The assessment of the suicidal patient is an ongoing process that comprises many interconnected elements (Table 1). In addition, there are a number of points during patients' evaluation and treatment at which a suicide assessment may be indicated (Table 2).

The ability of the psychiatrist to connect with the patient, establish rapport, and demonstrate empathy is an important ingredient of the assessment process. For suicidal patients who are followed on an ongoing basis, the doctor-patient relationship will provide the base from which risk and protective factors continue to be identified and from which therapeutic interventions, such as psychotherapies and pharmacotherapies, are offered.

At the core of the suicide assessment, the psychiatric evaluation will provide information about the patient's history, current circumstances, and mental state and will include direct questioning about suicidal thinking and behaviors. This evaluation, in turn, will enable the psychiatrist to identify specific factors and features that may increase or decrease the potential risk for suicide or other suicidal behaviors. These factors and features may include developmental, biomedical, psychopathologic, psychodynamic, and psychosocial aspects of the patient's current presentation and history, all of which may serve as modifiable targets for both acute and ongoing interventions. Such information will also be important in addressing the patient's immediate safety, determining the most appropriate setting for treatment, and developing a multiaxial differential diagnosis that will further guide the planning of treatment.

Although the approach to the suicidal patient is common to all individuals regardless of diagnosis or clinical presentation, the breadth and depth of the psychiatric evaluation will vary with the setting of the assessment; the ability or willingness of the patient to provide information; and the availability of information from previous contacts with the patient or from other sources, including other mental health professionals, medical records, and family members. Since the approach to assessment does vary to some degree in the assessment of suicidal children and adolescents, the psychiatrist who evaluates youths may wish to review the American Academy of Child and Adolescent Psychiatry's Practice Parameter for the Assessment and Treatment of Children and Adolescents With Suicidal Behavior (4). In some circumstances, the urgency of the situation or the presence of substance intoxication may necessitate making a decision to facilitate patient safety (e.g., instituting hospitalization or one-to-one observation) before all relevant information has been obtained. Furthermore, when working with a team of other professionals, the psychiatrist may not obtain all information him- or herself but will need to provide leadership for the assessment process so that necessary information is obtained and integrated into a final assessment. Since the patient may minimize the severity or even the existence of his or her difficulties, other individuals may be valuable resources for the psychiatrist in providing information about the patient's current mental state, activities, and psychosocial crises. Such individuals may include the patient's family members and friends but may also include other physicians, other medical or mental health professionals, teachers or other school personnel, members of the patient's military command, and staff from supervised housing programs or other settings where the patient resides.

The psychiatric evaluation is the core element of the suicide risk assessment. This section provides an overview of the key aspects of the psychiatric evaluation as they relate to the assessment of patients with suicidal behaviors. Although the factors that are associated with an increased or decreased risk of suicide differ from the factors associated with an increased or decreased risk of suicide attempts, it is important to identify factors modulating the risk of any suicidal behaviors. Additional details on specific risk factors that should be identified during the assessment are discussed in Sections II.E, "Estimate Suicide Risk", and III.H, "Reassess Safety and Suicide Risk". For further discussion of other aspects of the psychiatric evaluation, the psychiatrist is referred to the American Psychiatric Association's Practice Guideline for Psychiatric Evaluation of Adults (5) (included in this volume). Additional information on details of the suicide assessment process is reviewed elsewhere (6, 7).

It is important to identify specific psychiatric signs and symptoms that are correlated with an increased risk of suicide or other suicidal behaviors. Symptoms that have been associated with suicide attempts or with suicide include aggression, violence toward others, impulsiveness, hopelessness, and agitation. Psychic anxiety, which has been defined as subjective feelings of anxiety, fearfulness, or apprehension whether or not focused on specific concerns, has also been associated with an increased risk of suicide, as have anhedonia, global insomnia, and panic attacks. In addition, identifying other psychiatric signs and symptoms (e.g., psychosis, depression) will aid in determining whether the patient has a psychiatric syndrome that should also be a focus of treatment.

A history of past suicide attempts is one of the most significant risk factors for suicide, and this risk may be increased by more serious, more frequent, or more recent attempts. Therefore, it is important for the psychiatrist to inquire about past suicide attempts and self-destructive behaviors, including specific questioning about aborted suicide attempts. Examples of the latter would include putting a gun to one's head but not firing it, driving to a bridge but not jumping, or creating a noose but not using it. For each attempt or aborted attempt, the psychiatrist should try to obtain details about the precipitants, timing, intent, and consequences as well as the attempt's medical severity. The patient's consumption of alcohol and drugs before the attempt should also be ascertained, since intoxication can facilitate impulsive suicide attempts but can also be a component of a more serious suicide plan. In understanding the issues that culminated in the suicide attempt, interpersonal aspects of the attempt should also be delineated. Examples might include the dynamic or interpersonal issues leading up to the attempt, significant persons present at the time of the attempt, persons to whom the attempt was communicated, and how the attempt was averted. It is also important to determine the patient's thoughts about the attempt, such as his or her own perception of the chosen method's lethality, ambivalence toward living, visualization of death, degree of premeditation, persistence of suicidal ideation, and reaction to the attempt. It is also helpful to inquire about past risk-taking behaviors such as unsafe sexual practices and reckless driving.

A review of the patient's treatment history is another crucial element of the suicide risk assessment. A thorough treatment history can serve as a systematic method for gaining information on comorbid diagnoses, prior hospitalizations, suicidal ideation, or previous suicide attempts. Obtaining a history of medical treatment can help in identifying medically serious suicide attempts as well as in identifying past or current medical diagnoses that may be associated with augmented suicide risk.

Many patients who are being assessed for suicidality will already be in treatment, either with other psychiatrists or mental health professionals or with primary care physicians or medical specialists. Contacts with such caregivers can provide a great deal of relevant information and help in determining a setting and/or plan for treatment. With patients who are currently in treatment, it is also important to gauge the strength and stability of the therapeutic relationships, because a positive therapeutic alliance has been suggested to be protective against suicidal behaviors. On the other hand, a patient with a suicide attempt or suicidal ideation who does not have a reliable therapeutic alliance may represent an increased risk for suicide, which would need to be addressed accordingly.

Identifying family history is particularly important during the psychiatric evaluation. The psychiatrist should specifically inquire about the presence of suicide and suicide attempts as well as a family history of any psychiatric hospitalizations or mental illness, including substance use disorders. When suicides have occurred in first-degree relatives, it is often helpful to learn more about the circumstances, including the patient's involvement and the patient's and relative's ages at the time of the suicide.

The patient's childhood and current family milieu are also relevant, since many aspects of family dysfunction may be linked to self-destructive behaviors. Such factors include a history of family conflict or separation, parental legal trouble, family substance use, domestic violence, and physical and/or sexual abuse.

An assessment of the patient's current psychosocial situation is important to detect acute psychosocial crises or chronic psychosocial stressors that may augment suicide risk (e.g., financial or legal difficulties; interpersonal conflicts or losses; stressors in gay, lesbian, or bisexual youths; housing problems; job loss; educational failure). Other significant precipitants may include perceived losses or recent or impending humiliation. An understanding of the patient's psychosocial situation is also essential in helping the patient to mobilize external supports, which can have a protective influence on suicide risk.

In estimating suicide risk and formulating a treatment plan, the clinician needs to appreciate the strengths and vulnerabilities of the individual patient. Particular strengths and vulnerabilities may include such factors as coping skills, personality traits, thinking style, and developmental and psychological needs. For example, in addition to serving as state-dependent symptoms, hopelessness, aggression, and impulsivity may also constitute traits, greater degrees of which may be associated with an increased risk for suicidal behaviors. Increased suicide risk has also been seen in individuals who exhibit thought constriction or polarized (either-or) thinking as well in individuals with closed-mindedness (i.e., a narrowed scope and intensity of interests). Perfectionism with excessively high self-expectation is another factor that has been noted in clinical practice to be a possible contributor to suicide risk. In weighing the strengths and vulnerabilities of the individual patient, it is also helpful to determine the patient's tendency to engage in risk-taking behaviors as well as the patient's past responses to stress, including the capacity for reality testing and the ability to tolerate rejection, subjective loneliness, or psychological pain when his or her unique psychological needs are not met.

In general, the more an individual has thought about suicide, has made specific plans for suicide, and intends to act on those plans, the greater will be his or her risk. Thus, as part of the suicide assessment it is essential to inquire specifically about the patient's suicidal thoughts, plans, behaviors, and intent. Although such questions will often flow naturally from discussion of the patient's current situation, this will not invariably be true. The exact wording of questions and the extent of questioning will also differ with the clinical situation. Examples of issues that the psychiatrist may wish to address in this portion of the suicide assessment are given in Table 3.

Inquiring about suicidal ideation is an essential component of the suicide assessment. Although some fear that raising the topic of suicide will "plant" the issue in the patient's mind, this is not the case. In fact, broaching the issue of suicidal ideation may be a relief for the suicidal patient by opening an avenue for discussion and giving him or her an opportunity to feel understood.

In asking about suicidal ideas, it is often helpful to begin with questions that address the patient's feelings about living, such as, "How does life seem to you at this point?" or "Have you ever felt that life was not worth living?" or "Did you ever wish you could go to sleep and just not wake up?" If the patient's response reflects dissatisfaction with life or a desire to escape it, this response can lead naturally into more specific questions about whether the patient has had thoughts of death or suicide. When such thoughts are elicited, it is important to focus on the nature, frequency, extent, and timing of them and to understand the interpersonal, situational, and symptomatic context in which they are occurring.

Even if the patient initially denies thoughts of death or suicide, the psychiatrist should consider asking additional questions. Examples might include asking about plans for the future or about recent acts or thoughts of self-harm. Regardless of the approach to the interview, not all individuals will report having suicidal ideas even when such thoughts are present. Thus, depending on the clinical circumstances, it may be important for the psychiatrist to speak with family members or friends to determine whether they have observed behavior (e.g., recent purchase of a gun) or have been privy to thoughts that suggest suicidal ideation (see Section V.C, "Communication With Significant Others"). In addition, patients who are initially interviewed when they are intoxicated with alcohol or other substances should be reassessed for suicidality once the intoxication has resolved.

If suicidal ideation is present, the psychiatrist will next probe for more detailed information about specific plans for suicide and any steps that have been taken toward enacting those plans. Although some suicidal acts can occur impulsively with little or no planning, more detailed plans are generally associated with a greater suicide risk. Violent and irreversible methods, such as firearms, jumping, and motor vehicle accidents, require particular attention. However, the patient's belief about the lethality of the method may be as important as the actual lethality of the method itself.

If the patient does not report a plan, the psychiatrist can ask whether there are certain conditions under which the patient would consider suicide (e.g., divorce, going to jail, housing loss) or whether it is likely that such a plan will be formed or acted on in the near future. If the patient reports that he or she is unlikely to act on the suicidal thoughts, the psychiatrist should determine what factors are contributing to that expectation, as such questioning can identify protective factors.

