Inpatient psychiatric care in the 21st century is defined by ultrashort lengths of stay. In the last two decades of the 20th century, length of stay for psychiatric inpatient care decreased from months to days. The sole focus of psychiatric inpatient treatment has become safety and crisis stabilization (1).
This Open Forum addresses the need to reconsider the current model of ultrashort inpatient hospitalization in order to maximize positive outcomes and emphasize appropriate transition to the community and less intensive levels of care. The patient population on which we focus includes those who most clinicians would agree require a 24-hour inpatient stay, not those who can be treated in partial hospitalization or in residential or other outpatient settings. We recognize that there is a body of literature on alternatives to the hospital, but for most patients in an acute psychiatric crisis, hospital stays are the only option (2). As this nation implements health care reform, we cannot fail to address the necessary changes in the mental health care system, including hospital care, that will make treatment more effective, efficient, and recovery oriented.
The President's New Freedom Commission report (3), the Surgeon General's report (4) that preceded it, and a variety of other public policy directives have steered psychiatric services in the United States clearly and emphatically toward the goals of the recovery movement. These goals require that psychiatric services address the patient's aspirations for a life as a member of the community and focus on the patient's residential, occupational, vocational, social, and spiritual needs in order to further that aim. It is not paradoxical to note that although a diminished role for institutional care is consistent with recovery, ultrashort hospitalizations may diminish opportunities for a sustained recovery.
Our evidence base would be vastly enhanced by controlled studies of inpatient procedures, length of stay, and outcomes. A modest literature of controlled studies of best treatment for particular disorders exists in this regard from the 1970s and 1980s, but few studies have been published since then. Most of that literature suggested that shorter rather than longer stays were more effective for most psychiatric disorders (5) (however, the shorter stays in these studies would qualify as longer stays today). The only study with findings that differed was by Glick and Hargreaves in the 1970s that compared short hospital stays (21—28 days) with long hospital stays (90—120 days) (5). Patients with schizophrenia who had good functioning before hospitalization and who received the extra days of hospital treatment showed better outcomes at six- and 18-month postdischarge follow-ups. This outcome resulted from a combination of postdischarge psychotherapy and medication compliance (5). There are no data from controlled studies on outcomes of ultrashort-stay hospitalizations to guide clinicians or public policy. Of course, in the absence of data, less expensive programs are preferred by payees. Notably, a recent review of the literature from Europe (6) made the point that "there has been some unsettling evidence to suggest that shortening hospital stays may not be a general panacea."
In the current prevalent hospitalization model, the average length of stay is five or six days. This model requires that the problem occasioning admission be formulated within 24 hours whether admission occurs on a weekday or weekend. Moving the patient through this hospital assembly line requires that diagnostic assessment be completed within 24 but no more than 48 hours. However, admission requirements that focus on dangerousness as the only criterion for medical necessity of an inpatient stay ignore the realities of mental illness. Many admitted patients are taking multiple medications or illicit drugs and have comorbid medical illnesses. Before completing the diagnostic process, psychiatric residents of a generation ago were taught the value of observing patients when they are off all drugs and medications. This is no longer possible. It is worth noting that patients admitted late in the week (Friday or Saturday) may spend the entire critical period of assessment, diagnostic formulation, treatment planning, and treatment initiation in the care of covering physicians, who are often "moonlighters."
Ultrashort stays have severely eroded the interpersonal connectedness of staff, patients, and families. At the same time, the emphasis on safety has deconstructed the physical environment of many hospital psychiatric units, lending them a prisonlike atmosphere. The overall effect is a dehumanized physical, psychological, and social environment for patients when they are in most acute need. We believe that ultrashort inpatient hospitalization may do more harm than good.
Given the lack of studies of outcomes of ultrashort stays, we propose a decision model based on a careful phenomenologic and psychosocial diagnosis of the problem using DSM-IV, as well as an evaluation of patient and family strengths, available resources, achievable goals, and interventions that cannot be safely undertaken in an outpatient setting. In short, we recognize the need to individualize treatment depending on the patient's condition, his or her previous experience with treatment, and the family and other resources available to support the treatment.
