OBJECTIVE: To investigate whether the course of depression changes when (a) follow-up duration is longer and (b) in addition to depression other mood and anxiety disorders are considered as outcome measures. DESIGN: Longitudinal observational cohort study. METHOD: We selected patients from the Netherlands Study of Depression and Anxiety (NESDA) who had active depression at baseline (n=903) and for whom data from the 2, 4 and/or 6 year measurements were available. Using DSM-IV diagnoses and data from the 'Life chart interview', we divided participants in one of the following four course categories: (1) recovered (no diagnosis at 2-year measurement or later), (2) recurring without chronic episodes, (3) recurring with chronic episodes or (4) consistent chronic depression since baseline. We looked at the distribution of patients over the course categories from a short, diagnostically narrow perspective (over 2 years, only looking at depression) to a long, diagnostically broad perspective (over 6 years, looking at depression, dysthymia, hypomania, mania and anxiety). RESULTS: In the short, diagnostically narrow perspective, 58% of participants had recovered and 21% met the criteria for a chronic episode. In the long, diagnostically broad perspective however, only 17% had recovered while 55% had chronic episodes. CONCLUSION: Monitoring patients with depression over a longer period and with broader outcome measures (depression and related psychiatric disorders belonging to the mood disorder spectrum) shows that the course of depression is unfavourable and chronic for the majority. Conceptualising depression as a defined episodic disorder underestimates the severity of the prognosis for many patients and, as a consequence, the type of care indicated.