In order to set the context for our discussion, this paper will begin by introducing the multifactorial causes of suicide worldwide and then present a broad overview of current strategies for prevention and treatment. We will then discuss how research into suicide has previously been conducted. Crucially, little research into the active suicidal crisis has been conducted, leaving clinicians with few tools to treat acute ideation or behavior. The remainder of the manuscript discusses practical and ethical issues surrounding research with participants with active suicidal ideation or recent suicidal behavior. This includes issues of consent, issues particular to experimental therapeutics trials, the myth of the iatrogenic potential of suicide research, and regulatory and policy related concerns. Throughout, we explore current discussions in the bioethical literature and synthesize results from empirical research that have relevant ethical implications for research on suicide. Our overarching conclusion is that there is a moral imperative to conduct suicide research, and that this research can be performed safely and ethically.
Causes of suicide
In 1897, Emile Durkheim authored a groundbreaking treatise of the phenomenon of suicide from a socio-cultural perspective [
16]. Durkheim conceptualized suicide as stemming from four different factors encompassing ideas of community integration, sacrifice, moral confusion, and desperation [
16]. Although suicides certainly occur for these reasons, his model failed to predict the trends in suicide deaths seen in modern times and may also not adequately capture the contributions of mental illness [
17]. In this context, a retrospective study of suicide-related emergency room visits found that 82.7% of presenting patients had a concurrent mental disorder, most commonly mood disorders, substance or alcohol related disorders, and anxiety disorders [
18].
Although psychological theories exist about the underlying causes of suicide (see, for instance, the review by Klonsky and colleagues [
19]), most current thinking from the neurobiological literature on suicide characterizes self-harm as occurring from a stress-diathesis model, whereby life stressors precipitate a suicidal crisis in individuals with a pre-existing diathesis that encompasses aggressive and impulsive personality traits as well as pessimism [
20‐
22]. In addition, authors such as Klonsky and May have argued for a three-step theory, with the causes of suicidal ideation rooted in pain and hopelessness and with social connectedness serving as a third—and protective— factor against the escalation of ideation to behavior [
23]. Models such as this and others [
24] decouple the processes behind ideation and behavior, as first suggested by Joiner [
25].
Although suicide has largely been medicalized in the research literature, societal and cultural aspects undoubtedly contribute to suicide rates. This is especially evident when examining minority groups that may suffer from discrimination. For example, suicide rates are higher in those who identify as lesbian, gay, bisexual, or transgender (LGBT), likely due at least in part to hostility and/or marginalization of this community [
26]. Another example is that of indigenous populations. For instance, the suicide rate for American Indian/Alaskan Native (AI/AN) individuals between 2007 and 2009 was 18.5 per 100,000, 1.6 times greater than the all-race suicide rate in the US of 11.6 for 2008 [
27]. The disparity is particularly striking for males in the 15–24-year old group, where AI/AN suicides occur at a rate of 58.7 per 100,000; in comparison, the all-race US suicide rate for males of the same age group is 16.0 per 100,000.
Geopolitical factors may also contribute to suicide rates. For example, the highly publicized cluster of 12 suicides by Chinese workers at the Foxconn factory (which made Apple iPod and iPad devices, among other electronic products) were thought to be primarily attributable to employee abuse, unethical labor practices, and failure to admit culpability [
28]. Worldwide, economic desperation is recognized as a frequent cause of suicide. The 2009 global financial crisis may have accounted for 5000 suicide deaths, and the US economic downturn in 2007/2008 is estimated to have increased suicide deaths in those with low education levels by 1.22 deaths per 100,000 [
29]. Evidence also suggests that the strict austerity measures in Greece and other European nations during this time, along with rising unemployment, contributed to an increase in suicide rates [
30].
