A retrospective analysis of the clinical records of 149 inpatient suicide subjects and the same number of control subjects representative of the entire admitted clinical population was carried out using a larger set of demographic, psychosocial, and clinical variables. The results of the univariate analysis of the data indicated a serious and incapacitating course of the mental illness in the suicide group. The multivariate stepwise logistic regression analysis yielded a set of five variables including sex, Research Diagnostic Criteria diagnosis, previous and recent suicidal behavior, and social exits that discriminated best between both groups. These findings have to be taken into consideration when estimating the suicidal risk. The assessment of the suicidal danger, however, remains an individualized procedure.