The Path To and From Suicide

 

From Rethinking Suicide: Why Prevention Fails, and How We Can Do Better - Craig J. Bryan 

Suicide isn’t a conventional problem, though; it’s a complex and wicked problem with no “right” way to define or conceptualize the problem, and no “right” solutions.

 

After introducing myself, I pointed at his chin and asked him if he would be willing to tell me the story of that injury. The Marine described an ordinary day by most combat zone standards: He woke up and prepared for the day’s mission, which involved a routine patrol in armored vehicles through the city. The weather that day was also typical by most summer-in-Iraq standards: sunny and hot, with the high temperature exceeding 110 degrees Fahrenheit (43 degrees Celsius). After several hours of an uneventful patrol, his convoy started its return trip to base. At some point during this hour-long leg of the journey, the Marine reported that he “saw this image in my mind of me shooting myself, and I just couldn’t stop thinking about it.” He attempted to push the image out of his mind and to think of other things, but was unable to do so for the remaining 30 to 60 minutes of the patrol. When the convoy finally returned to base, they headed to the motor pool to park, at which point the Marine hopped off the vehicle, placed the rifle under his chin, and pulled the trigger. 

The results of my evaluation indicated this Marine had no prior history of suicidal thoughts or behaviors and no prior history of mental illness or mental health treatment. The first time he had ever experienced suicide-related thoughts or imagery was on that day, less than an hour before he made the suicide attempt and only a few hours before I met him. He hadn’t been drinking alcohol or using any other substances or narcotics at the time—the toxicology lab results confirmed this—and he wasn’t having any particular problems with his sleep in the preceding weeks. He wasn’t experiencing symptoms of depression or traumatic stress, he wasn’t expressing any hopelessness or negative thoughts about himself, and he wasn’t experiencing any significant problems in life; no marital or relationship conflict, no financial hardship, and no pending or recent disciplinary actions. When I asked the Marine about his day, he described a routine, benign day. Nothing bad or out of the ordinary had happened, and he hadn’t been feeling especially stressed out. He had no discernible risk factors for suicide other than being a young, White male. I checked in with his supervisor and commander, who had come to the hospital with him, and they confirmed, to the best of their ability, all of the information he had provided. The case was utterly bewildering; it just didn’t make any sense to me. It didn’t make any sense to the Marine, either. “I don’t know why I did it,” he explained, shrugging his shoulders. “It just happened.”

While living in Utah, for instance, I learned that around one third of suicides that were precipitated by an argument occurred during the argument and involved a firearm. “It just happened,” a patient once told me when I asked him to share the story of his own close brush with death under very similar circumstances. He and his wife were having a heated argument, he explained, and “I just wanted it to stop,” so he grabbed his loaded handgun and held it to his head. Thankfully, he stopped himself before pulling the trigger. When I asked him what “it” was that he wanted to stop, he described intense emotional distress: “I just felt completely overwhelmed in that moment.”

Over and over again, some version of the refrain, “it just happened,” was offered as an explanation—in many cases accompanied by a confused shrug of the shoulders and resigned shake of the head. Over and over again, I’ve been struck by how many of these cases did not involve a history of mental illness or mental health treatment, no known or reported history of suicidal thoughts and behaviors, and little to no evidence of a current mental illness.

If these people had not engaged in suicidal behavior, their reactions to precipitating events would typically be considered “normal” by most standards and would not have reasonably led to the diagnosis of a mental illness, not even an adjustment disorder.

Only 46% of U.S. citizens who died by suicide between 1999 and 2016 had a known mental health diagnosis at the time of death.

Research studies suggest mental illness is only weakly correlated with suicide.

 

Paths that Help Heal

To date, only two psychological treatments have been shown to reduce the incidence of suicidal behaviors in multiple treatment studies: dialectical behavior therapy (DBT) and cognitive behavioral therapy for suicide prevention (CBT-SP).

First, suicide-focused treatments consistently reduce suicide attempts by 50% or more as compared to status quo treatments.

Many different types of mental health treatments reduce depression, but only a handful of treatments have also been shown to reduce the probability of suicidal behaviors.

If mental illness caused suicidal behavior, we should expect that treatments that reduce depression and/or other symptoms of mental illness would be just as good at reducing suicide attempts, but that’s not what we see at all.

  

Mental health diagnoses are often unreliable

Misdiagnosis is a problem that has plagued the mental health professions for decades. A major contributor to this issue is the tendency for clinicians to make diagnoses based on informal methods and procedures that often use one or more nonsystematic and untested “rules of thumb.”

