STATIC AND DYNAMIC PREDICTORS OF THE COUR SE OF MAJOR DEPRESSION - ANANALYSIS USING GENERALIZED ESTIMATING EQUATIONS

The present article examines the importance of static and variable pre dictors for the course of depression using the method of generalized e stimating equations. This approach allows to analyze rime dependent pa nel data with qualitative response variables. The sample of 90 depress ed inpatients was assessed at three points in time (admission, dischar ge and one year after discharge) using psychopathological variables (B DI, HAMD, LIFE) and life stress, The response variable is defined by t he presence of a major depression according to DSM-III-R during interv als of three months. Results show that the BDI and the KAMD are signif icant in predicting relapse, while the number of previous depressive e pisodes is not. The influence of stressful life events and the total s core of life stress taking place in the same quarterly interval is fou nd to be significant as well. In a rime lagged analysis, the relevance of the stressful life events on symptomatology in the following quart erly interval disappears, while the total score remains to be signific ant. These results indicate the importance of concurrent stressful lif e events and discharge symptomatology on the course of depression.