Keywords

Introduction: Emergency Psychiatry

Over the last 30 years, the psychiatric approach to individuals in the emergency setting with suicidal ideation and risk factors for suicide has undergone a paradigm shift, from one favoring triage and hospitalization to one favoring treatment and hospital diversion. In the best hands, the goal has always been to collaborate with the patient in resolving the crisis and selecting the most appropriate level of care. But this new emphasis has taken over, evolving from a best practice into the standard of care.

In part, this evolution has been facilitated by a growing range of nonhospital disposition options, including those listed in Table 8.1.

Table 8.1 Nonhospital options for psychiatric care
Full size table

At the same time, the availability and perceived desirability of hospitalization have decreased. Far from always being the gold standard for the psychiatric crisis, hospitalization sometimes exacerbates a crisis by confirming an individual’s perceptions of helplessness and inability to cope. Regional differences in criteria for acute hospitalization continue to exist, and out of financial necessity, public-sector mental health has embraced hospital diversion more aggressively than either private or Veterans Affairs systems. But if a general psychiatrist of the 1980s were transported to the present, today’s practice landscape would be almost unrecognizable.

It is a striking about-face, driven by a mixture of science, patient empowerment, economic policy, and social change. Elements of this mixture include improved differentiation between acute and subacute risk; innovations in crisis and outpatient treatment technique; a stronger emphasis on patient-centered care, with an overall healthier acceptance of risk; less irrational fear about medicolegal liability; managed care’s redefinition of medical necessity; severe cuts in hospital beds; surging numbers of mental health referrals to emergency departments (EDs); patient reports of disappointing experiences with inpatient psychiatric treatment [2]; and an increased appreciation of the potential negative effects of involuntary hospitalization, such as increased stigma, exacerbated patient helplessness, and damaged therapeutic alliance [3].

Many psychiatric emergency service (PES) units combine these contemporary thoughts and practices in offering interventions, lasting from a few hours to a day or two. The PES model achieves hospital diversion rates around 70% [4, 5]. Such a safety net for temporary psychiatric regression enables outpatient and subacute care practitioners to manage sicker patients in the outpatient setting. Unfortunately, a PES is rarely available to emergency medicine (EM) practitioners. This chapter applies clinical principles from that setting to the general medical ED.

The Challenge for Emergency Medicine

EM practitioners must often manage psychiatric emergencies with limited training and access to consulting services. A recent report on ED presentations from 2006 to 2014 indicates a fourfold increase in patients with suicidal ideation and a 44% increase in psychiatric cases [6]. EM staff has had increased external demands placed on their performance as well. In response to reports of suicides during hospitalization [7] and soon after an ED visit [8], the Joint Commission has set new standards for hospitals to identify and care for the suicidal patient [7]. This expectation is giving rise to routine screening of all psychiatric patients for suicidal risk in the ED and the hospital and, therefore, potential identification of even more patients with suicidal risk factors in need of evaluation and referral.

EM physicians, in particular, want an efficient, empirically validated tool or scale for suicide screening and suicide evaluation. The need is for a screening tool with very high sensitivity and specificity that would facilitate rapid disposition without causing the admission of false positives or the discharge of false negatives. Boudreaux et al. report some success [9], but successful screening remains very much a work in progress. The Suicide Prevention and Resource Center has created a screening tool that tries to distinguish when an individual needs a psychiatric evaluation before being discharged [10]. However, screening tools rely on patients’ self-reports of suicidal ideation, and self-reports may be unreliable [11] unless elicited by a skilled evaluator [12]. In fact, people both underreport and overreport suicidal thoughts for a variety of complex reasons [13, 14].

Standardized instruments for in-depth suicide assessment do not fare much better than screening tools. A recent comprehensive literature review concluded that while these instruments may contribute something to the overall clinical picture, none have enough scientific validity to give them independent value [15].

Despite the seriousness of suicide from a public health perspective, the incidence of suicide is very low. Statistics from the Centers for Disease Control and Prevention indicate that 99.5% of people with suicidal ideation do not kill themselves [16]. From a practical standpoint, picking out the truly high-risk person can be like finding a needle in a haystack.

