IntroductionThere is little knowledge of predictors for involuntary hospitalizations in acute psychiatric units. MethodThe Multi-center study of Acute Psychiatry included all cases of acute consecutive psychiatric admissions in twenty acute psychiatric units in Norway, representing about 75% of the acute psychiatric units during 2005–2006. Data included admission process, rating of Global Assessment of Functioning and Health of the Nation Outcome Scales. ResultsFifty-six percent were voluntary and 44% involuntary hospitalized. Regression analysis identified contact with police, referral by physicians who did not know the patient, contact with health services within the last 48h, not living in own apartment or house, high scores for aggression, level of hallucinations and delusions, and contact with an out-of office clinic within the last 48h and low GAF symptom score as predictors for involuntary hospitalization. Involuntary patients were older, more often male, non-Norwegian, unmarried and had lower level of education. They more often had disability pension or received social benefits, and were more often admitted during evenings and nights, found to have more frequent substance abuse and less often responsible for children and were less frequently motivated for admission. Involuntary patients had less contact with psychiatric services before admission. Most patients were referred because of a deterioration of their psychiatric illness. ConclusionInvoluntary hospitalization seems to be guided by the severity of psychiatric symptoms and factors “surrounding” the referred patient. Important factors seem to be male gender, substance abuse, contact with own GP, aggressive behavior, and low level of social functioning and lack of motivation. There was a need for assistance by the police in a significant number of cases. This complicated picture offers some important challenges to the organization of primary and psychiatric health services and a need to consider better pathways to care.