In recent years we have witnessed a markedly increased sensitivity to the potential for abuse of the so-called police powers of physicians. The criteria for involuntary admission have shifted from a treatment model to a dangerousness model, while the philosophical shift toward treatment in the least restrictive setting has been accelerated by hard economic reality. Between 1970 and 1994, the number of episodes of psychiatric care more than doubled, while the number of inpatient beds was cut by more than half (1). Payment for psychiatric hospital care has also become entwined with dangerousness. Lack of access may now be more of a rights issue than is deprivation of liberty.

Consequently, the concentration of aggressive patients in the hospital has risen (2), and hospitals have become increasingly dangerous places. Concern has also heightened about violence committed by mentally ill persons in the community. Emergency services are an increasingly important component in a process with very serious consequences for the consumer and the community. In this complex situation where emergency mental health professionals are asked to weigh a number of clinical, legal, and economic issues, debate now arises about the use of physical and chemical restraint or seclusion.

Although restraint may well be justifiable in many instances in the psychiatric emergency service, its ultimate value remains unclear. In an extensive review of the literature, Fisher (3) underscored the utility and clinical efficacy of restraint and seclusion in maintaining patient and staff safety in a variety of psychiatric treatment settings. However, that review and others also convincingly point to deleterious effects of restraint and seclusion on patients, who perceive them to be coercive and traumatic (4,5).

Early scrutiny of use of restraint involved nursing home populations. As a result of legislation in the Omnibus Budget Reconciliation Act of 1987 and the Commission on Accreditation of Rehabilitation Facilities standards of 1993, use of restraint in long-term-care facilities decreased substantially. Psychiatric and medical facilities received less scrutiny until recently. In 1994 the New York State Commission on Quality of Care reported 111 patient deaths over the ten-year period ending in 1993, which led to a statewide review of restraint and seclusion practices (6). Authoritative statements from the commission and, more recently, from the National Association of State Mental Health Program Directors now question the therapeutic value of restraint and seclusion and emphasize their traumatic nature (6,7).

Other work demonstrates a wide variability across sites in use of restraint and seclusion that can be accounted for by institutional norms but not by patient characteristics (8). Finally, in 1998 a five-part series published in the Hartford Courant documented 142 deaths of patients in restraint or seclusion in the United States over a ten-year period and estimated that 50 to 150 such deaths occur each year (9). This influential series was a direct antecedent of the recently introduced Health Care Financing Administration (HCFA) conditions of participation for facilities receiving Medicare and Medicaid payments. These rules, which address patients' rights in general, address restraint and seclusion practices specifically, and they appear to equate chemical and physical forms of restraint.

Few data are available on the extent of restraint and seclusion practices in emergency settings. One recent study of 50 psychiatric emergency services in the United States showed that 37.2 percent of patients presented involuntarily, but that only 8.5 percent of all patients were restrained at any point in their emergency stay. The mean duration of restraint was 3.3±2.9 hours. The mean± SD annual rate of assaults of staff by patients was 8±17.4, and, as with studies in other settings, restraint rates were not correlated with staff assault rates or with volume of patients treated (10). Less is known about the use of restraints in medical and surgical emergency settings, although recent adverse outcomes have been reported (11).

Both HCFA and the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) use different sets of restraint and seclusion guidelines for psychiatric and medical settings. Most psychiatric emergency service facilities in general hospital settings operate under the more permissive medical rules. As of this writing, JCAHO has not yet determined whether to apply medical-surgical or psychiatric standards to the psychiatric emergency service.

Virtually all other sites of psychiatric service delivery are included in the new HCFA standards, which went into effect on August 2, 1999. Under these rules, nonphysician "licensed independent practitioners," including nurses, psychologists, and social workers, may order restraints, but a face-to-face examination by a physician must occur within one hour. Restraint orders are limited to four hours for adults, two hours for persons aged nine to 17, and one hour for children under age nine.

The documentation standards for restraint and seclusion episodes cover nine topics, including circumstances leading to use, monitoring requirements, and staff debriefing. Use of restraint for "managing behavioral emergencies is allowed only when all less restrictive measures have failed and unanticipated severely aggressive or destructive behavior places the patient or others in imminent danger of self-harm."

