Strategies for the Prevention and Treatment of Suicidal Behavior

In 2005, the year of the latest Centers for Disease Control and Prevention statistics, there were approximately 32,000 suicides in the United States. This is the equivalent of 89 suicides per day or one suicide every 16 minutes. The number has fluctuated for the past 6 years, showing a significant increase in 2004 and a small decrease in 2005.

The unnecessary loss of life to suicide is a staggering problem for the U.S. population. Suicide is the second leading cause of death for people 25–34 years of age and the third leading cause of death for those 15–24 years of age. In 2005, the rate of suicide among adults older than 65 years was 14.3/100,000. Suicide accounts for 1.4% of all deaths in the United States. Although suicide is highest in white males older than 75 years (37.4/100,000), because there are many other reasons people of this age die, it is not among the leading causes of death for this group. Suicide rates are illustrated in Figure 1, which reflects data up to 2004, as the 2005 data are still preliminary. At all ages, the suicide rate is nearly four times higher in men than in women. The overall suicide rate in 2005 was 11.05/100,000, compared with the homicide rate of 5.9/100,000.

Psychological autopsy studies (using systematic, detailed interviews and records about the deceased, with close relatives, friends, and caretakers of the persons who died by suicide) done in various countries over 50 years report remarkably similar outcomes. These studies show that more than 90% of people who die by suicide have one or more psychiatric disorders, most commonly major depression or bipolar depression, with alcohol or substance abuse reported as the second most common diagnoses (17–8). Individuals with other diseases, such as schizophrenia, anorexia, and some anxiety disorders, also have elevated rates of suicide compared with the appropriate age-matched populations (9). The personality disorders of antisocial and borderline are also prominent (10, 11). The surprising fact is that although there have been at least 117 psychological autopsy studies (12), post-traumatic stress disorder is seldom mentioned as an associated illness. This finding probably reflects the omission of the appropriate questions from the diagnostic interviews used in these studies. Although there is criticism about these studies, as they could not conceal the cause of death and in most cases there were no controls, the slightly variable but mostly consistent outcomes, as well as the fact that many of them were consecutive samples of those who died by suicide, probably give credence to the findings (13).

There are ample data to confirm the fact that frequently people tell others they are contemplating suicide, but the percentage who communicate their intent to their doctors varies by age (58% of older people tell their primary care doctor in the month before the suicide and 77% in the prior year) (17). In a study of suicidal deaths in hospitals, 77% denied their intent on their last communication with the staff (18). Still, for every patient with depressive illness, there should be a record of depressive symptoms and suicidal intent on every visit to the doctor. This may best be accomplished by the use of a depressive symptom questionnaire such as the PHQ9 (Physician Health Inventory Self- Report) or some similar instrument. Certainly every medical student should be taught about the necessity of asking the questions related to suicidal intent in the appropriate patients (those with mental illness and with medical disorders who are depressed).

There are many risk factors associated with an outcome of suicide, although in any individual patient, it is impossible to predict. Individuals of all races, creeds, incomes, and educational levels die by suicide. There is no typical victim. In addition to the known psychiatric disorders, the other well-confirmed risk factor is a history of a past suicide attempt or a past serious plan and/or preparation for suicide (19). Unfortunately, there are many long-term follow-up studies of suicide attempters seen in an emergency setting, and they demonstrate that 7%–10% of these people eventually (up to 22 years later) die by suicide (20). It is true that if the patient is admitted to a hospital for at least 24 hours after an attempt, the likelihood of death by suicide increases substantially over that for those not admitted (21).

Other factors that are important when considering risk for suicide include a family history of suicide (22–25); the symptoms of hopelessness, anxiety, psychic anxiety, panic attacks, desperation, or being a burden in a person who is depressed (26, 27); a recent hospitalization for depression; drinking or drug use or abuse; and aggressive or impulsive personality traits (28). There are clear data showing that after a hospitalization for depression and a suicide attempt, careful aftercare, especially in the first week up to 1 month, reduces suicide rates. Other important factors to note are a history of trauma, abuse, or being bullied, chronic physical pain especially in elderly individuals, certain serious medical illnesses, and being a smoker (9, 29, 30). Environmental factors that must be part of the assessment are the access to lethal means (72% of elderly individuals kill themselves with guns) (31) and recent occurrence of other community or notable suicides that might create a “contagious” effect (32).

Despite the fact that we have many proven interventions (antidepressants and manual-based short-term psychotherapies), suicide rates are not falling. If we assume that therapies are efficacious, they may not be sought because there is still a stigma to admitting to mental illnesses or to seeking treatment. Another possibility is that illnesses associated with suicide are either not recognized by the affected people and their close relatives, friends, and physicians or the treatments being given are not optimal.

Interventions in the community include education and awareness for the general public and for general physicians. There are two successful programs, both from the armed service, here and in Norway showing that an intervention from the top down can decrease suicide or suicide attempts (33, 34). However, these programs have not been successfully transplanted to any civilian institution or group.

There are now three studies from Europe, two from Sweden (35, 36) and one from Hungary (37), showing that educating general physicians to recognize depression and to treat the illness with antidepressants lowers the suicide rate in the population that is being served. In another study from Germany, with a similar program, the suicide rate did not decrease, although the recorded suicide attempt rate did (38). Although there is much emphasis on education for mental health professionals and gatekeepers such as school teachers, clergy, and policemen, the outcome of such programs in preventing suicide has not been proven. A recent study (39) with aggregate analysis showed that there is a strong state by state correlation between suicides and access to health care, which the authors interpret to mean that clinical intervention is the crucial element in the prevention of suicide.

