Background. A previous study of women 9 months after discharge from in-patient treatment for depression found that low self-esteem as measured by a self-report questionnaire at admission was associated with a worse prognosis. This study investigates the womens' outcome more than 4 years after discharge and the continued prognostic relevance of self-esteem and other variables.

Method. Fifty-six women who had been admitted to hospital with major depression and studied in detail during their admission and 9 months following discharge were traced 4–5 years later. Possible prognostic factors were investigated in relation to time to recovery and recurrence, and to a global outcome measure.

Results. In keeping with other published work only 16 out of 52 (31%) women had recovered and remained well. Fifteen out of 52 (29%) subjects had experienced depressive symptoms for more than 70% of the follow-up time or died from unnatural causes. Low self-esteem scores recorded at the initial admission correlated with slow recovery but not subsequent recurrence of depression. The occurrence of one or more life events in the year preceding admission was associated with a better prognosis.

Conclusions. The poor long-term prognosis of many women with severe depression was confirmed. Social factors, such as social support and marital relationships were less important for prognosis than in previous studies of less severely ill subjects. The prognostic value of self-esteem warrants further investigation and appears to have therapeutic implications.