Although a consensus appears to be emerging that restraint is an extraordinary practice, each feature of the new rules has provoked controversy. The categories of providers who are licensed as independent practitioners vary by state, which has the effect of creating more limitations in some states than in others. Consumer advocates argue for more stringent limits, such as questioning any use of restraint or seclusion for patients younger than nine years. The American Medical Association (AMA) objects to the requirement that a face-to-face examination by a physician occur within one hour of initiating restraint or seclusion, citing difficulties in providing such coverage in rural settings. The American Psychiatric Association (APA) also has objections to the one-hour rule, and supports the permissibility of use of restraint and seclusion in clinical situations that may not involve imminent safety issues.

Both AMA and APA object to language stating that the less restrictive measures must always precede restraint or seclusion, citing the safety risks inherent in delaying appropriate intervention where clearly necessary. This narrowly constructed indication for restraint also precludes its use to maintain an orderly therapeutic milieu, which has been permissible in some jurisdictions, including Colorado. The failure to restrain some patients may have adverse effects on other patients, for which practitioners and institutions will also be liable.

Common elements in a number of published statements include an emphasis on training providers in alternative methods and on careful assessment, reassessment, and treatment planning involving the consumer. However, psychiatric emergency services are a poorly defined entity even in theory and a heterogeneous mix of practice settings in reality. A few states have defined psychiatric emergency services in terms of staffing requirements for a given volume or have provided support for model programs. To our knowledge, however, only New Jersey has a uniform system of screening centers statewide (1). The law in most states is driven by the due process requirements of civil commitment rather than by a rational plan for the care of severely disturbed patients.

The concept of chemical restraint hinges on whether an agent is given as a part of the treatment of the patient's condition or simply to control the patient's behavior. A recent bulletin from HCFA seems to indicate that it is the process of prescribing rather than the agent prescribed that distinguishes treatment from restraint (12). If a medication is prescribed as part of an assessment and rational plan of care, whether on a scheduled or an as-needed basis, it is a treatment. If prescribed simply as a reaction to the patient's behavior, it is a restraint. Hence the same medication administered to the same patient might be a treatment in some circumstances and a restraint in others.

The American Association for Emergency Psychiatry (AAEP) has advocated treatment rather than triage in emergency settings. Treatment of the underlying condition is predicated on appropriate assessment, diagnosis, and evidence of the treatment's value for that condition. The question then becomes one of the extent to which appropriately trained providers are available on a 24-hour basis to perform that assessment and what form of assessment should be available in circumstances grave enough to warrant restraint or seclusion.

As a matter of policy, a triage or screening examination hardly seems proportional to an intervention as controversial as physical restraint. Medical screening and a thorough psychiatric assessment should be performed if the situation is indeed serious enough to warrant any form of involuntary treatment. However, availability of this level of service depends on the volume of mental health cases seen by a given agency and on whether it has a training mission.

In general, hospitals that have 2,000 to 3,000 emergency mental health visits a year can support an organized psychiatric emergency service with appropriately trained mental health staff. An appropriate physical plant is also important, for both safety and privacy. It is useful to organize such services under a psychiatry department to facilitate training and cross-coverage by mental health rather than medical-surgical staff. One of the authors (MHA) has advocated the designation of level I psychiatric emergency services with 24-hour psychiatrist coverage as a first step in addressing these problems (1).

Given an appropriate setting and staff, what services should be provided? There is a remarkable lack of consensus about the scope of emergency assessment, although the American Psychiatric Association Practice Guideline for the Psychiatric Evaluation of Adults calls for a complete assessment including collateral contacts (13). Particular elements related to the assessment of dangerousness are even less well defined in practice.

Finally, although the risks of restraints may be thought by some to outweigh their benefits, medications have clear indications in the treatment of various mental illnesses. The most common medication strategy in psychiatric emergency settings is the use of haloperidol and lorazepam in combination. This approach is often applied before assessment has been made, with the goal of reducing agitated behavior to a level permitting safe assessment. Although this practice is generally regarded as safe and effective, the evidence supporting it is remarkably thin, with only two randomized, controlled studies totaling 118 subjects (14,15).

AAEP is committed to developing an evidence-based guideline for management of the behavioral emergency. It is hoped that regulatory attention to restraint and seclusion will provide some impetus to clarify the organizational status of psychiatric emergency services and improve understanding of the technical problems in the care of agitated patients.

Dr. Currier is affiliated with the departments of psychiatry and emergency medicine at the University of Rochester. Dr. Allen is with the department of psychiatry at the Denver Health Medical Center and the University of Colorado. Send correspondence to Dr. Currier, Director of Emergency Psychiatry, University of Rochester Medical Center, 601 Elmwood Avenue, Rochester, New York 14642 (e-mail, glenn_currier@urmc.rochester.edu). Douglas H. Hughes, M.D., is editor of this column.