Compulsory admissions of emergency psychiatric inpatients in Turin: the role of diagnosis.

Many studies have found high levels of compulsory admission (CA) among non-affective psychoses. Aims of the present study were to investigate whether there was a mere association between diagnosis and CA in a cohort of 848 patients referring to a Psychiatric Emergency Service in a catchment area in Turin during a 2-year period, independent of socio-demographic features, psychiatric history, and clinical status. Diagnosis as a risk factor for CA was assessed constructing a logistic regression model, using the following steps: first, assessing the association between diagnoses and CA, without controlling for confounding factors; second, entering socio-demographic factors; third, entering socio-demographic factors and psychiatric history; and fourth, entering socio-demographic, psychiatric history, and aspects of clinical presentation into the model. At step 1 patients with Non Affective Psychoses, Mania and Personality Disorders had a significantly higher CA risk, compared to patients with Depressive Disorders. At step 4 diagnosis was no longer associated with CA. History of CAs within past 5years and Brief Psychiatric Rating Scale (BPRS) hostile-suspiciousness were positively associated with CA. Length of illness, history of previous suicidal attempts and BPRS anxiety-depression were negatively associated with CA. Overall, the percentage of correctly predicted cases was 39.8%. The remaining 60% can be explained by inherent variability or unknown, lurking variables. Finally, the study was carried out at a single facility. Much of the increased CA likelihood in diagnostic categories might be attributable to specific symptom patterns, not to patients' severity of illness or diagnosis per se.