Spatial Characteristics

Overall, the region of residence was found to be statistically significantly associated with CS delivery in the simple regression. However, not all categories are statistically different from the reference category. The probability of CS in three regions (Kalimantan, Sulawesi, and Maluku and Papua) were statistically different from that in the reference region (Java). In the multivariate regression, this significance of this relationship holds. However, only Sumatera region that has statistically different CS from Java region. As for the place of residence, the simple association between residing in an urban area was a positive and statistically significant one. This association, however, became attenuated and no longer statistically in the final multivariate model.

Maternal Characteristics

The first maternal characteristic variable is maternal age. It is found that maternal age is statistically associated with the outcome with all categories statistically different from the reference category (20-24 years old). This relationship holds when other covariates are included in the final model.
The second maternal variable was parity. This was found to be statistically associated with the outcome. Multiparous women with previous CS delivery have a higher probability of having CS for their latest delivery compared to referent women. In contrast, multiparous women with no previous CS delivery have a lower probability of having CS for their latest delivery compared to nulliparous women. Moreover, the third maternal variable was whether the women had any complications during their pregnancy. High-risk pregnancy was found to be positively associated with the probability of CS delivery. This relationship holds in the final multivariate model.
Furthermore, the penultimate maternal variable was the number of PNC visits consisting of four categories with null visits as the reference category. All the categories of PNC visits are statistically different from the reference category with increasing probabilities as the visits increase. This association weakened but still highly significant when other covariates were included in the final multivariate model.
The last maternal variable, place of delivery, was found to be statistically related to the outcome variable. Women who gave birth in a private facility was a higher probability of having CS method than those who gave birth in public facilities or at home. This association barely changed in the final multivariate model.

Socio-Demographic Characteristics

The first socio-demographic variable was women’s years of schooling. The simple relationship between this variable and CS delivery was a positive and statistically significant one. In the final multivariate model, this relationship holds, although the coefficient weakened. Moreover, the second socio-demographic variable, father’s years of education, was found to be statistically and positively associated with CS in the simple regression. However, this variable was omitted in the final multivariate model due to its high correlation with mother’s years of education.
The second socio-demographic variable, health insurance cover, was found to be associated with CS delivery, with women having cover from the employer and private institutions were more likely to undergo CS for their latest delivery. When other covariates were included, the overall relationship is still significant. Furthermore, in terms of household wealth, quintiles of wealth index were found to be statistically significant in the simple regression with the coefficients getting larger as the quintile goes up. A similar pattern, albeit attenuated, was observed in the final multivariate model.