Suicide Risk Assessment and Management

Risk assessment is a key component of the overall psychiatric assessment. It is a repeating, four-part process that involves: (1) gathering data relevant to risk (mental status, risk factors, and protective factors); (2) synthesizing and interpreting the data; (3) intervening therapeutically; and (4) documenting the process, including one’s clinical decision-making. Screening can be thought of as a first, brief cycle of these components. Further, all evaluation schemes, such as the popular Suicide Assessment Five-Step Evaluation and Triage (SAFE-T) [17], follow this conceptual outline. The Collaborative Assessment and Management of Suicidality (CAMS) model also uses this framework and has been adapted to the emergency department setting [18].

The four component activities may and should occur simultaneously. For example, evaluations are most productive when patients collaborate, and cooperation is usually best obtained when patients are approached as an equal possibly needing some help, rather than as a “specimen of pathology being examined” [19]. There are also times that suicidal behavior is so emergent that immediate intervention is called for. In both scenarios, risk assessment and risk management are inseparable; thus, these terms are merely shorthand for a hybrid process.

Risk assessment in the emergency setting is often brief. Emergency practitioners typically truncate the psychiatric examination to focus on its most salient elements. These will vary depending on the case but must include the interview, the mental status exam, history of present illness, collateral history, past history of dangerousness, and an attempt to answer the question, Why now? Instruments that help to gather and synthesize data (e.g., empirically validated risk factors and protective factors) may be useful adjuncts to clinical judgment. The analogous field of violence risk assessment has termed the combination of the two approaches “Structured Professional Judgment” and now considers it to be best practice.

A consensus of opinion does not exist as to whether to conclude the assessment with an estimate of the level of risk. SAFE-T suggests deciding among three levels of risk: high, moderate, or low. The American Psychiatric Association Practice Guideline for Patients with Suicidal Behaviors describes four degrees of suicide risk [20]. One of us (JB) working in an established PES found that experienced emergency psychiatrists achieved more than a modicum of interrater reliability stratifying risk into five degrees: minimal-to-none, low, moderate, high, and imminent. When patients living with elevated risk are asked to rate their own risk, they often spontaneously use a scale of 1–10. Although some practitioners avoid documenting an explicit risk summary, it seems to us that a conclusion about risk is nevertheless being made and determining disposition. And if such an important opinion is being rendered, it should be stated.

Concepts and Guidelines

Suicide assessment remains both an art and a science. But emergency psychiatric practice has coalesced around a set of concepts and guidelines that is medicolegally safe and clinically sound.

  1. 1.

    Collaborate with Patient

    Collaborate with the patient and trustworthy significant others in selecting the most appropriate level of care. Consider all outpatient and subacute options. Little else but a note is needed when the parties involved reliably engage with one another and reach a shared, logical conclusion.

  2. 2.

    Assess and Stratify Risk

    1. (a)

      Perform the risk assessment and stratify short-term risk at a minimum into low, moderate, or high (cf. Kemp, SAFE-T). Key domains of information to review are support system, dangerous ideas and behaviors, engagement in treatment, clinical risk factors, and life stressors.

    2. (b)

      Consider use of rating scales and screening and assessment tools as adjuncts to clinical assessment, while acknowledging their limitations (cf. Crisis Triage Rating Scale, Bengelsdorf et al. [21]).

  3. 3.

    Treat

    1. (a)

      Triage as necessary, but also treat [22]. The goal of emergency psychiatric care is to turn an acute patient into an outpatient (adapted from Sederer) [23]. Initiate or carry out this process to its conclusion. Include targeted treatments aimed at suicidality (cf. CAMS [18]).

    2. (b)

      Repeat cycles of intervention and reassessment for suicide risk. Emergency work is an iterative process, and suicidal states are dynamic: They may improve or worsen in a short period of time. Do not assume that a risk assessment performed 8 hours ago is still valid.

    3. (c)

      For discharges, facilitate a good transition to, and ongoing partnership with, appropriately intensive and timely outpatient services. The higher risk the discharge, the better the aftercare needs to be. Attempt to restrict or limit access to lethal means until the crisis has passed.

