Each recommendation is identified as falling into one of three categories of endorsement, indicated by a bracketed Roman numeral following the statement. The three categories represent varying levels of clinical confidence regarding the recommendation:

In this guideline, the following terms will be used:

• Suicide—self-inflicted death with evidence (either explicit or implicit) that the person intended to die.
• Suicide attempt—self-injurious behavior with a nonfatal outcome accompanied by evidence (either explicit or implicit) that the person intended to die.
• Aborted suicide attempt—potentially self-injurious behavior with evidence (either explicit or implicit) that the person intended to die but stopped the attempt before physical damage occurred.
• Suicidal ideation—thoughts of serving as the agent of one's own death. Suicidal ideation may vary in seriousness depending on the specificity of suicide plans and the degree of suicidal intent.
• Suicidal intent—subjective expectation and desire for a self-destructive act to end in death.
• Lethality of suicidal behavior—objective danger to life associated with a suicide method or action. Note that lethality is distinct from and may not always coincide with an individual's expectation of what is medically dangerous.
• Deliberate self-harm—willful self-inflicting of painful, destructive, or injurious acts without intent to die.

A detailed exposition of definitions relating to suicide has been provided by O'Carroll et al. (1).

The psychiatric evaluation is the essential element of the suicide assessment process [I]. During the evaluation, the psychiatrist obtains information about the patient's psychiatric and other medical history and current mental state (e.g., through direct questioning and observation about suicidal thinking and behavior as well as through collateral history, if indicated). This information enables the psychiatrist to 1) identify specific factors and features that may generally increase or decrease risk for suicide or other suicidal behaviors and that may serve as modifiable targets for both acute and ongoing interventions, 2) address the patient's immediate safety and determine the most appropriate setting for treatment, and 3) develop a multiaxial differential diagnosis to further guide planning of treatment. The breadth and depth of the psychiatric evaluation aimed specifically at assessing suicide risk will vary with setting; ability or willingness of the patient to provide information; and availability of information from previous contacts with the patient or from other sources, including other mental health professionals, medical records, and family members. Although suicide assessment scales have been developed for research purposes, they lack the predictive validity necessary for use in routine clinical practice. Therefore, suicide assessment scales may be used as aids to suicide assessment but should not be used as predictive instruments or as substitutes for a thorough clinical evaluation [I].

Table 1 presents important domains of a suicide assessment, including the patient's current presentation, individual strengths and weaknesses, history, and psychosocial situation. Information may come from the patient directly or from other sources, including family members, friends, and others in the patient's support network, such as community residence staff or members of the patient's military command. Such individuals may be able to provide information about the patient's current mental state, activities, and psychosocial crises and may also have observed behavior or been privy to communications from the patient that suggest suicidal ideation, plans, or intentions. Contact with such individuals may also provide opportunity for the psychiatrist to attempt to fortify the patient's social support network. This goal often can be accomplished without the psychiatrist's revealing private or confidential information about the patient. In clinical circumstances in which sharing information is important to maintain the safety of the patient or others, it is permissible and even critical to share such information without the patient's consent [I].

It is important to recognize that in many clinical situations not all of the information described in this section may be possible to obtain. It may be necessary to focus initially on those elements judged to be most relevant and to continue the evaluation during subsequent contacts with the patient.

When communicating with the patient, it is important to remember that simply asking about suicidal ideation does not ensure that accurate or complete information will be received. Cultural or religious beliefs about death or suicide, for example, may influence a patient's willingness to speak about suicide during the assessment process as well as the patient's likelihood of acting on suicidal ideas. Consequently, the psychiatrist may wish to explore the patient's cultural and religious beliefs, particularly as they relate to death and to suicide [II].

It is important for the psychiatrist to focus on the nature, frequency, depth, timing, and persistence of suicidal ideation [I]. If ideation is present, request more detail about the presence or absence of specific plans for suicide, including any steps taken to enact plans or prepare for death [I]. If other aspects of the clinical presentation seem inconsistent with an initial denial of suicidal thoughts, additional questioning of the patient may be indicated [II].

Where there is a history of suicide attempts, aborted attempts, or other self-harming behavior, it is important to obtain as much detail as possible about the timing, intent, method, and consequences of such behaviors [I]. It is also useful to determine the life context in which they occurred and whether they occurred in association with intoxication or chronic use of alcohol or other substances [II]. For individuals in previous or current psychiatric treatment, it is helpful to determine the strength and stability of the therapeutic relationship(s) [II].

