27: Psychiatric Principles

Psychiatric problems may be the cause or the effect of many toxicologic ­presentations. Suicide attempts and aggressive behaviors are commonly associated with toxicity and can be uniquely difficult to assess and manage. Patient behaviors are often viewed dichotomously as either totally intentional and deliberate or totally “out of control” and irrational. The truth is usually more complex, with some aspects occurring within the awareness and control of the patient and other aspects either unknown or overwhelming to the patient. Neuropsychological conceptions of “behavioral disinhibition”­ as a baseline personality trait that is relatively fixed are complementary to presumed organic etiologies of disinhibition such as intoxication or brain injury; in the latter cases, frontal lobe dysfunction is directly implicated. Yet more subtle frontal lobe dysfunction is likely occurring in many patients without such a direct neurological insult: patients with psychosis or mania, personality disorders, attention-deficit/hyperactivity disorder, or even agitation in the context of psychological distress. A patient’s innate capacity for self-restraint can vary widely, and some will present with behavioral disturbances much more readily than others.¹¹²

Issues of capacity often arise with intoxicated or impaired patients, and the basic approach to capacity assessment will be covered. “Medical ­clearance” is usually requested prior to a patient’s transfer to psychiatric care, yet this has proven to be a rather vague concept that lacks a standard protocol; however, there are practice guidelines addressing this issue, which will be discussed. This chapter also explores the topics of suicide/self-harm and violence to enable the physician to adopt the appropriate role of both diagnostician and medical decision-maker. Finally, substance use disorders (SUDs) will be addressed, given that they are problematic in the emergency setting and highly comorbid with suicidality and violence.

Physicians are legally and ethically required to obtain informed consent before treating a patient to allow the patient to make a voluntary choice about his or her care. However, if a patient is cognitively impaired, the physician must make a careful assessment of capacity and proceed as medically necessary should the patient be found to lack capacity. It does not matter whether the cognitive impairment is due to psychiatric illness, medical illness, or a transient xenobiotic-induced delirium; if the patient is unable to make a logical and rational decision stemming from his or her personal value system, the physician can override the patient’s autonomy to provide necessary care.⁷ Traditionally, the term “competence” denotes a legal status determined by a court of law, whereas “capacity” is assessed by a physician in a clinical setting. The distinction between these terms has become blurred in both the medical and legal literature, but it is a helpful framework nonetheless.⁷

It is important to note that capacity assessments occur at a specific moment in time regarding a specific medical decision. The assessment is not a “global” determination that persists throughout the hospitalization or even the length of an emergency department (ED) visit, but rather a temporary determination that requires reassessment as factors change. Capacity may fluctuate with a change in the patient’s mental status, medical or psychiatric condition, the time of day, amount of pain, level of anxiety, perceived support, recent medication administration or in the context of withdrawal from a xenobiotic.

While disease status cannot be used to determine capacity, some correlations have been noted in the literature. Fifty percent of patients with schizophrenia who are hospitalized with an acute psychotic episode will not have capacity; more than 50% of patients with mild-to-moderate dementia will also lack capacity, and this increases to 100% with severe dementia. If the Mini-Mental State Exam (MMSE) (Table 27–1) score is lower than 19 out of 30 possible points, it is highly unlikely the patient will have capacity.⁵¹ Most patients with mild-to-moderate depression do not have impaired capacity, yet 20% to 25% of hospitalized patients admitted with severe depression do lack capacity.⁷ Nonetheless, capacity should always be carefully assessed, since active psychiatric symptoms or long-standing cognitive impairment do not invariably lead to a lack of capacity.

Depriving a patient of his decision-making rights is a serious infringement on liberty. While physicians are implicitly assessing patients’ capacity on a regular basis, they often have little formal training and may not have a systematic approach. This limitation leads to difficulty determining capacity in murky ethical situations, confusing clinical scenarios or when a patient is making “bad” decisions (such as refusing treatment of clear benefit or refusing life-sustaining measures). Patients refusing treatment in the setting of a recent suicide attempt, such as refusing an antidote for a potentially lethal ingestion, is one frequently encountered scenario (see Suicide).

Intricacies of capacity assessment may be dictated by individual jurisdiction and it is important to know the laws in the area within which you practice. For example, some states require two separate assessments by physicians that are separated in time. The assessment should be done by a physician familiar with the procedure and able to discuss risks and benefits, and should include assessment of the patient’s mental status, educational level, and any cultural factors that may be relevant. A well-researched scale for determining decisional capacity is the MacArthur Competence Assessment Tool for Treatment (MacCAT-T), a semi-structured interview that requires 15 to 20 minutes to administer and is designed for use in a clinical setting (Table 27–2).⁵⁹

The general legal standard for capacity requires assessment of four basic skills, which should be clearly documented:⁷

1. Communicate a Choice.

The patient must be able to indicate a choice, and to be consistent long enough for the choice to be implemented. A patient may alter the decision over time or with new information, but repeated reversals of a decision suggest a lack of capacity. Difficulties in this first measure can occur in patients who have a high degree of ambivalence, patients with poor short-term memory or a thought disorder, or patients who are unconscious or delirious.

2. Understand Relevant Information.

As a first step, the physician should provide all the relevant information in a patient’s primary language and in an understandable manner. In addition to describing the medical condition, the proposed intervention(s) and the risks and benefits, the physician should be certain to address the possibility of “doing nothing” and the associated risks/benefits with this decision.

Physicians often feel that this step is the key to capacity assessment, and that if a patient can verbally repeat the details of the procedure and the risks/benefits/alternatives that capacity has been established. However, the assessment is more nuanced than this. The patient should be able to understand and manipulate the information independently, which can be demonstrated by asking the patient to paraphrase the physician’s explanations rather than repeat verbatim. The patient should understand that probabilities, not certainties, are being discussed, and that he or she has an autonomous role as a ­decision-maker. Difficulties in this domain can arise from deficits in attention, intelligence, or memory.

3. Appreciate the Situation and Its Consequences.

The patient may be able to comprehend the relevant information, yet it is important to ensure that the patient is also able to apply the facts to the individual’s particular situation. The medical condition in question must be acknowledged, and the individual must demonstrate awareness that the risk and benefits being discussed apply to this unique personal situation. Courts have clarified that patients who do not acknowledge their illness do not have capacity. Difficulties in this area often arise from pathologic denial, mistrust of the motivation of the physician, cognitive or affective impairment, or delusional thinking.

4. Reason about Treatment Options.

This final step describes the patient’s ability to comprehend and incorporate the relevant information as applied to this particular situation in order to logically reach a decision about his or her treatment. The physician should ask the patient to compare treatment options and consequences, and should inquire about how the choice was selected; the patient should be able to explain the reasoning involved and discuss the major factors considered. If a patient is unable or unwilling to explain the personal reasoning process, as sometimes can occur in the setting of anger or irritability, the capacity assessment is not complete and this suggests a lack of capacity until proven otherwise. Patients are allowed to make “unreasonable” choices, but they should be able to explain their reasoning process. Difficulties in this domain often occur in patients with delirium, dementia, extreme phobia, panic, anxiety, psychosis, depression, or anger. Patients with SUDs might be experiencing withdrawal or intense craving and therefore demanding to leave and unwilling to ­participate in a discussion about their reasoning; addressing their ­withdrawal or cravings directly can often result in significant relief and rapid re-establishment of capacity.

Once a patient is deemed not to have capacity at a specific time to make a specific decision, the evaluation should proceed to determine the reason for this lack of capacity and the route through which it might be restored. Every effort should be made to assist in the restoration of patient capacity in order to protect autonomy. Individual state laws may mandate when court proceedings should be pursued, often depending on the level of invasiveness of a procedure: For example, a venipuncture is much less invasive than a surgical procedure, and would not require a special court order. However, for a procedure such as a limb amputation, the physician should likely pursue a court decision. In unclear circumstances, the physician can consult with the hospital’s risk management department. Every effort to find an alternative decision-maker should be made. When there is no designated health care proxy and family members are not in agreement about the best treatment choice, most states have determined a specific hierarchy to determine which family member should make the decision.⁸

Complexities of Capacity Determination in Clinical Settings

Decisional capacity is often part of a larger question about how to proceed with treatment for a particular patient. If a patient is found to lack capacity, ethical questions arise about the appropriateness and practicality of imposing treatment; such as the consideration of sedating a patient to obtain a blood sample. Sedating a patient has its own inherent risks, and psychological distress to the patient cannot be discounted. In these situations it is often helpful to have a discussion involving the psychiatrist, and a general consensus can usually be reached about the balance of costs and benefits to proceeding with treatment against a patient’s will. Life-threatening refusal of treatment is always clear, but the physician often quickly encounters a “gray area” in which the medical necessity of a treatment is less urgent. The physician has a responsibility to petition the court for invasive procedures such as surgery if the patient’s medical condition is stable enough to allow this, but if the situation becomes urgent, the physician can proceed without involving the court. In these cases, it is often helpful to have a team approach involving psychiatry, risk management, and a bioethics consultation if available.