Marked differences in suicide rates also exist across different countries. It should be noted that attitudes towards suicide continue to vary across cultures and religions, and reporting may therefore be more or less accurate depending on geopolitical region. According to the World Health Organization (WHO), the global average suicide rate was 10.7 per 100,000 in 2015 [
8]. The Eastern Mediterranean region, encompassing the Middle East and Northern Africa, had the lowest reported rate at 3.8 per 100,000, while Europe had the highest rate at 14.1 per 100,000 [
8]. A study examining differences in suicide rates across European nations found that both economic and climatic variables had significant effects [
31]. Another relevant factor may be cultural attitudes towards suicide. For instance, a recent study found that Chinese psychiatrists showed more stigmatizing attitudes and less empathy towards individuals with mental illness than non-physician urban community members [
32]. Importantly, evidence suggests that stigma against suicide may increase suicide risk in vulnerable individuals [
33].
It should be clear at this point that the etiology of suicide is complex and multifactorial. While many suicides occur in the context of mental illness, other social, cultural, economic, and even political factors are frequently involved. Thus, it is clear that prevention strategies will likely need to address multiple domains outside the medical model of suicide.
Treatment and prevention
Currently, no validated biological markers and few demographic or behavioral markers exist that predict suicide with high specificity and sensitivity [
22]. In this context, the most potent predictor of future suicidal behavior is past suicidal behavior [
21]. Repeated suicidal behavior is highly prevalent; in one study of 28,700 children, adolescents, and young adults in Ireland, 19.2% of patients engaged in another act of self-harm in the first year following the initial incident [
34]. Although the precise genetic factors underlying a tendency towards self-harm are unknown, around 50% of the risk for suicidal behavior appears to be heritable [
22]. Predicting and preventing suicidal behavior is fortunately a growing field of research, and several potential candidate markers have been identified for further study. One recent investigation evaluated a large array of biomarkers, then verified the most promising candidates in an independent sample. Apolipoprotein E (ApoE) and interleukin-6 (IL-6) emerged as markers, potentially indicating the involvement of inflammation and accelerated aging [
35]. Potential epigenetic and genetic markers of suicide also exist, although gene expression interacts with life events, and further replication is needed [
36]. Numerous studies have also identified altered sleep architecture as a biomarker of suicidal thoughts and behavior [
37‐
39].
In general, prevention strategies can be divided into two categories: a “high-risk” approach that targets individuals at high risk, and a “population” approach that targets social and environmental factors [
40]. Zalsman and colleagues [
41] recently published a comprehensive overview of both risk-based and population-based suicide prevention strategies studied over the past 10 years. Strikingly, they concluded that the most robust evidence supported practical population-based measures, such as reducing access to drugs, toxins, and jumping sites. Although research into possible links between accessibility to firearms and suicide risk has been very limited in the US, the CDC nevertheless reported that, in 2017, 60% of deaths by firearms were suicides and only 37% were homicides [
42]. Research into restricting the means of suicide—and potential interactions with cultural factors—can be difficult, and research assessing the actual impact of policy changes is even more so. In addition to restricting access to the means of suicide, adequate treatment of psychiatric disorders—either through pharmacological means or psychotherapy—has also been shown to reduce suicide rates [
41]. The study by Zalsman and colleagues also found sufficient evidence to support the implementation of school-based education programs, but noted that despite individual positive studies on primary care screening programs, gatekeeper training, and telephone or internet interventions, the evidence remains inadequate to support large-scale deployment [
41]. Nevertheless, it is important to acknowledge that despite the use of current evidence-based prevention strategies, suicide rates, at least in the US, have increased rather than declined [
1]; it should also be noted here that the strategies discussed above are, in general, primary prevention strategies.
As noted above, laws, regulations, and structural changes can also alter suicide rates. For instance, in the former Soviet Union, an anti-alcohol campaign initiated by former president Mikhail Gorbachev between 1985 and 1988 strikingly reduced the number of suicide deaths; following the collapse of the Soviet Union in 1991, rates began to sharply increase again [
43]. Another, and perhaps more surprising, example is that suicide rates by gas inhalation in the United Kingdom decreased dramatically as the percentage of carbon monoxide in domestic gas (used for home heating and cooking) decreased between 1955 and 1975, as utilities transitioned from coal gas to natural gas [
44]. Surprisingly, although non-gas inhalation suicides increased somewhat in younger men, overall suicide rates decreased substantially, again emphasizing that removing easily accessible means of attempting suicide can be an effective prevention method.
