A diagnosis test is carried out to establish the presence of health or illness. In the latter it could grade the severity. Due to its importance in clinical decisions, the diagnosis test is evaluated by mathematical strategies. We estimate the sensitivity and specificity once we know the existence or not of the disease, but we act in the reverse direction; with the presence "X" test positive or negative we estimate the presence of the disease, therefore, we use the positive and negative predictive values. Mathematical strategy allow us to quantify the observation, but it requires judgment to determine the quality making use of a minimum of features: a) selection under the same criteria for cases and controls; b) the inclusion of the full spectrum of disease severity (from mild to the most serious, ensuring that all levels have an enough number of subjects); c) the interpretation of both, the gold standard and the new tool of diagnosis, it must be blind and conducted by experts; d) the interpretation of results should show us what is their application in everyday clinical practice; e) the reproducibility must be checked. Do not forget that usually, we treat only one patient at once, what enforce us to have full knowledge of the performance of the diagnostic test, and to consider all clinical aspects for its proper implementation.