Implicative statistical analysis in the identification of prognostic factors for cervical cancer mortality

Introduction: the statistical techniques used for the identification of prognostic factors are multivariate; one of the most frequent is logistic regression. In this work another technique is proposed and to test it, cervical cancer is used as a health problem due to its high incidence and mortality. Objective: to evaluate the usefulness of implicative statistical analysis in the identification of prognostic factors and to identify prognostic factors for mortality in cervical cancer. Method: a case-control study was conducted on a population of women with clinical and histological diagnosis of cervical cancer attended at the Oncological Hospital of Santiago de Cuba from 2014 to 2017. Implicative statistical analysis was applied along with binary logistic regression, which was considered as gold standard to evaluate the performance of the proposed technique. Results: both techniques identified age as a poor prognostic factor and chemotherapy as a good prognostic factor. Implicative statistical analysis identified metastasis as a poor prognostic factor, not detected by logistic regression, and supported the analysis with a series of graphs that helped to better understand the results obtained. Conclusions: the usefulness of the statistical analysis is recognized and its routine use is proposed to improve the quality of these investigations.