PURPOSE: To evaluate the methodological and statistical design evolution of the publications in the Brazilian Journal of Gynecology and Obstetrics (RBGO) from resolution 196/96. METHODS: A review of 133 articles published in 1999 (65) and 2009 (68) was performed by two independent reviewers with training in clinical epidemiology and methodology of scientific research. We included all original clinical articles, case and series reports and excluded editorials, letters to the editor, systematic reviews, experimental studies, opinion articles, besides abstracts of theses and dissertations. Characteristics related to the methodological quality of the studies were analyzed in each article using a checklist that evaluated two criteria: methodological aspects and statistical procedures. We used descriptive statistics and the χ2 test for comparison of the two years. RESULTS: There was a difference between 1999 and 2009 regarding the study and statistical design, with more accuracy in the procedures and the use of more robust tests between 1999 and 2009. CONCLUSIONS: In RBGO, we observed an evolution in the methods of published articles and a more in-depth use of the statistical analyses, with more sophisticated tests such as regression and multilevel analyses, which are essential techniques for the knowledge and planning of health interventions, leading to fewer interpretation errors.
Statistics; Ethics, research; Ethics Committees, research; Helsinki Declaration; Scientific misconduct; Periodicals as topic; Periodicals as topic; Clinical trials as topic; Gynecology; Obstetrics