In the present study, we find that urban women have higher CD rates than rural women, which is similar to the findings in other developing countries. For example, research from Pakistan and 26 countries in Southern Asia or Sub-Saharan Africa both declare the same fact that a rising trend of CD presented in urban areas [8,9,10]. Firstly, the increasing availability and accessibility of CD in urban areas may contribute to the widely utilization [11]. Secondly, women in urban areas usually have a better socio-economic status, for example, from better educational background, with higher employment rates, and with better economic level, which can influence the affordability of CD [2]. However, we noticed that the proportion of women using CD in rural areas was also very high (53.5%). A possible explanation is that, with recent economic development in rural areas, there has been a great change in attitudes towards CD among rural women. Moreover, people have been able to get imbursement for the medical cost of CD with the implementation of NCMS. All of the above factors contribute to the high rates of CD in rural areas. Unnecessary CD raised potential risks both for mothers and infants and also resulted in unnecessary expenditure on medical care. High rates of CD, especially in rural areas, have become a potentially alarming phenomenon worthy of attention.
We also analyzed the association of social determinants and the mode of delivery. According to statistical analysis, the results showed that women in middle Jiangsu were more likely to use CD. We are, however, limited in ability to explain this finding, as, to our knowledge, there is very limited research on spatial variation. One likely explanation is that, the middle Jiangsu in Jiangsu (Yangzhou, Taizhou and Nantong), especially in poorer areas, have a tendency to use CD, which is partly due to the local cultural influence [12]. In addition, compared to the number of health providers in south Jiangsu, there are staffing and physical limitations of the public facilities in maternal care in middle Jiangsu, resulting in the higher use of CD services which is an efficient way to move patients through the systems [13].
CD is more prevalent among richer women in our findings, which is in accordance with the results of studies conducted in Asia and Africa [8, 9]. The possible reason for such difference is that richer women may be delivering in more expensive or highly rated institutions, which may be more likely to perform CD for economic incentives considering that the rich women can afford the extra cost associated with delivery mode. Moreover, the rich women are more likely to be fear about the painful experience caused by natural delivery which may prompt them to consider CD [9, 14]. It is implied that developing means of improving the childbirth experience may relieve maternal anxiety aiming at reducing unnecessary CD [15].
We also find that the more use of prenatal care visit, the more use of CD. The prenatal care is important for both mothers and babies, and viewed as an essential measure for a normal pregnancy. It is likely that many underlying complications are found by prenatal care, and which increase the possibilities of undergoing CD. And the women who made more prenatal care visits may tend to adopt CD [16]. This finding has significant public health implication: if necessary measures can be taken to prevent the incidence of complications during the prenatal period, it may dramatically reduce the frequency of unnecessary CD.
Our data suggest that the socio-economic status of residence region is a more important determinant of delivery mode than the women’s individual socio-economic characteristics in urban area [17]. We found that the CD rate was almost four times higher among women in middle Jiangsu than in south Jiangsu, irrespective of individual variations. Hence, it makes sense to identify these non-individual social determinants and to analyze how they influence the delivery mode. Supply-side factors may be a more important determinant of CD than individual factors in urban area [18].
There are two limitations that should be noted. Firstly, the study design was cross-sectional and in consequence the ability to adequately address issues such as causality may have been impaired. This issue could potentially be resolved by the collection of data longitudinally or by using path analysis and structured equation modeling on cross-sectional data as a means of better understand the underlying relationships. Secondly, the possible existence of confounding factors (such as maternal disease and access to health care) ought to be taken into account in future studies.