The International Journal of Psychosocial Rehabilitation



Peter Micheels, Ph.D., Clinical Psychologist - Bellevue Hospital Center

Louis F. Cuoco, D.S.W., Associate Executive Director - The Bridge Inc.

Frank Lipton, M.D., Deputy Commissioner - NYC Human Resources Administration

Arthur J. Anderson, Ph.D., Consulting Clinical Psychologist - U.S. Mission, Honduras

. . . . . . .

Citation:  Anderson, A. J., Micheels, P., Cuoco, L., Byrne, T. (1998)  Criteria Based Voluntary and Involuntary
      Psychiatric Admissions Modeling.  International Journal of Psychosocial Rehabilitation. 2(2), 176-188.

Abstract: This exploratory study examined acute care admissions criteria for voluntary and involuntary patients admitted to a large municipal hospital. Symptom presentation for both voluntary and involutary admissions were analyzed along with the mode of patient arrival, domicile information and psychiatric hospitalization history; using the Bellevue Psychiatric Audit, Structured Clinical Interview, mental status examination and psychiatric interview. As expected, it was found that symptom presentation upon arrival was the primary basis for admission. However, it was determined thathospitalization judgements are not solely based on the full range of diagnostic information. Intrapsychic distress and objective symptom severity are not always taken into account when determining the need for involuntary hospitalization. Those patients who presented with symptoms that could be behaviorally rated and assessed were hospitalized more often than those who suffered from more internalized distress, without a strong behavioral component. This has a critical impact on patient's course of treatment and strong implications for admissions policy in both secondary and tertiary treatment settings.

INTRODUCTION : Municipal hospitals have long been utilized by various sources including the police to determine whether an individual was an imminent danger to himself or to others for the purpose of acute hospital admission or related treatment. Bittner (1) has argued that "the decision to invoke the law governing emergency apprehension is not based on an appraisal of objective features of the case. Rather, the decision is a residual resource, the use of which is determined largely by the absence of other alternatives. "To warrant official police action a case must also present a serious police problem..." That is to say a situation which is considered to be a danger to life, to physical health, to property and to the order in public places.

The added burden of deinstitutionalization has severely taxed the resources of facilities such as these in. They struggle with the sheer volume of these cases. New York City's Health and Hospitals Corporation, particularly, Bellevue Hospital Center has been widely used in this manner due to the existence of a 24 hour psychiatric emergency room in addition to the general hospital emergency room. HHC facilities perform more voluntary and involuntary admissions assessments than any other hospitals of their kind. The number of Emotionally Disturbed Patients (EDPs) brought to HHC facilities by police and emergency services personnel between 1976 and 1988 have increased by 1600%. In 1976 1030 cases were treated, while in 1988 17,617 cases were seen in the psychiatric emergency room.

During this same period there was a corresponding reduction of 28.5% of State Psychiatric beds between 1978 1988; 22% less beds available to New York City and a 52% reduction in HHC transfers to New York State Psychiatric Centers (SOMH facilities) from 1981 1988. In fiscal year 1988, only 49% of the HHC patients requiring intermediate or long term psychiatric care were transferred to SOMH facilities.(2)

In the past twenty five years, inquiries have been made into compiling a possible profile of those individuals that would inflict bodily harm upon themselves or others and require involuntary admission and those who do not. Studies concerning the clinician's ability to predict violent behavior have generally agreed that dangerousness is difficult to predict with any degree of accuracy.(3,4)

Rubin and Mills (5) performed a retrospective examination on the prehospitalization behavior of voluntary and involuntary patients in order to determine the precipitating factors in each instance. They found that patients admitted involuntarily had higher incidences of "dangerous acts directed towards others" than voluntary patients. No significant differences were discerned when comparing the degree of harm caused by either group. Further, the overwhelming majority of the patients' "harmful acts" were either threats or acts which caused no harm to the victim.

Yesavage et al (6) examined the relationship between civil commitment for dangerousness to others and violent acts with behavioral ratings made immediately after commitment. The hostility scale of the BPRS (Brief Psychiatric Rating Scale) resulted in the sole statistically significant difference between subjects who were and those who were not considered dangerous to others. Specifically these were patients who were admitted as "dangerous to others" (p<.05) and those patients who "had one or more assaultive events" (p<.001). The other BPRS scales did not demonstrate significant differences between the groups.

Similarily, McNeil and Binder (7) found that over two thirds of their sample had a violent episode within the first 72 hours of admission to an acute psychiatric unit. Their findings suggest that there is a relatively high degree of short term predictive validity when assessing an emergency commitment situation.