Whether or not a plan is present, if a patient has acknowledged suicidal ideation, there should be a specific inquiry about the presence or absence of a firearm in the home or workplace. It is also helpful to ask whether there have been recent changes in access to firearms or other weapons, including recent purchases or altered arrangements for storage. 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. Such discussions should be documented in the medical record, including any instructions that have been given to the patient and significant others about firearms or other weapons.

Regardless of whether the patient has developed a suicide plan, the patient's level of suicidal intent should be explored. Suicidal intent reflects the intensity of a patient's wish to die and can be assessed by determining the patient's motivation for suicide as well as the seriousness and extent of his or her aim to die, including any associated behaviors or planning for suicide. If the patient has developed a suicide plan, it is important to assess its lethality. The lethality of the plan can be ascertained through questions about the method, the patient's knowledge and skill concerning its use, and the absence of intervening persons or protective circumstances. In general, the greater and clearer the intent, the higher the risk for suicide will be. Thus, even a patient with a low-lethality suicide plan or attempt may be at high risk in the future if intentions are strong and the patient believes that the chosen method will be fatal. At the same time, a patient with low suicidal intent may still die from suicide by erroneously believing that a particular method is not lethal.

Although a number of suicide assessment scales have been developed for use in research and are described more fully in Part B of the guideline, their clinical utility is limited. Self-report rating scales may sometimes assist in opening communication with the patient about particular feelings or experiences. In addition, the content of suicide rating scales, such as the Scale for Suicide Ideation (8) and the Suicide Intent Scale (9), may be helpful to psychiatrists in developing a thorough line of questioning about suicide and suicidal behaviors. However, existing suicide assessment scales suffer from high false positive and false negative rates and have very low positive predictive values (10). As a result, such rating scales cannot substitute for thoughtful and clinically appropriate evaluation and are not recommended for clinical estimations of suicide risk.

In conceptualizing suicide risk, it is important for the psychiatrist to develop a multiaxial differential diagnosis over the course of the psychiatric evaluation. Studies have shown that more than 90% of individuals who die by suicide satisfy the criteria for one or more psychiatric disorders. Thus, the psychiatrist should determine whether a patient has a primary axis I or axis II diagnosis. Suicide and other suicidal behaviors are also more likely to occur in individuals with more than one psychiatric diagnosis. As a result, it is important to note other current or past axis I or axis II diagnoses, including those that may currently be in remission.

Identification of physical illness (axis III) is essential since such diagnoses may also be associated with an increased risk of suicide as well as with an increased risk of other suicidal behaviors. For some individuals, this increase in risk may result from increased rates of comorbid psychiatric illness or from the direct physiological effects of physical illness or its treatment. Physical illnesses may also be a source of social and/or psychological stress, which in turn may augment risk.

Also crucial in determining suicide risk is the recognition of psychosocial stressors (axis IV), which may be either acute or chronic. Certain stressors, such as sudden unemployment, interpersonal loss, social isolation, and dysfunctional relationships, can increase the likelihood of suicide attempts as well as increase the risk of suicide. At the same time, it is important to note that life events have different meanings for different individuals. Thus, in determining whether a particular stressor may confer risk for suicidal behavior, it is necessary to consider the perceived importance and meaning of the life event for the individual patient.

As the final component of the multiaxial diagnosis, the patient's baseline and current levels of functioning are important to assess (axis V). Also, the clinician should assess the relative change in the patient's level of functioning and the patient's view of and feelings about his or her functioning. Although suicidal ideation and/or suicide attempts are reflected in the Global Assessment of Functioning (GAF) scoring recommendations, it should be noted that there is no agreed-on correlation between a GAF score and level of suicide risk.

The goal of the suicide risk assessment is to identify factors that may increase or decrease a patient's level of suicide risk, to estimate an overall level of suicide risk, and to develop a treatment plan that addresses patient safety and modifiable contributors to suicide risk. The assessment is comprehensive in scope, integrating knowledge of the patient's specific risk factors; clinical history, including psychopathological development; and interaction with the clinician. The estimation of suicide risk, at the culmination of the suicide assessment, is the quintessential clinical judgment, since no study has identified one specific risk factor or set of risk factors as specifically predictive of suicide or other suicidal behavior.

Table 4 provides a list of factors that have been associated with increased suicide risk, and Table 5 lists factors that have been associated with protective effects. While risk factors are typically additive (i.e., the patient's level of risk increases with the number of risk factors), they may also interact in a synergistic fashion. For example, the combined risk associated with comorbid depression and physical illness may be greater than the sum of the risk associated with each in isolation. At the same time, certain risk factors, such as a recent suicide attempt (especially one of high lethality), access to a firearm, and the presence of a suicide note, should be considered serious in and of themselves, regardless of whether other risk factors are present.

The effect on suicide risk of some risk factors, such as particular life events or psychological strengths and vulnerabilities, will vary on an individual basis. Risk factors must also be assessed in context, as certain risk factors are more applicable to particular diagnostic groups, while others carry more general risk. Finally, it should be kept in mind that, because of the low rate of suicide in the population, only a small fraction of individuals with a particular risk factor will die from suicide.

Risk factors for suicide attempts, which overlap with but are not identical to risk factors for suicide, will also be identified in the assessment process. These factors should also be addressed in the treatment planning process, since suicide attempts themselves are associated with morbidity in addition to the added risk that they confer for suicide.

In epidemiologic studies, a number of demographic factors have been associated with increased rates of suicide. However, these demographic characteristics apply to a very broad population of people and cannot be considered alone. Instead, such demographic parameters must be considered within the context of other interacting factors that may influence individual risk.

Suicide rates differ dramatically by age. In addition, age-related psychosocial stressors and family or developmental issues may influence suicide risk. The age of the patient can also be of relevance to psychiatric diagnosis, since specific disorders vary in their typical ages of onset.

Between age 10 and 24 years, suicide rates in the general population of the United States rise sharply to approximately 13 per 100,000 in the 20- to 24-year-old age group before essentially plateauing through midlife. After age 70, rates again rise to a high of almost 20 per 100,000 in those over age 80 (Figure 1). These overall figures can be misleading, however, since the age distribution of suicide rates varies as a function of gender as well as with race and ethnicity. For example, among male African Americans and American Indians/Alaska Natives, suicide rates rise dramatically during adolescence, peak in young adulthood, and then fall through mid- and later life. Thus, in adolescence and young adulthood, the suicide rates of African American men are comparable with those of white men, although overall, African American males are half as likely to die from suicide as white males. While suicide rates in many age groups have remained relatively stable over the last 50 years, the rate among adolescents and young adults has increased dramatically, and the rate among the elderly has decreased. Among the 14- to 25-year-old age group, suicide is now the third leading cause of death, with rates that are triple those in the 1950s (12).

Figure 1. Number and Rate of Deaths by Suicide in Males and Females in the United States in 2000, by Age Groupa aIncludes deaths by suicide injury (ICD-10 codes X60–X84, Y87.0). From the Web-Based Injury Statistics Query and Reporting System, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention (11).

Add to 'My Saved Images'

Suicide rates are higher in older adults than at any other point in the life course. In 2000 in the United States there were approximately 5,300 suicides among individuals over age 65, a rate of 15.3 per 100,000. Whereas older adults made up 12.6% of the population, they accounted for 18.1% of suicides. In addition, the high suicide rate in those over age 65 is largely a reflection of the high suicide rate in white men, which reaches almost 60 per 100,000 by age 85. While rates in Asian men also increase after age 65 and rates in Asian women increase dramatically after age 80, the rate for all other women is generally flat in late life.

Thoughts of death are also more common in older than in younger adults, but paradoxically, as people age they are less likely to endorse suicidal ideation per se (13). Attempted suicide is also less frequent among persons in later life than among younger age groups (14). Whereas the ratio of attempted suicides to suicides in adolescents may be as high as 200:1, there are as few as one to four attempts for each suicide in later life (15). However, the self-destructive acts that do occur in older people are more lethal. This greater lethality is a function of several factors, including reduced physical resilience (greater physical illness burden), greater social isolation (diminished likelihood of rescue), and a greater determination to die (15). Suicidal elders give fewer warnings to others of their plans, use more violent and potentially deadly methods, and apply those methods with greater planning and resolve (15, 16). Therefore, compared with a suicide attempt in a younger person, a suicide attempt in an older person confers a higher level of future suicide risk.

In virtually all countries that report suicide statistics to the World Health Organization, suicide risk increases with age in both sexes, and rates for men in older adulthood are generally higher than those for women (17). One exception is China, where the suicide rate of women is much greater than that of men (18). In the United States, death by suicide is more frequent in men than in women, with the suicide rate in males approximately four times that in females (Figure 1). In the psychiatric population, these gender differences are also present but are less prominent. In terms of murder-suicide, the male predominance is more pronounced, with identified typologies including young men with prominent sexual jealousy and elderly men with ailing spouses (19, 20). From age 65 on, there are progressive increases in suicide rates for white men and for Asian men as well as for men overall. With the exception of high suicide rates in Asian women over age 80, women in the United States are at highest risk in midlife (11).

A number of factors may contribute to these gender differences in suicide risk (21). Men who are depressed are more likely to have comorbid alcohol and/or substance abuse problems than women, which places the men at higher risk. Men are also less likely to seek and accept help or treatment. Women, meanwhile, have factors that protect them against suicide. In addition to their lower rates of alcohol and substance abuse, women are less impulsive, more socially embedded, and more willing to seek help. Among African American women, rates of suicide are remarkably low, a fact that has been attributed to the protective factors of religion and extended kin networks (22). At the same time, women have higher rates of depression (23) and respond to unemployment with greater and longer-lasting increases in suicide rates than do men (24).

Overall, for women in the general population, pregnancy is a time of significantly reduced suicide risk (25). Women with young children in the home are also less likely to kill themselves (26). Nonetheless, women with a history of depression or suicide attempts are at greater risk for poor outcomes postpartum. Although suicide is most likely to occur in the first month after delivery, risk continues throughout the postpartum period. Teenagers, women of lower socioeconomic status, and women hospitalized with postpartum psychiatric disorders may be at particularly increased risk postpartum (27, 28).

Women tend to choose less lethal suicide methods than men do (e.g., overdose or wrist cutting versus firearms or hanging). Such differences may in part account for the reversal in the gender ratio for suicide attempters, with women being reported to attempt suicide three times as often as men (29). This female predominance among suicide attempters varies with age, however, and in older adults the ratio of women to men among suicide attempters approaches 1:1 (11, 30). Rates of suicidal ideation and attempts are also increased in individuals with borderline personality disorder and in those with a history of domestic violence or physical and/or sexual abuse, all of which are more common among women (31–36). In addition, the likelihood of suicide attempts may vary with the phase of the menstrual cycle (37, 38).