In addition to the dearth of prevention strategies, few evidence-based medical practices exist to treat suicide risk. In terms of medications, the antipsychotic clozapine is the only FDA-approved drug for the treatment of suicide risk, but it is specific to patients with schizophrenia. Evidence also exists that the mood stabilizer lithium [
45] and the N-methyl-D-aspartate (NMDA) modulator and rapid-acting antidepressant ketamine [
46,
47] may exert anti-suicidality effects, along with electroconvulsive therapy (ECT) [
48]. Indeed, ketamine has been used off-label in the clinic to treat active suicidal ideation or behavior, and the use of esketamine (recently approved by the FDA for treatment-resistant depression in adults) [
49] may not be far behind, although there are relatively few prospective trials. A recent meta-analysis of studies that examined ketamine as an agent for reducing suicidal ideation found evidence to support its use in the clinic but emphasized the need for further randomized, controlled trials of adequate power [
50]. Psychotherapeutic interventions have also been investigated, with evidence supporting the efficacy of cognitive behavioral therapy and dialectical behavioral therapy [
51‐
53] in addition to other suicide-targeted therapies [
54]. It should be noted that these treatment strategies could be considered both secondary or tertiary prevention strategies, designed to prevent both relapse and recurrence of suicidal ideation or behavior.
Any discussion of suicide treatment and prevention would be incomplete without addressing access to care. According to the WHO, low-and middle-income countries have fewer than 0.5 psychiatrists per 100,000 people [
55]. In the US in 2011, there was an average of 10.9 psychiatrists per 100,000 people, though access varied substantially by region [
56]. Financial access to care is also an issue. In 2010, 16% of US citizens lacked health insurance, although that number had declined to 9.1% by 2015, in large part due to government programs designed to increase access to care [
57]. Nevertheless, one study found that 76.9% of individuals who attempted suicide had contacted a health care provider within the last 3 months, potentially indicating missed opportunities for prevention [
58]. Transitional access is also important; while an emergency room visit or brief psychiatric hospitalization can forestall a death due to suicide, without further transition plans to community care, patients are at risk for future attempts. In fact, the weeks after discharge from psychiatric hospitalization are the period of highest risk for suicide [
59,
60].
Presently, suicide risk is difficult to assess with specificity and sensitivity, although numerous biomarkers are being pursued. In addition, while treatment strategies exist, their effectiveness is not always well established, and population-level interventions may have the greatest impact, at least on overall suicide rates. Although a full discussion of this topic is well beyond the scope of this manuscript, the existence of better screening and prevention methods introduces an ethical question regarding the appropriate degree of paternalism that could be exerted in an effort to reduce suicide rates. This is related to the idea of quaternary prevention, whereby prevention strategies should be formulated in a manner that prevents undue overmedicalization of a population [
61]. While mandatory screening and treatment may lower suicide rates in the future, this may be an unacceptable intrusion on the rights and liberties of private citizens.
Current research into suicidal behavior
The statistics cited above suggest that current research into suicidal behavior is inadequate. The last five decades of research have had no tangible impact on suicide rates in the US overall; in fact, suicide rates have increased. Although policy and practical approaches can reduce the number of suicides, these measures do not directly address the underlying causes. However, basic science investigations are hampered by the fact that no animal model for suicide currently exists, although work is underway to develop such models [
62]. In addition, while post-mortem studies have revealed numerous isolated abnormalities, no overarching mechanism has been identified. Replication also remains a problem, as does the extremely limited number of post-mortem brains available for research [
63].
As noted above, little research exists regarding the treatment of an active suicidal crisis. Most neurobiological suicide studies typically compare individuals who have a history of suicide attempt with individuals who have no such history, such that the observation time point is far removed (in many cases, by years) from any neurobiological or behavioral changes taking place at the time of the attempt (for a review of neuroimaging studies, see [
22]). While this research may be able to reveal trait-type differences between individuals, very little can be concluded about what precipitates a suicidal crisis from a biological perspective or how it can be treated. In addition, most studies investigate suicide only in the context of a mental illness, usually major depressive disorder, bipolar disorder, or schizophrenia.