Aim Cesarean section rates are increasing worldwide and Robson's classification system allows a practical approach to study this phenomenon. C-section in Chile has been indicated as unexpectedly high, with important variability within the country and payment systems. The aim was to report our data using Robson's system and the evolution of local C-section rate in a public hospital during a 9-year period. Methods Retrospective analysis (2005-April 2014), in a metropolitan hospital in Santiago. All deliveries were classified into Robson groups. Time changes were analyzed with Pearson's correlation. P value <0.05 was considered significant. A 'relevance index' (RI) for each group was calculated as 100 x C-S rate x relative contribution. Results The overall C-section rate increased from 24 to 27% (P < 0.05) in 53 571 deliveries, with a greater increase in groups 1 (nulliparous, single, term cephalic, spontaneous labor), 3 (multiparous, single, no previous C-S, term cephalic, spontaneous labor) and 4 (multiparous, single, no previous C-S, term cephalic, induced or no labor). Despite no increase in Group 5 (women with one or more previous C-S) this group had the highest RI (20.3), which defined priority for intervention over others. Conclusion C-S rate was lower than that reported in other centers from Chile and Latin America. Robson's classification and the RI allowed prioritization. Although increase in groups 1, 3 and 4, group 5 needs attention because of stronger impact on overall C-S rate. This analysis allowed to define how to lower C-S rate in our institution.