Therefore, it is necessary to recognize that rational treatment for psychiatric patients differs from that for medical or surgical patients. A focus on ensuring only safety leads to an overemphasis on the biological aspects of care (generally psychopharmacologic) to reduce aggressive behavior and leaves far too little time to address the psychosocial aspects critical to understanding and intervening in the larger context and changing the course of illness. A frequent result is rudimentary discharge plans that do not account for many predisposing issues and all too often lead to recurrence and readmission. Factors that commonly compound the need for acute treatment—for example, cognitive impairment, comorbid disorders, denial of illness, and severe functional impairment—often go unaddressed or, worse, unrecognized. Issues such as the need for stable housing, which is critical to continued remission of illness, can barely be discussed. If we are to give anything more than lip service to recovery, we need to rethink the current model.
We cannot rethink the model without considering the goals of recovery. Even rudimentary knowledge of the recovery movement suggests that the ultrashort-stay model of hospitalization is a medical model—paternalistic and not patient centered. With the focus on safety, most patients are admitted involuntarily. The rapidity with which treatment decisions must be made makes a mockery of patient participation in multidisciplinary treatment planning and psychoeducation. Finally, the larger goals of recovery that involve the social-interpersonal, residential, occupational, vocational, and spiritual contexts of patients' lives and illnesses can rarely be adequately addressed.
We suggest a model of care based on established principles that have been lost in the length-of-stay crush. Our model proceeds in three stages: assessment, implementation, and resolution of hospitalization. The assessment phase is what needs to be done before or as soon as the patient arrives in the inpatient unit from the emergency department. As much as possible should be accomplished before the patient arrives, which often requires an intake clinician. The implementation phase (that is, the period of active medication trial or of detoxification) varies in length according to the specifics of the case, and the resolution phase is also variable depending on goals. The resolution phase focuses on solidification of gains made during the implementation phase and on ensuring an effective transition to the community and the next level of care, including discharge planning tasks that should begin during the assessment phase.
Ideally, a substantial portion of the assessment phase should occur before actual admission. We readily acknowledge that the reality of unknown patients in pressured emergency departments makes this an ambitious goal. Certainly, for patients who are known to the hospital system, financial screening, critical history gathering, and preliminary diagnostic assessment should be done before admission.
The assessment phase is a critical time to observe, gather information, and formulate a plan before active treatment (except for safety measures, which need immediate implementation). Effective gathering and communication of information are emphasized. Old and current medical records (electronic preferred) must be obtained, and contact should be made with treatment providers and the patient's significant others. The model requires relying on the intake clinician for more than the usual admission decision and provision of a rudimentary history. The intake clinician may have access to sources that are not available to inpatient staff (for example, a detaining police officer) and must capitalize on this unique opportunity to gather and transmit critical information. Poor handoff from the emergency department to the inpatient unit is a common source of wasted time and energy. The intake clinician must accurately convey the history, diagnostic assessment, and reasons for admission. In other areas of medicine or surgery, the admission diagnosis may directly set the course of the treatment plan. Treatment algorithms for acute appendicitis or myocardial infarction clarify the next steps of care for providers and payers alike. However, few such algorithms are in place for psychiatric care. In their absence, goals of the assessment phase include a thorough history, a plan for proposed treatments, a sense of achievable goals, and likely prognosis.
The implementation phase is the core of the actual inpatient stay and should accomplish more than merely ensuring safety during a crisis. The crucial objective is to further define the issue or issues that led to hospitalization and implement interventions that change the illness trajectory. For example, if a patient with major depressive disorder has relapsed and his or her condition is resistant to other treatments, the task may be to implement electroconvulsive therapy. If a young adult with schizophrenia has stopped taking medication and is acutely psychotic, the primary task may be psychoeducation for patient and family and resuming medications, with consideration of intramuscular injections. The objective, which is lost in the ultrashort model, is to treat the current episode but, equally important, to put measures in place that will prevent subsequent episodes.
