Hamilton Scale: study of the psychometric characteristics in a sample from Southern Brazil*

Objective To investigate the psychometric characteristics of a translated version of the scale, proposing a reviewed version in order to attend the transcultural adaptation criteria. Methods This study included 231 subjects – depressed (45.5%), bipolar (7.8%), and healthy (46.7%) – who participated of an epidemiological research in southern Brazil. The evaluation of mental disorders was made through Clinical Interview for DSM-IV (SCID) and a translated version of the Hamilton Scale (HAM-D), usually utilized in Brazil without adaptation studies. Results The ROC curve analysis identified the cutoff (9 points) to discriminate the presence or absence of depression, resulting in a sensibility and specificity of 90 and 91%, respectively. The internal validity was investigated by the factorial analysis and consistence of the items. It was observed that all 17 original items, except “Consciousness”, presented psychometric quality to evaluate general depression, and that there were five dimensions underling those 16 items: Depressed humor, Anorexia, Insomnia, Somatization, and Anxiety, and all of them, excepting the last, showed homogeneity in their constructs (alpha coefficients between 0.66 and 0.78). On content analysis, five specialists suggested editing changes in seven items. Conclusion This study determinates a different cutoff and psychometric evidences favourable to the use of HAM-D in Brazil.