If the patient reports a specific method for suicide, it is important for the psychiatrist to ascertain the patient's expectation about its lethality, for if actual lethality exceeds what is expected, the patient's risk for accidental suicide may be high even if intent is low [I]. In general, the psychiatrist should assign a higher level of risk to patients who have high degrees of suicidal intent or describe more detailed and specific suicide plans, particularly those involving violent and irreversible methods [I]. If the patient has access to a firearm, the psychiatrist is advised to discuss with and recommend to the patient or a significant other the importance of restricting access to, securing, or removing this and other weapons [I].

Documenting the suicide assessment is essential [I]. Typically, suicide assessment and its documentation occur after an initial evaluation or, for patients in ongoing treatment, when suicidal ideation or behaviors emerge or when there is significant worsening or dramatic and unanticipated improvement in the patient's condition. For inpatients, reevaluation also typically occurs with changes in the level of precautions or observations, when passes are issued, and during evaluation for discharge. As with the level of detail of the suicide assessment, the extent of documentation at each of these times varies with the clinical circumstances. Communications with other caregivers and with the family or significant others should also be documented [I]. When the patient or others have been given specific instructions about firearms or other weapons, this communication should also be noted in the record [I].

Suicide and suicidal behaviors cause severe personal, social, and economic consequences. Despite the severity of these consequences, suicide and suicidal behaviors are statistically rare, even in populations at risk. For example, although suicidal ideation and attempts are associated with increased suicide risk, most individuals with suicidal thoughts or attempts will never die by suicide. It is estimated that attempts and ideation occur in approximately 0.7% and 5.6% of the general U.S. population per year, respectively (2). In comparison, in the United States, the annual incidence of suicide in the general population is approximately 10.7 suicides for every 100,000 persons, or 0.0107% of the total population per year (3). This rarity of suicide, even in groups known to be at higher risk than the general population, contributes to the impossibility of predicting suicide.

The statistical rarity of suicide also makes it impossible to predict on the basis of risk factors either alone or in combination. For the psychiatrist, knowing that a particular factor (e.g., major depressive disorder, hopelessness, substance use) increases a patient's relative risk for suicide may affect the treatment plan, including determination of a treatment setting. At the same time, knowledge of risk factors will not permit the psychiatrist to predict when or if a specific patient will die by suicide. This does not mean that the psychiatrist should ignore risk factors or view suicidal patients as untreatable. On the contrary, an initial goal of the psychiatrist should be to estimate the patient's risk through knowledgeable assessment of risk and protective factors, with a primary and ongoing goal of reducing suicide risk [I].

Some factors may increase or decrease risk for suicide; others may be more relevant to risk for suicide attempts or other self-injurious behaviors, which are in turn associated with potential morbidity as well as increased suicide risk. In weighing risk and protective factors for an individual patient, consideration may be given to 1) the presence of psychiatric illness; 2) specific psychiatric symptoms such as hopelessness, anxiety, agitation, or intense suicidal ideation; 3) unique circumstances such as psychosocial stressors and availability of methods; and 4) other relevant clinical factors such as genetics and medical, psychological, or psychodynamic issues [I].

It is important to recognize that many of these factors are not simply present or absent but instead may vary in severity. Others, such as psychological or psychodynamic issues, may contribute to risk in some individuals but not in others or may be relevant only when they occur in combination with particular psychosocial stressors.

Once factors are identified, the psychiatrist can determine if they are modifiable. Past history, family history, and demographic characteristics are examples of nonmodifiable factors. Financial difficulties or unemployment can also be difficult to modify, at least in the short term. While immutable factors are important to identify, they cannot be the focus of intervention. Rather, to decrease a patient's suicide risk, the treatment should attempt to mitigate or strengthen those risk and protective factors that can be modified [I]. For example, the psychiatrist may attend to patient safety, address associated psychological or social problems and stressors, augment social support networks, and treat associated psychiatric disorders (such as mood disorders, psychotic disorders, substance use disorders, and personality disorders) or symptoms (such as severe anxiety, agitation, or insomnia).

Psychiatric management consists of a broad array of therapeutic interventions that should be instituted for patients with suicidal thoughts, plans, or behaviors [I]. Psychiatric management includes determining a setting for treatment and supervision, attending to patient safety, and working to establish a cooperative and collaborative physician-patient relationship. For patients in ongoing treatment, psychiatric management also includes establishing and maintaining a therapeutic alliance; coordinating treatment provided by multiple clinicians; monitoring the patient's progress and response to the treatment plan; and conducting ongoing assessments of the patient's safety, psychiatric status, and level of functioning. Additionally, psychiatric management may include encouraging treatment adherence and providing education to the patient and, when indicated, family members and significant others.