Variations in suicide rates across racial and ethnic groups have been mentioned earlier in the discussion of the influences of age and gender on suicide risk. Overall, however, in the United States, age-adjusted rates for suicide in whites and in non-Hispanic Native Americans are approximately double those observed in Hispanics, non-Hispanic African Americans, and Asian-Pacific Islanders (12.1 and 13.6 per 100,000 versus 6.1, 5.8, and 6.0 per 100,000, respectively) (11). For immigrant groups, in general, suicide rates tend to mirror the rates in the country of origin and converge toward the rate in the host country over time (39–41).

In the United States, racial and ethnic differences are also seen in the rates of suicide across the lifespan, with the highest suicide rates occurring in those over age 65 among non-Hispanic whites, Hispanics, and Asian-Pacific Islanders (11). In contrast, among Native Americans and African Americans, the highest suicide rates occur during adolescence and young adulthood (11). Such figures may be deceptive, however, since each of these groups exhibits a striking degree of heterogeneity that is rarely addressed in compilations of suicide rates.

Racial and ethnic differences in culture, religious beliefs, and societal position may influence not only the actual rates of suicide but also the views of death and suicide held by members of a particular group. For some groups, suicide can be considered a traditionally accepted way of dealing with shame, distress, and/or physical illness (42). In addition, cultural values about conveying suicidal ideas may differ; in some cultures, for example, suicidal ideation may be considered a disgraceful or private matter that should be denied. Cultural differences, particularly in immigrants and in Native Americans and Alaska Natives, may generate acculturative stresses that in turn may contribute to suicidality (43, 44). Thus, knowledge of and sensitivity to common contributors to suicide in different racial and ethnic groups as well as cultural differences in beliefs about death and views of suicide are important when making clinical estimates of suicide risk and implementing plans to address suicide risk.

Suicide risk also varies with marital status, with the suicide rate of single persons being twice that of those who are married. Divorced, separated, or widowed individuals have rates four to five times higher than married individuals (45, 46). Variations in suicide rates with marital status may reflect differing rates of baseline psychiatric illness but may also be associated with psychological or health variations. The presence of another person in the household may also serve as a protective factor by decreasing social isolation, engendering a sense of responsibility toward others, and increasing the likelihood of discovery after a suicide attempt. For women, the presence of children in the home may provide an additional protective effect (26, 47). It is also important to note that although married adults have lower rates of suicide overall, young married couples may have increased risk, and the presence of a high-conflict or violent marriage can be a precipitant rather than a protective factor for suicide.

Although no studies have examined rates of suicide among gay, lesbian, and bisexual individuals, available evidence suggests that they may have an increased risk for suicidal behaviors. Many recent studies involving diverse sample populations and research methods have consistently found that gay, lesbian, and bisexual youths have a higher risk of suicide attempts than matched heterosexual comparison groups (48–53). The female-to-male ratio for reported suicide attempts in the general population is reversed in lesbian and gay youths, with more males than females attempting suicide (48). While some risk factors leading to suicide, such as psychiatric and substance use disorders, are shared by both gay, lesbian, and bisexual youths and heterosexual youths, others are unique to being gay, lesbian, or bisexual (e.g., disclosure of sexual orientation to friends and family, experience of homophobia and harassment, and gender nonconformity). Aggressive treatment of psychiatric and substance use disorders, open and nonjudgmental support, and promotion of healthy psychosocial adjustment may help to decrease the risk for suicide in gay, lesbian, and bisexual youths and adults.

Occupational groups differ in a number of factors contributing to suicide risk. These factors include demographics (e.g., race, gender, socioeconomic class, marital status), occupational stress (54, 55), psychiatric morbidity (56), and occupationally associated opportunities for suicide (56, 57). Physicians have been consistently found to be at higher risk for suicide than persons in other occupations including professionals (57, 58). After basic demographic correlates of suicide across 32 occupations were controlled, risk was found to be highest among dentists and physicians (with multivariate logistic regression odds ratios of 5.43 and 2.31, respectively) and was also increased among nurses, social workers, artists, mathematicians, and scientists (54). Although evidence is more varied, farmers may be at somewhat higher risk, whereas risk in police officers generally does not appear to differ from that of age- and sex-matched comparison subjects (54, 57).

The presence of a psychiatric disorder is probably the most significant risk factor for suicide. Psychological autopsy studies have consistently shown that more than 90% of persons who die from suicide satisfy the criteria for one or more psychiatric disorders (59, 60). The psychological autopsy method involves a retrospective investigation of the deceased person, within several months of death, and uses psychological information gathered from personal documents; police, medical, and coroner records; and interviews with family members, friends, co-workers, school associates, and health care providers to classify equivocal deaths or establish diagnoses that were likely present at the time of suicide (61–63).

In addition to there being high rates of psychiatric disorder among persons who die by suicide, almost all psychiatric disorders with the exception of mental retardation have been shown to increase suicide risk as measured by standardized mortality ratios (SMRs) (64) (Table 6). An SMR reflects the relative mortality from suicide in individuals with a particular risk factor, compared with the general population. Thus, the SMR will be equal to 1.0 when the number of observed suicide deaths is equivalent to the number of expected deaths by suicide in an age- and sex-matched group in the general population. Values of the SMR for suicide that are greater than 1.0 indicate an increased risk of suicide, whereas values less than 1.0 indicate a decreased risk (i.e., a protective effect). It is also important to note that SMRs do not correspond precisely to the incidence or prevalence of suicide and may vary in their reliability depending on the number of suicides in the sample, the time period of the study, and the representativeness of the study population. Thus, SMRs should be viewed as estimates of relative risk and not as reflections of absolute risk for individuals with a particular disorder. It is equally necessary to appreciate distinctions in risk across disorders and variations in risk at differing points in the illness course in the effort to differentiate high-risk patients within an overall at-risk population identified in terms of standardized mortality.

Study after study has confirmed that the presence of a major mood disorder is a significant risk factor for suicide. Not surprisingly, mood disorders, primarily in depressive phases, are the diagnoses most often found in suicide deaths (59, 65–67). Although most suicides in individuals with bipolar disorder occur during depressive episodes, mixed episodes are also associated with increased risk (68–70). Suicidal ideation and attempts are also more common during mixed episodes than in mania (71).

When viewed from the standpoint of lifetime risk, mood disorders are associated with an increased risk of mortality that has been estimated to range from a 12-fold increase in risk with dysthymia to a 20-fold increase in risk with major depression (64). Lifetime suicide risk in bipolar disorder has generally been found to be similar to that in unipolar major depression (69, 72). However, several longitudinal studies of patients followed after an index hospitalization have demonstrated suicide risks in patients with major depressive disorder that are greater than those in patients with either bipolar I disorder or bipolar II disorder (73–75).

Particularly for younger patients, suicides are more likely to occur early in the course of illness (68, 73, 75, 76). Nonetheless, risk persists throughout life in major depressive disorder as well as in bipolar disorder (73, 74). Suicide risk also increases in a graduated fashion with illness severity as reflected by the level of required treatment. Lifetime suicide rates in psychiatric outpatients ranged from 0.7% for those without an affective disorder to 2.2% for those with affective disorders, whereas lifetime suicide rates for individuals requiring hospitalization ranged from 4% for those whose admission for depression was not prompted by suicidal behavior or risk to 8.6% for those whose admission was the result of suicidality (77). Illness severity may also be an indicator of risk for suicide attempts (75, 78).

Among patients with mood disorders, lifetime risk also depends on the presence of other psychiatric symptoms or behaviors, some of which are modifiable with treatment. For example, patients with mood disorders who died by suicide within 1 year of initial evaluation were more likely to have panic attacks, severe psychic anxiety, diminished concentration, global insomnia, moderate alcohol abuse, and severe loss of pleasure or interest in activities (79). At later time points, hopelessness has been associated with increased suicide risk in mood disorder patients (78, 79). Suicidal ideation and a history of suicide attempts also augment risk (74, 79). Comorbid anxiety, alcohol use, and substance use are common in patients with mood disorders and may also increase the risk of suicide and suicide attempts (see Sections II.E.2.f, "Alcohol Use Disorders" and II.E.2.g, "Other Substance Use Disorders"). Although a greater risk for suicide or suicidal behaviors among patients with psychotic mood disorders has been seen in some studies (80–83), this relationship has not been found in other studies (84–88).

Compared to the risk in the general population, the risk of suicide in patients with schizophrenia is estimated to be about 8.5-fold higher (64), with even greater increments in risk in patients who have been hospitalized (89). Although earlier research suggested a 10%–15% lifetime risk of suicide among patients with schizophrenia (90–93), such estimates were likely inflated by biases in the patient populations and length of follow-up. More recent estimates suggest a lifetime risk of suicide in schizophrenia of about 4% (94).

Suicide may occur more frequently during the early years of the illness, with the time immediately after hospital discharge being a period of heightened risk (83, 89, 90, 95–98). However, risk continues throughout life (99, 100) and appears to be increased in those with a chronic illness course (83, 89, 101), multiple psychiatric hospitalizations (89, 95), or a previous suicide attempt (89, 90, 95, 100). Other consistently identified factors that confer an increased risk of suicide in patients with schizophrenia include male sex (83, 89, 90, 95, 102, 103), younger age (<30 years) (83, 90, 102), and social isolation (97, 104).

In individuals with schizophrenia or schizoaffective disorder, psychotic symptoms are often present at the time of a suicide attempt or suicide (105–107). However, command hallucinations seem to account for a relatively small percentage of suicides, and there is limited evidence on whether they increase suicide risk. Nonetheless, they may act as a precipitant to a suicide attempt or to suicide in some individuals (106, 108) (see Section II.E.3.c, "Command Hallucinations"). Suicide in patients with schizophrenia may be more likely to occur during periods of improvement after relapse or during periods of depressed mood (83, 89, 90, 95, 100, 109–111), including what has been termed postpsychotic depression (112, 113). Also, patients with schizoaffective disorder appear to be at greater risk for suicide than those with schizophrenia (114).

Suicide risk may paradoxically be increased in those who have insight into the implications of having a schizophrenic illness, particularly if this insight is coupled with a feeling of hopelessness. Suicide risk is also increased in those who recognize a loss of previous abilities and are pessimistic about the benefits of treatment in restoring those abilities (93, 101, 115). This pattern is consistent with the increased risk of suicide observed in individuals with schizophrenia who had a history of good premorbid and intellectual functioning (83, 89, 103) as well as with the decreased risk of suicide in patients with prominent negative symptoms (83, 89, 103, 116).

Suicidal ideation and suicide attempts are common among individuals with schizophrenia and need to be identified and addressed in the assessment process. In series of hospitalized or longitudinally followed patients with schizophrenia, 40%–53% reported having suicidal ideation at some point in their lives and 23%–55% reported prior suicide attempts (80, 93, 108, 117). For individuals with schizoaffective disorder, these figures are likely to be even higher (80). Patients often reported that suicide attempts were precipitated by depression, stressors, or psychotic symptoms (108). In addition, suicide attempts among individuals with schizophrenia or schizoaffective disorder were often medically serious and associated with a high degree of intent (108), both of which would confer greater future risk for suicide.