Another important concept is the sliding scale of capacity, which has been supported by both expert consensus guidelines and the legal ­system. This “sliding scale” concept indicates that the threshold for deeming a patient to lack capacity changes depending upon the seriousness of the consequences. For example, a patient refusing treatment in a life- or limb-threatening situation would be held to a higher standard of capacity than for a patient refusing a venipuncture. A life-threatening situation requires the patient to clearly demonstrate a thorough understanding of the relevant information, its application to the situation and the individual’s reasoning process in order to be deemed to have capacity. On the other hand, even if a patient doesn’t technically have capacity to refuse a blood draw, if it is not medically necessary at that moment, the patient might be allowed to refuse. This philosophy stems in part from our conception of a modern democracy in the United States, in which the vast majority of persons are considered capable to make their own decisions. We would like to preserve this autonomy when it is not dangerous to the patient’s health.⁷

Not all patients who present with psychiatric symptoms have mental illness. Intoxication, withdrawal syndromes, medical illness, and organic brain disease can often mimic mental illness. Patients with behavioral disturbances present unique challenges to diagnostic assessment given that they are unlikely to be cooperative and it can be difficult to achieve a comprehensive medical history. Emergency physicians are usually asked to “medically clear” the patient before transfer to psychiatric care, yet there is no standard process for this.

In general, psychiatrists requests that ED physicians: (1) determine if presenting symptoms are caused or exacerbated by a medical illness, (2) assess and treat any acute medical issue necessitating urgent or emergent intervention, and (3) determine if the patients symptoms are primarily related to a xenobiotic, which would preclude an accurate psychiatric evaluation.⁸⁰ Obvious signs of need for further evaluation include abnormal vital signs, delirium, altered cognition, or an abnormal physical examination. Four other groups are identified as high risk for medical instability: the elderly, patients with substance abuse, patients without any prior psychiatric history, and those with preexisting medical disorders or current medical complaints.⁵⁸

While routine laboratory testing is often requested by the psychiatrist for medical clearance, a number of studies demonstrate that selective testing based on history and physical examination is probably the correct strategy. There is a strong consensus among ED physicians that routine laboratory testing is unnecessary, and without any clinical suspicion, the probability of false positive laboratory results begins to outweigh true positives.^16,42 The American College of Emergency Physicians (ACEP) provides a level B recommendation that diagnostic evaluation be directed by the history and physical examination, and that routine laboratory testing is low yield and not necessary as part of the ED assessment.⁸⁰

ACEP also provides a level C recommendation against routine urine toxicology testing and blood alcohol concentrations in alert, awake and cooperative patients, and specifies that transfer to psychiatric care should not be delayed to await collection of samples for toxicologic analysis.⁸⁰­However, from the psychiatrist’s perspective, this testing is time-sensitive and can change psychiatric management and disposition considerably. If the emergency physician knows the patient will be transferred for psychiatric care and suspects substance abuse, it would be helpful to obtain toxicologic results as early as possible in the hospital course, but also can be done after transfer if necessary.

A final issue is if the patient is intoxicated with alcohol. At what point is it appropriate to “medically clear” the patient for a psychiatric evaluation? There are no evidence-based data to support that patients regain decision-making capacity at a particular blood alcohol concentration. Depending upon tolerance, cognition can vary widely, and patients often develop ethanol withdrawal while consequential blood alcohol concentrations persist.³⁹ Therefore, ACEP provides a level C recommendation that the patient’s cognitive abilities should be the basis on which the psychiatric assessment is begun, and a period of observation can be used to see if psychiatric symptoms resolve with intoxication and a psychiatric assessment is no longer needed.⁸⁰

Suicide was the tenth leading cause of death in the United States, accounting for 38,364 deaths in 2010.²⁸ It was the third leading cause of death among 15 to 24 year olds, and accounts for 20% of deaths in this age group annually. Although suicide may be attempted or accomplished in a variety of settings and by a variety of means, it is most typically associated with psychiatric ­disorders and/or substance abuse, especially ethanol, and is frequently accomplished with psychoactive xenobiotics alone or in combination.

Terminology

The terms and definitions of suicide and suicidal behaviors used in this chapter have been outlined in several reports on the subject. The term ­suicide refers to self-inflicted death with either explicit or implicit evidence that the person intended to die. The term suicidal ideation refers to thoughts of serving as the agent of one’s own death, and suicidal ideation may vary in seriousness depending on the degree of suicidal intent and the specificity of suicide plans. Suicidal intent refers to the subjective expectation and desire for a self-destructive act to end in death, and suicidal plan refers to the ­formulation of a specific method through which one intends to die.

A suicide attempt is a self-injurious behavior with a nonfatal outcome accompanied by either explicit or implicit evidence that the person intended to die. The related term suicide gesture generally refers to an attempt which is not felt to be medically serious, and often implies that the patient is making a “cry for help” rather than truly attempting to die; such a gesture can still result in death, even if unintentional. As defined in the literature, “serious” suicide attempts meet one of the following criteria: (1) treatment in a specialized unit (such as an intensive care unit), (2) surgery under general anesthesia, (3) medical treatment beyond gastric lavage, activated charcoal, or routine neurologic observation, (4) method with high risk of fatality (such as hanging or firearm), or (5) hospital admission longer than 24 hours.¹²

Self-harm or self-injurious behavior refers to the self-infliction of painful, destructive, or injurious acts without the intent to die. An example is the superficial cutting that occurs in a patient with a borderline personality disorder, which is actually a coping mechanism that provides psychological relief; one hypothesis is that relief is mediated by the endogenous release of opioids.⁹⁹ However, the distinction of suicidal and “parasuicidal” behaviors such as these is outside of the scope of the medical toxicologist, and a mental health professional should be consulted to help make such a distinction.

Finally, the lethality of suicidal behavior is an objective measure of the “danger to life” associated with self-injurious behavior. It is important to note that objective lethality may be distinct from subjective lethality. This occurs when a patient does not truly expect a behavior to be as medically dangerous as it turns out to be, or vice versa. For example, a patient may not be aware that an acetaminophen (APAP) overdose can be fatal, believing that its status as a “nonprescription” medication indicates that it is less likely to be fatal in overdose than a prescribed medication.

Epidemiology

Suicidal thoughts and behaviors are common. An estimated 8.3 million adults in the United States reported having suicidal thoughts in 2008, and approximately 1 million actually made an attempt that year.³⁵ Studies have consistently found that self-poisoning accounts for more than 70% of all serious suicide attempts.⁸³ While there is reportedly one suicide completion for every 25 attempts, this ratio dramatically decreases for adolescents and young adults. Among 15 to 24 year-olds, there are approximately 100 to 200 attempts for every completed suicide.⁴³

Ninety percent of suicide completers and attempters meet criteria for a diagnosable psychiatric illness, 70% have mood disorders, and 30% have alcohol abuse or dependence.⁶⁴ On average, one-third of suicide decedents test positive for alcohol, and 65% of them have a blood alcohol concentration higher than 80 mg/dL.⁷³ Many individuals who complete suicide do so on the first attempt. Although a prior suicide attempt is still the strongest predictor of a future completed suicide, only about one in 10 attempters is successful.^62,88 Multiple attempters also appear to have higher risk than those who make a single attempt.^57,103 Medically serious attempts, defined as involving treatment in a specialized unit, surgery, or medical treatment beyond stomach evacuation, may be a more robust marker of risk. Those who make ­serious attempts tend to share with completers a higher rate of serious mental illness such as schizophrenia, bipolar disorder, or severe major depressive disorder.¹²

Suicidal or self-injurious patients are often encountered in EDs. Approximately 70% of all self-inflicted, nonfatal injuries seen in an ED are the result of a suicide attempt, and an ED visit for self-harm is predictive of a completed suicide in the future.^23,94 Anywhere from 1% to 2.7% will commit suicide within the next year or two.^65,94 The average annual number of ED visits for attempted suicide and self-injury has nearly doubled over the last 15 years: from 244,000 in the early 1990s to 538,000 in the mid-2000s.¹⁰⁷ Self-poisoning was the most common method at 68%, followed by cutting or piercing at 20%.⁴⁴ Increases occurred in all major demographic groups: males, females, Caucasians and African Americans. ED visits were most common among those aged 15 to 19 years.¹⁰⁷

