A major proportion of patients admitted to psychiatric acute and emergency departments have symptoms, behaviors or conditions influenced by implicit or explicit suicidal intention (tendency towards self-harm, or suicide attempt). There is an increased risk of suicide in the acute phase during in-patient stay and after discharge. At present there is no satisfactory tool to predict post-discharge suicide risk. Multidimensional rating is expected to have better predictive properties than one-dimensional rating for suicidal behavior. Multidimensional rating assesses anxiety, panic, agitation, suicidal thoughts and ideations, as well as therapist’s reactions during interviews. The project’s overall objective is to find parameters that can predict severe affective impulsivity during hospitalization and the coming two years after index hospitalization. The primary aim is to try two different psychometric instruments: “Multidimensional Assessment of Suicide Risk” (MARIS) and an extended version of the PANSS Excitation Component Scale (PANSS-EC +). The secondary aim is to examine whether specific immunological, endocrine, genetic or organic cerebral conditions can predict severe affective impulsivity during admission or within the next two years.