Gerson and Bassuk (8) concluded that "the essential task in the emergency room is to delineate those factors which can be readily translated into a dispositional choice". They call for "a model aimed at a relatively rapid evaluation, containment, and referral of the patient in crisis." A pragmatic model which integrates "an evaluation of both the patient's and the community's adaptive resources and competence' and minimizes the more subtle diagnostic considerations. In their review of the literature they found that symptoms and not diagnosis are related to judgments about the nature of the emergency room visit. (9,10)However, before such a model can be developed, a phenomenologic evaluation to determine what intrapsychic and behavioral factors exist which are commonly used by ER staff to make an admissions determination. By determining the diagnostic and other biopsychosocial biases that exist in current involuntary and voluntary admissions practices a clearer understanding of the full current admissions criteria can be more fully understood. This could then lead to more effective and diagnostically appropriate use of the psychiatric emergency room and acute hospitalizations in general.

How do potential psychiatric inpatients present when they arrive at the hospital's emergency room? Does the symptomatology of involuntarily admitted patients differ from that of those admitted voluntarily? Can a symptom profiles be constructed to facilitate the decision making process on the part of the clinician and or other social agent (police officer, protective services worker, community based social services...) to make more adequate and effective clinical and fiscal use of the hospital emergency room and inpatient facilities?

By phenomenologically examining current practice with regard to admission criteria for voluntary and involuntary admissions trends should appear which can then be used as a baseline for policy change and more effective admissions criteria modeling.


This investigation examined the symptom profiles of two patient samples admitted to Bellevue Hospital Center. Each of these samples was differentiated into two groups those patients who were either admitted voluntarily (with the patient's consent) or admitted involuntarily (determined by psychiatric evaluation to be at risk to self or community). Analysis was performed between and within samples to distinguish the major symptom constellations for voluntary and involuntary admissions to Bellevue Hospital Center inpatient psychiatric units.


Voluntary and Involuntary Chart Review Sample

This sample consisted of 50 voluntary and 50 involuntary adult male and female patients selected at random from Bellevue Hospital Center admissions in 198l. The selection procedure consisted of 10 random days selected for each of 5 randomly chosen months during a 1 1/2 years period (198081). From each day one voluntary and one involuntary admission was arbitrarily chosen. Symptom descriptions were obtained from the evaluation of the admitting physician and from the Bellevue Psychiatric audit BPA (BPA, ). The BPA is a symptom checklist which identifies major symptomatic disturbance and assigns a numerical severity weight (03) to each symptom. The criteria for inclusion of symptoms into the present investigation depended on a BPA symptom weight of greater than zero.

Of the 100 patients studied 37 were male and 63 were female. Mean age was 30.85; S.D.=11.23 years; range 2166 years. Between the voluntary and involuntary admissions groups there were no significant differences with regard to age, sex, race, educational level. Patient diagnosis included: schizophrenia (56), acute psychosis (4), alcohol abuse (4), drug abuse (2), major affective disorder (3), personality disorders (14), other (7). In addition to symptom identification, voluntary and involuntary patients group within this sample were evaluated for current housing situation, psychiatric history, and mode of arrival; and no significant differences were established for housing situation or psychiatric history. However, significant differences were identified in mode of arrival.

SCI Tested Sample

This sample consisted of 20 voluntary and 20 involuntary adult male and female psychiatric patient admitted to Bellevue Hospital Center in 1987. Patients were selected at random as they entered the hospital for psychiatric evaluation. The Structured Clinical Interview (11) was administered by trained independent interviewer to both the voluntary and involuntary groups. The SCI is a symptom rating scale administered by trained examiner which rates patients along 10 subscales and an overall severity scale to yield standard scores. Mean standard scores for the 10 SCI symptom subscales and the SCI symptom severity scale for each group was determined and analyzed both within and between groups.

This second sample included 26 males and 14 females. Mean age was 31.775 years; S.D.=11.11 years; Range 1960 years. Between the voluntary and involuntary groups there were no significant differences with regard to age, sex, race educational level, or diagnosis. In addition, patient diagnostic distribution did not significantly differ from that of the first sample. Current patient housing situation, psychiatric history and mode of patient arrival were also examined for both the voluntary and involuntary admission groups; with significant differences being established only for mode of arrival.


Symptom levels for voluntary and involuntary patients were analyzed within and between samples to establish gross symptom differences between voluntary and involuntary patient admissions. Chi Squared statistics were obtained to demonstrate differences between the levels and constellations of symptoms of voluntary and involuntary patients in both samples. Analysis of variance was performed on the voluntary and involuntary SCI measured sample to demonstrate intergroup symptom differences. Mode of arrival was analyzed for these groups within and between samples. All significant statistical differences between the voluntary and involuntary groups of both samples are detailed in the results.