  4. 4.

    Set a High Bar for Hospitalization

    1. (a)

      Set a high bar for hospitalization and a higher bar for involuntary hospitalization. Inpatient treatment is not for most patients with risk factors, but for some it remains a crucial endpoint on the care continuum. Consider admission for: (i) high, short-term risk for serious harm [20]; (ii) failure or inadequacy of the most intensive community-based crisis services; and (iii) new onset of severe mental illness, when risk potential, underlying diagnosis, and receptivity/responsiveness to treatment are largely unknown. (For more thoughts on when to hospitalize, see Chap. 20).

    2. (b)

      Develop an appropriately high degree of risk tolerance for low, moderate, and chronically high risk. Understanding the difference between ongoing chronic risk and acute risk is crucial. “Sometimes the acceptance of a chronic risk of suicide is the price of outpatient treatment …” [24].

  5. 5.

    Medicolegal Risk Management

    Documentation and consultation are the two pillars of medicolegal risk management [25].

    1. (a)

      When in doubt, obtain a consultation. Sometimes, a curbside consult may be sufficient.

    2. (b)

      Document one’s attention to protective factors, risk factors, risk mitigation, consults, and clinical decision-making (see the Columbia-Suicide Severity Rating Scale (C-SSRC)/SAFE-T combination form [26] or Appendix 1: Berlin-Stefan form—Brief Documentation of Release).

    3. (c)

      Foster a good doctor–patient relationship with patient and family.

  6. 6.

    Avoid Excessive Risk Tolerance

    Be on guard for excessive risk tolerance and nonadmission driven by bed shortages, cost capitation, or other factors, an overcorrection more likely to occur at present in PESs than EDs. Hospitalization should never be the default, but is sometimes the most appropriate choice.

Clinical Correlations

The greatest pitfall seen in ED practice today is excessive fear of considering discharge for a patient with suicide risk factors, even when the risk is remote and mitigating measures are readily available.

Case Example 1: Low Risk

Mr. A. was a schoolteacher brought into the ED by his wife for a citalopram refill. She was concerned about his being low-key at a party that evening. He confessed to making the mistake of going off his antidepressant. Neither of them regarded the situation as an emergency, but his psychiatrist was on vacation, and they decided to visit a local ED for a refill.

On exam, Mr. A. presented as likable and relaxed. He acknowledged some down moods recently and admitted to a remote history of nondangerous, fleeting suicidal thoughts, but he believably denied any suicidal ideation in the past 3 years or any hospitalization. He was committed to his family and work. He had no problems with sleep, high anxiety, or emotional turmoil. He had no family history of suicide. Wife corroborated his history. The physician promised discharge and left to retrieve his prescription pad. Instead, he called the police to have Mr. A. forcibly taken to the local PES without explanation.

Discussion

The ED physician’s information gathering was excellent, but his medical decision-making was outdated. At a minimum, he should have told Mr. A. that he needed a psychiatric consult and regretted that there was only one way to obtain it. However, when used unnecessarily, coercive intervention is humiliating, stigmatizing, and wasteful of resources. It may also prevent future help-seeking. Much better would have been to call the PES psychiatrist for a telephone consult. He would have been told something like this:

This is a very low-risk case. I’m hearing two risk factors—remote history of suicidal ideation and possible early recurrence of a clinical depression—and an abundance of protective factors: good support system, absence of serious suicidal thoughts or behavior past or present, willingness to accept professional help, a history of good response to treatment, an absence of acute clinical risk factors such as severe anxiety, insomnia, or despair, and a coherent story, which the wife corroborates. Just to be on the safe side, you ought to ask about substance abuse, firearms, and major life stressors. I’ll see him if you really want me to, but bottom line: If you think he and his wife are telling the truth, I don’t need to—you can let him go. Just document this consultation and use my name. (The consultant might also have asked what was making the doctor uneasy about discharge.)