Although studies of lifetime suicide risk in anxiety disorders are more limited than for mood disorders, evidence suggests anxiety disorders are associated with a six- to 10-fold increase in suicide risk (64, 118, 119). Among persons who die from suicide, rates of anxiety disorders appear to be lower than rates of mood disorders, with one psychological autopsy study identifying an anxiety disorder in only 11% of persons who died from suicide (120). However, the prevalence of anxiety disorders may be underestimated because of the masking of anxiety by affective disorders and by alcohol use (121).

Of the anxiety disorders, panic disorder has been studied in the most detail. In psychological autopsy studies, panic disorder is present in about 1% of persons who die from suicide (120), whereas other studies of panic disorder show an SMR for suicide that is about 10 times that of the general population (64). As with anxiety disorders in general, comorbid depression, alcohol use, or axis II disorders are often present in individuals with panic disorder who die by suicide (122, 123).

Suicidal ideation and suicide attempts are common in individuals with anxiety disorders, but their rates vary with the patient population and with the presence of comorbid diagnoses. In panic disorder, for example, reported rates of prior suicide attempts range from 0% to 42% (124–129). In other anxiety disorders, the relative risks of suicidal ideation and suicide attempts also appear to be increased (118, 130). In addition, in patients with major depression, the presence of a comorbid anxiety disorder appears to increase the risk of suicidal ideation or suicide attempts (131, 132). Less clear, however, is whether anxiety disorders are associated with an increased risk for suicide and other suicidal behaviors in the absence of comorbid diagnoses (130, 132–136) or whether the observed increases in risk can be accounted for solely on the basis of comorbid disorders (127, 137). Nonetheless, suicide risk may be diminished by identifying masked anxiety symptoms and anxiety disorders that are misdiagnosed as medical illness as well as by explicitly assessing and treating comorbid psychiatric diagnoses in individuals with anxiety disorders.

Eating disorders, particularly anorexia nervosa, are a likely risk factor for suicide as well as being associated with an increased risk of mortality in general (64, 138, 139). Exact risk is difficult to determine, however, as data on rates of suicide in eating disorders may be subject to underreporting bias (140). Suicide attempts are also common, particularly in individuals with bingeing and purging behaviors and in those with comorbid mood disorders, aggression, or impulsivity (141, 142). Conversely, suicide attempters may have increased rates of abnormal eating behaviors (142). The role of comorbid diagnoses in increasing the risk of suicidal behaviors remains to be delineated. It is also not clear whether the self-imposed morbidity and mortality associated with severe caloric restriction or bingeing and purging should be viewed as a self-injurious or suicidal behavior. Regardless, clinicians conducting a suicide risk assessment should be attentive to the presence of eating disorders and especially the co-occurrence of eating disorders with behaviors or symptoms such as deliberate self-harm or depression.

The relationship between attention deficit hyperactivity disorder (ADHD) and suicidal behavior is unclear, with some studies indicating an association between the diagnosis of ADHD and suicide attempts or completions (143, 144) and other studies indicating no such connection (145, 146). However, individuals with ADHD, combined type, may be at greater risk than those with ADHD, inattentive type, perhaps because of an increased level of impulsivity in the combined type of the disorder (144). In addition, the presence of ADHD may increase suicide risk through comorbidity with conduct disorder, substance abuse, and/or depressive disorder (143).

Alcoholism is associated with an increased risk for suicide, with suicide mortality rates for alcoholics that are approximately six times those of the general population (64, 94). In fact, abuse of substances including alcohol may be the second most frequent psychiatric precursor to suicide (147). Although suicide rates among alcoholics are higher in Europe and older literature indicated a lifetime risk for suicide in the 11%–15% range, recent literature suggests the lifetime risk of suicide among alcoholics in the United States is as low as 3.4% (148). In addition, in psychological autopsy studies, alcohol abuse or dependence is present in 25%–50% of those who died by suicide (59, 149–151).

Several factors, including recent or impending interpersonal losses and comorbid psychiatric disorders, have been specifically linked to suicide in alcoholic individuals. The loss or disruption of a close interpersonal relationship or the threatened loss of such a relationship may be both a consequence of alcohol-related behavior and a precipitant to suicide (110, 152–154). Suicide is also more likely to occur among alcoholics who suffer from depressive episodes than in persons with major depression or alcoholism alone. In addition, studies have found major depressive episodes in half to three-fourths of alcoholics who die by suicide (67, 120, 149, 152, 155–157). As a result, psychiatrists should systematically rule out the presence of a comorbid depressive disorder and not simply assume that depressive symptoms result from alcohol use or its psychosocial consequences.

Whereas full-time employment appears to be a protective factor in alcoholics, factors that increase suicide risk include communications of suicidal intent, prior suicide attempts, continued or heavier drinking, recent unemployment, living alone, poor social support, legal and financial difficulties, serious medical illness, other psychiatric disorders, personality disturbance, and other substance use (64, 149, 152, 154, 156, 158, 159). In terms of gender, alcoholic men are more likely to die by suicide, but female alcoholics appear to have a greater standardized mortality due to suicide than men (64), indicating an increased risk of suicide in alcoholics regardless of gender. While the likelihood of a suicidal outcome increases with the total number of risk factors (149, 160), not all of these factors suggest an immediate risk. In fact, in contrast to suicide in depressed and schizophrenic patients, suicide in alcoholics appears to be a relatively late sequela of the disease (161), with communications of suicidal intent usually being of several years' duration and health, economic, and social functioning showing a gradual deterioration (149).

In addition to being associated with an increased risk of suicide, alcohol use disorders are associated with a greater likelihood of suicide attempts (162, 163). For suicide attempts among alcoholics, greater rates are associated with female sex, younger age, lower economic status, early onset of heavy drinking and alcohol-related problems, consumption of greater amounts of alcohol when drinking, and having a first- or second-degree relative who abused alcohol (164–167). The risk of suicide attempts among alcoholics is also increased by the presence of a comorbid psychiatric diagnosis, particularly major depression, other substance use disorders, antisocial personality disorder, or an anxiety disorder (165–171).

Thus, individuals with alcohol use disorders are at increased risk for suicide attempts as well as for suicide. Family histories of alcoholism and comorbid psychiatric disorders, particularly mood disorders and other substance use disorders, are frequent in alcoholics who die by suicide and who attempt suicide. Interpersonal loss and other adverse life events are commonly noted to precede suicide in alcoholics. These factors may act as precipitants, or, conversely, alcohol use disorders may have a deteriorating effect on the lives of alcoholics and culminate in suicide. Together, however, these findings suggest the need to identify and address comorbid psychiatric diagnoses, family history, and psychosocial factors, including recent interpersonal losses, as part of the suicide assessment of persons with alcohol use disorders.

Although the role of alcoholism in suicide has been widely studied and recognized, abuse of other substances is also associated with increased rates of suicide (172). Substance use disorders are particularly common among adolescents and young adults who die by suicide (110, 145, 173, 174). In fact, it has been suggested that the spread of substance abuse may have contributed to the two- to fourfold increase in youth suicide since 1970 (147). For many individuals, substance abuse and alcoholism are co-occurring, making it difficult to distinguish the contributions of each to rates of suicide (153, 172, 173). In addition, other comorbid psychiatric disorders, particularly mood disorders and personality disorders, may add to the risk of suicide in patients with substance use disorders (145, 173–175).

Substance use disorders also seem to make an independent contribution to the likelihood of making a suicide attempt (176, 177). In addition, a history of suicide attempts is common among individuals with substance use disorders (31, 178–180). Even after other factors, including comorbid psychiatric disorders and demographic characteristics, are controlled, it is the number of substances used, rather than the type of substance, that appears to be important (176). As with suicide in individuals with alcohol use disorders, the loss of a significant personal relationship is a common precipitant for a suicide attempt (179). Suicide attempts are also more likely in individuals with substance abuse who also have higher childhood trauma scores for emotional neglect (180, 181). Moreover, a substance use disorder may complicate mood disorders (182), increasing susceptibility to treatment resistance, increasing psychological impairment, and contributing to an elevated risk for suicide attempts. Thus, it is important to identify patterns of substance use during the psychiatric evaluation and to note comorbid psychiatric diagnoses or psychosocial factors that may also affect the likelihood of suicidal behaviors among individuals with substance use disorders.

Diagnoses of personality disorders have been associated with an increased risk for suicide, with estimated lifetime rates of suicide ranging from 3% to 9% (183–185). Compared with the general population, individuals with personality disorders have an estimated risk for suicide that is about seven times greater (64). Specific increases in suicide risk have been associated with borderline and antisocial personality disorders, with possible increases in risk associated with avoidant and schizoid personality disorders (186). Psychological autopsy studies have shown personality disorders to be present in approximately one-third of those who die by suicide (174, 183, 186, 187). Among psychiatric outpatients, personality disorders are present in about one-half of patients who die by suicide (78, 188).

In individuals with personality disorders, suicide risk may also be increased by a number of other factors, including unemployment, financial difficulty, family discord, and other interpersonal conflicts or loss (189, 190). In individuals with borderline personality disorder, in particular, impulsivity may also increase suicide risk (185).

Although comorbid diagnoses do not account for the full increase in suicide risk with personality disorders (184, 185), comorbid diagnoses are frequent and augment suicide risk. In fact, for individuals with personality disorders, concurrent depressive symptoms or substance use disorders are seen in nearly all individuals who die by suicide (187).

Among individuals who attempt suicide, diagnoses of personality disorders are also common, with overall rates of about 40% (31, 177, 184). Individuals with personality disorders tend to attempt suicide more often than individuals with other diagnoses (191–193), with 40%–90% of individuals with personality disorders making a suicide attempt during their lifetime (184). Comorbid psychiatric diagnoses, including mood disorders and substance use disorders, are quite prevalent among suicide attempters with personality disorders and independently contribute to risk (131, 184, 191, 192, 194, 195). Impulsivity has also been shown to increase the risk of suicide attempts in some (196, 197) but not all studies (191). Rates of suicide attempts in those with personality disorder may also vary with treatment setting, with greater risk in individuals who are receiving acute inpatient treatment (198).

Of personality disorder diagnoses, borderline personality disorder and antisocial personality disorder confer an added risk of suicide attempts (31, 177, 191, 193). In individuals with borderline personality disorder, there is some evidence of increased risk being associated with the number and severity of symptoms (195). Among female suicide attempters, rates of borderline personality disorder are higher than among male suicide attempters (199, 200). These findings suggest that personality disorders, particularly borderline personality disorder and antisocial personality disorder, should be identified and addressed as part of the suicide assessment process.

As discussed in preceding sections, comorbid psychiatric diagnoses (most commonly, major depression, borderline and antisocial personality disorders, and alcohol and other substance use disorders) increase suicide risk and are often present in individuals who die by suicide (13, 59, 120, 174, 201). Comorbid medical diagnoses may also increase suicide risk, as will be discussed in Section II.E.5, "Physical Illness". In general, the greater the number of comorbid diagnoses that are present, the greater will be the increase in risk. Furthermore, even in the absence of a formal comorbid diagnosis, suicide is more likely to occur when there are high levels of additional psychiatric symptoms (67, 185, 202–204).