If someone who attempts suicide is not diagnosed with certain mental health conditions, the mental health professional may not be paid for any treatment provided. Under these circumstances, mental health professionals are incentivized to diagnose someone with a mental illness even if that person doesn’t actually meet all of the criteria for a diagnosis or is experiencing a reasonable and otherwise normal stress response.

 

Why Suicide isn’t Easy to Understand

I found myself thinking about that Marine and all the other cases I’ve seen and heard about who were similar to him. It occurred to me that in each of these cases, a diagnosable mental illness did not seem to exist prior to the behavior’s occurrence. If the individual had not made a suicide attempt, their reported symptoms and behaviors would not have met the threshold for diagnosing a mental illness at all.

In each of these cases, there was little evidence that a mental illness existed before the suicidal behavior occurred. Furthermore, the suicidal behavior served as the only source of evidence supporting the presence of a mental illness. This pattern suggests our traditional assumptions about the causal role of mental illness for all (or nearly all) suicidal behaviors warrants further consideration.

Cavanaugh reported that the percentage of suicide cases across all 76 studies who were diagnosed with a mental illness ranged from 23% to 100%. This means that at least one of the studies diagnosed only one in four suicide decedents with a mental illness and at least one of the studies diagnosed every single suicide decedent with a mental illness. That’s a pretty wide range, to say the least.

  

Poor Reasoning

Because we believe all or nearly all suicidal behaviors are associated with mental illness, we assume that suicidal behavior is a sign or symptom of mental illness. Because we assume that suicidal behavior is a sign or symptom of mental illness, we diagnose individuals who attempt suicide with a mental illness. Because we diagnose individuals who attempt suicide with a mental illness, we later find that all or nearly all suicidal behaviors are associated with mental illness. This loop is not only self-sustaining, it actually may be contributing to the adoption of even more extreme views on the matter, specifically the increasingly popular perspective that the true statistic is actually 95% or even 100%.

First, mental illness is only weakly correlated with suicidal behaviors. Second, a much larger percentage of suicides than we may have traditionally recognized occur in the absence of mental illness.

The unexpected nature of many suicide deaths stands in sharp contrast to the suicide prevention community’s assertion that suicidal behaviors are preceded by warning signs or indicators that signal the behavior’s emergence.

The majority of suicide decedents may therefore show one or more warning signs for suicide but very few individuals who show these warning signs will actually die by suicide. 

In other words, 99.97% of those who said they were having no thoughts of death or self-harm and 99.7% of those who said they were having thoughts of death or self-harm nearly every day were still alive one year later.

Thoughts about suicide or wanting to die should be taken seriously because they reflect being in a vulnerable state, but they do not necessarily lead to suicide. 

99.6% of those who say they are having no thoughts of death or self-harm as compared to 96% of those who say they are having these thoughts nearly every day did not make a suicide attempt in the year after they reported having thoughts about death or self harm. 

What this all reveals is a key challenge for suicide prevention: any given warning sign for suicide will be wrong far more often than it will be right.

Approximately 1 in 7,692 individuals die by suicide and 1 in 303 individuals attempt suicide in the United States each year. If someone reports suicidal thoughts, the likelihood of dying by suicide in a given year increases to approximately 1 in 6,250, and the likelihood of attempting suicide in a given year increases to 1 in 204. Knowing that someone is thinking about suicide is therefore better than not knowing, but not by much.

A significant percentage, therefore, of individuals who attempt or die by suicide deny having any suicidal thoughts at all.

Conversely, an individual may have fleeting thoughts about suicide multiple times per day, such that frequency is high but intensity is low. If thoughts about death and suicide, and warning signs more generally, have such low accuracy and precision as indicators of suicidal behaviors, how can we possibly know when someone is about to attempt suicide?

 

Key to Understanding The Course of Suicide

 Key Assumptions of Emergence

1. Change is constant. 
2. The whole is greater than the sum of its parts. 
3. Each component of a system depends on the other components to function. 
4. Complex systems behave in nonproportional ways

This is, in essence, the central implication of Franklin’s meta-analysis: a lot of things are correlated with suicidal behaviors but none of those things has an especially strong correlation and none of those things is sufficient to explain suicide. An important implication of this assumption is that suicide can result from many different combinations and configurations of these same variables. In other words, there is no single “right” combination of suicide risk and protective factors; many, many, many different combinations of these factors can lead to suicide.

If we assume that suicidal behaviors result from linear cause and effect relationships, we can mistakenly assume that suicide warning signs are always apparent before the occurrence of suicidal behavior when in reality this may not always be the case.