Patients with suicidal thoughts, plans, or behaviors should generally be treated in the setting that is least restrictive yet most likely to be safe and effective [I]. Treatment settings and conditions include a continuum of possible levels of care, from involuntary inpatient hospitalization through partial hospital and intensive outpatient programs to occasional ambulatory visits. Choice of specific treatment setting depends not only on the psychiatrist's estimate of the patient's current suicide risk and potential for dangerousness to others, but also on other aspects of the patient's current status, including 1) medical and psychiatric comorbidity; 2) strength and availability of a psychosocial support network; and 3) ability to provide adequate self-care, give reliable feedback to the psychiatrist, and cooperate with treatment. In addition, the benefits of intensive interventions such as hospitalization must be weighed against their possible negative effects (e.g., disruption of employment, financial and other psychosocial stress, social stigma).

For some individuals, self-injurious behaviors may occur on a recurring or even chronic basis. Although such behaviors may occur without evidence of suicidal intent, this may not always be the case. Even when individuals have had repeated contacts with the health care system, each act should be reassessed in the context of the current situation [I].

In treating suicidal patients, particularly those with severe or recurring suicidality or self-injurious behavior, the psychiatrist should be aware of his or her own emotions and reactions that may interfere with the patient's care [I]. For difficult-to-treat patients, consultation or supervision from a colleague may help in affirming the appropriateness of the treatment plan, suggesting alternative therapeutic approaches, or monitoring and dealing with countertransference issues [I].

The suicide prevention contract, or "no-harm contract," is commonly used in clinical practice but should not be considered as a substitute for a careful clinical assessment [I]. A patient's willingness (or reluctance) to enter into an oral or a written suicide prevention contract should not be viewed as an absolute indicator of suitability for discharge (or hospitalization) [I]. In addition, such contracts are not recommended for use with patients who are agitated, psychotic, impulsive, or under the influence of an intoxicating substance [II]. Furthermore, since suicide prevention contracts are dependent on an established physician-patient relationship, they are not recommended for use in emergency settings or with newly admitted or unknown inpatients [II].

Despite best efforts at suicide assessment and treatment, suicides can and do occur in clinical practice. In fact, significant proportions of individuals who die by suicide have seen a physician within several months of death and may have received specific mental health treatment. Death of a patient by suicide will often have a significant effect on the treating psychiatrist and may result in increased stress and loss of professional self-esteem. When the suicide of a patient occurs, the psychiatrist may find it helpful to seek support from colleagues and obtain consultation or supervision to enable him or her to continue to treat other patients effectively and respond to the inquiries or mental health needs of survivors [II]. Consultation with an attorney or a risk manager may also be useful [II]. The psychiatrist should be aware that patient confidentiality extends beyond the patient's death and that the usual provisions relating to medical records still apply. Any additional documentation included in the medical record after the patient's death should be dated contemporaneously, not backdated, and previous entries should not be altered [I]. Depending on the circumstances, conversations with family members may be appropriate and can allay grief [II]. In the aftermath of a loved one's suicide, family members themselves are more vulnerable to physical and psychological disorders and should be helped to obtain psychiatric intervention, although not necessarily by the same psychiatrist who treated the individual who died by suicide [II].

In developing a plan of treatment that addresses suicidal thoughts or behaviors, the psychiatrist should consider the potential benefits of somatic therapies as well as the potential benefits of psychosocial interventions, including the psychotherapies [I]. Clinical experience indicates that many patients with suicidal thoughts, plans, or behaviors will benefit most from a combination of these treatments [II]. The psychiatrist should address the modifiable risk factors identified in the initial psychiatric evaluation and make ongoing assessments during the course of treatment [I]. In general, therapeutic approaches should target specific axis I and axis II psychiatric disorders; specific associated symptoms such as depression, agitation, anxiety, or insomnia; or the predominant psychodynamic or psychosocial stressor [I]. While the goal of pharmacologic treatment may be acute symptom relief, including acute relief of suicidality or acute treatment of a specific diagnosis, the treatment goals of psychosocial interventions may be broader and longer term, including achieving improvements in interpersonal relationships, coping skills, psychosocial functioning, and management of affects. Since treatment should be a collaborative process between the patient and clinician(s), the patient's preferences are important to consider when developing an individual treatment plan [I].

Evidence for a lowering of suicide rates with antidepressant treatment is inconclusive. However, the documented efficacy of antidepressants in treating acute depressive episodes and their long-term benefit in patients with recurrent forms of severe anxiety or depressive disorders support their use in individuals with these disorders who are experiencing suicidal thoughts or behaviors [II]. It is advisable to select an antidepressant with a low risk of lethality on acute overdose, such as a selective serotonin reuptake inhibitor (SSRI) or other newer antidepressant, and to prescribe conservative quantities, especially for patients who are not well-known [I]. For patients with prominent insomnia, a sedating antidepressant or an adjunctive hypnotic agent can be considered [II]. Since antidepressant effects may not be observed for days to weeks after treatment has started, patients should be monitored closely early in treatment and educated about this probable delay in symptom relief [I].

