delivering babies through cesarean section in South India. Thus, medical justications need to be taken to deal with this concern. On the other hand, Women should also be driven towards regular ANC check-ups for the well-being of maternal and newborn health that can also help to increase the cesarean delivery for women who need C-section delivery especially in low-income states.

However, during the last decade, the increase in cesarean section deliveries has become a matter of serious concern for public health experts globally. 8 India is not an exception. The rate of cesarean delivery is dramatically increasing from 10% in 2005-06 to 17% in 2015-16. The Southern Indian states have recorded substantially higher levels of cesarean deliveries in comparison to the other parts of the country. 9,10 Studies have indicated that socio-demographic, economic, cultural and psychological factors determine cesarean delivery in India. 8 It is evidenced that education, wealth status, occupation, place of delivery and nutritional status of the mother are signi cant predictors of cesarean delivery. 6,11 Previous studies revealed that educational attainment and household wealth status are positively associated with a cesarean birth. A study conducted in Bangladesh investigated higher educated women almost 4 times more likely to have cesarean birth. 6 In Brazil, on the contrary, women with higher education are less likely to have a cesarean birth compared to uneducated women. 12 In regard to cultural factors, the healthseeking behavior of women during pregnancy has a key role in cesarean delivery. 7,13 Additionally, various psychological causes, such as fear related to prolonged labor and vaginal delivery pain reinforce women's preferences for cesarean delivery. 7 The Indian government has initiated several important safe motherhood programs to reduce the incidence of maternal mortality and to improve the overall well-being of mothers and babies. India has made remarkable achievements in institutional delivery over recent decades. The Indian government has started Janani Suraksha Yojana (JSY) to improve institutional delivery in the country, particularly among the vulnerable section of society by providing economic incentives to poor mothers. Despite several maternal healthcare initiatives, the incidence of maternal mortality for some of the states is still alarming.
For instance, maternal mortality is substantially higher in Bihar compared to the national average. The United Nations (UN) set Sustainable Development Goal (SDG) to reduce the incidence of maternal mortality as low as below 70 per 100,000 live births by 2030. A study on maternal mortality in India observed the highest maternal mortality rate in rural areas of poorer states (397 per one million live births), in which 82% maternal deaths were attributed to direct obstetric causes. 14 It is noteworthy to mention that maternal mortality could be averted by cesarean delivery. Women who had obstetric complications should undergo cesarean delivery.
Studies also indicated that the geographical region also has a signi cant in uence on cesarean delivery. 11,15 According to the NFHS-4, the Southern states, namely Telangana (58%), Andhra Pradesh (40%) and Tamil Nadu (34%) constitute higher cesarean delivery rate, whilst central and eastern states like Bihar (6%), Jharkhand (10%) and Orissa (13%) have a lower proportion of cesarean delivery.
In this present study, we assess the socio-demographic factors of cesarean delivery among Indian women using the most recent nationally representative database. Furthermore, our study focuses on two socio-culturally distinct states such as Bihar and Tamil Nadu to understand the differential in uence of various factors in cesarean delivery.

Methods
We used the data from the fourth round of the National Family Health Survey (NFHS-4), conducted in 2015-2016. Data were collected from 601,509 households and 699,686 women in the age group of 15-49 years, with a response rate of 97% . The samples were collected using a strati ed two-stage sampling   technique covering 28586 clusters, where 20,059clusters are from rural areas, 8,397 clusters are from urban areas, and 130 clusters selected from the slum list provided by Municipal Corporation O ce (MCOs). In the rst stage, clusters have been chosen using probability proportional to cluster size. In the second stage, 22 households were selected from each cluster with an equal opportunity systematic selection from the household listing. A detailed description of the sampling design and survey procedure has been provided in the national report of NFHS-4. Our study participant restricted to most recent birth during the ve years preceding the survey (N=190,898).

