A 5-year follow-up study was performed on 82 homeless men, with mental problems, who had been contacted by an outreach team run by the Social welfare administration of Stockholm 1995/1996. Data have been collected from the Cause of Death Register, death certificates, forensic autopsy reports, hospital medical reports, Hospital Discharge Register, interviews with social workers and with those men who were able to participate. The standardized mortality ratio (SMR) was 4.7 times higher than expected. The highest mortality was found in the group where drug addiction was dominant; 46% had died. In the group of men with severe psychiatric disorders, with diagnosis such as schizophrenia, none had died. Compared with the others, they had spent less time in homelessness. Among the survivors, 75% were still homeless at the follow-up in spite of considerable treatment interventions from the social services and health authorities. Residential institutions or treatment seemed to have some protective effect concerning misuse, diseases and injuries. Among the still homeless, the mental health problems combined with substance use problems had increased with 17%. The life and housing situation for the whole group seemed not to have improved, even if fewer of them were staying in hostels for homeless people.