Most of the data on education, interventions, and treatments come from studies involving patients who have made suicide attempts. As already stated, the relationship between changes in attempt rates with changes in suicide rates is tenuous.

There are now several reports of simple interventions that either lower the suicide rate at follow up or lower the suicide attempt rate. The first was from the work of Motto and Bostrom (40). A group of 3,005 patients who had been hospitalized because of a depressive or suicidal state were contacted 30 days after discharge about follow-up treatment. They were asked if they had enrolled in and continued in the posthospital therapy plan for 30 days. Only 843 had not. These patients were randomly assigned to either a contact group or a no contact group. The contact group received a monthly letter for 4 months, then every 2 months for 8 months, and after the first year, a letter every 3 months for the next 4 years and were followed for up to 15 years. The letters were from the research staff member who had interviewed them in the hospital. They were short, expressing concern that the person was getting along all right, and invited a response. They were always worded differently, were individually typed, and included comments if the person had responded. In the 2,782 patients who could be followed for 5 years, the accumulated percentage of suicides was significantly lower for the first 2 years in those who got the letters compared with those who did not or those who at discharge claimed they were in follow-up treatment. Although the rates remained lower than those for the others for all 5 years, the difference was only significant in the first 2 years. Those who said they were getting treatment had the highest suicide rates at all points. Severity of illness at discharge was not assessed, but it may be that those who indicated they were engaged in the posthospital treatment plan had more severe illness. There is no information, however, about whether they actually were receiving treatment.

Carter et al. (41) reported on a somewhat different study from Australia where, after hospital admission for a suicide attempt by self-poisoning, 394 patients got treatment as usual and 378 got treatment as usual as well as eight postcards simply inquiring about their health over the following year. Although the percentage of repeaters was similar in the two groups, the number of suicide attempts greatly decreased in the group who received postcards compared with those who did not (101 compared to 192) because the repeats greatly decreased in those women with three or more attempts.

Although percentages vary by study, probably only 20% of medicated depressed patients are adequately treated with antidepressants (42). There are a multitude of reasons proposed (43), which include failure/stoppage because of side effects, because the medications did not work, or because of cost, undermedication of patients, patients not taking the medication for fear of drug addiction, insufficient duration of treatment, patients not convinced that depression is an illness treatable by drugs, and undetected substance use or abuse. Additional problems arise because the physician does not combine the treatment with psychotherapy, does not use other medications, or does not consider all the options such as electroconvulsive therapy (ECT) or other somatic therapies.

The STAR*D study is an excellent example of several of the necessary options that should be pursued by patients and their doctors (44–48). It included 2,876 patients with depression (80% of whom had chronic or recurring depression) who were seen by a psychiatrist or a family care physician at multiple sites across the country. Patients could be included in the study if they had suicidal thoughts and/or a past suicide attempt. If they were found to have alcohol dependence, they needed to get substance abuse treatment before they entered the study. Everyone got the first medication, citalopram. The medication was given for at least 12 weeks and the dose was maximized. The outcome had to be remission, implying that the patient was almost well.

Only 33% of the patients achieved remission with the first medication. After that the patient was allowed to choose the next step(s) in treatment. The treatments varied by “steps” in the protocol but always allowed either an augmentation, which included psychotherapy or a switch in medication or to psychotherapy alone. There were four levels of treatments, each for 12 weeks unless the patient dropped out. By the end of the study 67% of the patients had achieved remission, although the relapse rate varied at each level. There were no differences between those treated by a psychiatrist or a primary care doctor. In all settings a depression rating scale was used to monitor treatment. The patients were called by staff to keep them involved in the protocol. The study was profound because it taught us if the first medication is pushed to its limits and the treatment is extended to 12–14 weeks, 33% of patients will achieve remission, half by 6 weeks and another half in the next 6–8 weeks. If the second step is taken and the patient stays in treatment, 50% of the patients achieve remission by 20 weeks. Although some may be disappointed by these results, it should be reassuring that with close care and monitoring half of outpatients will be in recovery. Returning to the data (35–37) showing that educating primary care physicians to recognize and treat depression with antidepressants lowers suicide rates, it is assumed that a treatment regimen in the STAR*D tradition could lower suicide rates.

It should be noted that Weissman (49), as part of the STAR*D collaboration, showed that if mothers are treated for their depressions and achieve remission, their children do better than those whose mothers did not achieve remission. In addition, from the STAR*D study, a genetic variation in the HTR2A gene and a new marker in the GR1K4 gene, which is a glutamate receptor, are reported to correlate with response to selective serotonin reuptake inhibitors (50) and with a host of other findings. A recent STAR*D report in 2007 (51) indicated that there is a genetic marker for suicidal ideation. In this study, 1,915 outpatients with major depressive disorder were given optimal doses of the antidepressant citalopram. Within 12–14 weeks of treatment, 120 of these outpatients presented with treatment-emergent suicidal ideation. Two significant markers were found on the inotropic glutamate receptors, genes GRIA3 and GRIK2 in this sample, and the risk was more significant for those patients who carried both risk alleles. Two patients attempted suicide during this phase, and the one patient who participated in the gene collection, although he continuously denied suicidal ideation, carried the high-risk alleles for both genes.