Opening of Psychiatric Observation Unit Eases Boarding Crisis

Objectives: The objective of this study was to evaluate the effect of a psychiatric observation unit in reducing emergency department (ED) boarding and length of stay (LOS) for patients presenting with primary psychiatric chief complaints. A secondary outcome was to determine the effect of a psychiatric observation unit on inpatient psychiatric bed utilization.

Methods: This study was a before-and-after analysis conducted in a 1,541-bed tertiary care academic medical center including an adult ED with annual census over 90,000 between February 2013 and July 2014. All adult patients (age > 17 years) requiring evaluation by the acute psychiatry service in the crisis intervention unit (CIU) within the ED were included. Patients who left without being seen, left against medical advice, or were dispositioned to the pediatric hospital, hospice, or court/law enforcement were excluded. In December 2013, a 12-bed locked psychiatric observation unit was opened that included dedicated behavioral health staff and was intended for psychiatric patients requiring up to 48 hours of care. The primary outcomes were ED LOS, CIU LOS, and total LOS. Secondary outcomes included the hold rate defined as the proportion of acute psychiatry patients requiring subsequent observation or inpatient admission and the inpatient psychiatric admission rate. For the primary analysis we constructed ARIMA regression models that account for secular changes in the primary outcomes. We conducted two sensitivity analyses, first replicating the primary analysis after excluding patients with concurrent acute intoxication and second by comparing the 3-month period postintervention to the identical 3-month period of the prior year to account for seasonality.

Results: A total of 3,501 patients were included before intervention and 3,798 after intervention. The median ED LOS for the preintervention period was 155 minutes (interquartile range [IQR] = 19-346 minutes), lower than the median ED LOS for the postintervention period of 35 minutes (IQR = 9-209 minutes, p < 0.0001). Similar reductions were observed in CIU LOS (865 minutes vs. 379 minutes, p < 0.0001) and total LOS (1,112 minutes vs. 920 minutes, p = 0.003). The psychiatric hold rate was statistically higher after intervention (before = 42%, after = 50%, p < 0.0001), however, coupled with a statistically lower psychiatric admission rate (before = 42%, after = 25%, p < 0.0001).

Conclusions: Creation of an acute psychiatric observation improves ED and acute psychiatric service throughput while supporting the efficient allocation of scare inpatient psychiatric beds. This novel approach demonstrates the promise of extending successful observation care models from medical to psychiatric illness with the potential to improve the value of acute psychiatric care while minimizing the harms of ED crowding.