DSM-III-R; level ofoverall psychiat nc impairment at the time of admis sion, based on the treating clinician's rating on the Global Assessment Scale (GAS) (6); the extent of recent suicidal behavior, which was coded from chart notes according to an Un published scale developed by Han greaves, LeGoullen, and Gaynor in 1983; a rating by the treating clii cian ofwhcthcr drug or alcohol prob lems were present; and the presence or absence ofviolcnt behavior during the two weeks before admission to the psychiatric emergency service. Accounts of violent behavior in the chart notes were coded using a modified version (7) of the rating form developed by Lagos and as sociates (8). The form includes three levels ofincreasingly severe violence: no violence, fear-inducing behavior (verbal attacks, threats to attack per sons, and attacks on objects), and physical attacks on persons. If a record indicated that more than one type of violence had occurred, only the most serious type was coded. We also evaluated the presence or ab sencc of violence during the first 24 hours that the patient was in the psychiatric emergency service. In addition, we extracted infor mation from the charts on legal status, previous treatment in the psy chiatric emergency service, disposi non (whether the patient was hospi talized), and whether the patient was secluded or restrained in the psychi atnic emergency service. Comparisons between police re fernals and patients referred from other sources were made using t tests for continuous variables, such as age, and chi square analyses, corrected for continuity if appropriate, for cate gorical variables, such as gender. Results Demographic characteristics. A total of33.6 percent (N= 108) of the 32 1 patients evaluated during the study period were referred to the psy chiatnic emergency service by the police. Of the remaining 213 pa tients, 18.8 percent (N=40) were self-referred, 10.8percent(N=23) were referred by family members, 26.3 percent (N= 56) by primary medical care facilities, 17.4 percent (N=37) by mental health outpatient will not accept disruptive mentally ill persons who appear dangerous or who have concomitant substance abuse problems, police will arrest them rather than taking them to a psychiatric facility, resulting in “¿ cniminalization ofmental disorder.― On the other hand, emergency room staffoften feel that police bring in the “¿ wrong― cases, that is, patients who do not meet the legal criteria for admission on who are unlikely to ben efit from treatment (4,5). There have been few studies, however, of the characteristics of persons referred by police to the psychiatric emergency room. This study uses retrospective chart review to compare and contrast the demographic and clinical charac teristics of a large sample of police referrals and patients referred by other sources to an urban psychiatric emergency room. Methods The setting was the psychiatric emergency service ofthe San Francis co General Hospital, a university-af filiated hospital funded to serve the city and county ofSan Francisco. The sample consisted of all patients (N=321)evaluated in the psychiatric emergency service during a four week period in August 1988. If a patient had multiple visits to the psychiatric emergency service during the study period, only the first visit was included. The patients' charts were ne viewed retrospectively to determine whether they had been referred by the police and to collect data on dc mographic characteristics and several clinical variables that had been re corded in routine clinical practice by the treating physicians and nurses. Clinical variables included primary psychiatric diagnosis, based on the Dale E. McNiel, Ph.D. Chris Hatcher, Ph.D. Harriet Zeiner, Ph.D. Harriet L. Wolfe, M.D. Richard S. Myers, M.D. The police frequently become in volved when mentally disordered persons exhibit disturbed or disrup tive behavior in the community. When police have detained someone for whom they believe mental health intervention is needed, the psychiat nc emergency room typically repre sents the point ofentry into the men tal health system (1). Police referrals to the psychiatric emergency room represent an impor tant and controversial group, about whom few empirical data exist. Police often express frustration that they have difficulty obtaining treat ment for people they perceive as oh viously needing hospitalization (2,3). Teplin (3) suggested that be cause police believe that hospitals Dr. McNiel is associate adjunct professor of psychology and Dr. Hatcher is clinical professor of psychology in the department of psychiatry at the University of California. Dr. Zeiner is clinical instructor in the department of surgery at Stanford University. When this research was done, she was a fellow in clinical psychology in the department ofpsychiatry at the University of California, San Francisco. Dr. Wolfe is associate clinical professor of psychiatry and Dr. Myersis assistant clinical professor ofpsychiatry at the Ui versity of California, San Francis co. Address correspondence to Dr. McNiel at 401 Parnassas Aye nue, San Francisco, California 94143. 425 Hospital and Community Psychiatry April 1991 Vol. 42 No.4 Characteristics of Persons Referred by Police to the Psychiatric Emergency Room