Reflections on Suicidal Ideation

According to the Substance Abuse and Mental Health Service Administration (SAMHSA) in the United States, 10.6 million American adults have serious thoughts of ending their lives by suicide each year (SAMHSA, 2018). In that same year, we further know that 1.4 million American adults attempted suicide while approximately 47,000 across all ages died by suicide (Drapeau & McIntosh, 2018). While suicidologists and public health officials are understandably preoccupied with suicides and suicide attempts, we have recently begun to reflect on those with suicidal ideation who too often escape the focused attention of our suicide prevention research, clinical treatments, and even national health-care policies. Upon reflection, the prevalence of suicidal ideation in the United States is truly staggering: 10,600,000 people experiencing thoughts of ending their lives is more than the population of the US state of Georgia. From an international perspective, this figure is roughly the size of the population of the Czech Republic.

As suicide prevention researchers, we understand the appeal of observable suicidal behaviors with implications for morbidity and mortality. However, the morbidity of suicidal ideation should not be underestimated. As a focus of research, suicidal ideation tends to be a more elusive, ephemeral, and often fluid construct. But the proportion of people who experience serious suicidal thoughts represents the larger mass of the suicide iceberg below the surface of the water. Suicide deaths and attempts represent the tip of this iceberg, which is dwarfed by the much larger problem, at least with regard to numbers, of all the people beneath the surface who are experiencing suicidal misery, often in silence.

The definition of suicidal ideation offered by the Centers for Disease Control and Prevention in the United States (Crosby, Ortega, & Melanson, 2011), echoing the US National Strategy for Suicide Prevention, is: "…Thoughts of engaging in suicide-related behavior." This is an appropriately broad definition for a phenomenon that includes, but is not limited to, specific plans to die and explicit intent to die imminently. As we argue here, all aspects of suicidal ideation deserve attention; these two specific instances certainly do, since they signal imminent danger for self-inflicted death.

In a meta-analysis conducted by Franklin and colleagues (2017), the number-one risk factor for future episodes of suicidal ideation was prior suicidal ideation. While this finding is unsurprising, it highlights the recurrent and chronic nature of suicidal ideation, and underscores key aspects of its morbidity. In terms of predicting death by suicide, the same meta-analysis found that suicidal ideation was the third most potent predictor of future death by suicide, behind only prior psychiatric hospitalizations and prior suicide attempts. It should be added that in the Franklin et al. meta-analysis, all predictors were relatively weak (e.g., odds ratios between approximately 2 and 4, even for those in the top five). It is important to note that prior psychiatric hospitalizations were the leading predictor of later suicide death; notably, suicide ideation is one of the most common reasons for hospitalizations (e.g., Bowers, 2005). The same logic can be applied to risk for suicide attempt in the Franklin et al. meta-analysis; suicidal ideation was not among the top five predictors of future attempt, but psychiatric hospitalizations were. Again, hospitalizations are often prompted by suicidal ideation.

It is peculiar, upon reflection, to in any way diminish an ideational morbidity. Behavioral morbidity deserves its due. But ideational morbidity is a regular emphasis in mental health, regarding, for example, worry in generalized anxiety disorder, obsessions in obsessive-compulsive disorder, grandiosity in the manic phase of bipolar conditions, and delusions in psychotic disorders. One may counter that these ideational factors have behavioral consequences, to which we reply that so do suicidal ideational factors.

Some believe that suicidal behavior can occur in the absence of prior ideation. We are skeptical, for at least two reasons. First, it is not at all clear that those who attempt suicide, survive, and are then queried about their prior ideation are accurate in their reporting. The demand to report a lack of ideation after the attempt can be considerable, stemming from such sources as stigmatizing attitudes from health professionals and others, embarrassment, confusion, etc. Second, given the recurrent nature of suicidal ideation, planning for a suicide that occurred in a prior episode may, as it were, be stored on a mental "shelf" for later use. Then, during a later suicidal episode, the already-developed plan is available for quick retrieval; all that is needed is a quick psychological "reach for the shelf." Although a potentially rapid-fire self-destructive behavioral process may ensue, it is hardly lacking a requisite ideational component.

With death by suicide occurring as a relatively low base-rate behavior, studies of interventions to reduce suicide deaths often require prohibitively large sample sizes. It follows that clinical researchers have pragmatically settled on suicide attempts as the next best proxy to target for intervention. To this end, we have a number of remarkable clinical interventions shown to reliably decrease suicide attempt behaviors within replicated randomized controlled trials (RCTs). Indeed, dialectical behavior therapy (DBT; Linehan et al., 2015), cognitive therapy for suicide prevention (CT-SP; Brown et al., 2005), brief cognitive behavior therapy (BCBT; Rudd et al., 2015), attempted suicide short intervention program (ASSIP; Gysin-Maillart, Schwab, Soravia, Megert, & Michel, 2016), and crisis response planning (CRP; Bryan et al., 2017) have all shown an impressive ability to significantly reduce suicide attempts in 1–2-year follow-up assessments, depending on the study. Similarly, a large cohort comparison trial of safety plan intervention (SPI; Stanley et al., 2018) also showed a robust reduction in suicide attempts associated with safety planning. Interestingly, while there is some evidence that DBT can impact suicidal ideation (DeCou, Comtois, & Landes, 2019), the remaining effective clinical treatments and interventions for suicide attempts (relative to control care) have limited to no impact on suicidal ideation – a population that is 7.6 times larger than the population who make suicide attempts and 225.5 times larger than the population of people who die by suicide according to the figures we presented at the beginning of this editorial. From an international perspective, there is some between-country variability in terms of relative proportions of suicidal ideation versus suicide attempt versus suicide death populations (c.f., Nock et al., 2008). Nevertheless, at the population-level around the world, suicidal torment is universally dominated by millions upon millions of people with suicidal ideation.