Although the prevalence of suicide and suicidal behaviors shows significant overlap, one notable area of difference is the consistent pattern of higher rates of suicide attempts among women and of completed suicides among men.^22,26 Men account for 80% of completed U.S. suicides at a rate highest among men greater than 75 years of age. Men are more likely to use a firearm, at rate of 57% versus 34% in women; some authors suggest this difference in method leads to the disparity in completed suicide rates as compared with women, who are more likely to self-poison.^28,73

Self-Poisonings

According to the National Electronic Injury Surveillance System, which surveys a select number of nationally representative hospitals, an estimated 170,000 people were treated in ED for poisonings in 2000. Sixty percent of these were classified as “probable” suicide attempts, 10% were “possible” suicide attempts, and the intent for the remaining 30% was “unclear/unknown.” Women were more likely to self-poison than men (72% vs 55%), and the highest rate of self-inflicted injury occurred in adolescents aged 15 to 19 years (especially women), followed by young adults aged 20 to 24 years. Ethnically, rates were highest among non-Hispanic Caucasians of both sexes.²³

In the 16 states participating in the National Violent Death Reporting System in 2009, self-poisoning was the third leading method of suicide (following firearms and hanging/suffocation).⁷² Detailed data from 2005 to 2007 indicates that 75% of individuals poisoned themselves with a substance (either illicit or licit, including alcohol) as opposed to other types of poison such as carbon monoxide. Seventy percent of individuals ingested just one “drug type,” defined as prescription drugs versus nonprescription drugs versus alcohol/illicit drugs. Out of poisoning deaths due to a single “drug type,” 80% were due to prescription drugs such as opioids, benzodiazepines, or antidepressants, and 10% were due to nonprescription drugs such as APAP. When individuals consumed more than one “drug type,” 45% involved alcohol or illicit drugs as part of the combination.^25,28,72 It was estimated that for every self-poisoning death in the United States, there were 33 intentional overdoses reported to poison centers (Chap. 136), of which two ultimately went untreated.²⁹

There is a large body of literature about self-poisoning in the United Kingdom, and their data indicate proportionally higher death rates among poisonings with tricyclic antidepressants (TCAs) versus selective serotonin reuptake inhibitors (SSRIs), which is not surprising given the cardiac toxicity of TCAs. Venlafaxine, a serotonin-norepinephrine reuptake inhibitor (SNRI), is more toxic in overdose than SSRIs. Poisoning deaths with atypical antipsychotics such as clozapine, olanzapine, and quetiapine were increasing, especially as of the early 2000s.⁴⁹

Given that the act of suicide is a statistically rare event in the overall population, it is virtually impossible to predict who will actually commit suicide. Therefore, it is critical to identify risk factors that increase the likelihood that any individual might attempt suicide, and to identify risk factors that are modifiable (Tables 27–3 and 27–4). The identification of modifiable risk factors provides opportunities for interventions that may decrease suicide risk. Additionally, there are protective factors that mitigate the risk for suicide, and it is important to assess for the presence or absence of these factors in determining the overall risk for suicide in a patient.

Table 27–5 depicts a case of suspected self-poisoning from the starting point of prehospital care through the completion of a comprehensive assessment and treatment planning.

Initial Psychiatric Management

In any case of suspected self-poisoning, a thorough psychiatric assessment is warranted. It can be helpful to call a psychiatric consult immediately after medical stabilization, even if the patient is unable to communicate, since important information such as pill bottles, ambulance reports, and the ability to call the patient’s family and outside providers may be lost if the consult is delayed. Although not required for an ED assessment unless clinically indicated, it should be helpful for the psychiatrist if urine toxicology and blood alcohol concentration is obtained as early as possible.

An early, focused assessment is necessary to ascertain elopement risk and decisional capacity. Subacute residual central nervous system (CNS) effects of ingestions, such as confusion, fatigue, and fear, can dispose patients to wander or elope. The patient should be searched for weapons, pills, and other potentially toxic materials to prevent additional self-harm or ingestion in the hospital. Given that the patient’s intentions remain unclear at this point, the question of unintentional versus intentional exposure to a xenobiotic cannot be completely resolved. For this reason, a high level of supervision should be maintained, and a patient should not be allowed to leave or be left unattended until an adequate assessment of the patient’s mental status is completed. Depending on the architecture and organization of the ED and its personnel, it may be sufficient to place the patient in an open area in the direct line of sight of the medical staff. If such an arrangement is not possible, or if the patient is agitated and disruptive, it may be necessary to separate the patient from the general population. Under these circumstances, an individual aide should be assigned to observe the patient on a one-to-one basis. Safe physical and/or chemical restraints may be necessary to prevent further injury to both the patient and the staff.

At a relatively early point in the patient’s course, when the patient can be cooperative, a more detailed psychiatric assessment is critical to address specific clinical concerns. Both history-taking and collateral contacts can help to establish the patient’s preingestion mental status and baseline, as it may be difficult in an initial, time-limited assessment to differentiate a ­xenobiotic-induced delirium from mental illness and therefore ensure that any delirium has fully resolved. The determination that the patient is stable is not solely established on the basis of blood concentrations of a xenobiotic or ancillary medical tests, but rather when the emergency physician or medical toxicologist with an understanding of toxicokinetics and toxicodynamics deems it appropriate.⁶⁶ Although a thorough psychiatric examination is not possible until an altered mental status has cleared, it is reasonable to have the psychiatrist involved earlier for the above-mentioned reasons.

The physician should not unequivocally attribute altered mental status to poisoning or toxicity until the signs of altered consciousness have resolved and cognitive functions have returned to normal. Until that time, other toxic-metabolic and structural conditions that might coexist with, or masquerade as, toxicity cannot be excluded. If the patient’s cognitive functioning is impaired by xenobiotics, then critical historical details may be unreliable.³³ It should be understood that much of what the patient reports may be ephemeral, caused by the predictable temporary and reversible effects on mood of these xenobiotics.¹⁴ However, any patient utterances about intended overdose or self-harm should be documented in the record, since it is important information to consider if the patient becomes guarded and evasive after returning to baseline mental status.

In addition to routine laboratory studies, such as complete blood count, electrolytes, renal function, and liver function tests, psychiatrists usually request thyroid function tests, B12, folate, HIV testing, and syphilis immunoglobulin G antibody or RPR; abnormalities in these laboratory values can be associated with medical disorders producing psychiatric manifestations that would affect management and disposition. However, these laboratory tests have proven to be low yield when obtained routinely and without clinical suspicion, and are not necessary as part of the ED work-up.⁸⁰

Special Issues of Capacity in Suspected Self-Poisonings.

Patients may request to be discharged, refuse care, or become aggressive. Aggression may arise from lingering effects of ingestion or withdrawal, severe anxiety, fear, anger at the loss of autonomy, or the discomfort associated with unpleasant procedures.⁸⁶ Although patients may respond to verbal limit-setting and repeated explanations of their care, they may also require pharmacologic or physical restraint and involuntary treatment. Patients are not allowed to make poor health care decisions if their ability to weigh the risks and benefits of the proposed care is limited by cognitive deficits or mental illness. In the setting of toxicity, appropriate care may be provided under the doctrine of implied consent.

The emergency exception to the doctrine of informed consent may also apply in circumstances where self-injury is suspected. The emergency exception permits forcible detention, restraint, medication over objection, and necessary medical care until psychiatric assessment can be accomplished. This includes collecting information from collateral sources without the patient’s consent. After the management of the immediate medical emergency and resolution of toxicity, suspected self-injury is sufficient evidence of impaired decisional capacity for the emergency physician to hold a patient for further psychiatric assessment. The emergency physician should document the patient’s objections in the medical record and indicate the basis for the determination of diminished capacity.

After the intentionally self-poisoned patient is stabilized, there may be a need for a more thorough assessment of decisional capacity. Psychiatric consultation is appropriate at this stage to help document the degree of impairment, determine the etiology, and predict the likely course.

Immediate Risk of Self-Harm.

After these safety considerations are addressed, the aim of the focused psychiatric assessment moves toward a determination of immediate suicide risk. This examination should answer the following questions: What is the patient’s attitude toward lifesaving care? What are the patient’s current wishes with regard to living or dying? What are the patient’s thoughts about his or her rescue and likely recovery?

These questions can only be answered in the course of a frank discussion between the patient and the emergency physician. The physician should not be concerned about “provoking” further self-injurious impulses by having this vital discussion; many patients will be relieved that the health care professional is speaking directly about their distress.

