诊断技术与生物医学分析杂志

Are Both Antidepressant Drug Effects and Test Scores Unspecific?

Ivar Aursnes, Ingunn F Tvete, Jorund Gaasemyr, Bent Natvig and Marianne Klemp

Are Both Antidepressant Drug Effects and Test Scores Unspecific?

Background: Disappointingly small clinical effects of antidepressant drugs have recently been reemphasized. It is also unclear to what degree the 17 questions in the Hamilton Rating Scale of Depression and other questionnaires represent true aspects of depression or rather can be said to measure unspecific symptoms.

Methods: Within all clinical study reports concerning citalopram submitted to the Danish Health Authorities before registration in 1988 we identified among 49 studies two studies with placebo groups reporting individual patient scores. A multiple regression analysis was performed by treating questionnaire item no 1 (depressed mood) as the dependent variable and the scores for all other items as potentially explanatory variables. The questionnaire items were sorted in two groups: items with p-value less than 0.05 considered related to depression and items with p-value above 0.05 considered unrelated. By pooling the data together we performed an analysis of the effect size (standardized differences in means; citalopram versus placebo treatment, baseline score minus end score) for each of the two questionnaire groups, as well as for the total HAM-D 17 score.