In patients with a mood disorder, either bipolar disorder or major depression, the risk of suicide is particularly increased in the presence of comorbid alcohol or substance use (68, 205–207), with some studies suggesting that males are at additional risk (68, 205). Comorbid alcohol use may also increase suicide risk in patients with schizophrenia (107). In addition, suicide in schizophrenia may be more likely to occur during periods of depression (83, 90, 109–113). In anxiety disorders and particularly in panic disorder, individuals who die by suicide often have experienced comorbid depression, alcohol use, or axis II disorders (122, 123). Similarly, when suicide occurs in individuals with eating disorders, it is often associated with a comorbid mood disorder or substance use disorder (138).

For individuals with alcohol use disorders, major depression is found in half to three-fourths of individuals who die by suicide (67, 120, 149, 152, 155–157), and alcoholics who suffer from depressive episodes are more likely to die from suicide than persons with major depression or alcoholism alone. Serious medical illness and other psychiatric disorders, including personality disturbance and other substance use disorders, also increase suicide risk in alcoholics (64, 149, 152, 154, 156, 158, 159). For many individuals, substance abuse and alcoholism are co-occurring, making it difficult to distinguish the contributions of each to rates of suicide (153, 172, 173, 208). Furthermore, it appears to be the number of substances used, rather than the specific substance, that determines risk (176).

Individuals who die by suicide and who abuse or are dependent on substances other than alcohol are typically adolescents or young adults. Comorbid mood disorders are commonly seen in both males and females (66, 145, 204). In addition, borderline personality disorder is relatively frequent in females with substance use disorders (175), whereas young males with substance use disorders who die by suicide more commonly have comorbid antisocial personality disorder (120, 159, 173, 204). The presence of ADHD may increase suicide risk through comorbidity with conduct disorder, substance abuse, and/or depressive disorder (143). For individuals with personality disorders, concurrent depressive symptoms or substance use disorders augment suicide risk (184, 185, 209, 210) and are seen in nearly all suicides (187).

Comorbid diagnoses are also essential to identify and address because of their role in increasing the risk of suicide attempts (199). Furthermore, the likelihood of a suicide attempt appears to increase with an increasing number of comorbid diagnoses (166, 176, 177, 211). In addition, the number and severity of symptoms may play a role in increasing risk, regardless of whether the full criteria for a separate diagnosis are met. The specific comorbid disorders that augment the risk of suicide attempts are similar to those that are commonly seen to augment the risk of suicide and include comorbid depression (129, 131, 193, 195, 197, 211), alcohol and other substance use disorders (31, 129, 167, 168, 170, 180, 182, 191, 199, 211–214), anxiety disorders (127, 130–135, 137, 211, 215), and personality disorders (184, 191), particularly borderline personality disorder (31, 195, 200) and antisocial personality disorder (165, 204, 216). Thus, given the evidence that comorbidity increases the risks for suicide and for suicide attempts, the suicide risk assessment should give strong consideration to all current and previous psychiatric diagnoses.

Anxiety appears to increase the risk for suicide (79, 217, 218). Specifically implicated has been severe psychic anxiety consisting of subjective feelings of fearfulness or apprehension, whether or not the feelings are focused on specific concerns. Clinical observation suggests that anxious patients may be more inclined to act on suicidal impulses than individuals whose depressive symptoms include psychomotor slowing. Studies of suicide in patients with affective disorders have shown that those who died by suicide within the first year after contact were more likely to have severe psychic anxiety or panic attacks (79, 219). In an inpatient sample, severe anxiety, agitation, or both were found in four-fifths of patients in the week preceding suicide (218). Similar associations of anxiety with suicide attempts have been noted in some (212) but not all (220) studies. Since severe anxiety does seem to increase suicide risk, at least in some subgroups of patients, anxiety should be viewed as an often hidden but potentially modifiable risk factor for suicide (109). Once identified, symptoms of anxiety can be addressed with psychotherapeutic approaches and can also respond rapidly to aggressive short-term treatment with benzodiazepines, second-generation antipsychotic medications, and possibly anticonvulsant medications.

Hopelessness is well established as a psychological dimension that is associated with increased suicide risk (10, 78, 79, 217, 221–223). Hopelessness may vary in degree from having a negative expectation for the future to being devoid of hope and despairing for the future. In general, patients with high levels of hopelessness have an increased risk for future suicide (78, 221–225). However, among patients with alcohol use disorders, the presence of hopelessness may not confer additional risk (226, 227). For patients with depression, hopelessness has been suggested to be the factor that explains why some patients choose suicide, whereas others do not (222). Hopelessness also contributes to an increased likelihood of suicidal ideation (192, 228) and suicide attempts (197, 212, 229–231) as well as an increased level of suicidal intent (197, 232, 233).

Hopelessness often occurs in concert with depression as a "state-dependent" characteristic, but some individuals experience hopelessness on a primary and more enduring basis (221). High baseline levels of hopelessness have also been associated with an increased likelihood of suicidal behaviors (234). However, patients experiencing similar levels of depression may have differing levels of hopelessness (222), and this difference, in turn, may affect their likelihood of developing suicidal thoughts (228). Whatever the source or conceptualization of hopelessness, interventions that reduce hopelessness may be able to reduce the potential for suicide (10, 222, 235–237).

Command hallucinations, which order patients to carry out tasks or actions, can occur in individuals with psychotic disorders, primarily schizophrenia (238). Evidence for the association of command hallucinations with suicide is extremely limited (102, 239). The presence of auditory command hallucinations in inpatients does not appear to increase the likelihood of assaultiveness or of suicidal ideation or behavior over that associated with auditory hallucinations alone (240). Furthermore, in patients who do experience auditory command hallucinations, reported rates of compliance with commands vary widely from 40% to 84% (106, 241–244). Variables that have been associated with a propensity to obey command hallucinations include being able to identify the hallucinatory voice, having more severe psychotic disturbance, having a less dangerous command, and experiencing the commands for the first time or outside of a hospital environment (241, 242, 245). Thus, at least for some individuals, suicidal behaviors can occur in response to hallucinated commands, and individuals with prior suicide attempts may be particularly susceptible (106). Consequently, in the psychiatric evaluation, it is important to identify auditory command hallucinations, assess them in the context of other clinical features, and address them as part of the treatment planning process.

Impulsivity, hostility, and aggression may act individually or together to increase suicide risk. For example, many studies provide moderately strong evidence for the roles of impulsivity and hostility-related affects and behavior in suicide across diagnostic groups (89, 217, 246–248). Multiple other studies have also demonstrated increased levels of impulsivity and aggression in individuals with a history of attempted suicide (31, 193, 197, 212, 220, 249–252). Many patients with borderline personality disorder exhibit self-mutilating behaviors, and, overall, such behaviors are associated with increased impulsivity (251). However, for many self-mutilating patients, these behaviors are premeditated rather than impulsive (253). Consequently, self-mutilatory behaviors alone should not be regarded as an indicator of high impulsivity. Moreover, measures of aggression and impulsivity are not highly correlated (253), making aggression a poor marker of impulsivity as well. Thus, impulsivity, hostility, aggression, and self-mutilating behaviors should be considered independently in the psychiatric evaluation as well as in estimating suicide risk.

In addition to the increased suicide risk conferred by alcohol abuse or dependence, intoxication itself appears to play a role in alcoholic as well as nonalcoholic populations (254). Autopsies have found alcohol to be present in 20%–50% of all persons who die by suicide (121, 255). Those who consume alcohol before suicide are more likely to have experienced a recent breakup of an interpersonal relationship but less likely to have sought help before death (255). They are also more likely to have chosen a firearm as a suicide method (151, 256, 257). Alcohol intoxication at the time of suicide may also be more common in younger individuals (154, 255, 258), in men (121, 255), and in individuals without any lifetime history of psychiatric treatment (154). Among suicide attempters who later died by suicide, alcohol appeared to contribute to death in more than a third (259). In addition, a study of the interaction of employment and weekly patterns of suicide emphasizes the role of intoxication in suicides and indicates that employment may be a stabilizing factor that curbs heavy drinking during the work week (260), thereby decreasing rates of suicide. Consequently, in some subsets of patients, alcohol consumption appears to contribute to the decision to die by suicide (255).

Alcohol use is also a common prelude to suicide attempts (258). Some estimates show that more than 50% of individuals have used alcohol just before their suicide attempt. Among alcoholics, heavier drinking adds to risk (64, 149, 165). Suicide attempts that involve alcohol are more likely to be impulsive (258). Indeed, the majority of acutely intoxicated alcoholics either did not remember the reason for their attempt or had done it on a sudden impulse (258). Thus, alcohol consumption may make intervention more difficult by simultaneously limiting the communication of intent (255, 261), increasing impulsivity, decreasing inhibition, and impairing judgment (262).

Alcohol use in conjunction with attempted suicide is more common in men than in women (258), although among younger attempters, females may be more likely than males to consume alcohol (258). Alcohol use in conjunction with a suicide attempt has also been associated with repeated suicide attempts and future suicide (263). In some individuals, intentionally drinking to overcome ambivalence about suicide may signify serious suicidal intent. Thus, since intoxication is a risk factor for suicide attempts as well as for suicide, the clinician should inquire about a patient's drinking habits and consider the effect of alcohol intoxication when estimating suicide risk.

Individuals who have made a suicide attempt constitute a distinct but overlapping population with those who die by suicide. As with individuals who die by suicide, a high preponderance of suicide attempters have one or more axis I or II diagnoses, with major depression and alcohol dependence observed most commonly for axis I and borderline personality disorder observed most commonly for axis II (199, 200, 264). However, suicide attempts are about 10–20 times more prevalent than suicide (265), with lifetime prevalence ranging from 0.7% to 6% per 100,000 in a random sample of U.S. adults (2). Although a substantial percentage of individuals will die on their initial suicide attempt (266), a past suicide attempt is one of the major risk factors for future suicide attempts (164, 267) and for future suicide (64, 78, 79, 266, 268–271).

After a suicide attempt, there can be significant mortality from both natural and unnatural causes (259, 272). A suicide attempt by any method is associated with a 38-fold increase in suicide risk, a rate that is higher than that associated with any psychiatric disorder (64). Depending on the length of the follow-up, from 6% to 27.5% of those who attempt suicide will ultimately die by suicide (64, 273), and similar results have been suggested for acts of deliberate self-harm (274). Some studies have found that suicide risk appears to be particularly high during the first year after a suicide attempt (259, 275). An additional increase in risk may be associated with aborted suicide attempts (276, 277) or repeated suicide attempts (64, 259, 263, 272, 274, 278). Thus, the increase in suicide mortality subsequent to attempted suicide emphasizes the need for aftercare planning in this heterogeneous population.

