The objective of this article was to determine a 7-year naturalistic progression of depression as well as a number of potential prognostic factors among Finnish primary care and psychiatric care patients. Depression-screened patients from primary care and psychiatric care, aged 18-64, were interviewed in 1991-92 with the Present State Examination (PSE) as the diagnostic instrument. The patients were re-contacted in 1998-99, and their depression at final assessment (FinalA) and during the follow-up period (F-up) was assessed by telephone interview using the Composite International Diagnostic Interview--Short Form (CIDI-SF). 250 primary care (58.1%) and 170 (40.2%) psychiatric care patients were successfully followed. Of the primary care patients with severe depression at baseline, 42.4% had had depression during F-up and 21.2% had depression at FinalA. For the patients with mild depression at baseline, the corresponding figures were nearly the same, but for the patients with depressive symptoms clearly lower. Of the psychiatric care patients with severe depression at baseline, 61.0% had had depression during F-up and 26.2% had depression at FinalA. As with primary care patients, the corresponding figures were nearly the same for mild depression at baseline but clearly lower for depressive symptoms. Experienced lifetime mood elevation was associated with having depression during F-up in both primary care and psychiatric care patients. High Depression Scale (DEPS) score at baseline was associated with having depression at FinalA in primary care patients, but in psychiatric care patients, it was the high Hamilton Rating Scale for depression (HAM-D) and drinking problems. Severe depression and mild depression are predictive for subsequent depression at both levels of care. The long-term prognosis for depression is better in primary care. DEPS and HAM-D are useful, prognostic instruments.