Several factors call for a longer length of stay in this model. They include a return to more thorough diagnostic assessments, which require complete history gathering and periods of patient observation. The time allowed for trials of medication should comport with what we know about how psychotropic agents actually work. Assessments of precipitating factors and interventions to address them, family interventions, psychoeducation, and the establishment of the therapeutic alliance must all be given their due and are critical factors in the patient's future adherence to treatment.
The resolution phase is absolutely critical in this redefined model of psychiatric hospitalization in order to consolidate the gains made during the implementation phase and ensure effective continuing treatment in day programs, intensive outpatient alternatives, and residential and community settings. Here again, the special needs of psychiatric patients with cognitive and functional impairments, comorbidities, and denial of illness may require interventions that demand more time in this phase. Patients (and family members) are adjusting to the changes produced by the interventions made during the implementation phase, and they need ongoing support to understand, tolerate, and adhere to them. Issues of financial support and posthospitalization living circumstances must be resolved in order to establish effective coordination and continuity in follow-up care. A mere generation ago, it was considered critical to make an effective handoff to the next level of care. Inpatients were routinely sent on passes to meet their caregivers and to be introduced to follow-up programs. We stopped doing this not because we had evidence to demonstrate that it was an ineffective practice but because insurers would no longer pay for the added time in the hospital. Better transitions are more consistent with the goals of recovery and are ultimately more cost-effective.
The inpatient hospitalization provides a unique opportunity to marshal resources that otherwise would not be available for the treatment of an episode of mental illness. Precise criteria for determining an appropriate length of stay are inexact at best, although a number of core principles may be applied. An overarching principle is that length of stay should be driven by clinical need and determined by clinicians involved in the patient's care. Clinical need should be measured against the ongoing effectiveness of the inpatient intervention to ensure safety, produce stability (and remission when possible), and set the stage for successful reintegration to life outside the hospital (reducing the likelihood of readmission). Focusing primarily on safety issues requires a parsing of clinical judgments that is too narrow and specific for the ambiguous clinical realities we so often face. As a result, risk may actually be increased while the utility of hospitalization to fully address the episode of illness and to prevent recurrence is diminished.
The following specific principles are notable equally for being both obvious and frequently ignored. Their practice may increase length of stay. They may be thought of as criteria for determining length of stay in that discharge should not occur until each has been accomplished.
◆ Practice established principles of psychopharmacology. These include withdrawal of ineffective or toxic medications over appropriate periods, observation periods after withdrawal, reintroduction or addition of one medication at a time, and appropriate periods for medication trials.
◆ Treat comorbid factors (for example, substance abuse and general medical problems) that contribute to the need for hospitalization and that may increase the risk of recurrence.
◆ Address the intrapsychic life of the patient, along with family issues and other social and environmental factors. Such factors must be considered in order to understand precipitating factors and response to treatment and to ensure successful disposition planning.
◆ Adhere to recovery principles, especially the role of the patient in treatment planning. The patient's meaningful participation in treatment planning implements the patient's right to choose. This may increase length of stay but may also increase adherence and the likelihood of a positive outcome.
◆ Provide psychoeducation and build a therapeutic alliance.
In addition to the above principles, our model has a number of staffing requirements. We suggest the following:
◆ An experienced preadmission or intake specialist at the "inpatient door" who understands the dynamics of the community, the emergency room, and the inpatient unit and who can work easily with physicians and other clinicians, patients, families, and significant others to facilitate the patient's entry into the care system. Such a person must be skilled at negotiating the precertification systems established by payers, which so often seem like roadblocks to care.
◆ A transition to "psychiatric hospitalists," psychiatrists, and advanced-practice registered nurses who are very experienced in treating severe mental illness in inpatient settings and are very competent in psychopharmacology. The time is short and the complexity high.
◆ Specialized nurses, analogous to operating room nurses, who are experienced in acute care, safety, and the procedures necessary for achieving the objectives described above.
