Deconstructing depression: A 3D perspective

Major Depressive Disorder (MDD) is a common disorder and acknowledged to be one of the largest contributors to the global burden of disease. However, understanding the underlying mechanisms and optimal treatment of MDD has been shown to be difficult. This is because the current diagnostic system allows for high levels of heterogeneity. MDD patients differ to a large extent with respect to their clinical presentation, etiology, course, severity and comorbidity. Research has been focused on identification of more homogeneous subtypes of MDD, in the hope that such subtypes could be linked to more specific clinical decisions (e.g. course, treatment) and underlying mechanisms. However, subtypes have so far been simple, mostly differentiating between patients based on a few criterion symptoms (e.g. appetite gain vs. loss), but unable to capture more complex variations across patients in how different symptoms develop and interact over time. Therefore, this dissertation presents a new method to subgroup patients based on their course-trajectories on different symptoms, using a data-driven statistical approach that can handle data in which multiple sources of variation exist and does justice to the complexity of depression. The results indicated that this approach allows us to gain a much better view of the many possible forms that depression can take and the variations that exist among patients in terms of their characteristics, course and prognosis. Such insights are essential to gain a better understanding of the etiology of depression and to optimize the diagnostics and treatment of depression.