Outcome variable:
The outcome variable of this study is cesarean birth. In NFHS-4, women were asked, "was the baby delivered by cesarean section, that is, did they cut your belly open to take the baby out?" Women who reported having cesarean delivery for most recent birth were classi ed as yes (coded as 1) or no (coded as 0).

Explanatory variables:
In this study, the socio-demographic characteristics of women were included as explanatory variables. Demographic characteristics include women's age (15-24, 25-34 and 35+ years), age at rst marriage (<18, 18-24 and ≥24 years), birth order ( rst, second, third and fourth or more). Body mass index (BMI) of mothers was also included as an independent variable. Women's BMI was classi ed as thin (BMI < 18.5kg/m 2 ), normal (BMI of 18.5-24.9 kg/m 2 ) and obese (BMI ≥25 kg/m 2 ) as per the WHO standards.
The selected socio-economic variables are women's education (no education, primary, secondary and higher), religion (Hindu, Muslim, Christian and others), social group (Scheduled Caste [SC], Scheduled Tribe [ST], Other Backward Classes [OBC] and others) and wealth index (poorest, poorer, middle, richer, and richest). It is noteworthy to mention that the household wealth index has been used as a proxy indicator of economic status. Wealth index was computed by the ownership of household assets including consumer items and housing characteristics using principal component analysis (PCA).
Place of delivery was categorized as public facility and private facility. Public facilities include government/municipality hospitals, government dispensaries, urban health clinics/urban health posts (UHP)/urban family welfare centers (UFWC), community health centers (CHC)/rural hospitals/block primary health centers (BPHC), PHC/additional PHC, sub-centers and other public sector health facilities, while private facilities include hospital/maternity home/clinics, other private sector health facilities and NGOs or trust hospital/clinics. Antenatal care (ANC) is an important predictor of cesarean delivery. ANC visits variable was classi ed as no visit, 1-3 visits and four or more visits. Pregnancy loss was dichotomized as yes (includes abortion, miscarriage and stillbirth) and no. Similarly, delivery complications were grouped as binary responses, such as yes (includes prolonged labor, breech presentation and excessive bleeding) and no.
Women's exposure to mass media was measured from the frequency of listening radio, reading newspapers/magazines and watching television. All these three variables on mass media have been dichotomized (yes/no). Place of residence was classi ed as rural and urban. Indian states and union territories grouped into six regions based on geographical contiguousness and cultural similarities. These regions are North, Central, East, Northeast, West and South.

Statistical analysis:
Descriptive statistics were carried out to understand the distribution of study participants. Bivariate analysis was conducted to examine the nature of the association between cesarean delivery and sociodemographic characteristics of women. The sample weight was used to estimate the percentages. Furthermore, Pearson's chi-square was performed to test the level of signi cance in the association.
Finally, binary logistic regression was applied to examine the factors associated with cesarean delivery. The regression results are presented by an estimated odds ratio (OR) with 95% con dence interval (CI). In case of Tamil Nadu, the rate of CS delivery rapidly increased (from 7.1% in 1992-93 to 34.1% in 2015-16), while comparatively slower progress in C-section delivery had been observed in Bihar (from 1.1% in 1992-93 to 6.2% in 2015-16). We found a large gap in cesarean delivery between Tamil Nadu and Bihar in all four survey years.
Cesarean delivery according to the states and union territories of India: Table 1 displays the prevalence of CS deliveries in all the states and union territories of India. On average, the rate of cesarean birth was 19.2% in India, ranging from 7.1% in Nagaland to 59.7% in Telangana. It was observed that Kerala, Tamil Nadu, Andhra Pradesh, and Telangana together comprised 24.5% of all C-section delivery in India. In nine states, the rate of CS delivery was lower than 10% namely, Rajasthan (9.9%), Meghalaya (9%), Bihar (7.4%), and Nagaland (7.1%). Among union territories, the prevalence of CS delivery was ranging from 39.9% in Lakshadweep to 17.1% in Daman and Diu.