Reliability and Confidentiality.

Mention should be made here about the difficult issues of reliability and confidentiality with regard to gathering history. Evasiveness, lack of detail, inconsistency, and improbability may lead to an unreliable history. It is appropriate to confront the patient with the implausible aspects of this history and offer an opportunity for the patient to rethink the history. This is often successful, although subsequent reports are, of course, equally suspect.

The most important step from the standpoint of both clinical care and risk management is to locate other sources of information to clarify the patient’s situation. A careful review of any previous medical and psychiatric records is critical. Any pattern to a patient’s presentations such as increasing frequency, more aggravated behavior, or disheveled appearance should be noted.

Collateral contacts are another important source of information, although the level of involvement, sophistication, and reliability of the collateral contacts must also be taken into account. In the interest of providing necessary medical and psychiatric care for a patient in an emergent situation, the ED staff is legally permitted to solicit information from collateral contacts without the patient’s consent. However, an effort should be made to obtain consent for any broader discussion of the patient’s situation with family, friends, or other physicians. The patient may express concern about the ED staff contacting a family member or counselor. Any information to be imparted to third parties can be discussed in advance with the patient. The patient may restrict consent to receiving information only and may withhold consent to impart certain information. More caution is indicated in contacting an employer. Although disclosing information about the patient without the patient’s consent is a breach of confidentiality, a physician may do so in the interest of protecting the patient (Chap. 141).⁵

Comprehensive Psychiatric Assessment

The goal of comprehensive psychiatric assessment is to: (1) characterize the nature of the attempt and any ongoing suicidal ideation that might be present, (2) explore risk factors for another suicide attempt, and (3) formulate a diagnostic impression. These three elements help to determine the level of risk and guide immediate treatment and disposition planning.¹¹⁰

The best understanding of suicide at this time is that it results from intrinsic vulnerability factors interacting with external circumstances, which can be termed the “stress vulnerability.” Intrinsic vulnerability may be conferred by a variety of traits such as impulsivity or conditions such as depression, anxiety, low self-esteem, low self-efficacy, loneliness, and hopelessness. External factors include stressful life events, access to lethal means, and a host of other factors, positive and negative. Poor interpersonal problem-solving skills and a perceived lack of problem-solving ability also appear to increase risk.⁴⁰

Assessing Suicide Attempts and Suicidal Ideation.

The core of the suicide risk assessment is a detailed discussion of the patient’s ­suicide attempt and any ongoing suicidal thoughts and urges. It is important to establish rapport and introduce these topics in an appropriate context in order to improve the patient’s candor. This evaluation requires significant time and skillful interviewing, for which there is no substitute. This approach will enhance both the therapeutic quality of the interview and its reliability.

The clinician should explore the exact details of the attempt, including precipitating factors that may have begun days or weeks prior to the actual act. It is critical to determine the level of actual and intended lethality, along with the seriousness of the intent. Why did the patient make an attempt on that day or time? Understanding the patient’s thought process can help to gauge the extent of impulsivity versus planning involved in the attempt. Signs of premeditation and planning are concerning, such as organizing one’s affairs (giving away possessions, ensuring a will is updated) or writing goodbye letters (such as a phone call, internet posting or text message). It is crucial to determine if the patient expected to be found, and if any effort was made to notify someone about the impending attempt such as a phone call, internet posting, or text message. How was the patient discovered, and by whom? A patient who overdoses alone in a hotel room is very different from someone who overdoses in the bedroom while the family is at home. Current feelings about surviving should also be assessed: Is the patient relieved or upset to be alive? Is there currently active or passive suicidal ideation? How forthcoming does the patient appear to be when discussing this? Other important information includes prior suicide attempts and their lethality and circumstances, the frequency and duration of suicidal ideation in the past, prior psychiatric treatment, prior and current medication trials, and a detailed substance abuse history. Psychological factors such as reactivity to positive and negative external events and subjective distress are also important to explore. The social history should focus on interpersonal conflict, stressors within romantic or family relationships, and employment or financial concerns. Current support systems or lack thereof are important to note, as are feelings of isolation and abandonment, which can all be contributing factors.⁶⁷

The communication of suicidal ideas either directly or indirectly should not be misconstrued as a “cry for help” and hence evidence of lower risk. Communication is probably related to the degree of preoccupation with morbid thoughts and to personality characteristics that dispose individuals to revealing their thoughts to various degrees.⁷⁴ In psychological autopsy studies, approximately 50% to 70% of those who completed suicides gave some warning of their intention, and 30% to 40% disclosed a direct and specific intent to kill themselves.^10,91

Risk Factors for Suicide

The goal of assessing suicide risk factors is to identify factors that may increase or decrease the level of suicide risk in a particular patient, which will enable the clinician to develop a plan that addresses the modifiable factors. For example, hopelessness is one such modifiable risk factor, which would likely improve with time and treatment; when risk factors such as hopelessness can be modified on the inpatient psychiatric ward, an inpatient admission is warranted to help decrease the risk of suicide. Unfortunately, to date no study has ever identified one specific risk factor or set of risk factors as specifically predictive of suicide or other suicidal behavior; therefore, the assessment is ultimately based on clinical judgment. Despite the lack of such predictive factors for suicide, there is a large body of evidence on the multiple risk factors that contribute to suicide risk, and a growing body of ­evidence on protective factors (Tables 27–3 and 27–4). Knowledge of this evidence is critical to informing the clinical determination of suicide risk.

Risk factors are additive, with suicide risk increasing with the number of risk factors that are present, but certain risk factors interact synergistically to increase suicide risk. The combined risk of concomitant depression and alcohol intoxication may be greater than the sum of the risk associated with each in isolation. Certain risk factors, such as a recent suicide attempt associated with a high degree of lethality or the presence of a suicide note, should be considered serious on its own, regardless of whether other risk factors are present.³⁴

A number of avenues of inquiry suggest that violent suicide attempts are associated with a persistent deficiency in brain serotonin concentrations. Impulsive types of aggression and impulsive suicidal behavior have been linked to serotonergic dysfunction in prefrontal cortical regions of the brain.³⁶ This deficiency has been measured in the postmortem brains and ­spinal fluid of suicide victims and survivors of violent attempts as compared to nonviolent attempts and to other patients. Hopelessness has also received a significant amount of study as a potential predictor of suicide; unfortunately, it appears to have a high sensitivity but a low specificity.¹³ However, identifying hopelessness does provide for an intervention.

While much is known about the risk of suicide for various groups over time, little can be said with certainty about an individual patient at a particular point in time. The risk of suicide increases 50 to 100 times within the first 12 months after an episode of self-harm as compared to the general population risk. About one-half of all people who commit suicide have a history of self-harm, and this increases to 60% in juveniles.⁶ Unfortunately, there is no “typical” suicidal patient or clinically useful test or rating scale at this time. Albeit, while one investigator was able to prospectively identify almost all of those who ultimately died by suicide (97% sensitivity), the investigator overpredicted suicide by almost one-half (56% specificity).⁸⁸ However, there is also no patient in distress for whom the risk of suicide is so remote that it need not be considered.

Ultimately, most persons belonging to a high-risk group do not commit suicide, and some individuals with no apparent risk factors do. Many risk factors are not modifiable. This type of information, then, weighs most heavily in the assessment in the absence of other more specific data, early in the hospital course, or in the case of the uncooperative or hostile patient. The best foundation for treatment planning and clinical decision making is direct examination and clinical diagnosis.⁴³

Psychiatric Illness and Suicide.

