The aim of this study was to understand which of a number of factors are most associated with psychiatric inpatient length of stay (LoS). We hypothesized that a longer LoS would be predicted by: older age, male gender, unmarried marital status, foreign nationality, more than one hospitalization, being hospitalized involuntarily, psychotic symptoms and behavioral dyscontrol at admission, discharge diagnosis of psychotic and personality disorders, not having a substance use disorder, treatment with more than one class of medications, and being discharged to a community residential facility. All admissions to the Psychiatric Inpatient Unit of Santa Maria della Misericordia, Perugia Hospital, Umbria, Italy, from June 2011 to June 2014, were included in a medical record review. Bivariate analyses were performed and a multiple linear regression model was built using variables that were associated (p < .05) with LoS in bivariate tests. The study sample included 1236 patients. In the final, most parsimonious regression model, five variables independently explained 18 % of variance in LoS: being admitted involuntarily, being admitted for thought disorders, not having a substance-related disorder, having had more than one hospitalization, and being discharged to a community residential facility. LoS on this inpatient psychiatric unit in Umbria was associated with a number of sociodemographic and clinical characteristics. Knowledge of these and other predictors of LoS will be increasingly important to, when possible, reduce the length of restrictive, costly hospitalizations and embrace community-based services.
Keywords: Inpatient psychiatry; Involuntary treatment; Length of stay; Substance-related disorders; Thought disorders.