Validity of the definite and semidefinite questionnaire version of the Hamilton Depression Scale, the Hamilton Subscale and the Melancholia Scale. Part I

Abstract

Instruments for self-rating in depression are available, but their psychometric properties have not been fully explored; discrepancies with clinician ratings have been identified. This study was longitudinal with 85 patients fulfilling the DSM-III-R diagnosis of Seasonal Affective Disorder. Self-reporting versions (definitely and semidefinitely anchored) corresponding to the Hamilton Depression Scale (HAMD), the Hamilton Subscale (HAM₆), and the Bech-Rafaelsen Melancholia Scale (MES) were compared to each other and the clinician-rated version. The unidimensional property of the sum score in each scale was tested by the item-response theory model ad modum Rasch. The scales were also tested for their sensitivity to discriminate between placebo and citalopram therapy. The sum scores and the sum score variances of the definite self-rating versions did not differ significantly from the sum scores of the corresponding observer scales at any of the five time points. The semidefinite scales significantly over-scored at all time points. The convergent validity between corresponding definite self-ratings and observer ratings was very high with correlations exceeding 0.90. Only item responses from the MES, the HAM₆, and their corresponding definite versions of the self-rating questionnaires DMQ and DHAM₆ were accepted by the Rasch analysis, and only these four valid scales discriminated significantly between the effect of citalopram and placebo treatment. Our results are limited to patients with moderate depression. Two new self-report scales with unparalleled construct validity, reliability, sensitivity, and convergent validity have been identified (DMQ and DHAM₆). We have also identified a crucial importance of format for the means and variances of self-rating scales. These findings are of high practical and scientific value.