One major consideration in suicide risk assessment is the occurrence of severe mental illness. Suicide risk for individuals with severe mental illness is 20 to 40 times higher than it is for the general population.⁷⁵ Psychological autopsy studies, which focus historically on the decedent’s intentions and mental state prior to death, have consistently revealed major psychiatric illness to be a factor in suicide and present in 93% of adult suicides.^68,75,89,92 This is also true of those who make medically serious suicide attempts.^13,68 In particular, prospective cohort studies and retrospective case control investigations have revealed clinical depression and bipolar disorder to dramatically increase suicide risk.^60,74,109 For mood disorders, factors correlated with current suicidality include current depression, severe anxiety, anhedonia, panic, insomnia, ambivalence, and acute alcohol abuse.⁷⁴

After mood disorders, chronic alcoholism is the most commonly reported disorder and is present in approximately 20% of cases. Moreover, alcoholic patients who also experience episodes of depression are at a higher risk for suicide than patients who present with either disorder separately. There are considerable data that other types of substance use such as heroin, cocaine, or polydrug use also increase the risk of suicidality when psychiatric illness is present, and this seems to be especially true in depressive or dysphoric mood disorders (unipolar, bipolar II, and mixed types of bipolar I disorders).^15,108 As a result, any assessment conducted on a patient with a substance use history must include an examination of symptoms of major depression or bipolar illness.^47,111

Patients with schizophrenia are at risk for suicide at rates comparable to major depression and are 20 times more likely to attempt suicide than the general population.¹⁰⁶ Approximately 50% of patients with schizophrenia will attempt suicide and 13% of schizophrenic patients will successfully complete suicide.^17,106 Additionally, between 5% and 18% of patients with severe borderline personality disorder (especially those patients with comorbid depression) ultimately kill themselves.^52,68,102

The ability to treat psychiatric disorders such as mood disorders, schizophrenia, borderline personality disorder, and alcoholism suggests that most suicides are preventable. Indeed, a suicide prevention program designed for general practitioners in Sweden demonstrated evidence for prevention based on the detection and treatment of depression.⁹⁰ The Centers for Disease Control and Prevention reported that psychiatric problems in US EDs represented approximately 3% of mental illness visits, which is significantly lower than the national psychiatric rate of 20% to 28%.²⁷ This suggests that significant psychiatric underdiagnosis is occurring in the ED. Consequently, emergency physicians must enhance the comprehensive nature of their psychiatric screening to identify suicidality and concomitant mental disorders in patients presenting with self-injury.

Treatment

Following a comprehensive psychiatric assessment, the next step is deciding on treatment alternatives. Any patient who has made a suicide attempt must be considered to be at risk for another attempt and some further intervention is warranted. The risk of a subsequent lethal attempt is approximately 1% per year over the first 10 years. The risk is highest during the first year. Suicide is most commonly a symptom of an underlying disease process, so the goal is to diagnose and treat the underlying disease in the setting that is the least restrictive while also ensuring safety for the patient. The treatment alternatives available will depend on the psychiatric sophistication of staff available to the ED at any given time. This section describes the commonly used interventions in the ED; they can be employed singly or in combination.

Psychotropic medications can be used acutely in the treatment of severe anxiety or psychosis; however, in the case of antidepressants, several weeks are required for therapeutic effect, so their immediate use is not indicated in the ED. However, if the patient is to be discharged to the community with follow-up, it is reasonable for a psychiatrist to start an antidepressant in the ED setting. There are concerns about prescribing medications with relatively high potential for lethality in overdose, such as the TCAs and nonselective monoamine oxidase inhibitors, to persons who have recently attempted suicide. Medications for medical illnesses, such as insulin, should also be considered for risk of overdose. However, newer antidepressants, particularly the SSRIs, can be used as first-line drugs for treatment of most depressions and are relatively safe in overdose.

In 2007, the US Food and Drug Administration (FDA) ordered that all antidepressants should include a black box warning stating that there is an increased risk of suicidality in children, adolescents, and adults younger than 24 years of age. This risk was not increased in adults 25 to 64 years of age, and was actually decreased in adults older than 65 years of age. Further studies have confirmed this age-related difference.^18,101­Proposed explanations include an ascertainment bias, activation and/or akathisia as an adverse event in the first few weeks, or increased energy resulting in increased ability to carry out suicidal plans. Nonetheless, the FDA and other authors emphasize that untreated depression also carries a risk of suicide, and treatment options should be carefully weighed with regard to their risks and benefits. Therefore, the initiation of antidepressant therapy by the nonpsychiatric physician is not recommended unless a tight linkage can be made between discharge and immediate (within days) aftercare by either a community outreach team or a crisis clinic.

Patients with depressive disorders may suffer from significant anxiety, as may patients with overwhelming situational stressors such as job loss, new financial hardship, bereavement, or divorce. The prescription of a short course (days to weeks) of a benzodiazepine may provide significant relief to the patient in crisis. Yet again, close psychiatric follow-up is essential.

After the patient’s immediate symptoms have been treated in the ED, the next treatment decision is determining the setting in which further treatment may safely be provided. Not all patients with suicidal ideation or even significant attempts necessarily require hospitalization, and there is still a substantial stigma attached to psychiatric hospitalization. In general, hospitalization should be used if less restrictive measures cannot ensure the patient’s safety. If significant doubt exists about the safety of outpatient treatment, then the patient should be observed in the ED for further evaluation, admitted to a general hospital with close nursing supervision, or admitted to a psychiatric unit. “Holding beds,” now available in some larger psychiatric EDs, are ideal for this purpose. Some localities may also have crisis outreach services that follow the patient after discharge from the ED and can provide appropriate monitoring and continuity of care.

Patients most likely to respond to interventions in the ED are individuals who have been stable, until recently but now, as a result of some external event, may find their way of life threatened. This acute change results in a painful state of anxiety and the mobilization of some combination of adaptive and maladaptive coping strategies. Finally, a second event, the precipitant, intensifies the anxiety to the point that the patient cannot tolerate the instability and is thrown into crisis. The patient then feels desperate and may be completely immobilized or vulnerable to various strong impulses including the impulse to run away, strike out at someone else, or kill him or herself. Reality testing is preserved, and no major psychiatric syndrome is present. The patient accurately perceives his or her situation, understands that the current reaction is a psychological problem, and is highly motivated to obtain help. The crisis may last for a matter of hours or weeks prior to the ED presentation and will ultimately resolve. Such patients respond well to crisis intervention and may actually undergo some positive development in the course of treatment.

By contrast, patients whose condition has been deteriorating for some time in the absence of significant stressors, and who appear on examination to be suffering from severe depressive symptoms, are unlikely to benefit rapidly from supportive techniques. If such patients present with suicidal ideation or attempts, it will be difficult, although not impossible, to manage them outside the hospital.

Outpatient settings have the advantage of maintaining the patient’s functioning as much as possible. Work and childcare responsibilities, financial obligations, and social relationships are not disrupted. Unnecessary regression is halted. The patient can assume more responsibility for his or her outcome, and independence helps preserve self-esteem. These individuals remain closer to and more engaged with the people and situations with whom and with which they must learn to cope. Their morale may be rapidly improved by the combination of support, planning, and modest early treatment successes.

However, in some cases, these same factors may be disadvantageous. Routine tasks may seem overwhelming. High levels of conflict may render major relationships at least temporarily unworkable. Inpatient settings offer the advantage of respite, high levels of structure, more intensive professional and peer support, constant supervision, and, usually, more rapid pharmacologic and psychosocial intervention.

The choice of inpatient or outpatient setting will depend on the balance of strengths and weaknesses of the patient, the involvement and competence of family or friends, the availability of a therapist in the community, and the ongoing stresses in the patient’s life. This decision is best made by a psychiatrist. Because a psychiatrist is not always present in many facilities, a trained mental health professional should optimally be on call to every ED. This may be a psychiatric social worker, nurse clinician, or psychologist. When such services are not available, it is appropriate to detain patients in the ED until a practitioner with specific competence is available or to transfer the patient to another facility for evaluation. Every U.S. state has laws that provide for the involuntary commitment of the mentally ill under circumstances that vary from state to state (Chap. 141). Any acute, deliberately self-injurious behavior would generally qualify. Chronic, repetitive dangerous behavior that is not deliberate, such as frequent unintentional opioid, alcohol, ­sedative-hypnotic, or illicit “recreational” psychoactive drug overdoses, warrants careful evaluation. In the absence of psychiatric illness, involuntary treatment is usually not necessary, but should be considered carefully if a patient appears unable to achieve self-care. The practitioner should be familiar with the criteria for commitment and the classes of health care professionals so empowered under state law.

There are other treatment interventions that can be provided in the emergency setting, including crisis intervention, substance abuse counseling, and family therapy. A single session in the ED may be sufficient to defuse a crisis or to spur the drug-abusing patient to seek help. ­Alternatively, the intervention may be initiated in the ED and continued as an outpatient.

Crisis intervention is a brief, highly focused therapy that seeks to deconstruct how a crisis occurred, with the intent of examining the patient’s role. Often, patients have distorted perceptions of the crisis, and a gentle “correction” of catastrophic thinking can be extremely helpful. The crisis is presented to the patient as an unfortunate and perhaps tragic experience that the patient can overcome. Ideally, the patient should have a relief of symptoms and learn how crises may be avoided in the future. This insight intervention will likely fail for patients with severe depression because of the presence of profound hopelessness. It is most successful for patients who give a history of high functioning just prior to the crisis.

