Structural validity of MADRS during antidepressant treatment

In order to assess the internal structure of MADRS, a principal component analysis was performed in a group of 137 patients suffering from DSM-III major depression (MADRS score at baseline > 20) receiving a 4 week treatment in three double-blind drug trials (medifoxamine vs clomipramine, imipramine or amineptine). Prior to antidepressant treatment, three factors could be identified with 32 %, 12 % and 11.5 % of explained variance. The first factor corresponded to the global severity of depression whereas the second, with sleep and appetite items scoring high, could be postulated as a somatic factor. After 28 days of antidepressant treatment (mean MADRS total score = 13.3), only one factor could be substantiated, representing 66% of total variance. The study was primarily based on data obtained without rotation, which ensures the independence of the different factors. The good internal consistency of MADRS accounts for the high correlation of all items with the first factor (r = 0.95). MADRS appears as a unidimensional scale more oriented towards psychic as opposed to somatic aspects of depression.