Mental health care reform has brought an increasing emphasis on community care, with concomitant reductions in inpatient psychiatric resources. Hospitalization remains a necessary and integral component of the mental health care system, but it is taking on a more specialized role. Examining the circumstances in which hospitalization is indicated can help clarify emergency psychiatric practices and determine whether patients' needs are being met within this changing environment. This pilot study examined the impact of selected patient and contextual characteristics on the decision to admit patients to inpatient psychiatric units and assessed the utility of the Severity of Psychiatric Illness (SPI) scale for monitoring clinical practice in emergency psychiatric services.
Crisis workers in two emergency psychiatric services crisis teams in Toronto, Canada, used the SPI in the assessment of 205 visitors to the services during the winter of 1998-1999. Contextual characteristics, including bed availability, service site, and the admitting physician's level of training, were recorded. Multivariate logistic regression was used to assess the relative contribution of patient and contextual variables in the admission decision.
The severity of axis I symptoms and difficulties with self-care were significantly associated with the decision to admit. Site, bed availability, and the admitting physician's level of training did not appear to be associated with clinical decisions.
Patients with the most need are being admitted to inpatient units despite significant systemic pressures on inpatient services. The SPI is a useful and discriminating tool for evaluating clinical practice in emergency services.