Three factors are commonly involved in the hospitalization of children and adolescents: caregiver problems, acting out, and self-harm (1). Significant variation exists from location to location in the relative importance of these factors, suggesting that other issues are in play as well. For example, in a study of a children's psychiatric service of a New York City inner-city hospital, a high frequency of violence to self or others was found to be the presenting problem that accounted for hospitalization (2). In contrast, problem behaviors, academic problems, and family stress were associated with mental health hospitalization service need and crisis use in the Netherlands (3). In some settings psychiatric hospitalization of adolescents is particularly impractical. For example, in rural areas—because of difficulties of continuity of care, family involvement, and visitation—outpatient alternatives often are developed to try to provide treatment in community settings, even for situations that might have precipitated hospitalization elsewhere.

Because of the meager resources in our two-county area in a semirural community in Ohio, we felt that is was important to find alternatives to hospitalization whenever possible. Our impression was that hospitalizations, in many instances, might be avoided if a more effective evaluation method could be established. We describe our program for assessing adolescents in a semirural area with a standardized interview protocol. We hope that others in resource-poor areas will find our procedures for emergency evaluation of interest and useful.

Professional resources in our rural two-county area are limited. We have one urgent care center and do not have a general hospital or inpatient psychiatric facility in either of the two counties. One clinic with three satellite offices serves the mental health needs of all adults. The clinic also has outpatient substance abuse services for adults and adolescents. Another clinic with two satellite offices serves the mental health needs of children and adolescents. All outpatient clinics collaborate with the emergency hotline by alerting the crisis staff who are on call to any potential crises, so clients who seem to be at risk can be served after clinic hours. These clinics are staffed by part-time general psychiatrists and part-time nurses. The clinics employ full-time clinical social workers, who are supervised by licensed independent clinical social workers with master's degrees, and mental health and substance abuse case managers (community service providers). The standardized protocol for emergency evaluation of youths is implemented by social workers who are supervised by one of the general psychiatrists. The general psychiatrists live at such a distance as to be unable to attend to most crises in person.

We report on 48 adolescents who were admitted consecutively for emergency hospitalization evaluation. These evaluations were specifically requested by members of the community (police, urgent care physicians, nurses, teachers, or family). The adolescents were aged 12 to 18 years, with a modal age of 16.5 years. Twenty-two males and 26 females were assessed in the study reported here. The adolescents were evaluated by master's-level clinical social workers, who were supervised over the telephone by a licensed independent social worker and a general psychiatrist. All adolescents who had taken overdoses were first treated medically in the urgent care facility. The mental health assessments were done in the urgent care facility at the request of the physician. In the facility, adolescents can be helped for up to 24 hours, if medically necessary. Semistructured interviews and mental status examinations were conducted for all individuals by the clinical social workers employed by the agency.

The interviews included a suicidal risk assessment tool that was developed by a hospital utilization management advisory workgroup for the Ohio Department of Mental Health (unpublished scale, ODMH, 1998). The suicidal risk assessment tool used 5-point Likert scales to rate suicidal plan, support, previous suicide attempts, stressors, reality testing, acceptance of help, disorganization of thought processes, and presence of medical problems. A homicidal risk assessment tool (unpublished scale, ODMH, 1998), provided as part of the department's recommended forensic assessment materials, was also completed. The homicidal risk assessment tool used 5-point Likert scales to rate verbal threats of violence, temper outbursts, availability of interventions, and history of arrests. Both ODMH instruments can be obtained by writing to ODMH, Office of the Medical Director, 30 East Broad Street, Columbus, Ohio 45215-3430.A Global Assessment of Function score was determined (4). The Brief Psychiatric Rating Scale also was completed (5). The Crisis Triage Rating Scale was then completed by the clinical social workers employed by the agency (6,7). This scale consists of three subscales—dangerousness, support system, and ability to cooperate—which have been empirically validated for use among adults and adolescents as a supplement to help with the decision about whether or not to hospitalize a psychiatric patient.After consulting with one of the general psychiatrists by telephone after the initial evaluation, the clinical social worker attempted a supportive, reality-based, present-focused therapeutic intervention at the time of the emergency assessment. In some cases the clinical social worker returned a second time to the urgent care facility to reevaluate the status of the adolescent within 24 hours. After this intervention the adolescent was either hospitalized psychiatrically or referred to one of several types of outpatient programs with immediate intensive case management.

Of the 48 individuals for whom emergency hospitalization assessments were requested, only five were hospitalized (10 percent). Of the 43 who were referred to an outpatient program, all received intensive case management (8), were engaged in treatment with a therapist within the next week, and were evaluated by a psychiatrist within the next one to three weeks. None of the adolescents were hospitalized within a month of the emergency assessment, but two were hospitalized within the next six months, both after taking overdoses.

