Resumen
This article presents an application for the statistical methods for classified variables, such as association measures and logistic regression, of a comparative study whose objective is to identify risk factors for the hepatitis B or C condition, correcting deficiencies and distortions in the process of clinical screening of candidates to blood donors at Hemocentro Regional de Maringá, in the period from 1993 to 1996. For the logistic regression the detected risk factors were: to have more than three sexual partners and have had unknown partners, homosexual or prostitute. In spite of the model being significant it restricted as for prognosis, once for the classification rule the percentile of successes among the donors with at least one positive serology of hepatitis is low. It was observed that the medical interview in the clinical screening provided a significant contribution to the inaptitude identification.