The results illustrated in Tables 13 and Figure 1, listed above, demonstrated significant differences between the symptom constellations of patients admitted voluntarily and those admitted involuntarily to Bellevue Hospital Center. These symptom differences have an effect on the mode of arrival to the hospital and the probability of admission to an inpatient psychiatric units.

Voluntarily admitted patients, regardless of their diagnosis, tended to possess an internally focused constellation of symptoms and differed significantly from the symptoms of those patients who were admitted involuntarily. Patients who were admitted involuntarily tended to present symptoms that were either externally focused or projected.

The results in Tables 1 and 3 and Figure 1 illustrate the internal and external constellations for the two patient groups and variance between the groups. Patients seeking assistance and admitted voluntarily possessed significantly greater levels of fear, anxiety, self depreciation, physical complaints, depressed mood, suicidal ideation, sleep disturbance and perceptual dysfunction than involuntarily admitted patients. Those patients who were admitted against their will were determined to possess significantly higher levels of grandiosity, conceptual (cognitive) dysfunction, anger/hostility, incongruous ideation/behavior, antisocial attitudes/acts, motor and affective agitation, and suspiciousness than the voluntarily admitted groups.

The differential constellation of symptoms between the two groups was consistent for each sample, though symptom evaluation procedures and instruments differed for each sample. In general, the voluntarily admitted patients possessed internally focused symptoms. This constellation could be conceptualized as an intrinsic constellation. Symptoms in this constellation tend to have internal origins and are focused on the internal state of the patient. Involuntarily admitted patients present symptoms that could be thought of as an extrinsic symptom constellation; with an external origin and focus as construed from the patients point of view.

The differential symptom constellation correlated with and was confirmed by the mode of arrival. Table 2 demonstrates that voluntarily admitted patients possessed an intrinsic constellation of symptoms. These patients were more likely to recognize their need for medical/psychiatric assistance and would seek out such assistance. Those patients presenting more of an extrinsic constellation were not as likely to recognize their need for assistance and were more likely to be brought to the hospital by a friend or the police.

Finally, while overall symptom severity did not differ significantly between the two groups in the SCI measured sample, patients admitted voluntarily tended have a marginally greater level of symptom severity (Table 3). This finding leads to the speculation that patients possessing an intrinsic constellation of symptoms require a greater degree of symptom severity in order for the symptoms to be noted and used in the psychiatric admissions process. The involuntarily admitted patient's extrinsic constellation was more readily noted in the intake evaluation and thus, a lower overall severity level was necessary to confirm the need for hospitalization.


The results of this investigation are consistent with the previously noted Gerson and Bassuk analysis. Presenting symptoms are generally used as a basis for determining the type of emergency room visit. Overlaying the mode of arrival and results of structured behavioral and symptom assessments, it is clear that current practice and emergency room hospitalization judgements do not take diagnostic nor intrapsychic distress and objective symptom severity into account when determining the need for involuntary hospitalization. To determine the conditions, factors and symptom severity necessary to hospitalize patients against their will, more comprehensive objective measures need to be employed then are commonly used in emergency room practice today.

Current emergency room clinical assessment techniques are biased in their approach and may account for at least part of the inaccurate assessment of need for involuntary hospitalization. Most admissions are currently made on the basis of an unstructured clinical interview and historical information provided by the patient and external, familial and social sector service staff. Because of this, all three informational contexts represent sources of bias that color the admitting clinicians judgement.

Psychiatric Residents, Psychological Interns, Social Workers and Nurse Practitioners often use a loose, unstructured interview that generally reflects their particular level of training and competency. Often essential intrinsic, intrapsychic symptom and diagnostic information is inconsistantly noted by these professionals. Patients presenting with more overt, extrinsic, behavioral disturbances are often assumed to be more disturbed and require hospitalization. Such clinicians may overlook other patients who present with less dramatic, internally based symptoms. This is clearly the case in this study where patients with severely dysfunctional intrinsic symptoms tended not be involuntarily hospitalized as readily as those who possessed more extrinsic symptoms.

Depending on the level of reality testing, presenting patients have an active stake in the outcome of the psychiatric admissions evaluation. If they desire hospitalization they will exaggerate and confabulate their type and degree of symptom severity. This is particularly true with the seriously and persistently mentally ill who have histories and multiple hospitalizations. Thus, these patients have learned to present severe extrinsic symptomatology and welcome voluntary hospitalization. Patients who do not desire hospitalization often present more realistic, negative symptom constellation and are probable less likely to confabulate their symptom type or severity. In fact, these patients may often attempt to mislead interviewers as to the true nature of their symptoms. This is an attempt to lead psychiatric interviewers into believing they do not require acute hospitalization. Again, this is consistent with the outcomes of this investigation and may lead to inappropriate treatment and release of these individuals from emergency rooms.

