Advancing Our Understanding of the Who, When, and Why of Suicide Risk

Abstract

Suicide is one of the most devastating and perplexing of all human behaviors. Whereas the mortality rate for many leading causes of death (eg, tuberculosis, pneumonia, and influenza) has declined over the past century, the suicide rate is virtually identical to what it was 100 years ago.1 Our lack of progress in suicide prevention is in large part owing to our limited understanding of this problem. Suicidal thoughts and behaviors (STBs) rarely occur in a research laboratory where they can be carefully probed, and we have not had the technology to study them in situ. As a result, we lack a firm understanding of the fundamental properties of STBs, and when, why, and among whom they unfold. The study by Henson et al2 in this issue of JAMA Psychiatry helps to advance the understanding of suicide in several important ways. The authors examined the population of more than 4 million people in England who received a diagnosis of cancer between 1995 and 2015 and found that 2491 of the 3 078 843 individuals (0.08%) who died during the study period died by suicide. They observed a 20% increase in the risk of suicide in those with a cancer diagnosis compared with those in the general population. Suicide risk was especially high in the first 6 months after cancer diagnosis, as well as among those with one of several specific forms of cancer (including mesothelioma, pancreatic, esophageal, and lung cancer). The population-based nature of this study, the documentation of patient characteristics associated with increased risk, and the long study period all strengthen the inferences that can be drawn from this excellent study. This study2 adds to a growing body of research that has identified segments of the population at elevated risk for STBs, and in doing so, it also highlights important lacunae in our understanding. Some of the most consistent findings in studies of STBs are that 90% to 95% of those who die by suicide have a diagnosable mental disorder before their death3 and that the presence of certain types of physical conditions, such as multiple sclerosis and cancer (as in the study by Henson et al2), also are associated with increased risk.4 The fact that most people who die by suicide have one of these conditions has led some to suggest that the conditions offer an explanation of the suicide.5 However, the explanatory power of such an association is limited, given that most patients with mental and physical conditions never even consider suicide. We will achieve much greater progress toward understanding and preventing suicide when we answer several additional questions about people with such conditions. Who Among Those With Mental and Physical Disorders Are at Greatest Risk of Suicide? A recent meta-analysis examining all studies designed to test the prediction of STBs over the past 50 years revealed that (1) no one category of risk factors (eg, mental illness, physical illness) is substantially stronger than the others, (2) most studies have examined 1 risk factor at a time, and (3) the magnitude of effects for known risk factors has not grown over time.6 Taken together, these findings suggest that researchers will not make significant progress in this area if we continue to examine 1 broad risk factor at a time. Henson et al2 did not stop at 1 risk factor but rather tested whether the associations were stronger when considering additional patient characteristics (eg, sex, age at death, and typeofcancerdiagnosed).Severalrecentstudieshaveshownthat simultaneously examining dozens (or more) of putative risk factors using machine learning methods can significantly enhance the assessment of which patients with mental or physical illness are at greatest risk for suicide.7 The accuracy of such an approach is still far from perfect, and the rate of false positives is unacceptably high. This represents an important point of departure for future research.