Respondent characteristics: Table 2 presents socio-demographic characteristics of the women. A substantial proportion of women were in the age group of 20-24 years (56%), married between 18-24 years of age (54%). Nearly 60% of women belong to average BMI. Almost a quarter of the respondents had no formal education and about 46% of them had a secondary level of education. Majority of the women were living in rural areas, a liated to Hindu religion, and belonged to OBC social group. The distribution of study participants decreased from bottom to upper quintile of household wealth. Around 71% of the women in India were depends on public hospitals to conduct the delivery. A substantial proportion (47%) of participants has received 4 or more than 4 times ANC visits during their pregnancy. Only 15% and 11% of respondents had ever experienced pregnancy loss and delivery complications respectively. Most of the women were watching television compared to reading newspapers or magazines and listening to radio in their daily life. The study participants from central India were comparatively higher than any region in India.
Prevalence of cesarean delivery by socio-demographic characteristics: Table 3 presents the proportion of cesarean delivery by socio-demographic characteristics of women. The results show that the prevalence of cesarean delivery was substantially lower among the older women (aged 35-49 years) than the younger women (aged 15-24 years and 25-35 years) at the national level.
Similarly, in Bihar, the rate of CS delivery decreased with the increasing age of women. Contrastingly, in Tamil Nadu, the incidence of C-section delivery increased among the older age group of women. For instance, in Tamil Nadu, about 44% of women aged 35-49 years had given birth to the C-section as compared to 31% of younger women (aged 15-24 years). The prevalence of CS delivery also increased with increasing age at marriage. A greater percentage of women who married at 25 years or above had their most recent birth at C-section. Birth order had a negative association with C-section delivery where the incidence of CS delivery was found to be decreased with higher-order children in Bihar and Tamil Nadu and as well as at the national level. The occurrence of C-section delivery was signi cantly higher among overweight/obese women as compared to thin and normal women. It is noticed that CS delivery had a rising trend with an increasing level of education in which respondents with higher education had a higher proportion of CS delivery. The prevalence of CS delivery was lower among Muslim women compared to Hindus, Christians, and others at the national level and Bihar. On the other hand, in Tamil Nadu, the rate of CS delivery was lower among Hindu women than the Muslims, Christians, and others. The prevalence of C-section delivery was lowest among Scheduled Tribe women, followed by Scheduled Caste women, whilst the General caste women had the highest rate of CS delivery in both the states and India. Furthermore, the rate of C-section delivery increased among women from bottom to upper quintile of household wealth at the national level as well as two other selected states.
In regard to public-private distribution, the incidence of CS birth was remarkably higher in private hospitals than the public hospitals at the national level (43 vs. 13%). In Bihar, the proportion of C-section delivery was 34.1% in private hospitals, while it was only 2.9% in public hospitals. In Tamil Nadu, over half (52.8%) of the deliveries in private hospital shad occurred in C-section.
There was a positive relationship between the number of ANC visits and the prevalence of CS deliveries where mothers who received four or more ANC visits had higher CS delivery compared with women who did not have su cient ANC. A signi cantly higher percentage of women who had experienced pregnancy loss undergo C-section delivery as compared with those who did not experience pregnancy loss. Similarly, mothers who had faced delivery complications at the time of childbirth were more likely to experience CS deliveries in India and Bihar. In Tamil Nadu, contrasting results were found where women who experienced delivery complications considerably were less likely to deliver in C-section than those who did not experience delivery complications.
Exposure to mass media had a positive impact on CS delivery. Women who had access to any mass media reported higher experience of C-section delivery in India and Bihar. In Tamil Nadu, women's exposure to newspaper/magazine only made signi cant difference CS birth. Our study found that the prevalence of CS delivery signi cantly differed across the rural-urban residence and geographical regions. Women residing in rural areas were less frequent in CS birth compared with women living in urban areas. Regarding geographical region, the highest rate of C-section delivery was observed in South region (38.2%), followed by West region (21.5%), while the lowest the prevalence of CS birth was found in Central region (10.6%), followed by East region (14.6%).