Interestingly, sustained contact with patients via letters, postcards, or telephone calls can reduce suicidal behavior in the months after their ­presentation. A toxicology service in Australia found that a “postcard intervention” significantly reduced the rate of repeated self-poisoning by patients who presented to their ED for self-poisoning.²¹ The postcard was mailed to patients in a sealed envelope eight times over the 12 months after their initial presentation, and simply stated: “It has been a short time since you were here, and we hope things are going well for you. If you wish to drop us a note we would be happy to hear from you. Best wishes.” Although the proportion of patients who self-poisoned again did not differ in the experimental versus control group, the total number of self-poisoning episodes were halved for the experimental group on follow-up 24 months later.²¹­Follow-up data at 5 years indicated that this benefit was sustained, and that psychiatric admissions were reduced by one-third.²⁰ Other studies examining letter-writing or postcard interventions in various clinical settings have also shown some benefit when examining rates of attempted or completed ­suicide or psychiatric emergency department visits.^63,71,84 Given that such interventions do not require many resources and are quite cost effective, they appear promising.

“Contracting for safety” was a popular technique in the past, and consists of patients being asked whether they can remain safe and agree not to engage in self-harm. However, there are no empirical data to show that it is effective, and in court cases it has not been protective in terms of liability for the clinician. It is not recognized as part of the standard of care for the suicidal patient, and if anything, it seems to provide the clinician with a false sense of security. Its use is not recommended except as a technique to engage in a larger discussion of a safety plan that delineates scenarios the patient might face after leaving the hospital, and possible coping strategies.⁵³

Aggression presents unique challenges to the emergency physician, the psychiatrist, the nurses, and the other staff. Moreover, aggression is intimately related to suicidal behavior. Chronic aggression and impulsivity are risk factors for suicidality. Like suicidal patients, aggressive patients are difficult to treat and they tend to elicit strong negative reactions in ED personnel.⁹⁶ In one study of violence in the ED, directors of emergency medicine residency programs were surveyed as to the frequency of verbal threats, physical attacks, and the presence of weaponry in the area. Of the 127 institutions surveyed, 74.7% of the residency directors responded; 41 (32%) reported receiving at least one verbal threat each day; moreover, 23 (18%) reported that weapons were displayed as a threat at least once each month. Fifty-five program directors (43%) noted that a physical attack on medical staff members occurred at least once a month.⁷⁷

Another study involved a retrospective review of university police log records and ED staff incident reports to examine the problem of violence in the ED setting. Almost 75% of the incidents occurred during the evening or night. Of the 686 episodes of violence in this study, more than 25% required physical restraint or removal from the premises. In addition, it was found that the hospital security responded to the ED nearly twice daily.⁸⁷ These studies underscore the need for timely identification of the violent patient, as well as appropriate management for this diagnostically heterogeneous group.³⁸ The assessment and management of the violent patient should include provisions for patient and staff safety as well as a thorough search for the cause of violent behavior.^69,96

This section addresses the differential diagnosis of violent behavior, the pharmacotherapy of aggressive and/or agitated behavior (often termed “chemical restraint”), and the use of seclusion and physical restraint. It also provides an overview of potential risk factors for violent behavior.

Stress-Vulnerability Model of Aggression

As in the case of suicide, there are many and varied causes of violent ­behavior, some more social and some more medical in nature. The stress-vulnerability model suggests that violence should be considered as the outcome of a dynamic interaction among numerous factors both intrinsic and extrinsic to the individual, some of which promote and some of which ameliorate the potential for violent behavior at any given moment. Education may provide alternatives to violence, but a xenobiotic-induced delirium will render any education ineffective since delirium prevents patients from reasoning or exercising impulse control. Once confused, the patient may misinterpret health care efforts in a paranoid manner, and become violent under circumstances that would not normally be sufficient to provoke a violent outburst. Some patients, on the other hand, come from cultures in which aggressive behavior is more acceptable, and these patients require little stress or provocation before responding in what can be perceived as aggression by Western cultural standards.

In the ED, likely medical sources of vulnerability include metabolic derangements, exposure to xenobiotic (both licit and illicit) withdrawal syndromes, seizure disorders, head trauma, psychosis, and personality ­disorders. Additionally, patients with severe pain, delirium, or extreme anxiety can respond to the efforts of emergency personnel with resistance, hostility, or frank aggression.

Prediction of Violence

Research on risk factors for community violence may not apply to the prediction of inpatient violence. Some researchers have postulated that violence committed outside the hospital may not be predictive of inpatient violence and that hospital violence may result from the interaction of patients with specific factors found in the hospital environment.^41,96 Other studies are contradictory.⁹⁸ Consequently, prior violence is not a perfect predictor of future violence. Other factors, such as mental illness and substance abuse, need to be examined to make meaningful predictions of inpatient violence for each individual case.⁹⁶ One study found that the most common types of hospital violence were incidents of aggression against objects in the hospital (57%), violence directed against the hospital staff (28%), and violence directed against other patients (14%).⁹⁷ In this study, men did not commit significantly greater incidents of violence than women. Other ­studies concur that men are not necessarily more of a risk for inpatient violence than women. For example, researchers examining inpatient violence found that close to half of the violent incidents were committed by women, and the number of violence-related injuries committed by men and women inpatients were almost proportional to the ratio of men and women inpatients on the unit.^76,89 The conclusion was that gender should not be considered a risk factor for inpatient violence. Long hospitalization was not considered a factor predictive of violence for the majority of inpatients. As with outpatient violence, the correlation of violence with younger age appears to hold true in the inpatient setting.^82,89,90

Substance Use

The association between substance use and violence is well established. Alcohol is found in the offender, the victim, or both in one-half to two-thirds of homicides and serious assaults.^30,89 Substance use is seldom the sole cause, but it may contribute to violence in a number of ways. The direct pharmacologic effects include disinhibition and misinterpretation, suspiciousness, or paranoia. Psychological effects of substance use include cultural expectations of appropriate behavior under the influence and the ability to excuse or disavow inappropriate behavior that occurs while intoxicated. Substance use then interacts with other physiologic, cognitive, psychological, situational, and cultural factors including any underlying mental illness. A tripartite model has been described: (1) systemic violence related to drug distribution, (2) economic compulsive violence associated with the criminal activity necessary to sustain a drug habit, and (3) psychopharmacologic violence resulting from the direct effects of the particular xenobiotic.^55,56

Mental Illness

The relationship between mental illness and violence is also complex. Efforts made to destigmatize mental illness have confused the issue, but it seems clear that mental illness is associated with an increased risk for violence.⁸⁹ In one large epidemiologic study, the prevalence of violence for those with no mental illness was 2%. Schizophrenia was associated with an 8% rate of violent behavior, and other mental disorders all had similar prevalence of approximately 12%. But of all respondents reporting violent behaviors, 42% had an SUD. Substance use more than tripled the rate of violence for individuals with schizophrenia. Mental illness appears to reduce the threshold for aggression, and the more comorbid conditions present, the greater the risk.^85,104,105

Researchers consistently find a greater prevalence of personality ­disorders among violent inpatients than among nonviolent inpatients.²⁴ However, antisocial personality is the condition most strongly associated with both substance use and aggression. In one study, when the history of juvenile deviant behavior was controlled for alcohol, the drug most commonly associated with violence, it accounted for only 2% of the violent behavior.

In conclusion, some aggressive behavior is attributable to the direct pharmacologic effects of xenobiotics, but probably represents only a modest fraction. Substance use is often implicated in violent behavior in the community, it can occur as a coincidental part of the lifestyle of violent individuals, and both substance use and violence are related to common underlying characteristics such as a personality disorder.

Additional Factors in Aggressive Behavior

Many of the factors correlated with aggression are easy to observe and monitor in the hospital, yet some additional factors may not be as easy to detect. For example, one study found that most violent incidents in the hospital occur on Mondays and Fridays, with very few incidents on weekends. Furthermore, researchers have postulated a seasonal variation of violence.³² There is an increase in the frequency of assaults by inpatients during the winter months, and it has been hypothesized that increased population density, cold temperature, and less sunlight during the day could account for the increased violence. This finding is in contrast to the literature on outpatient violence, which has reported greater incidence of violence during the warmer months.⁶ However, this same review conceded that any extreme temperature could evoke aggressive feelings and frustration. Yet another study examined the relationship between temperature and violence and found that more aggressive acts occur during the summer months, both in the hospital and in the community.⁵⁰ They cited several explanations, one of which was that the high rate of staff turnover (eg, vacations) disrupts the social networks that the patients have established, and evokes aggressive feelings.

Although it is unclear whether cold temperatures can provoke aggression as much as it has been established that heat can, it does seem clear that overcrowding and social stressors can lead to violent behavior. If the effects of temperature and social stressors (eg, holidays) correlate so drastically with violence in the community, it is likely that such effects would have even more impact when comorbid with severe mental illness, substance use, or any of the other risk factors of aggression.

