Predictive model Archives - Revista Brasileira de Ginecologia e Obstetrícia

PURPOSE: to create a predictive model for cesarean section at the "Professor Monteiro de Morais Maternity" after evaluation of antepartum risk factors of the pregnant women who delivered from September 1, 1999 to August 31, 2000, and then, to verify the efficacy of indication for cesarean section. METHODS: a longitudinal, case control study with 3.626 pregnant women was performed to identify the antepartum risk factors for cesarean section in the period from September 1, 1999 to August 31, 2000. Thereafter an ideal model able to quantify the risk for cesarean section for each patient in the presence of one or more risk factor was created. Then, the model was applied to the patients of the study in order to verify the efficacy of indication for cesarean section. RESULTS: the baseline risk for cesarean section was 15.2%. The concordance between the percentage estimated through logistic model and cesarean delivery was 86.6%. CONCLUSIONS: the logistic model was able to identify the baseline risk for cesarean section and to quantify the increase in risk for cesarean section in each patient when risk factors were introduced in the model. The model can be considered efficient and able to predict cesarean section because the agreemant between the prediction and the correct indication was 86.6%, and 53.6% of the patients who had vaginal delivery did not have any risk factor for cesarean section.