Multivariate analysis: Table 4 presents the results of the multivariate logistic regression models showing the sociodemographic factors associated with cesarean delivery for India, Bihar, and Tamil Nadu.
Educational attainment of women was found to be positively associated with C-section delivery in which women with higher education were 42% (OR=1.42; 95% CI: 1.04-1.20) more likely to have cesarean delivery compared with uneducated women. Compared to Hindu women, the probability of CS delivery was signi cantly higher among Muslims (OR=1.05; 95% CI: 1.00-1.11), whilst the odds of CS delivery were 14% (OR= 0.76; 95% CI: 0.70-0.82) lower for Christian women. Social group also had a signi cant in uence on C-section delivery. Women belonged Scheduled Tribe (OR= 0.79; 95% CI: 0.74-0.84) and Other Backward Classes (OR= 0.86; 95% CI: 0.82-0.90) were less likely and women belonged to General caste (OR=1.07; 95% CI: 1.01-1.12) were more likely to have cesarean birth compared to Scheduled Caste women. Wealth quintile of the household also had a strong positive correlation with C-section delivery. Compared to the poorest women, the likelihood of cesarean birth was 57% higher among the richest quintile of household wealth (OR= 1.57; 95% CI: 1.35-1.59).
Place of delivery had a strong association with CS delivery. It was found that women delivered in private hospitals were associated with almost four-fold increased odds of C-section delivery (OR=3.90; 95% CI: 3.77-4.03). Antenatal care visits acted as enabling factors of C-section delivery. Women received four or more ANC visits were 1.6 times (OR=1.6; 95% CI: 1.50-1.70) more likely to have cesarean birth compared to those who did not receive ANC service. Women who had ever pregnancy loss (OR=1.22; 95% CI: 1.17-1.27) were more likely to undergo cesarean delivery than those who did not experience pregnancy loss. Women who had delivery complications were less likely to experience C-section delivery (OR= 0.18; 95% CI: 0. 16-0.20). Furthermore, it is observed that exposure to mass media had a positive association with cesarean births. For instance, women who had exposure to radio and television were 15% (OR=1. 15 Bihar and Tamil Nadu: The odds of cesarean delivery increased with the increasing age of women in both Bihar and Tamil Nadu. However, the association of age on cesarean delivery appears to be stronger in Tamil Nadu as compared to Bihar. The impact of age at marriage on cesarean delivery was found to be signi cant in Tamil Nadu only. The results show that women who married at 25 years or later were 35% more likely to have Csection delivery (OR= 1.35; 95% CI: 1.07-1.72) compared to those who married before 18 years. Birth order had a negative relationship with cesarean delivery in both states. The odds of cesarean delivery decreased with the increasing birth order. For instance, mothers with four or more order-child were 68% and 78% lower likelihood of having cesarean delivery in Bihar and Tamil Nadu, respectively. Regarding mothers' BMI, overweight condition of women increased the odds of cesarean delivery in both states. However, the likelihood of cesarean birth was slightly higher in Tamil Nadu (OR=2.26; 95% CI: 1.86-2.75) as compared to Bihar (OR=1.92; 95% CI: 1.51-2.43). Although we found a strong association between educational attainment of women and cesarean delivery at the national level, a marginal association was found in Bihar and the association became insigni cant in case of Tamil Nadu. In Bihar, women who belonged to forward caste were signi cantly more likely to have a cesarean birth (OR= 1.28; 95% CI: 1.00-1.63) as compared to SC women. We found a decreasing trend in cesarean delivery from bottom to upper quintile of household wealth in Tamil Nadu. However, wealth index was found to be less important in Bihar. Place of delivery was found to be a strong predictor of cesarean delivery. In Bihar, women delivered in private hospitals were associated with almost 13 fold increased odds of cesarean delivery (OR=12.86; 95% CI: 10.92-15.15) as compared to public hospitals. In Tamil Nadu, the likelihood of cesarean birth in private hospitals increased by nearly three-fold (OR=2.74; 95% CI: 2.40-3.13). The number of ANC visits was positively correlated with cesarean delivery in Bihar. Women who received four or more ANC visits were 1.8 times more likely to have a cesarean birth (OR= 1.82; 95% CI: 1.49-2.13) compared with women who did not receive ANC service. In Tamil Nadu, women who had pregnancy loss were at 24% increased likelihood of having cesarean birth (OR= 1.24; 95% CI: 1.06-1.46) compared with women who had no pregnancy loss. Likewise, women having delivery complications were associated with 63% higher odds of cesarean delivery (OR= 1.63; 95% CI: 1.41-1.88) in Bihar, whereas women having delivery complications were associated with 67% decreased odds of cesarean delivery (OR= 0.33; 95% CI: 0.28-0.38) in Tamil Nadu. We found no signi cant association between women's exposure to mass media (newspaper/magazine, radio, and television) and cesarean delivery in both states. Place of residence had a signi cant relationship with cesarean birth in Bihar where women living in rural areas were 20% less likely to have a cesarean birth (OR= 0.20; 95% CI: 0.65-0.99) as compared to urban women.

