October 2011
Reviewer: Jeff C. Huffman, MD
Suicide categories by patterns of known risk factors: a latent class analysis
Logan J, Hall J, Karch D
Arch Gen Psychiatry 2011; 68(9):935-941
Background: Suicide is a major public health problem, with 35,000 Americans per year dying from suicide. Many people who commit suicide have multiple suicide risk factors, but it is not well-known whether there are clusters of risk factors that tend to co-exist in different categories of suicide completers; such information about risk profiles could allow better and more specific suicide prevention strategies.
Methods: The authors used 2003-2008 data from the National Violent Death Reporting System (NVDRS), from 12 states, which identifies data about suicide completers (e.g., circumstances leading to death, mechanism of death), and they selected only persons who had more than 1 known risk factor for suicide. The risk factors used included history of mental health treatment/conditions, alcohol use disorder, other substance use disorder, medical conditions, psychosocial stressors (including interpersonal, financial, or legal problems), history of suicide attempt, probable use of alcohol at time of attempt, and disclosure of intent to commit suicide. The authors then used latent class analysis to classify suicide completers by combinations of risk factors to develop profiles/classes of suicide risk.
Results: Overall, 28,703 suicides were reviewed and met study criteria. Nine total classes of suicide risk factor profiles were generated. Caucasian men were more likely to commit suicide, major depression/dysthymia was the most common mental health condition associated with suicide, comorbid substance use was common in those with mental health conditions who committed suicide, and firearms were the most common cause of death. Of note, three-quarters of patients with mental health problems who committed suicide were in mental health treatment around the time of the suicide. The nine classes spanned a range from those with mental health conditions only (the most populated class), to those generally without mental health conditions (e.g., recent life crises plus either criminal legal problems or interpersonal problems), to those with combinations of these factors—crises, mental health conditions, and substance use disorders.
Commentary: This was an interesting examination of a large cohort. In general, it does not raise major new emphasis on a particular risk factor or combination of risk factors, but it does underscore a few points for those of us who do suicide risk assessments at the bedside. First, it is a reminder that patients with multiple mental health problems who have no other major risk factors and who are currently/still in treatment can still complete suicide and still require careful assessment—in fact this was the largest subgroup in this study. At the same time, people without a clear mental health disorder can also complete suicide, such as those persons in crisis, those with legal problems, and those with interpersonal problems; indeed, lack of a mental health disorder does not entirely protect someone, and if patients have these risk factors (especially in combination and/or especially with ongoing use of alcohol/substances) it is worth careful assessment.
Most patients had combinations of risk factors, many of which cannot be ameliorated by standard mental health treatment alone (indeed 75% of the patients with known mental health problems were in treatment yet still completed suicide). This suggests (1) that we need to continue to find interventions in mental health that specifically can reduce risk of suicide and (2) it is optimal to better coordinate efforts between mental health providers and personnel with expertise in managing legal/financial/medical issues or other crises. Again, none of this may be groundbreaking, but this article emphasizes the importance of these issues in reducing the risk of suicide in vulnerable patients.