Thirty-seven of the 48 adolescents (77 percent) already had cases open at the agency. Twenty (42 percent) had taken overdoses (three had taken antidepressants, one had taken methylphenidate, and 16 had taken over-the-counter medications), two (4 percent) had cut themselves, and one had injected chlorine bleach into the arm. Sixteen (33 percent) had reported suicidal thoughts to community members, and four (8 percent; three boys and one girl) had evidenced violent behavior. In contrast to the findings of a previous study of rural populations (9), none of the adolescents reported seeking help from any informal network or support system, such as the clergy. Acute family problems of some type were involved in 31 cases (65 percent), and 25 of these 31 cases (81 percent) involved arguments or conflict with a stepfather, usually over his attempt to discipline the adolescent. Situations involving conflict or breakup with a boyfriend or girlfriend were involved in 12 cases (25 percent). Conflict with peers was involved in three cases (6 percent).Thirty-two of the 48 adolescents (67 percent) had affective disorders (unipolar depression, bipolar depression, or dysthymia or adjustment disorder with depressed mood), eight (17 percent) had evidence of a conduct disorder, seven (15 percent) were psychotic, three (6 percent) had anxiety disorders, two (4 percent) had anorexia, and one (2 percent) had a principal diagnosis of a substance use disorder. A substance use disorder also was a complicating factor in 11 other cases (23 percent).

A majority of adolescents for whom emergency hospital screening was requested had affective disorders. This finding is comparable to those of previous studies of hospital admission, in which dangerousness to self or others was the principal cause of hospitalization (1,2,3). In the study reported here, most adolescents referred for evaluation were managed on an outpatient basis, with immediate intensive case management and intensive wraparound services. Most of the adolescents already had cases open at the outpatient agency, so detailed data were available when the adolescents required crisis assessments and most adolescents already had relationships with one or more of the agency staff members. Also, clinic staff at the outpatient agency routinely submitted crisis alerts to the on-call crisis staff, so clients who appeared to be at risk could be served after outpatient business hours. Finally, an urgent care center was available where patients could be medically evaluated and treated for overdoses locally, so that patients did not have to be transferred out of the county for emergent medical care.

Most of the adolescents who were referred for emergency hospitalization evaluation but not hospitalized had relatively low scores on the Crisis Triage Rating Scale and were at relatively low risk of self-harm. We cannot say whether these adolescents were less or more impaired than those who are hospitalized elsewhere, because a crisis interview protocol with standardized rating scales is not commonly used in most settings. However, it was apparent that many of the adolescent mental health emergencies for which our crisis team was called were not caused by psychopathology itself but rather by interpersonal issues and emotional turmoil.Development of a standardized protocol of evaluation, which used empirically validated scales such as the Crisis Triage Rating Scale and the Brief Psychiatric Rating Scale, was helpful for several reasons. The standardized scales gave the clinicians anchor points for comparison purposes when they reevaluated an adolescent during a follow-up interview, especially when a therapeutic intervention had been done. The standardized format improved telephone communication with the consulting psychiatrist and ensured that all pertinent information had been obtained, including medical evaluations, blood tests, and vital signs after overdoses. The protocol was useful for quality assurance purposes during later peer reviews and treatment team meetings.The rural community in this study has supported a local mental health tax levy for a number of years, and these funds support extra outpatient services for adolescents, such as respite care and psychotherapy services. However, we found much less of an informal support network in this community than in previous studies conducted in rural settings (9). Our rural community is in transition from an agricultural economy to one that is based on light industry and service. Traditional social supports, based on informal social networks, are less available as the community becomes less homogeneous and more diverse and as more residents commute to jobs outside the community. Family breakdown, illustrated by the large number of remarriages and stepparents among the adolescents' families, is related to the transitional nature of the community, where loss of family farms and disrupted intergenerational traditions exert a toll on marital stability.

In a setting where crisis services and hospitalization decisions must be provided by the psychiatrist in collaboration with other professionals in the field, we find it extremely helpful to use a standardized protocol for evaluation, which includes evidence-based decision-making tools, such as the Crisis Triage Rating Scale and the Brief Psychiatric Rating Scale. This protocol helps ensure a common frame of reference for decision making and also guides repeat assessments of the same patient. The protocol incorporates a clinical interview that includes critical questions about self-harm and harm to others and about specific medical issues and medication history, if any. The semistructured nature of the interview allows the clinician in the field discretion to conduct the interview in a way that seems natural to the clinician's training and to the setting.

In such situations, it is essential that the consulting psychiatrist is a well-trained, experienced clinician who has worked in public settings, knows and respects the outreach staff, and participates in the clinic and in the in-service training of the staff on site during the week. The crisis service is supported by intensive emergency case management, 24-hour hotline availability, and timely outpatient appointments with the psychiatrist in the clinic. In our rural setting this protocol seems to provide adequate care for adolescents, who otherwise would have to be transferred to distant inpatient facilities, then reintegrated into family, school, and community after inpatient treatment.

Dr. Gillig is professor of psychiatry at Wright State University, P.O. Box 927, Dayton, Ohio 45401 (e-mail, paulette.gillig@wright.edu). Charles Huffine, M.D., is editor of this column.