In the context of a suicide attempt, a number of other factors are associated with increases in suicide risk. For example, risk is augmented by medical and psychiatric comorbidity, particularly comorbid depression, alcohol abuse, or a long-standing medical illness (64). Low levels of social cohesion may also increase risk (64). Risk of later suicide in males, particularly younger males, appears to be two to four times greater than that in females after a suicide attempt (275). In addition, serious suicide attempts are associated with a higher risk of eventual suicide, as are having high intent (164), taking measures to avoid discovery, and using more lethal methods that resulted in physical injuries (263).

Given this increased likelihood of additional suicide attempts and suicide deaths after a suicide attempt or aborted suicide attempt, psychiatric evaluation should be incorporated into emergency medical assessments of suicide attempters (279) and the importance of follow-up should be emphasized (2, 280).

A history of childhood abuse has been associated with increased rates of suicidal behaviors in multiple studies. Rates of suicide in individuals with a history of childhood abuse have not been widely studied, but available evidence suggests that suicide rates are increased at least 10-fold in those with a history of childhood abuse (36). In addition, a number of studies have demonstrated that individuals with a history of childhood abuse have an increased risk of suicide attempts (230, 281–283), suggesting that risk of later suicide will also be increased. Rates of suicide attempts are increased in individuals who report experiencing childhood physical abuse (196, 250, 284–290) as well as in individuals who report experiencing childhood sexual abuse (33, 35, 36, 164, 196, 250, 284, 285, 288–294). Rates of suicidal ideation are similarly increased in individuals with a childhood history of abuse (284).

Since many traumatized individuals have experienced both sexual and physical abuse during childhood, it is often difficult to establish the specific contributions of each form of abuse to the risk of suicide and other suicidal behaviors. In addition, the duration and severity of childhood abuse vary across individuals and can also influence risk. It appears, however, that the risk of suicide attempts is greater in individuals who have experienced both physical and sexual abuse in childhood (288) and that greater levels of risk are associated with increasing abuse severity (285, 286, 291).

Childhood trauma can also be associated with increased self-injurious behaviors, including self-cutting and self-mutilation, without associated suicidal intent. Sexual abuse may be a particular risk factor for such behaviors, which can often become repetitive (164). Indeed, deliberate self-harm is common in patients with posttraumatic stress disorder and other traumatic disorders and serves to reduce internal tension and provide nonverbal communication about their self-hate and intense distress (295). As a result, inquiring about the motivations of self-injurious behavior may help to inform estimates of suicide risk.

Gender may also influence the risk of suicidal behaviors in those with a history of childhood abuse. This influence, in part, relates to differences in the prevalence of childhood abuse between men and women, with rates of childhood physical abuse being higher in men and rates of childhood sexual abuse being higher in women (288). However, in individuals who have a history of childhood sexual abuse, the risk of a suicide attempt may be greater in men than in women (33).

Given the significant rates of childhood physical and/or sexual abuse, particularly among psychiatric patient populations (35, 284, 288, 292), and the increased risk for suicidal behaviors that such abuse confers, it important to assess for a history of physical abuse and sexual abuse as part of the psychiatric evaluation. In addition, the duration and severity of childhood abuse should be determined, as these factors will also influence risk.

Domestic partner violence has been associated with increased rates of suicide attempts and suicidal ideation; however, there is no information about its effects on risk for suicide per se. The risk for suicide attempts in individuals who have experienced recent domestic partner violence has been estimated to be four- to eightfold greater than the risk for individuals without such experiences (34, 296–300). Conversely, among women presenting with suicide attempts, there is a severalfold increase in their risk for experiencing domestic partner violence (230, 301).

Although much more commonly experienced by women, domestic partner violence is also experienced by men and can increase their risk for suicide attempts (302). Men with a history of domestic violence toward their partners may also be at increased risk for suicide (303). Furthermore, domestic violence in the home may increase the risk for suicide attempts among children who are witnesses to such violence (281).

Given the clear increase in risk for suicide attempts in individuals experiencing domestic partner violence and the likely association of suicide attempts with an increased risk for suicide, it is important to specifically ask about domestic partner violence as a part of the suicide assessment. Such inquiry may also help to identify individuals in addition to the identified patient who may be at increased risk for suicidal behaviors.

Multiple studies have shown that greater treatment intensity is associated with greater rates of eventual suicide (64, 77, 198). Although hospitalization generally occurs because a patient has a more severe illness and is deemed to be at increased risk for suicide, for some patients, hospitalization could conceivably result in increased distress and thus an increase in suicide risk. Thus, as a general rule, a past history of treatment, including a past history of hospitalization, should be viewed as a marker that alerts the clinician to increased suicide risk.

Temporally, the risk for suicide appears to be greatest after changes in treatment setting or intensity (304), with recently admitted and recently discharged inpatients showing increased risk (64, 72, 91, 95, 305–308). This increase in rates of suicide after hospital discharge is seen across diagnostic categories and has been observed in individuals with major depressive disorder, bipolar disorder, schizophrenia, and borderline personality disorder. Rates decline with time since discharge but may remain high for as long as several years (91, 306, 309). Similar findings are seen with suicide attempts, which are also more frequent in the period after hospitalization (267, 305, 308). These observations suggest a need for close follow-up during the period immediately after discharge.

In some psychiatric disorders, suicide risk is greater at certain points in the illness or episode course. For example, in the course of major depressive disorder, suicidality tends to occur early, often before a diagnosis has been made or treatment has begun (304, 310–312). In patients with major depressive disorder (73, 313), as well as in those with bipolar disorder (73, 74, 305) or schizophrenia (83), suicide has been noted to be more likely during the first few episodes, early in the illness (314, 315). After a suicide attempt, the risk for suicide is also greatest initially, with most suicides occurring in the first year after the attempt (275). Although risks of suicide and suicide attempts later in the illness course are less than they are earlier on, these risks remain greater than those for the general population (74, 100, 316–318). These findings highlight the need for early identification of these disorders and for therapeutic approaches that will treat the illness while simultaneously promoting longer-term treatment adherence.

Risk may also vary with severity of symptoms. For example, higher levels of depression have been associated with increased risk of suicide in at least one study (319), whereas greater numbers of symptoms of borderline personality disorder have been associated with an increased risk for suicide attempts (195). In addition, higher levels of suicidal ideation and subjective hopelessness also increase risk for suicide (78) and suicide attempts (31). In contrast, higher levels of negative symptoms have been associated with decreased suicide risk in individuals with schizophrenia (320). It is also important to recognize that other factors such as age will modulate the effects of symptom severity on risk. With older adults, for example, milder symptoms may be associated with greater risk than moderate symptoms in younger adults (207, 321). Consequently, clinicians should consider the severity of a patient's illness and psychiatric symptoms in the context of other patient-specific factors when assessing suicide risk.

Identification of medical illness (axis III) is also an essential part of the assessment process. Such diagnoses will need to be considered in developing a plan of treatment, and they may influence suicide risk in several ways. First, specific medical disorders may themselves be associated with an increased risk for suicide. Alternatively, the physiological effects of illness or its treatment may lead to the development of psychiatric syndromes such as depression, which may also increase suicide risk. Physical illnesses are also a source of social and/or psychological stress, which in turn augments risk. Physical illnesses such as hepatitis C or sexually transmitted diseases may signal an increased likelihood of impulsive behaviors or comorbid substance use disorders that may in turn be associated with greater risk for suicidal behaviors. Finally, when physical illness is present, psychiatric signs and symptoms may be ascribed to comorbid medical conditions, delaying recognition and treatment of the psychiatric disorder.

Data from clinical cohort and record linkage studies indicate clearly that medical illness is associated with increased likelihood of suicide (Table 7). Not surprisingly, disorders of the nervous system are associated with an elevated risk for suicide. The association between seizure disorders and increased suicide risk is particularly strong and consistently observed (64, 322–328). Presumably because of its close association with impulsivity, mood disorders, and psychosis, temporal lobe epilepsy is associated with increased risk in most (322, 327, 328) but not all (325) studies. Suicide attempts are also more common among individuals with epilepsy (329–331). Other neurological disorders that are associated with increased risk for suicide include multiple sclerosis, Huntington's disease, and brain and spinal cord injury (25, 323, 332–334).

Other medical disorders that have also been associated with an increased risk for suicide include HIV/AIDS (25, 335, 336), malignancies (especially of the head and neck) (25, 333, 337, 338), peptic ulcer disease (25), systemic lupus erythematosus (25), chronic hemodialysis-treated renal failure (339), heart disease (337), and, in men, chronic obstructive pulmonary disease and prostate disease (337). In contrast, studies have not demonstrated increased suicide risk in patients with amyotrophic lateral sclerosis (ALS), blindness, cerebrovascular disease, hypertension, rheumatoid arthritis, or diabetes mellitus (25, 337).

Beyond the physical illness itself, functional impairments (321, 333, 338), pain (340–342), disfigurement, increased dependence on others, and decreases in sight (333) and hearing increase suicide risk. Furthermore, in many instances, the risk for suicide associated with a medical disorder is mediated by psychiatric symptoms or illness (321, 342, 343). Indeed, suicidality is rarely seen in individuals with serious physical illness in the absence of clinically significant mood disturbance. Finally, the risk for suicide or suicide attempts may also be affected by characteristics of the individual patient, including gender, coping style, availability of social supports, presence of psychosocial stressors, previous history of suicidal behaviors, and the image and meaning to the individual of the illness itself.

In individuals with a history of suicide among relatives, the risk of suicidal behaviors is increased, apparently through genetic as well as environmental effects. An increased relative risk for suicide or suicide attempts in close relatives of suicidal subjects has been demonstrated repeatedly (31, 82, 202, 214, 312, 344–364). Overall, it appears that the risk of suicidal behaviors among family members of suicidal individuals is about 4.5 times that observed in relatives of nonsuicidal subjects (365–368; R. Baldessarini, personal communication, 2002). Furthermore, this increase in the risk of suicidal behaviors among family members seems, at least in part, to be independent of genetic contributions from comorbid psychiatric diagnoses (355, 361, 367, 368).

Twin studies also provide strong support for the role of a specific genetic factor for suicidal behaviors (365, 368, 369), since there is substantially higher concordance of suicide and suicide attempts in identical twins, compared with fraternal twin pairs (370–375). Adoption studies substantiate the genetic aspect of suicide risk in that there is a greater risk of suicidal behavior among biologic than among adopted relatives of individuals with suicidal behavior or depression (376–378).

Despite the fact that family, twin, and adoption methods provide highly suggestive evidence of heritable factors in risk of suicide as well as some evidence for nonlethal suicidal behavior, the mode of transmission of this genetic risk remains obscure. Thus far, molecular genetic approaches have not yielded consistent or unambiguous evidence of a specific genetic basis for suicide risk (16). In addition, genetic associations with suicide risk may be confounded by the heritability of other factors such as mood disorders or substance use disorder that are also associated with increased risk for suicidal behaviors.

