BACKGROUND Most studies which showed an excess mortality in psychiatric patients have been conducted on hospitalised samples.
METHOD This was a case register study. All South Verona patients with an ICD diagnosis who had psychiatric contacts with specialist services in 1982-1991 were included. Mortality was studied in relation to sex, age, diagnosis, pattern of care and interval from registration. Standardised Mortality Rates (SMRs) and Poisson regression analysis were calculated.
RESULTS The overall SMR was 1.63 (95% Cl = 1.5-1.8), which is the lowest value reported so far. Mortality was higher among men (SMR = 2.24; 95% Cl = 1.9-2.6), among patients who were admitted to hospital (SMR = 2.23; 95% Cl = 1.9-2.6), among younger age groups (SMR = 8.82; 95% Cl = 4.9-14.6) and in the first year after registration (SMR = 2.32; 95% CL = 1.8-2.9). Higher mortality was found in patients with a diagnosis of alcohol and drug dependence (SMR = 3.87; 95% Cl = 3.0-4.9). The SMR for suicide was 17.41. Using a Poisson regression model, diagnosis, pattern of care and interval from registration were all found to be significantly associated with mortality. When all these variables were entered together in the model, each maintained its predictive role.
CONCLUSIONS The overall mortality of psychiatric patients treated in a community-based system of care was higher than expected, but lower than the mortality reported in other psychiatric settings. The highest mortality risk was found in the first year after registration.