Discussion
Our study made a comprehensive assessment of factors in uencing cesarean delivery in India, with a particular focus on two states, namely Bihar and Tamil Nadu. In India, the rate of cesarean delivery is 17% in 2015-16. There is a large variation in the rate of C-section delivery across Indian states. The prevalence of cesarean delivery is considerably higher in Tamil Nadu (34%) than the national average, while a very low proportion of women have undergone cesarean delivery in Bihar (6%).
Multivariate analysis revealed that cesarean delivery increased with the increasing age of women where older women were more likely to have cesarean birth compared to younger women. This could be explained by the fact that older mothers are highly susceptible to pregnancy and delivery-related complications. 8,16,17 Therefore, older women may incline towards the cesarean section as a safe delivery option to protect their fetus after a long period of conception di culty. 8,18 The higher prevalence of cesarean birth was also found among women who married at a higher age. The possible reason could be adult married women may have higher decision-making autonomy in terms of their health-seeking behavior compared to child married women. In case of Bihar, we did not nd any signi cant association between age at marriage and cesarean delivery. Birth order was found to be signi cantly associated with a lower likelihood of cesarean delivery. The results show that the odds of cesarean birth decreased with higher older-birth. Mothers in rst birth are more likely to prefer cesarean delivery probably due to fear of labor pain. 7,19 On the other hand, mothers with a vaginal delivery experience at rst birth might reject cesarean delivery for the next children due to overcoming fear related to childbirth. Moreover, the nutritional status of mothers has a signi cant in uence on cesarean delivery. 20 Several studies have found that excessive weight gain and obesity increase relevance in obstetrics for cesarean delivery. 21 Vinturache et al., 22 in a study of the Canadian population investigated that obesity is associated with increased risk of pregnancy, delivery, and postpartum complications.
Our study found that education is signi cantly associated with an increased likelihood of cesarean delivery at the national level. This nding is consistent with several previous studies conducted in Bangladesh, 23,24 Brazil, 25 Pakistan, 7 Nepal 3 and China 26 . Women with higher education may have better decision-making ability to access obstetric care and more aware of the risk of childbirth-related complications. 8 In the present study, caste had a signi cant in uence on cesarean delivery. Consistent with other studies conducted in India, 17,27 the current study also found that the likelihood of cesarean delivery was higher among ST women. Moreover, the rate of home delivery is higher among tribal women. Due to impoverishment and marginalization of tribal communities, they do not prefer cesarean delivery. A study conducted in Madhya Pradesh reported that about 41% of maternal death occurred among tribal women. 28 Several studies from Bangladesh, 6 Pakistan, 7 China, 26 Mozambique 29 and Ghana 30 demonstrated a signi cant in uence of wealth status on cesarean delivery. In this study, the rate of cesarean delivery increased from bottom to upper wealth index groups. Women in wealthier quintiles may have no nancial constraints to have a cesarean delivery. Conversely, a negative relationship was found between household wealth index and cesarean delivery in Tamil Nadu. Similarly, Robelo et al., 12 found that women in higher quintiles are less likely to deliver in the C-section.