Unemployment has long been associated with increased rates of suicide (379, 380). In recent case-control and longitudinal studies, higher rates of unemployment have been consistently noted in suicide attempters (78, 149, 361, 381–383) and in persons who died by suicide (24, 190, 384, 385). Compared with individuals in control groups, unemployed persons have a two- to fourfold greater risk for suicide. Risk is particularly elevated in those under age 45 and in the years closest to job loss, with even greater and longer-lasting effects noted in women (24). Parallel increases in rates of suicide and suicide attempts are also seen in socioeconomically deprived geographical areas, which have larger numbers of unemployed people (386).

For many individuals, unemployment occurs concomitantly with other factors that affect the risk of suicidal behaviors. For example, with job loss, financial and marital difficulties may increase. Alternatively, factors such as psychiatric illness (380) or adverse childhood experiences (361) may affect rates of suicidal behaviors but also influence the likelihood of gaining and maintaining employment. Thus, while unemployment appears to be associated with some independent increase in risk, a substantial fraction of the increase in risk for suicidal behaviors among unemployed persons can be accounted for by co-occurring factors (361, 381, 384, 385).

Among individuals with alcohol use disorders, particularly those under age 45, unemployment is one of a number of stressors that is a common precipitant to suicide (149, 382, 387). Even in those without substance use disorders, unemployment may result in increased drinking, which in turn may precipitate self-destructive behavior (154). Conversely, in those with substance use disorders, full-time employment protects against suicidal behaviors, a finding that may in part relate to decreases in use of alcohol or other substances during the work week (260). Thus, unemployment may serve as a risk factor for suicide, whereas employment may have protective effects on suicide risks.

The likelihood of suicide may also vary with religious beliefs as well as with the extent of involvement in religious activities. In general, individuals are less likely to act on suicidal thoughts when they have a strong religious faith and believe that suicide is morally wrong or sinful. Similar findings of low suicide rates are found in cultures with strong religious beliefs that the body is sacred and not to be damaged intentionally. In the United States, Catholics have the lowest rate of suicide, followed by Jews, then Protestants (388). Among other religious groups, Islamic tradition has consistently regarded suicide as morally wrong, and some Islamic countries have legal sanctions for attempted suicide (389, 390). In some countries, suicide rates among Muslims appear to be greater than those among Hindus (391, 392), although suicide rates across countries do not appear to vary with the proportion of Muslims in the population (393).

Additional evidence suggests that it is the strength of the religious beliefs and not the specific religion per se that alters suicide rates (43, 394–398). In the African American community, for example, religion is viewed as a source of social solidarity and hope (22). Religious involvement may also help to buffer acculturative stress, which is associated with depression and suicidal ideation (43). The religious belief system itself and the practice of spiritual techniques may also decrease suicide risk by acting as a coping mechanism and providing a source of hope and purpose.

Although protective effects can be afforded by religious beliefs, this is not invariably the case. For example, suicide may be more likely to occur among cultures in which death by suicide is a traditionally accepted way of dealing with distress or in religions that deemphasize the boundaries between the living and the dead. Particularly for adolescents, belief in an afterlife may lead to suicide in an effort to rejoin a deceased loved one. Thus, it is important to gain an understanding of the specific religious beliefs and religious involvement of individuals and also to inquire how these religious beliefs relate to thoughts and conceptions of suicide.

The presence of a social support system is another factor that may reduce suicide risk (399, 400). Consequently, communicating with members of the patient's support network may be important in assessing and helping to strengthen social supports (see Section V.C, "Communication With Significant Others"). Although social supports typically include family members or friends, individuals may also receive support from other sources. For example, those in the military and those who belong to religious, community, or self-help organizations may receive support through these affiliations.

In addition to determining whether a support system is present, the clinician should assess the patient's perception of available social supports. Individuals who report having more friends and less subjective loneliness are less likely to have suicidal ideation or engage in suicidal behaviors (401). By the same token, if other social supports are not available, living alone may increase suicide risk (149, 385, 402), although this is not invariably true (343, 403, 404). Family discord, other relationship problems, and social isolation may also increase risk (403, 405, 406). Risk of suicidal behaviors may also increase when an individual rightly or wrongly fears that an interpersonal loss will occur (149). Thus, in estimating suicide risk, the clinician should assess the patient's support network as well as his or her perception of available social supports.

An additional protective factor against suicidal behaviors is the ability to cite reasons for living (231, 407), which reflects the patient's degree of optimism about life. A sense of responsibility to family, particularly children, is a commonly cited reason for living that makes suicide a less viable option to escape from pain. The presence of children in the home as well as the number of children appear to decrease the risk for suicide in women (26, 47). Although less well-studied, a smaller effect on suicide potential may also be present in men who have children under age 18 within the home (408). Thus, knowledge of the patient's specific reasons for living, including information about whether there are children in the home, can help inform estimates of suicide risk.

Estimates of suicide risk should also incorporate an assessment of the patient's strengths and vulnerabilities as an individual. For example, healthy and well-developed coping skills may buffer stressful life events, decreasing the likelihood of suicidal actions (409). Conversely, lifelong patterns of problematic coping skills are common among those who die by suicide (410). Such factors may be particularly important in patients with substance use or personality disorders, for whom heightened suicide risk may be associated with life stressors or interpersonal loss.

In addition to the diagnosis of categorical axis II disorders, as discussed elsewhere, dimensional and trait approaches to personality can also inform estimates of suicide risk. Although the positive correlation value of individual personality traits with suicide is low, increased suicide risk may be associated with antisocial traits (411) as well as with hostility, helplessness/dependency, and social disengagement/self-consciousness (246).

Extensive clinical literature and clinical consensus support the role of psychodynamics in assessing a patient's risk for suicidal behavior (409, 410, 412–419). Suicide may have multiple motivations such as anger turned inward or a wish of death toward others that is redirected toward the self. Other motivations include revenge, reunion, or rebirth. Another key psychodynamic concept is the interpretation of suicide as rooted in a triad of motivations: the wish to die, the wish to kill, and the wish to be killed (415). Other clinicians have conceptualized these motivations as escape (the wish to die), anger or revenge (the wish to kill), and guilt (the wish to be killed). The presence of one or several of these motivations can inform the psychiatrist about a patient's suicide risk.

Object relations theories offer important concepts for psychodynamic formulations of suicide. Suicidal behavior has been associated with poor object relations, the inability to maintain a stable, accurate, and emotionally balanced memory of the people in one's life (413). In some cases the wish to destroy the lives of the survivors is a powerful motivator (415, 420). For other individuals, a sadistic internal object is so tormenting that the only possible outcome is to submit to the tormentor through suicide (416, 417).

Other important psychodynamic concepts for the clinician to assess are shame, worthlessness, and impaired self-esteem. Early disturbance in parent-child relationships through failure of empathy or traumatic loss can result in an increased vulnerability to later injuries of self-esteem. These patients are vulnerable to narcissistic injuries, which can trigger psychic pain or uncontrollable negative affects. In these situations some patients may experience thoughts of death as peaceful, believing that their personal reality is emotionally intolerable and that it is possible to end pain by stopping consciousness.

Suicidal individuals are often ambivalent about making a suicidal action. As a result, suicide is less likely if an individual sees alternative strategies to address psychological pain (410). However, certain traits and cognitive styles limit this ability to recognize other options. For example, thought constriction and polarized, all-or-nothing thinking are characterized by rigid thinking and an inability to consider different options and may increase the likelihood of suicide (410, 421–423). Individuals who are high in neuroticism and low in "openness to experience" (affectively blunted and preferring the familiar, practical, and concrete) may also be at greater risk for suicide (424). Perfectionism with excessively high self-expectation is another factor that has been noted in clinical practice to be a possible contributor to suicide risk (425). As already discussed, pessimism and hopelessness may also act in a trait-dependent fashion and further influence individual risk.

In estimating suicide risk it is therefore important for the clinician to appreciate the contributions of patients' individual traits, early or traumatic history, ability to manage affects including psychological pain, past response to stress, current object relations, and ability to use external resources during crises. Identifying these issues may help the psychiatrist in assessing suicide risk. In addition, gaining an empathic understanding of the patient's unique motivations for suicide in the context of past experiences will aid in developing rapport as well as in formulating and implementing a psychotherapeutic plan to reduce suicide risk (410, 412, 421, 426).

Suicidal ideation is an important determinant of risk because it precedes suicide. Moreover, suicidal ideation is common, with an estimated annual incidence of 5.6% (2) and estimated lifetime prevalence of 13.5% (427). Since the majority of individuals with suicidal ideation will not die by suicide, the clinician should consider factors that may increase risk among individuals with suicidal ideation. Although current suicidal ideation increases suicide risk (78, 79), death from suicide is even more strongly correlated with the worst previous suicidal ideation (273, 428). Thus, during the suicide assessment, it is important to determine the presence, magnitude, and persistence of current as well as past suicidal ideation.

In addition to reporting suicidal ideation per se, patients may report thoughts of death that may be nonspecific ("life is not worth living") or specific ("I wish I were dead"). These reports should also be assessed through further questioning since they may serve as a prelude to later development of suicidal ideas or may reflect a sense of pessimism and hopelessness about the future (see Section II.E.3.b, "Hopelessness"). At the same time, individuals with suicidal ideation will often deny such ideas even when asked directly (218, 429–431). Given these associations of suicide with suicidal ideation, the presence of suicidal ideation indicates a need for aggressive intervention. At the same time, since as many as a quarter of suicide attempts occur impulsively (432), the absence of suicidal ideation does not eliminate risk for suicidal behaviors.

Determining whether or not the patient has developed a suicide plan is a key part of assessing suicide risk. For many patients, the formation of a suicide plan precedes a suicidal act, typically within 1 year of the onset of suicidal ideation (427). A suicide plan entails more than simply a reference to a particular method of harm and includes at least several of the following elements: timing, availability of method, setting, and actions made in furtherance of the plan (procuring a method, "scoping out" the setting, rehearsing the plan in any way). The more detailed and specific the suicide plan, the greater will be the level of risk. Plans that use lethal methods or are formulated to avoid detection are particularly indicative of high risk (433). Access to suicide methods, particularly lethal methods, also increases suicide risk. Even in the absence of a specific suicide plan, impulsive actions may end in suicide if lethal methods are readily accessible. Thus, it is important to determine access to methods for any patient who is at risk for suicide or displays suicidal ideation.

In the United States, geographic variations in rates of firearm suicide parallel variations in the rates of gun ownership (434). Although individuals may opt for a different suicide method when a particular method is otherwise unavailable, studies show some decreases in overall suicide rates with restrictions in access to lethal suicide methods (e.g., domestic gas and paracetamol) (435–437). Men are most likely to use firearms in suicidal acts, but other specific populations at increased risk of using firearms include African Americans, elderly persons, and married women. In adolescents and possibly in other age groups, the presence of firearms may be an independent risk factor for suicide (438). Consequently, 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.

In addition to addressing access to firearms, clinicians should recognize the potential lethality of other suicide methods to which the patient may have access. As with restrictions for firearms, it is important for the psychiatrist to work with the patient, family members, and other social support persons in restricting the patient's access to potentially lethal suicide methods, particularly during periods of enhanced risk. Removal of such methods from a patient's presence does not remove the risk for suicide, but it removes the potential for the patient to impulsively gain access to the means with which to carry out a suicidal wish.