Assessment of the Violent Patient

The comprehensive evaluation of the violent patient should include a complete physical examination. The examination may reveal the underlying cause of the violent behavior as well as ensuring the treatment of any secondary patient injuries. Laboratory analysis may include blood chemistries (glucose, electrolytes, including sodium and calcium), a complete blood count, liver function tests, thyroid function tests, lumbar puncture and/or neuroimaging as guided by the examination and clinical history.

Substance Intoxication and Withdrawal.

Illicit xenobiotic and alcohol use often present with symptoms of violence. Please refer to subsequent chapters for details on presentations of acute intoxication for specific ­substances.

Withdrawal syndromes can also promote aggressive behavior as a consequence of physical discomfort or anticipatory anxiety. Patients experiencing any of these symptoms may become aggressive, verbally abusive, or threatening. Prompt recognition of these syndromes and immediate treatment can prevent some aggressive outbursts. Because drug use is often concealed, is difficult to ascertain on clinical grounds, and frequently contributes to violent behavior, urine and blood toxicological studies may be useful to enhance the understanding and long-range treatment of some patients.⁸⁹

Alternative Etiologies.

Delirium can be a cause of aggression. Patients are often suddenly confused, frightened, or frankly psychotic as a result of impaired perception. Patients may require sedation or physical restraint in order to prevent injury to themselves as well as staff members.

Although persons with psychotic disorders are not generally aggressive, there are aspects of their psychosis that place them at risk for aggressive behavior. Paranoid ideation can serve to promote misperceptions of impending bodily harm (“They’re trying to kill me”), sexual victimization (“Men and women are raping me”), or humiliation (“Everyone is laughing at me”). It follows that these fearful perceptions might provoke violent reactions in a patient. Hallucinations can lead to aggression, either when patients explicitly follow the instructions of a command auditory hallucination or in reaction to the anxiety and fear that patients can experience with loud or persistent auditory hallucinations (“hearing voices”). Patients with either borderline or antisocial personality disorder are at risk for ­violent ­acting out as a result of poor impulse control and impaired frustration tolerance in the context of poor coping skills.

Violence risk is also associated with cognitive dysfunction. Both acute mental illness and chronic substance use can result in neurologic impairment. Psychiatric patients with compromised cognitive abilities such as impaired attention, memory, or executive functioning, such as reasoning and planning, are found to be at increased risk for violence.⁸⁹ Patients presenting with cognitive impairment may also be at increased risk for committing acts of violence in the ED.

Treatment

The acute pharmacotherapy of violent behavior is directed simply at reducing the level of arousal. If agitation and violent outbursts are viewed as transient disturbances of the usual treatment relationship between the physician and patient, then pharmacotherapy, seclusion, and restraint are to be used as needed to restore that relationship for the benefit of the patient as well as staff.² Sleep delays rather than promotes assessment, may further frighten or anger the patient, and does not guarantee elimination of the agitated state on awakening. There are two main approaches to controlling aggressive behavior: chemical restraint and physical restraint.

Chemical Restraint.

Chemical restraint is defined as “chemical measures for confining a patient’s bodily movements, thereby preventing injury to self or others and reducing agitation.”³⁷ Given that aggression can result from multiple etiologies, there is much debate about the specific sedative and route of administration that should be used. Overall, it has been found that both benzodiazepines and antipsychotics resulted in rapid control of agitation and aggression.^19,45,74,76

Haloperidol has been safely used in the treatment of agitation and aggression in patients with psychoses, acute alcohol intoxication, and delirium.^1,19,74,76 It can be administered orally, intravenously, or intramuscularly. Dosing intervals range from 30 minutes to 2 hours, with a usual regimen of haloperidol 5 mg given every 30 to 60 minutes; most patients respond after one to three doses. Older studies indicated that the dose of haloperidol needed to achieve sedation rarely exceeded a total of 50 mg in acute management, but it is unusual that more than 20 mg is used today.^78,79 Most psychiatrists would switch to a second-generation antipsychotic such as olanzapine or add a mood stabilizer such as valproic acid rather than continue to titrate haloperidol.

Benzodiazepines are also quite effective for sedation; their use has been examined in patients with psychoses, stimulant intoxication, sedative-hypnotic and alcohol withdrawal, and postoperative agitation.^54,81 Diazepam may be given as intravenous (IV) 5 to 10 mg, with rapid repeat dosing titrated to desired effect. Because diazepam is poorly absorbed from intramuscular (IM) sites, its preferred route of administration is either IV or oral. Lorazepam 1 to 2 mg or midazolam 5 to 10 mg may be given orally or parenterally and repeated at 30- or 15-minutes intervals, respectively, until the patient is calm (Antidotes in Depth: A23). Midazolam is frequently used in the ED due to its quick onset of action and short half-life of 1 to 4 hours, but it has a significant amnestic effect; lorazepam may be preferable since it is less amnestic and will provide longer anxiolysis given its slower elimination half-life of 10 to 20 hours.

If a patient is agitated in the context of alcohol intoxication, antipsychotics can be used but benzodiazepines should be avoided due to the potential to cause additive respiratory depression. Benzodiazepines may have a unique role in the treatment of agitation secondary to cocaine intoxication (Chap. 78). Antipsychotics, particularly low-potency antipsychotics (such as chlorpromazine), lower the seizure threshold in animals, so their use for patients with cocaine/amphetamine toxicity or alcohol/sedative-hypnotic withdrawal should be avoided. Some studies suggest that the combined use of lorazepam with antipsychotics in patients with known psychiatric illness and delirium afforded relief of psychotic symptoms while allowing for a reduced dose of antipsychotic medications.^1,31,95 On psychiatric services, the combination of haloperidol and lorazepam is still the most frequently used IM pharmacotherapeutic intervention, likely due to the relative lack of data for second-generation antipsychotics combined with lorazepam.⁶¹

Special concern has been raised about oversedation and respiratory depression with IM olanzapine, given that eight fatalities were reported in the European literature when olanzapine was used in excessive dosages or combined with benzodiazepines and/or other antipsychotics; although significant comorbidities were present in the patients who died, it is now recommended that IM olanzapine should not be coadministered with other CNS depressants.¹¹ In general, concerns regarding respiratory depression mandate careful observation and monitoring of patients receiving sedation with any xenobiotic.

Physical Restraint.

Isolation and mechanical restraints are also used in the treatment of violent behavior. Isolation or seclusion can help to diminish environmental stimuli and thereby reduce hyper-reactivity. However, a few aspects are worth mentioning: Because seclusion is defined by a condition of very limited interactive and environmental cues, it is not indicated for patients with unstable medical conditions, delirium, dementia, self-injurious behavior (cutting, head banging), or those who are experiencing extrapyramidal reactions as a consequence of antipsychotic medication (such as an acute dystonic reaction).³ Mechanical restraint is used to prevent patient and staff injury, although it does occasionally lead to patient and staff injury itself.^9,61 All facilities should have clear, written policy guidelines for restraint that address monitoring, documentation, and provisions for patient comfort. Frequent reassessment of the patient and documentation of the need for continued restraint are essential, and should be performed according to state law. The Joint Commission standards require a renewed restraint order every 4 hours, but the physician (or licensed independent practitioner) is only required to evaluate the patient within the first hour and then every 24 hours. Other staff members provide ongoing monitoring at time intervals specified by the institution. The Joint Commission notes that state laws should be followed when they are more restrictive. Although there are limited controlled data at this point, numerous interventions are being considered in psychiatric settings as alternatives to de-escalate a patient before restraints are employed, and there is a general trend to attempt to reduce the use of mechanical restraints, although it seems doubtful that such restraints can be eliminated entirely.¹⁰⁰

Training.

Finally, training in the management of aggression helps to reduce violence and injuries through the early identification of impending episodes of violence, use of verbal techniques to defuse incidents, and appropriate physical techniques to minimize injuries in those that occur. It is necessary for health care providers to maintain their skills through training and to advocate for continuing medical education on this topic at the workplace.^17,93 See Tables 27–6 and 27–7 for violence warning signs and the S.A.F.E.S.T. Approach.

SUDs are characterized by compulsive drug seeking and drug taking in a setting of loss of control, continued use despite negative consequences, and functional impairment in several areas of a person’s life. They are highly comorbid with suicidality, violence, and mental illness as described above in previous sections. Prevalence rates of substance use in the ED range from 9% to 47%, depending on the sample.⁴⁶ ED physicians and medical toxicologists are in a unique position to identify substance use disorders and refer patients to additional treatment services. There is an extensive array of ­evidence-based treatment for SUDs, but fewer than 10% of persons with SUDs in the United Sates receive substance abuse treatment.

