Objective — The analysis aims to study patterns of care of patients in contact with 5 Psychiatric Services in Lombardy. Four patterns have been identified :long term-high users, non long term-high users, long term-non high users, non long term- non high users. Design — Data were provided by the regional Psychiatric Information System. The cohort of patients have been composed by 5,670 patients included in 1994 one year prevalence. Setting — Five Psychiatric Services (Merate, Treviglio, Crema, Desio, Castano Primo) with a total population of 610,184 inhabitants aged over 14. Main utilised measures — Some sociodemographic and clinical variables have been taken into consideration for a descriptive analysis; a multinomial logistic regression model was used to identify the characteristics of patients associated with different patterns. Results — Long term-high users were 5.3%, i.e. a mean rate of 4.9/10,000 residents over 14, and absorbed 60% of resources; the absence of a partner was associated in regression analysis with this pattern. Non long term-high users were 1.2%, i.e. a mean rate of 1.1/10,000 residents over 14, and absorbed 7.8% of resources; age below 45, unemployment, absence of a partner, severe mental illness and first contact with Psychiatric Services in the period 1985-1989 were predictive variables. Long term-non high users were 23.4%, i.e. a mean rate of 21.6/10,000 residents over 14, and absorbed 18.1% of resources; age below 45, unemployment, living alone, absence of a partner, severe mental illness and first contact with Psychiatric Services before 1990 were predictive variables. Non long term-non high users were 70.1%, i.e. a mean rate of 64.8/10,000 residents over 14 and, and absorbed 18.1% of resources. Conclusions — Data show that on the whole the activity of Psychiatric Services is addressed to most serious patients, though considerable differences between Psychiatric Services utilisation may be found. This study highlights the importance of a regional Psychiatric Information System, that allows the monitoring in time and in the regional territory of patterns of care.