Suicidal intent refers to the patient's subjective expectation and desire to die as a result of a self-inflicted injury. This expectation may or may not correspond to the lethality of an attempt, which represents the medical likelihood that death will result from use of a given method. For example, some patients may make a nonlethal attempt with the intention of being saved and getting help, whereas others may make a nonlethal attempt, thinking it will kill them. From the standpoint of suicide risk assessment, the strength of the patient's intent to die and his or her subjective belief about the lethality of a method are more relevant than the objective lethality of the chosen method (439, 440). The presence of a suicide note also indicates intensification of a suicidal idea and/or plan and generally signifies premeditation and greater suicidal intent. Regardless of whether the patient has attempted suicide or is displaying suicidal ideation, the clinician should assess the timing and content of any suicide note and discuss its meaning with the patient. The more specifically a note refers to actual suicide or steps to be taken after death, the greater the associated increase in suicidal intent and risk. Factors separating suicide attempters who go on to make future fatal versus nonfatal attempts include an initial attempt with high intent (164, 441), having taken measures to avoid discovery (224), and having used more lethal methods that resulted in physical injuries (263), all of which indicate a greater degree of suicidal intent. Consequently, suicidal intent should be assessed in any patient with suicidal ideation. In addition, for any patient who has made a prior suicide attempt, the level of intent at the time of the suicide attempt should be determined.

Patients are often admitted to an inpatient unit in the midst of an acute suicidal crisis with either overt suicidal behavior or intense suicidal ideation. Even when a patient who is not in an acute suicidal crisis is admitted, the symptoms and disorders that typically lead to psychiatric hospitalization are associated with an increased suicide risk. There do not appear to be specific risk factors that are unique to the inpatient setting, with about half of inpatient suicides in a recent study involving individuals with prior suicide attempts and about half occurring in individuals with psychosis (218). Inpatient suicides also cannot be predicted by the reason for hospitalization, since fewer than half of the patients who die by suicide in the hospital were admitted with suicidal ideation and only a quarter were admitted after a suicide attempt. However, extreme agitation or anxiety (218) or a rapidly fluctuating course (442) is common before suicide. Thus, it is important to conduct a suicide risk assessment, as discussed earlier, when individuals are admitted for inpatient treatment, when changes in observation status or treatment setting occur, when there are significant changes in the patient's clinical condition, or when acute psychosocial stressors come to light in the course of the hospitalization. For patients with repeated hospitalizations for suicidality, each suicidal crisis must be treated as new with each admission and assessed accordingly.

An initial evaluation of a patient in an office-based setting should be comprehensive and include a suicide assessment. The intensity and depth of the suicide assessment will depend on the patient's clinical presentation. In following outpatients over time, the psychiatrist should be aware that suicidality may wax and wane in the course of treatment. Sudden changes in clinical status, which may include worsening or precipitous and unexpected improvements in reported symptoms, require that suicidality be reconsidered. Furthermore, risk may also be increased by the lack of a reliable therapeutic alliance, by the patient's unwillingness to engage in psychotherapy or adhere to medication treatment, or by inadequate family or social supports. Again, however, the frequency, intensity, and depth of the suicide assessment will depend on the patient's clinical state, past history, and other factors, including individual strengths, vulnerabilities, and stressors that will simultaneously influence risk. These factors will also be important in judging when family members or other significant support persons may need to be contacted.

Regardless of the patient's presenting problem, the suicide assessment is an integral part of the psychiatric evaluation in an emergency setting. As in the inpatient setting, substantial numbers of individuals present to emergency settings with suicidal ideation or after having made a suicide attempt (443–447). Even when suicidality is not a part of the initial presentation, the majority of individuals seen in emergency psychiatric settings have diagnoses that are associated with an increased risk of suicide (268, 269, 271, 275, 448).

As the suicide assessment proceeds, the psychiatrist should be alert for previously unrecognized symptoms of trauma or toxicity resulting from ingestions. Ambivalence is a key element in individuals presenting with suicidality, and individuals may simultaneously seek help yet withhold information about recent ingestions (449) or self-induced trauma. Thus, in addition to initially assessing the patient's vital signs, the psychiatrist should investigate any changes in the patient's physical condition or level of consciousness that may develop during the course of the evaluation. For patients who are administered medications in the emergency area or who have concomitant alcohol or substance use, serial monitoring of vital signs is important to detect adverse events or signs of substance withdrawal.

Simultaneous presentation with intoxication and suicidality is common in emergency settings (444, 450–454) and requires some modification in the assessment process. Depending on the severity of the intoxication, medical intervention may be needed before psychiatric assessment begins. Also, it is often necessary to maintain the patient in a safe setting until the intoxication resolves and a thorough suicide assessment can be done. In this regard, some institutions find it helpful to quantify the level of intoxication (with serum alcohol levels or breath alcohol measurements), since some individuals may not show physical symptoms of intoxication despite substantially elevated blood alcohol concentrations (455). At some facilities, short-term observation beds are available in the emergency area or elsewhere for monitoring and serial assessments of intoxicated individuals who present with suicidality. At other facilities, such observation may need to be carried out in a more typical medical or psychiatric inpatient setting.

Although obtaining collateral information is useful with all suicidal individuals, in the emergency setting such information is particularly important to obtain from involved family members, from those who live with the patient, and from professionals who are currently treating the patient. Patients in emergency settings may not always share all of the potentially relevant aspects of their recent symptoms and their past psychiatric history, including treatment adherence. In addition, most psychiatrists who evaluate patients in emergency settings do not have the benefit of knowing and working with the patient on a longitudinal basis. Corroboration of history is particularly important when aspects of the clinical picture do not correspond to other aspects of the patient's history or mental state. Examples include patients who deny suicidal ideas and request discharge yet who made a highly lethal suicide attempt with clear suicidal intent or those who request admission on the basis of command hallucinations while seeming relaxed and jovial and without appearing to respond to internal stimuli.

The process by which the patient arrived at the emergency department can provide helpful information about his or her insight into having an illness or needing treatment. Typically, individuals who are self-referred have greater insight than those who are brought to the hospital by police or who reluctantly arrive with family members. For individuals who are brought to the emergency department by police (or as a result of a legally defined process such as an emergency petition), it is particularly important to address the reasons for the referral in estimating suicide risk.

When evaluating patients in long-term care facilities, psychiatrists and staff should be aware of the varied forms that suicidality may take in such settings. In particular, it is important to recognize that indirect self-destructive acts are found among both men and women with chronic medical conditions (456–459) and are a common manifestation of suicide in institutional settings (460). Despite these occurrences, suicide rates in long-term care facilities are generally lower than expected (460, 461), perhaps as a result of greater supervision and residents' limited access to potentially lethal means and physical inability to carry out the act as well as underreporting or misattribution of self-destructive behaviors to accident or natural death (66).

Risk factors for suicide and other self-destructive behaviors are similar to those assessed in other settings of care. For example, 90% or more of randomly sampled residents of long-term care facilities have been shown to have a diagnosable psychiatric illness (462, 463), with the prevalence of depression in nursing homes estimated to range from 15% to 50% (66). Physical illness, functional impairment, and pain are associated with increased risk for suicide and are ubiquitous factors in long-term care facilities. Hopelessness (228) and personality styles that impede adaptation to a dependent role in the institutional setting also play a role (464).

When treating individuals in long-term care facilities, the psychiatrist should be mindful of the need for follow-up assessments, even when initial evaluation does not show evidence of depression or increased risk for suicide or other self-injurious behaviors. To facilitate early intervention, safety and suicide risk should be reassessed with significant changes in behavior, psychiatric symptoms, medical status, and/or level of functional disability. Psychiatrists can also play a critical role in educating long-term care providers about risk factors and warning signs for suicide in residents under their care.

In jails, prisons, and other correctional facilities, most initial mental health assessments are not done by psychiatrists (465, 466); however, psychiatrists are often asked to perform urgent suicide assessments for individuals identified as being at risk. The actual rates of suicide in jails and in prisons are somewhat controversial, and reported rates depend on the method by which they are calculated (467). The U.S. Department of Justice Bureau of Justice Statistics reported that the rate of suicide per 100,000 prison inmates was 14 during 1999, compared with 55 per 100,000 jail inmates (468). However, reported rates are generally based on the average daily census of the facility. Since jails are local facilities used for the confinement of persons awaiting trial and those convicted of minor crimes, whereas prisons are usually under state control and are used to confine persons serving sentences for serious crimes, jails have a much more rapid turnover of detainees than prisons. This turnover results in a higher reported rate of suicides per 100,000 incarcerated persons in jails relative to prisons, since annual jail admissions are more than 20 times the average daily jail census, whereas the annual number of persons admitted to prisons nationwide is about 50% of the average daily prison census. Reported suicide rates in jails are also elevated relative to those in prisons because the majority of suicides in jail occur during the first 24 hours of incarceration (469, 470).

The importance of identification and assessment of individuals at increased risk for suicide is underscored by the fact that suicide is one of the leading causes of death in correctional settings. For example, from July 1, 1998, to June 30, 1999, natural causes other than AIDS barely led suicide as the leading cause of death in jails. Between 1995 and 1999, suicide was the third leading cause of death in prisons, after natural causes other than AIDS and deaths due to AIDS (468). In relative terms, suicides among youths in juvenile detention and correctional facilities are about four times more frequent, suicide rates for men in jails are about nine to 15 times greater, and the suicide rate in prisons is about one-and-a-half times greater than the suicide rate in the general population (471).

Factors that increase risk in other populations are very prevalent and contribute to increased risk in correctional populations (472, 473). Persons who die by suicide in jails have been consistently shown to be young, white, single, intoxicated individuals with a history of substance abuse (470, 474–476). Suicide in correctional facilities generally occurs by hanging, with bed clothing most commonly used (470, 474, 476–478). It is not clear whether first-time nonviolent offenders (474, 476) or violent offenders (473, 477) are at greater risk. Most (473, 474, 476, 479) but not all (480) investigators have reported that isolation may increase suicide in correctional facilities and should be avoided. While inmates may become suicidal anytime during their incarceration, there are times when the risks of suicidal behavior may be heightened. Experience has shown that suicidal behaviors increase immediately on entry into the facility, after new legal complications with the inmate's case (e.g., denial of parole), after inmates receive bad news about loved ones at home, or after sexual assault or other trauma (471).

There is little doubt that successful implementation of suicide prevention programs results in a significantly decreased suicide rate in correctional facilities (469, 481–483). Consequently, the standards of the National Commission on Correctional Health Care (NCCHC) require jails and prisons to have a written policy and defined procedures for identifying and responding to suicidal inmates, including procedures for training, identification, monitoring, referral, evaluation, housing, communication, intervention, notification, reporting, review, and critical incident stress debriefing (484, 485). Other useful resources include a widely used instrument for suicide screening (486) and the detailed discussions of specific approaches to suicidal detainees that are provided in a later NCCHC publication (487).