[Clinical research XIV. From the clinical judgment to the statistical model]

A statistical test is incomprehensible when it is out of context, so it is necessary to identify the details of the phenomenon of causality in the clinical course of the disease and to integrate the statistical model. Thus, the statistical tests used will try to characterize baseline, maneuver and the outcome, and will show the relationship between them. When we read the results in clinical research, the first thing that the author describes are general characteristics of the population, starting with number of patients evaluated and selected, average age, gender, and number of subjects meeting the outcome. This is extremely important because with the same criteria two studies may contain populations completely opposite. Posterior description usually continues through tables that follow a logical sequence, which allow us to integrate the statistical model to clinical judgment: baseline characteristics of the population and its distribution in each of the maneuvers, characteristics of the main and peripheral maneuvers, main effect of the maneuver on the outcome, and the impact of principal maneuver in the outcome, but adjusted for any variable that can alter this impact.