A validation analysis of two self-reported HAM-D6 versions

The six items of the clinician-administrated Hamilton Depression Scale (HAM-D-6) cover the core items of depressive states reflecting the antidepressive effect of medication. In this study, the two self-reported versions of the HAM-D-6 have been psychometrically validated to ensure the unidimensionality of this administration form in patients with mild-to-moderate depression. The item response theory analysis of Mokken was used to test the unidimensionality of both the Interactive Voice Recording System (IVRS) version of the HAM-D-6 and a paper-and-pencil self-reported version (S-HAM-D-6). Patients with typical major depression and with seasonal affective disorder were included. The Mokken analysis showed that the two self-reported versions of the HAM-D-6 obtained coefficients of homogeneity above 0.40, similar to the clinician-rated HAM-D-6 and thus implying unidimensionality. By contrast, the full HAM-D-17 versions (self-reported as well as clinician-rated) obtained coefficients of homogeneity below 0.40, implying that the HAM-D-17 is a multidimensional scale. The analysis show that both the IVRS version and the S-HAM-D-6 version are unidimensional self-rating scales for the measurement of depressive states.