Place of delivery is found to be a strong predictor of cesarean delivery. At the national level, women delivered in private hospitals were 3.9 times more likely to have cesarean birth compared to public hospitals. In case of Bihar, private hospitals increased the odds of cesarean birth by almost 13 folds.
Similar ndings have been reported in other studies. 11,15,31 The higher rate of cesarean delivery in private hospitals could be due to su cient modern medical instruments, specialized treatment, su cient medical staff and caretakers, the demand of couple or mother's request. 32,33 In our study, women who received four or more ANC visits were more likely to deliver in the cesarean section which is in accordance with other studies. 8,10 Counseling of mothers during the ANC period could have a positive in uence on cesarean birth. Moreover, women having pregnancy complications could increase ANC visits and they are more likely to prefer C-section as a safe delivery option. 17 Likewise, women who had ever experienced pregnancy loss were more likely to have cesarean birth compared to those who had never experienced pregnancy complications. Women who had ever experience fetus loss prefer cesarean delivery to avoid any complications during childbirth, especially mothers who are married and have a pregnancy at a very late age might drive them quickly opt for a cesarean section. 18,34 Studies have indicated that women's exposure to mass media has a positive in uence on contraceptive use and maternal health-seeking behavior. [35][36][37] In our study, exposure to mass media had a signi cant impact on cesarean delivery. The study found that women who listened to the radio and watched television were more likely to have cesarean delivery at the national level. However, exposure to mass media had no signi cant in uence on cesarean delivery in Bihar and Tamil Nadu. Furthermore, place of residence had a signi cant impact on cesarean delivery in case of India and Bihar. Women residing in rural areas were less likely to have cesarean birth compared to urban women. This could be due to limited healthcare facilities, lack of knowledge, and impoverishment of rural women. Our study found that the rate of cesarean delivery considerably varies across the geographical region. For instance, women from the south region were associated with more than two-fold increased odds of having C-section delivery as compared to women from the north region. This nding is in line with previous studies. 11,17 Southern states are better in the condition in terms of socio-economic status and healthcare facilities compared to northern states. Moreover, women in southern states are highly educated, having a higher level of autonomy and more aware of the risk of reproductive health, which could lead to an increase in cesarean delivery in this region.

Strength and limitations:
The present study is based on a nationally representative sample survey of NFHS-4. Therefore, the results of this study could be generalized for the whole country. Additionally, our study focuses on two selected states, namely Bihar and Tamil Nadu, which account for the lowest and highest rate of cesarean birth in India, respectively. Therefore, the ndings of this study could be helpful for policymakers to design effective policies and programs to address the vulnerable section of society where the incidence of maternal mortality is widespread by increasing cesarean delivery among them.
The present study has several limitations. First, the collected information is self-reported; therefore, the data are prone to recall bias. Second, due to the cross-sectional nature of data, a causal relationship could not be established between the independent variables and the dependent variable. Third, cesarean delivery can be in uenced by many cultural, physiological and behavioral factors; however, we could not include these factors in the analysis due to the unavailability of information in the dataset. Furth, the data does not provide details of medical reasons for cesarean delivery. Further in-depth qualitative research is needed to understand the driving force behind cesarean section delivery.

Conclusion
In India, the rate of cesarean birth is 17%.This study has revealed that education, wealth quintile, ANC visits, pregnancy loss, place of residence and geographical region are found to be signi cant predictors of cesarean birth at the national level. It is also evident that the rate of cesarean delivery was substantially higher in private facilities compared to public facilities. Awareness regarding the importance of ANC visits can also help to reduce pregnancy complications and that can also help to increase the cesarean delivery for women who need C-section delivery. Women should also be driven towards regular ANC check-ups for the well-being of maternal and newborn health. The Government should intervene for rural and poor women for easy access to healthcare services.