Multilevel analysis of birthplace effect on the proportion of C-Section in Colombia".

Objectives: To determine the general contextual effect of the department in the variation of Cesarean section in Colombia in 2016, and their individual and contextual related factors. Methods: Cross-sectional study based on a birth cohort. We used the birth certificate database from January 1 to December 31, 2016. Three multilevel logistic models were constructed. Pregnant women were set at the first level and department at the second level. We fitted a null multilevel model followed by two multiple models, including individual and individual and department variables, respectively. Results: The overall prevalence of C-section was 45.5% (95% CI: 45.4-45.6), ranged from 5 to 76%. The variance partition coefficient was 15%. Individual factors did not explain the general contextual effect. However, the region to which these departments belong explained 71% of the variance. The Caribbean region was positively associated with C-section compared to the Andean region (OR:3.88, 95% CI: 2.65-5.67). Conclusions: Multilevel analysis of individual heterogeneity and discriminatory accuracy showed that the department plays an important role in the variation of C-sections in Colombia. Our results suggest that the proportional universalism approach should be applied to reduce the proportion of C-Section, focusing on departments with the highest risk and on the Caribbean and Insular region.