In the absence of specialty consultation, the EM physician’s use of risk or depression screening tools would have corroborated Mr. A.’s low suicide risk. By any measure, he would have scored as safe to discharge. His Crisis Triage Rating Scale score would have been a perfect 15, his Patient Health Questionnaire-9 (PHQ-9) would have put him in the minimal depression group, his Columbia Suicide Severity Rating Scale would have been 0 out of 6 positive responses indicating suicide concern [26], and his believable absence of suicidal ideation would not even warrant use of the Suicide Prevention Resource Center’s second-step screening tool [10]. Under the American Psychiatric Association’s guidelines for selecting a treatment setting for patients at risk for suicide, this man’s risk is too low even to make the low end of the chart [20].

What about the legal perspective? Haven’t individuals judged to be low risk gone on to kill themselves?

Yes, but … No one believes that clinicians can predict or always prevent suicide with any degree of certainty. They are held to a standard of care that expects them to assess and manage risk as well as possible under real-world conditions [14, 25, 27].

Liability is based on whether or not the practitioner conducts an appropriate and careful assessment of risk, which he clearly did in this case. Suicide cases are disfavored and rarely taken by malpractice lawyers [14, 27]. Lawyers require more than a bad outcome to take a case. They look for a gasp-worthy narrative of neglect or indifference to the patient’s circumstances.

Examples of malpractice-worthy cases include patients held involuntarily with literally no face-to-face evaluation at all [28,29,30]; individuals with extremely recent and lethal suicide attempts who wanted to be hospitalized being turned away [31]; strip-searching a woman in the presence of male security guards with no cause to believe she had contraband [30, 32]; soliciting psychiatric specialists’ advice and then ignoring it; failing to take into account the detailed information of credible family members; failing to read available records or to consult with readily available community treaters; and basing clinical decisions on nonclinical factors, such as insurance status.

A carefully done evaluation and thoughtful weighing of risk factors rarely bring litigation, even when the outcome is tragic. Many states have immunized the decisions of ED professionals to admit or discharge from liability if the evaluation is done professionally and according to the applicable commitment statute [30, 31, 33].

Documentation is essential in obviating medicolegal risk, but is also time-consuming. The authors developed a checklist called the Brief Documentation of Release and Mitigation of Risk (BDR) to supplement the charting on an elevated-risk patient deemed appropriate for release (or nondetention). It is a public domain tool regularly requested at conference presentations and published here for the first time.

ED clinicians can also take steps interpersonally to reduce their legal risk. In the remote likelihood of a bad outcome, patients and families can always retaliate, but fostering and maintaining a good working relationship with them reduces the risk of retaliatory malpractice lawsuits [25]. One of us (SS) has successfully represented a number of individuals who were initially only looking for an acknowledgment of error and an apology from a hospital, but who decided to sue because their complaints were met with defensive hostility. Doctors who are perceived as distant, cold, and uncaring are sued more often than those who are perceived as genuinely caring, regardless of the nature of the medical error involved [14]. One study showed a higher rate of being sued when the physician’s voice tone scores high for perceived dominance [34].

Case 2: Moderate and Chronic Risk

Ms. B. was a 45-year-old woman with a schizoaffective disorder and posttraumatic stress who presented to the ED requesting hospitalization for overwhelming thoughts of taking an overdose. She had also forgotten to take her psychotropic medication, which resulted in an increase in her self-denigrating auditory hallucinations. She had a history of suicide attempts and helpful psychiatric hospitalizations. The physician agreed with her self-assessment and wrote the order to proceed with admission.

Unfortunately, there were no open psychiatric beds anywhere in the city. Ms. B. agreed to wait in the ED overnight and take her usual medicines. However, when there were no beds available the next day, either, she requested to be released. This placed her EM physician in a quandary. There was no mental health consultation available, and he wondered if Ms. B. should be placed on a mental health hold and detained until a bed opened up, or would this only make her worse? Luckily, he decided to reevaluate the patient and obtain a collateral history.

