Well conducted investigations into the long-term outcome of depressive disorders are rare. This issue of Psychological Medicine publishes two important papers reporting on different cohorts of depressed patients, one series from Japan (Kanai et al. 2003) and the other from Cambridge UK (Kennedy et al. 2003). Both were ascertained in the early 1990s and have been followed for 5–6 and 8–10 years respectively. Each study demonstrates methodological advances. Both invite comparison with previous reports of the long-term outcome of depression whose follow-up periods span the last 40 years. Is the outcome of depressive disorders at last improving in the era of modern treatments? Is this merely an artefact of better research methods, or does it also reflect therapeutic advances? If there has been very little improvement in some aspects of outcome, what lessons can we learn for future research and practice development?