Characteristics of Repeat Utilizers of Psychiatric Emergency Services

Introduction/Background: Psychiatric emergency services (PES) are a gateway to mental health services such as inpatient hospitalization or intensive case management. Frequent use of PES is an inefficient use of mental health resources that can be better addressed by intensive and preventative case management intervention. Identifying characteristics of frequent PES utilizers would allow improved targeting of intensive case management services for individuals with serious mental illness. This study examined the client demographic and service use factors that predict readmission to PES.

Methods: Charts for three months (January 2005 – March 2005) of admissions to San Francisco's only PES were reviewed. Patient demographic and clinical characteristics were collected. Service utilization data for all individuals with a PES event during these three months were obtained from administrative records for two years prior and one year following the index PES admission. Only individuals who were involuntarily admitted were retained in the study sample (N= 1540). Time-to-event analysis was conducted to predict readmission to PES. Demographic, clinical and service use characteristics were used to predict the time-to-readmission using Cox Regression.

Results: The ethnic distribution of this sample was approximately representative of San Francisco for Whites (47%) and Latinos (11%), but African Americans (27%) were overrepresented and Asian Americans (12%) were underrepresented. The sample was primarily male (65%); the average age was 41. 51% of the sample had a readmission to PES in the year following the index PES admission. The median number of PES admissions for the 2 years prior to the index admission was 1, with 33% of the sample having 2 or more PES admissions. For the index PES admission, the mean Global Assessment Scale (GAS) score was 30 out of 100, indicating serious mental health impairment. Demographic characteristics were not predictive of time-to-readmission after controlling for other covariates, with the exception that being male increased the rate of readmission by 28%. Each 10-unit increase in GAS score decreased the rate of readmission by 20% and each previous admission to PES increased the rate of readmission by 7%. Self-referral to PES was associated with a 57% increase in the rate of readmission to PES.

Conclusions: In many communities, individuals with serious mental illness initially come into contact with the mental health system though PES. Understanding the utilization of PES can inform efforts to improve the efficiency of mental health services. This study found high rates of readmission to PES. Readmission was predicted by severity of mental illness and frequency of previous use of PES. Men and individuals who referred themselves to PES also had a higher rate of readmission the year following the index admission. Ethnicity, age, substance use, and inpatient admission following the index PES admission did not predict readmission. PES is a crisis intervention; this study hints that targeting intensive case management toward frequent utilizers of PES with serious mental illness, especially men and individuals who are referring themselves could be an efficient use of mental health resources and reduce the utilization of PES.