◆ The "treatment manager," who can be a social worker (and most often is), a psychologist, or even a nurse (not working within the usual nursing hierarchy). The person in this role conducts assessments, individual and family work, and dispositions (the traditional social work role). In other words, this is the patient's primary therapist-clinician, who works collaboratively with the physician. For some hospitals, this may be a new role.
◆ A part-time general internist to address the high level of general medical comorbidity inherent in many psychiatric illnesses.
◆ The family member, significant other, or outpatient caseworker-clinician who provides the glue to move from pre- to postadmission stages should also be considered a member of the team. The focus is involving and educating families about the illness—its recognition, its treatment, and access to services.
Our model calls for consideration of the culture of inpatient units and the treatment methods employed. We have lost our focus on the therapeutic power of the milieu. Many training programs are available for staff in the culture of recovery and related themes, such as trauma-informed care. Programs to reduce reliance on seclusion and restraint generally enhance the therapeutic milieu, increase safety, diminish regression, and facilitate the patient's progress.
Many would argue that outpatient care can accomplish the above goals without the need for a hospital if the patient is not at risk. Two categories of reasons argue against this assertion. The first set of indications for hospitalization is similar to those our colleagues in medical and surgical specialties follow. Medical and surgical patients are hospitalized (arguably for many reasons) when their conditions are life threatening, require procedures that can be done only in a hospital, require a large team that cannot be assembled in an outpatient setting, require extensive or in-depth diagnostic procedures, or require long periods of observation while the patient is receiving treatment to try stepwise procedures or alternative treatments.
The second set of reasons is more specific to psychiatry because psychiatric patients have problems that patients treated by other specialties do not have and that make outpatient treatment difficult, if not impossible. First, psychiatric patients usually have cognitive problems and psychotic symptoms that prevent them from being "full partners on the treatment team." These include denial of illness and fearfulness about seeking treatment. Second, these patients have a high rate of nonadherence in outpatient settings to complicated psychosocial and psychopharmacological treatment regimens. Third, psychiatric patients often lack family or significant others to facilitate treatment and lack resources, such as transportation, to access treatment. And finally, the stigma associated with psychiatric treatment works against the patient's getting adequate help in an outpatient setting as well as in continuing treatment once he or she is hospitalized. For all of these reasons, voluntary or involuntary hospitalizations may be indicated and may require longer stays (longer than ultrashort stays) to achieve stabilization—that is, to ensure that the patient can survive outside the hospital and to arrange long-term (posthospital) care in order to change the downward trajectory. The argument against this model is that it can be done just as well for less money outside of a high-tech hospital—that is, in a new version of the "hospital of the past," which encompasses specifically low-tech observation followed by social interaction and time spent in a truly therapeutic milieu outside of the hospital. Of course, we don't have such a model or such a setting at this point.
Our model is not new or revolutionary, nor is it evidence based in the true meaning of the term. However, it is our effort to address a problem that is vexing and enduring with an approach that is provocatively looking to the past for a way toward the future. The evidence base for various approaches to inpatient psychiatric care is sadly lacking. We would be greatly helped by moving beyond patient satisfaction surveys to objective measures of outcomes. But in the absence of an evidence base for ultrashort hospitalization, we have an ethical obligation to promote what we consider to be best practice. Health system reform means just that—reform of the system itself. In the inpatient psychiatric setting, it should start with providing treatment that is nuanced and, in the spirit of recovery, intended to make an effective impact (beyond the assurance of safety) on the life course of the patient with severe psychiatric illness.
Dr. Glick has received research support from Lundbeck and Pfizer; has served on advisory boards for Bristol-Myers Squibb, Merck, Novartis, and Organon; and owns stock in Johnson and Johnson. Dr. Sharfstein is president and chief executive officer of Sheppard Pratt Health System, a not-for-profit organization. Dr. Schwartz is vice-president of behavioral health and director of the Department of Psychiatry at Hartford Hospital and psychiatrist-in-chief at the Institute of Living.