DSM-5 Diagnoses

The psychiatrist’s diagnostic handbook, the Diagnostic and Statistical Manual of Mental Disorders (DSM), codifies various diagnoses involving the use of substances. The latest version is the 5th edition, DSM-5, which was released in 2013.⁴ While previous versions of the DSM differentiated “substance abuse” and “substance dependence,” the DSM-5 combines these categories into an overarching “substance use disorder” diagnosis with mild, moderate, and severe specifiers. “Substance abuse” was considered to be a milder or earlier phase of an addictive disorder, while “substance dependence” was the more severe manifestation; however, in practice, substance abuse criteria were sometimes quite severe, and the distinction was not found to be clinically helpful. In the combined “substance use disorder” criteria found in DSM-5, the criterion of recurrent legal problems has been eliminated, and a new criterion for “craving” has been introduced. Table 27–8 shows the criteria for an alcohol use disorder, and the reader can use this as a template to substitute other substances of abuse as well.

Neurobiology of Substance Use Disorders

It is helpful for the emergency physician to be able to recognize the signs and symptoms of an SUD. Although the DSM-5 criteria are straightforward, many patients may not be forthcoming and may not share the extent of their drug use and its effect on their lives. In those cases, one can look for inconsistencies in either the history or the presentation. Denial is often prominent. For example, a patient might present intoxicated with an arm injury, but is unconcerned and dismissive about the incident when no longer intoxicated. One would expect a patient to demonstrate appropriate concern, rather than just “wanting to leave.”

One of the more vexing aspects of substance-dependent patients is the frequency with which they minimize the dangerousness of their actions or the severity of their presentation. This can be interpreted as a sign of prefrontal cortex dysfunction due to the addiction. The patient is incorrectly processing risks/benefits, and is increasingly driven to compulsive drug use at the cost of natural rewards such as food, sex, love, or safety. Initial drug use results in relaxation, euphoria, stress relief, or a variety of other psychological effects, as mediated by the dopaminergic “reward pathway.” It was traditionally taught that dopamine release in this pathway was synonymous with experiencing “pleasure,” but it has become apparent that the function of dopamine is more accurately portrayed as indicating “salience” (or importance). A burst of dopamine in this pathway confers enhanced salience onto a particular stimulus or aspect of the environment, which means that this stimulus now appears more important and there is greater motivation to pay attention to it. One can see how this works well as a survival system, with natural rewards such as food and sex resulting in increased dopamine and subsequently greater salience. Drugs of abuse essentially “hijack” this pathway by directly producing supraphysiologic dopamine increases that would normally not occur.⁷⁰

However, activity in the reward pathway only explains the acute effects of drugs, and does not mediate any enduring behavioral changes. The cravings and compulsive behavior that occur in addiction result from glutamate-mediated circuits that develop in the prefrontal cortex and establish aberrant connections to the limbic system. Areas affected include the orbitofrontal cortex, which regulates salience attribution, and the anterior cingulate gyrus, which regulates inhibitory control. Late-stage addiction is characterized by derangements in these areas of the prefrontal cortex, which lend increased salience to drugs of abuse coupled with reduced ability to inhibit behavior. Environmental “triggers” established by classical conditioning can subsequently act to induce powerful states of craving and motivation to find and use a particular drug.⁷⁰

Despite our sophistication in describing the neurobiology of addiction, it can be challenging for an individual clinician to act accordingly when dealing with an addicted patient; we are used to assuming that behavior is voluntary, especially if a person is coherent and oriented. It can be difficult to recall that a patient who is hostile and in denial actually has true neurobiologic dysfunction, not unlike a patient with dementia or psychosis. Unfortunately, our current civil commitment laws and society’s perception of these patients have not “caught up” with this neurobiological understanding.

Suicidality

Substance use itself (not necessarily as part of an SUD) is frequently comorbid with self-injurious behavior. Based on data from 16 reporting states in the National Violent Death Reporting System in 2009, 33.3% of suicide decedents tested positive for alcohol, 23% for antidepressants, 20.8% for opiates (including heroin and prescription analgesics).⁷³ As described earlier, patients often combine alcohol or illicit substances with prescription medication when multiple substances are used to self-poison.^25,72 It is hypothesized that substance use can lead to disinhibition and the resultant ability to carry out an act of self-harm that may not have happened if the person had remained sober.

Treatment

Substance abuse treatment is ultimately an intermediate (weeks to months) to long-term (months to years) intervention. However, there are powerful initial steps that the emergency physician can take. Central among these is confronting the patient about the medical consequences of substance use. This can take the form of discussion only, or the physician can invite the patient to examine clinical laboratory results or view remarkable clinical diagnostic findings (hepatomegaly, repeated fractures from falls, liver enzyme abnormalities, or evidence of “silent” past myocardial infarction). There is little to be lost from a respectful but blunt confrontation of the patient’s deterioration, and the patient may listen to a physician rather than family or friends. “Physician advice” in this manner is shown to be useful for patients who do not yet meet criteria for substance dependence, and therefore still have sufficient control of their substance use to make a logical decision about the risks versus benefits of continued use.

Patients who are further along the addictive spectrum and who display behavioral or physiologic signs of substance dependence are much less likely to curb their substance use in the face of adverse medical sequelae. For these patients, more intensive intervention is warranted, often in the form of a consultation. For most hospitals, this type of consultation would be fulfilled by a psychiatrist, but some institutions have a separate substance abuse consult service. Ideally, all patients suspected to have an SUD should be referred to some type of aftercare when possible: inpatient or outpatient detoxification, inpatient or outpatient rehabilitation, or even follow-up with primary care.

In a busy ED this can be quite cumbersome logistically, depending upon the demographics of the patient population and the percentage of patients suspected to have an SUD. Regulatory agencies at the state and ­federal levels are increasingly mandating substance abuse screening, brief interventions, and referrals, so hopefully this will result in increased resources for additional ancillary staff to assist with such tasks.

It is helpful to note that peer counseling is particularly useful in addictive disorders, and community 12-step programs like Narcotics Anonymous (NA) and Alcoholics Anonymous (AA) are ubiquitous worldwide. AA has its own website and specific phone lines in some cities that are staffed by AA members with the express purpose of orienting potential new members and assisting them in getting to their first meeting. Even if there is insufficient infrastructure in the ED to provide the more intensive intervention that might be needed for a patient who is substance-dependent, the physician can at least refer directly to AA. Patients with repeated presentations for substance intoxication/withdrawal should be considered for a higher level of care, since they may be demonstrating an inability to care for themselves. Such a decision would be made in conjunction with the psychiatrist and would have to comply with the local civil commitment laws. Finally, it is important to note that substance abuse treatment is quite effective, and that coerced treatment has been found to be just as effective as voluntary treatment. Given the remarkably high comorbidity of SUDs in the ED setting, it behooves us to attempt to connect these patients with the appropriate treatment.

• Capacity assessment should be approached systematically, and such an approach is helpful in more complex situations. A “sliding scale” is a useful concept; when refusing an intervention would have serious consequences, a patient is held to a higher standard to demonstrate capacity.

• “Medical clearance” does not have a standard protocol, but in general the ED assessment should be guided by the history and physical examination.

• Both violent and suicidal behavior in the ED may be the cause or the effect of many toxicologic presentations.

• It is incumbent on all emergency physicians to screen patients for psychiatric emergency presentations as part of a comprehensive screening for self-harm.

• Identifying risk factors for suicide and aggression can aid the clinician in employing preventive or early intervention strategies in the ED.

• Important risk factors for both suicidal and violent behavior include past history of the behavior, comorbid mental illness, drug and alcohol intoxication, and young age.

• Mental status examination for suicidality should focus on extrinsic factors such as current ideation, intent, lethality of plan, current life stressors, as well as intrinsic vulnerability factors such as comorbid mental illness, feelings of hopelessness, and impulsivity.

• In terms of violence risk assessment, drug and alcohol intoxication, mental illness, and psychiatric medication noncompliance (alone or in combination) are robust predictors of aggressive behavior in the ED and other inpatient settings.

• Substance use disorders are highly prevalent in patients who present with psychiatric issues in the ED, and there are possibilities for identification and referral even in busy ED settings.

Acknowledgment

Kishor Malavade, MD, Mark R. Serper, PhD, Michael H. Allen, MD, Wendy Rives, MD (deceased), Brett R. Goldberg, PhD, and Cherie Elfenbein, MD, contributed to this chapter in previous editions.

References