The best we can ever hope for in a suicide-risk screening tool designed to distinguish between those who will die by suicide and those who will not is being right only 1% of the time.

No studies showed that screening led to reduced suicide attempts or deaths. 

In Miller’s study, routine screening of all patients who visited an emergency department for any reason had almost no impact on subsequent suicide attempt rates;

The Change of Pattern, Not the Content is a Predictor of Emotional Imbalance 

As people got closer and closer to killing themselves, for instance, they were increasingly likely to post content describing stressful life situations followed by content describing negative thoughts or expectations.

Social media content did not, therefore, meaningfully distinguish between those who would eventually die by suicide and those who would not, but change in social media content within individuals signaled the approach of suicidal behavior, just as a gymnast’s dramatic body movements can signal an approaching fall but the same body movements across multiple gymnasts may not be so helpful in determining who will fall and who will not.

 

A New Model of Suicide from Emergence, “Jumps” of emotions, not “Slides”

Individuals who have attempted suicide give us a very different picture, however: Most (if not all) were previously in a low risk state but only half experienced intermediate risk states, almost as if they had skipped over these intervening risk levels.

Because of nonproportionality—the fourth key assumption of emergence—change can sometimes be both smooth or continuous and other times be sudden and jarring, almost as if it came out of the blue. Both are possible, but only if we move beyond the unidimensional continuum model of suicide risk.

Individuals who died by suicide often seemed to be doing reasonably well right up until the moment they were attempting suicide, as if they had crossed over some sort of threshold or tipping point. 

Many (though not all) of my patients often said that they didn’t necessarily feel any better or any worse than they had the day before (or the week or the month before that). On the contrary, they often explained that they “just couldn’t take it anymore,” implying that they had crossed some boundary that lies between attempting suicide and not. When I would ask these patients what had changed in the time leading up to their attempt, they often couldn’t identify a precipitating event or point to a particular reason. “I don’t know,” many of them said. “It just happened.” A relatively small change—a slight nudge—had effectively pushed them past this tipping point, resulting in a catastrophic outcome.

  

Five Properties of the Cusp Catastrophe Model of Suicide 

1. Suicide risk has two distinct states. 
2. Change between suicide-risk states occurs suddenly. 
3. Certain suicide-risk states are highly improbable. 
4. Very small changes in risk and protective factors can lead to very different outcomes in suicide risk, specifically, attempting suicide or not. 
5. The conditions under which someone transitions from low suicide risk to high suicide risk do not necessarily coincide with the conditions under which someone transitions from high suicide risk to low suicide risk.

 

Half of suicide attempts occur within five to 20 minutes of deliberation over the act.

Millner interviewed 30 individuals who had recently attempted suicide and asked these individuals to reflect upon their experiences leading up to the attempt. On average, participants reported the following timeline: 

• Five years prior to the attempt, they first experienced suicidal thoughts and experienced these thoughts off and on for several years; 
• Two weeks prior to the attempt, these suicidal thoughts increased in intensity and frequency, such that they were occurring nearly continuously; 
• One week prior to the attempt, they started to think about where to attempt suicide; 
• Six hours prior to the attempt, they experienced an internal debate about whether or not to attempt suicide; 
• Two hours prior to the attempt, they decided upon which method to use; 
• Thirty minutes prior to the attempt, they decided where to make the attempt; and 
• Five minutes prior to the attempt, they made the final decision to act.

For most individuals, suicidal thinking remained relatively low in intensity until just a few hours before making the attempt, at which point the progression toward suicidal behavior rapidly accelerated.

While in a state of suicidal mulling, the very slightest of nudges in either direction could make the difference between returning to a lower risk state or transitioning to a higher risk state where the probability of suicidal behavior is greatly increased.

The tipping point between two discrete suicide-risk states provides the basis for the fourth property of the cusp catastrophe model: Very small changes in conditions can lead to very different outcomes. If this property sounds familiar, that’s because it is also one of the key assumptions of emergence: nonproportionality.

When someone is not at the tipping point, very large changes in stress could have hardly any impact on their suicide-risk level.

Patients who had a very strong wish to die combined with no wish to live—the highest score possible—were more than six times more likely to die by suicide than everyone else. Among patients who had at least some wish to live, however, the risk of suicide was significantly reduced, suggesting that even a small wish to live could offset a whole lot of wish to die. Conversely, a relatively small decline in the wish to live could precipitate the emergence of suicidal behavior.