Ms. B. readily engaged in conversation and smiled. Her mood was somewhat depressed but not hopeless. She believed that restarting her medication and staying in a safe place overnight had helped. On a scale of 1–10, the usual degree of suicidal ideation that she lived with on an outpatient basis was in the 3–7 range. It had increased to a 9 when she sought admission yesterday and was now back down to a 5. When asked, Ms. B. also disclosed the acute precipitant for her current trouble: Her stepfather who had molested her as a child had just been released from prison, triggering flashbacks and dissociation.

The doctor still wondered whether she might be downplaying her suicidal thinking in order to be released. She gladly gave him her case manager’s cell phone number. The case manager corroborated all of Ms. B.’s history, noting that she was a reliable historian and rarely minimized symptoms. If anything, Ms. B. was too quick to retreat to the safety of the hospital. She lived in supported housing for people with mental illness. She had a psychiatrist, nurse, case manager, therapist, and peer support specialist.

The doctor’s reevaluation was that an acute exacerbation of an ongoing illness and increased risk had subsided. Ms. B. was now subacute and no longer a high, short-term risk for serious harm to self or others. She had a good support system, no imminent suicidal thoughts or behaviors, engagement in treatment, partial relief of her most acute symptoms, and the ability to talk about her life stressors. Her deciding against hospitalization was positive, and involuntary treatment was contraindicated. She was discharged back to her group home and the care of her assertive community treatment (ACT) team.

Discussion

Discharging a chronically suicidal patient with moderate risk to return to treatment in the community was a new concept to the EM physician in Case 2, but there are classic writings on the subject [3, 35]. In his own mind, this doctor compared it to releasing patients with hard-to-control diabetes or hypertension, and high but nonemergent blood sugars or blood pressures.

A number of articles from different disciplines are suggesting that treatment in the ED is a crucial component of ED response to individuals with suicidality [36], and recent developments suggest that it may be legally required under the Emergency Medical Treatment and Labor Act (EMTALA) for psychiatric patients who are seriously suicidal [37].

Most of the concepts and guidelines above informed Ms. B.’s clinical management. There is one additional practice tip from a legal standpoint: Documentation should not only attempt to justify the discharge, it should also clearly delineate the risk factors—the triggering release of the stepfather, the suicidal ideation, and past hospitalizations—and the ways Ms. B’s wraparound supports would mitigate those risks. Documentation of the decision-making process, based on the knowledge reasonably available to the ED professional at the time, including consultation with the case manager who knew Ms. B well, serves as a protection from liability in the event that, unbeknownst to the ED staff, those community supports somehow fail the patient.

Exaggerated or Feigned Risk

There are other clinical scenarios of suicide risk that can safely proceed from an ED or PES to nonhospital management. For example, EM practitioners and psychiatric trainees are often challenged by the individual who exaggerates or feigns suicidal risk in order to obtain hospitalization [13, 38]. (See Chap. 14, “Malingering and Factitious Disorders in the Emergency Department.”) Briefly, we note that for patients not well known to the emergency service, an extended stay in the emergency setting may permit a more definitive assessment. Given time, patients may confide in a staff member and become less contentious when approached in a consistent, nonpunitive, therapeutic manner. In addition to attempts at engagement and identifying something to treat, consultation, referrals, risk tolerance, and documentation of clinical decision-making are all key. In the final analysis, it is perfectly acceptable to discharge a person threatening suicide if one’s careful assessment is that, based on all available information, the threat is not credible. However, practitioners should be prepared to contain explosive reactions in the malingering patient and negative feelings in themselves (e.g., indignation, fear, and self-doubt) at the point that the patient’s request for admission is denied.

Conclusion

At present, the Joint Commission goal of preventing suicides through better screening in the emergency setting [7] may be more aspirational than realistic. Moreover, without adequate preparation, this goal might inadvertently encourage reflexive, counterproductive treatments such as unwarranted involuntary hospitalization. However, there are multiple more realistic opportunities for improving care. Current trends in psychiatry and jurisprudence are guiding us toward positive, effective, less restrictive approaches for patients with suicide risk. In situations when patients cannot believably describe their own risk—unlike Mr. A. and Ms. B. who could—this chapter hopefully provides a framework for nuanced, evidence-based management of suicide risk in contemporary emergency settings.