If interventions can significantly decrease suicidal behavior but not suicidal ideation, one may question the value of decreasing ideation. We have two responses. First, to repeat, the decrease of any morbidity, whether ideational, behavioral, or otherwise, represents a meaningful decrease in human misery. Second, and more speculatively, it is conceivable that a reduction in suicidal ideation may delay later behavioral events. Just as a suicide plan from months or even years ago can exert a contemporary ideational influence on a current behavioral outcome (e.g., a suicide attempt), so may an elimination or reduction of an ideational morbidity from months or years ago have beneficial behavioral consequences in the here-and-now. In other words, current beneficial consequences may be observed that were not apparent early on, perhaps because a triggering crisis had not occurred until later. To adequately assess this possibility, intervention study designs would need to carefully track the impact of an intervention on suicidal ideational and related behavioral consequences – and how these critical constructs interact over time, long term – to help us unravel this important interplay.

A further complication is that interventions can have different trajectories of overall gain. Take, for example, the treatment of bulimia nervosa. At 4-month follow-up, cognitive-behavioral therapy (CBT) results appeared somewhat more favorable than those for interpersonal therapy (IPT; e.g., Agras, Walsh, Fairburn, Wilson, & Kraemer, 2000; Fairburn, Jones, Peveler, Hope, & O'Connor, 1993), but at 1-year follow-up, CBT and IPT were equivalent in their beneficial effects; this equivalence was not as readily apparent earlier on. Still a further issue is that treatments can vary within-intervention as to trajectories of ideational versus behavioral gain. In the bulimia nervosa literature, outcomes for behavioral indices such as objective bulimic episodes were better for CBT than IPT at posttreatment; by 1-year follow-up, IPT had "caught up." A similar pattern occurred with regard to ideational outcomes (e.g., attitudes toward shape and weight), but IPT caught up with CBT sooner (e.g., by a 4-month follow-up). Researchers conducting this work suggested that these differences involve causal versus consequential specificity (see Hollon, DeRubeis, & Evans, 1987). That is, the focus of CBT was specifically on eating-disordered behaviors and attitudes, and so its effects on outcomes were direct and thus causal. The focus of IPT was not on eating, shape, or weight – indeed trial investigators went to some lengths to train IPT therapists to avoid these topics and to focus instead on interpersonal themes – and so its effects were indirect (via improved interpersonal functioning) and thus consequential. One may quibble with the term consequential in that its referent can be confused with concepts like spuriousness or third variables. A plausible alternative phrasing for consequential specificity that avoids such confusion might be indirect causal specificity or mediated causal specificity.

Despite the abundant and pressing need to adequately address the suffering of this massive population of suicidal individuals, interventions and treatments that effectively reduce suicidal ideation with strong RCT support are remarkably limited. In terms of medications, there is evidence from RCTs that clozapine reduced suicidal ideation in suicidal people with schizophrenia (Meltzer et al., 2003) and intravenous infusion of ketamine has shown encouraging short-term impact (i.e., days) on suicidal ideation as well (Grunebaum et al., 2018; Wilkinson et al., 2018). Among RCTs of psychological interventions, the aforementioned CRP significantly reduced suicidal thoughts in a one-touch intervention and attachment-based family therapy (Diamond et al., 2010) effectively reduced suicidal ideation as well. The Collaborative Assessment and Management of Suicidality (CAMS; Jobes, 2016) has been shown to reliably reduce suicidal ideation across multiple RCTs in the United States and elsewhere (e.g., Comtois et al., 2011; Ryberg, Zahl, Diep, Landro, & Fosse, 2019). Of note, a recent critical review (Hoge, 2019) of a CAMS RCT with suicidal US soldiers (Jobes et al., 2017) essentially trivialized the elimination of suicidal ideation in six sessions (a reduction that was sustained at 3-, 6-, and 12-month follow-up assessments). In our view this critique reflects a ubiquitous professional bias that "only" reducing suicidal ideation is somehow deficient; again, the reduction of an ideational morbidity is otherwise a frequent clinical goal, and in other domains (e.g., generalized anxiety disorder) would be viewed and celebrated as a substantial accomplishment.

In the United Kingdom, patients who receive mental health care through the National Health Service (NHS) may not be afforded effective crisis or emergency department care for their suicidal suffering in the absence of a suicide attempt (in which case they might be eligible for DBT). Better defining the links between self-harm and suicide is a key requirement of NHS policy, but treating suicidal thinking in the absence of self-harm behavior is simply not a policy focus. We have colleagues in the United Kingdom who note that the national health-care system may inadvertently prompt some distressed suicidal patients to resort to self-destructive behaviors, attempting suicide as a means for receiving evidence-based care for their suicidal distress.

In our singular pursuit to prevent suicide deaths, we need to stop trivializing the obvious and vital importance of attending to suicidal ideation. We therefore assert that suicidal ideation must become an essential intervention target in and of itself. Indeed, it can be argued that better identification and more effective treatment of suicidal ideation upstream would invariably lead to many fewer suicide attempts and many fewer suicide deaths downstream (but see, e.g., Zuromski et al., 2019). To our way of thinking, it is inescapably true that at the moment of one's suicidal death, the idea of suicide is on the mind; suicidal ideation is the essential ingredient that differentiates a suicide death from an accidental death. And in the vast majority of suicides, suicidal ideation of some kind, whether disclosed or not, has been on the person's mind for hours, days, weeks, and years. It is time for us to pay attention and respond accordingly if we truly aspire to meaningfully decrease suicide-related suffering in all its forms.