Finally, the familial and social service sector staff who escort patient to emergency rooms may also be introducing a bias into the system well. These escorts have independently determined the need for inpatient hospitalization and are likely to be strong advocates for hospitalization. In addition, because of the advocacy position they are forced to adopt a strong admissions posture with emergency room staff regarding the need for admission. This introduces and extra symptomatic bias into the admissions process. Consequently these patients are more likely to be admitted regardless of their true symptomatology. The outcomes of this study (table 2) demonstrate these bias artifacts as well.

Underlying this entire investigation lies the questions of psychiatric prediction of level of dysfunction and relative probability of harm to self or others. In a global sense, one might assume that low GAF Axis V patients who present with extrinsic symptoms would be excellent candidates for involuntary hospitalization. The underlying assumption here is that these patients are least able to control their behavior, since their symptom constellations are of an overt, behavioral nature, and are more likely to inflict harm to self or others. This is essentially and behavioral, Axis V bias in the admissions process.

As demonstrated in the voluntary and involuntary groups for both measures and the results of the escorted patient assessment, severe intrapsychic distress and dysfunction is not related a particular level of overt function. The related symptoms appear to be internal and intrinsic and do lead to accurate GAF assessment, thus clouding and biasing the diagnostic and assessment in the emergency room. Thus the measures for assessing such dysfunction are biased against this population of dysfunctional patients. This is also confirmed by Schrader (11) who found that Axis V assessments were invalid predictors of prognosis and long term functioning, which was one of the key reasons for the Axis V development, and consequently an invalid predictor of involuntary hospitalization.

Due to the obtained results, it is clear that objective assessment tools that reflect true levels of both extrinsic and intrinsic symptomatology be developed and employed in emergency rooms to more adequately determine the need for hospitalization. Using present techniques and related assessment technology the least severe patients who voluntarily present for treatment are considered for hospitalization before the more severely effected patients.

In order to more fully understand the impact of underlying symptomatology of these two differential groups of patients and develop models of voluntary and involuntary hospitalizations that will facilitate more effective use of the psychiatric emergency rooms to accurately assess the need for costly hospitalization, there must be a greater reliance on objective criteria and measures used to assess patients symptomatology and severity of pathology.

These objective measures must include a detailed assessment of both intrapsychic, affective and behavioral dysfunction and distress and be rapidly applied and readily available to all emergency room professionals and paraprofessionals who assess patients for admission.

The results of this investigation underscore the need for the development of such tools and further investigation into psychiatric emergency room modeling for admissions. Without such tools and patient modeling both patients and society as a whole will continue to incur patient and social costs surrounding inappropriate admissions of marginally dysfunctional patients and failures to admit patients who require hospitalization to prevent them from further deterioration, distress and harm to self an others.


1. Bittner E : Police discretion in emergency apprehension of mentally ill persons. Social Problems 4:278292. 1967

2. Crisis in Mental Health : Issues Affecting the Health and Hospitals Corporation's Psychiatric Inpatient and Emergency Room Services Summary Data. January 1989.

3. Cocozza, J Steadman, H : The failure of psychiatric predictions of dangerousness: clear and convincing evidence. Rutgers Law Review 29: 10941101, 1976.

4. Rofman E, Askinazi C, Fant E : The prediction of dangerous behavior in emergency civil commitment. American Journal of Psychiatry 137:10611064.

5. Rubin L, Mills M : Behavioral precipitants to civil commitment. American Journal of Psychiatry 140:603606. 1983

6. Yesavage J, Werner P, Becker J, Mills M: Shortterm civil commitment and the violent patient. American Journal of Psychiatry 139:11451148. 1982.

7. McNeil D, Binder R: Predictive validity of judgment of dangerousness in emergency civil commitment. American Journal of Psychiatry 144:197200. 1987

8. Gerson S, Bassuk E: Psychiatric emergencies : an overview. American Journal of Psychiatry 137:111. 1980

9. Trier T, Levy R : Emergent, urgent, and elective admissions. Archives of General Psychiatry 21: 423430. 1969

10. Muller J, Chafetz M, Blane H: Acute psychiatric services in the general hospital: 111 statistical survey. American Journal of Psychiatry, October supplement 4653. 1967

11. Hardesty, A., Burdock E.

12. Schrader G., Gordon M., & Harcourt R. (1986) The Usefullness of DSMIII Axis IV and Axis V assessments. American Journal of Psychiatry (Jul), Vol 143(7), 904907.

(Return to Cover Page)