Critical analysis of the use of statistical tests in Brazilian publications related to digestive tract surgery

Abstract

To evaluate the evolution of scientific papers published on digestive tract surgery in Brazilian journals in 20 years and examine whether the level of evidence in the studies has improved, as well as the incorporation of statistical procedures and their proper application. We selected all original articles related to digestive tract surgery published in the years of 1987 and 2007 in 4 leading Brazilian surgical journals. Studies were divided according to the level of evidence (I: prospective, controlled and randomized, II: prospective without control or randomization and III: retrospective) and compared them to assess whether there was an improvement in the level of evidence between these two years surveyed. We also assessed whether there was increased use of analytical statistics and correct application of statistical procedures. Comparing the articles published in 1987 with those of 2007, we observed no improvement with respect to the level of evidence, with more than half of the articles published being case series and retrospective studies (56.14%). There has been a significant increase in the use of analytical statistics (70.4% in 2007 vs. 40% in 1987) in the 20 years, however 16.7% of the studies published in 2007 did not correctly apply or adequately describe the statistical analyses used. In this study, we observed no improvement in the level of evidence presented in publications on digestive tract surgery in the last 20 years. There was an increased use of statistical analysis, but there is a need to correctly apply statistical procedures.