To treat symptoms such as severe insomnia, agitation, panic attacks, or psychic anxiety, benzodiazepines may be indicated on a short-term basis [II], with long-acting agents often being preferred over short-acting agents [II]. The benefits of benzodiazepine treatment should be weighed against their occasional tendency to produce disinhibition and their potential for interactions with other sedatives, including alcohol [I]. Alternatively, other medications that may be used for their calming effects in highly anxious and agitated patients include trazodone, low doses of some second-generation antipsychotics, and some anticonvulsants such as gabapentin or divalproex [III]. If benzodiazepines are being discontinued after prolonged use, their doses should be reduced gradually and the patient monitored for increasing symptoms of anxiety, agitation, depression, or suicidality [II].

There is strong evidence that long-term maintenance treatment with lithium salts is associated with major reductions in the risk of both suicide and suicide attempts in patients with bipolar disorder, and there is moderate evidence for similar risk reductions in patients with recurrent major depressive disorder [I]. Specific anticonvulsants have been shown to be efficacious in treating episodes of mania (i.e., divalproex) or bipolar depression (i.e., lamotrigine), but there is no clear evidence that their use alters rates of suicide or suicidal behaviors [II]. Consequently, when deciding between lithium and other first-line agents for treatment of patients with bipolar disorder, the efficacy of lithium in decreasing suicidal behavior should be taken into consideration when weighing the benefits and risks of treatment with each medication. In addition, if lithium is prescribed, the potential toxicity of lithium in overdose should be taken into consideration when deciding on the quantity of lithium to give with each prescription [I].

Clozapine treatment is associated with significant decreases in rates of suicide attempts and perhaps suicide for individuals with schizophrenia and schizoaffective disorder. Thus, clozapine treatment should be given serious consideration for psychotic patients with frequent suicidal ideation, attempts, or both [I]. However, the benefits of clozapine treatment need to be weighed against the risk of adverse effects, including potentially fatal agranulocytosis and myocarditis, which has generally led clozapine to be reserved for use when psychotic symptoms have not responded to other antipsychotic medications. If treatment is indicated with an antipsychotic other than clozapine, the other second-generation antipsychotics (e.g., risperidone, olanzapine, quetiapine, ziprasidone, aripiprazole) are preferred over the first-generation antipsychotic agents [I].

ECT has established efficacy in patients with severe depressive illness, with or without psychotic features. Since ECT is associated with a rapid and robust antidepressant response as well as a rapid diminution in associated suicidal thoughts, ECT may be recommended as a treatment for severe episodes of major depression that are accompanied by suicidal thoughts or behaviors [I]. Under certain clinical circumstances, ECT may also be used to treat suicidal patients with schizophrenia, schizoaffective disorder, or mixed or manic episodes of bipolar disorder [II]. Regardless of diagnosis, ECT is especially indicated for patients with catatonic features or for whom a delay in treatment response is considered life threatening [I]. ECT may also be indicated for suicidal individuals during pregnancy and for those who have already failed to tolerate or respond to trials of medication [II]. Since there is no evidence of a long-term reduction of suicide risk with ECT, continuation or maintenance treatment with pharmacotherapy or with ECT is recommended after an acute ECT course [I].

Psychotherapies and other psychosocial interventions play an important role in the treatment of individuals with suicidal thoughts and behaviors [II]. A substantial body of evidence supports the efficacy of psychotherapy in the treatment of specific disorders, such as nonpsychotic major depressive disorder and borderline personality disorder, which are associated with increased suicide risk. For example, interpersonal psychotherapy and cognitive behavior therapy have been found to be effective in clinical trials for the treatment of depression. Therefore, psychotherapies such as interpersonal psychotherapy and cognitive behavior therapy may be considered appropriate treatments for suicidal behavior, particularly when it occurs in the context of depression [II]. In addition, cognitive behavior therapy may be used to decrease two important risk factors for suicide: hopelessness [II] and suicide attempts in depressed outpatients [III]. For patients with a diagnosis of borderline personality disorder, psychodynamic therapy and dialectical behavior therapy may be appropriate treatments for suicidal behaviors [II], because modest evidence has shown these therapies to be associated with decreased self-injurious behaviors, including suicide attempts. Although not targeted specifically to suicide or suicidal behaviors, other psychosocial treatments may also be helpful in reducing symptoms and improving functioning in individuals with psychotic disorders and in treating alcohol and other substance use disorders that are themselves associated with increased rates of suicide and suicidal behaviors [II]. For patients who have attempted suicide or engaged in self-harming behaviors without suicidal intent, specific psychosocial interventions such as rapid intervention; follow-up outreach; problem-solving therapy; brief psychological treatment; or family, couples, or group therapies may be useful despite limited evidence for their efficacy [III].