Predicción del número de linfocitos T CD4 en sangre periférica a partir de teoría de conjuntos y probabilidad en pacientes con VIH/SIDA

Abstract

Previously a refinement of a methodology was developed based on set theory and probability, with the aim of improving the predictive ability of CD4 T cells (LT-CD4) for HIV/AIDS from the total number of leukocytes and lymphocytes. In this work the clinical applicability of the method was shown by refining prediction in 150 samples. Taking data of leukocytes/mm³, lymphocytes/mm³, and LT-CD4 cells/mm³ of each patient, called triples, they were organized from highest to lowest based on the number of leukocytes, to evaluate ranges of 1,000 leukocytes. The membership of the triples to the set A∪C, B∪D was established, as well as the intersection between the two sets (A∪B)∩(B∪D), in which a prediction of the number of CD4 associated with specific values of leukocytes and lymphocytes is established. The number of elements belonging to each set was counted and the probability of belonging to each of the ranges tested was determined. A total of 7 out of the 9 ranges of leukocytes measured showed a probability of success equal to or greater than 0,76, achieving a probability of 1 in the ranks lower than 4.000/mm³ and 3.000/mm³, respectively. The ability of the methodology for determining the value of the LT-CD4 was demonstrated, achieving a higher predictive capacity than the unrefined methodology. The evidence shows that the applied methodology is effective for clinical use, thus leading to a reduction of costs and resources.