When it comes to the tipping point between suicidal behavior and the absence of it, the wish to live seems to carry somewhat more weight than the wish to die.

 

Moving from High Risk to Low Risk Takes a Lot of Time 

High-risk states can emerge rapidly, but returning to a low-risk state may take much longer.

Someone who has been psychiatrically hospitalized is more than 100 times more likely than someone who has never been hospitalized to die by suicide. This risk is highest in the weeks immediately following discharge and slowly declines over time, but remains elevated even a year after discharge. This slow, gradual decline in suicide risk after hospitalization stands in stark contrast to the rapid rise in suicide risk that precipitated the hospitalization in the first place. Someone can move from low risk to high risk very rapidly but take months or years to move back to a low-risk state.

I think a better explanation for the incredibly high suicide rate after psychiatric hospitalization is hysteresis. Just as memory foam quickly compresses under pressure but slowly recovers after the pressure has been removed, suicidal individuals may quickly transition from a low-risk to high-risk state when under pressure but experience a much slower recovery after that pressure has been removed. 

Several research studies have found that after someone has attempted suicide, they often experience a decrease in emotional distress.

 

Brain Studies

Those with greater levels of self-control also demonstrated higher levels of brain activity in the prefrontal cortex, a region of the brain involved in decision-making processes, reducing emotional distress, and inhibiting or controlling one’s actions.

The prefrontal cortex is located at the very front of the brain, right behind our forehead, and is involved in how we organize our thoughts and behaviors in ways that help us to achieve our goals. The ventral striatum is located in the center of the brain and is involved in the anticipation of future rewards, the perception and experience of reward, and learning from prior experience. In combination, these two regions of the brain are critical for decision-making, especially decisions that involve the inhibition or stopping of actions that are inconsistent with our desired goals and intentions. We have our prefrontal cortex and the ventral striatum to thank when we get angry but choose not to become violent or aggressive, when we feel overwhelmed by stress but then put together a plan to resolve the situation, and when we’re upset and want to do something destructive but decide against it. Together, these regions of the brain function as an internal braking system.

Among those who have attempted suicide, the part of the ventromedial prefrontal cortex that processes negative emotions such as anxiety, fear, and disgust tends to become overactivated when in stressful situations, especially stressful or uncomfortable interpersonal situations. At the same time, the part of the ventromedial prefrontal cortex that processes positive emotions such as joy, hope, and contentment tends to be underactivated. Because positive emotions and expectations serve to modulate or counteract the effects of negative emotions and expectations, the combination of overactivity in certain areas of the brain and underactivity in other areas can result in heightened sensitivity to stressful situations and increased emotional distress, especially situations with uncertain outcomes. In the presence of such intense, unmodulated emotional distress with no expectation of success or positive outcome, risky “all-in” or “Hail Mary” decisions become more probable.

 

What Helps Restore Balance to Someone Suicidal

Reducing emotional distress and mental illness is certainly helpful, but that strategy alone may not be enough to stop someone’s forward momentum toward suicide. The ability to quickly identify and utilize self-regulatory strategies is also essential, and these concepts are central to the treatments that are most effective for reducing the occurrence of suicidal behaviors. Considerable research supports the value of these approaches, but mental health professionals rarely use them when working with suicidal or high-risk patients. 

Two treatments in particular—dialectical behavior therapy (DBT) and cognitive behavioral therapy for suicide prevention (CBT-SP)1—have demonstrated the ability to reduce the probability of suicidal behaviors in multiple studies conducted by multiple research teams.

Yet another treatment called the Collaborative Assessment and Management of Suicidality (CAMS) also has repeatedly demonstrated the ability to reduce suicidal thinking faster and to a larger degree than status quo treatments.

Each of these treatments differs from status quo mental health treatments in one important way: They aim to directly reduce suicidal thoughts and behaviors, regardless of a patient’s mental illness, rather than indirectly reducing suicidal thoughts and behaviors by reducing mental illness, the approach taken by status quo treatments. Because DBT, CBT-SP, the crisis response plan (CRP), and CAMS directly focus on suicide risk instead of mental illness, they often are collectively referred to as “suicide-focused treatments.”

  

The ”Something” that Helps

If mental illness is unrelated to the reduction of suicidal behaviors in suicide-focused treatments, what is the “something else” that these treatments are acting upon? Based on the research reviewed in the previous chapters, I would argue that the “something” involves decision-making processes related to self-regulation. Specifically, DBT and CBT-SP seem to reduce the probability of suicidal behaviors because they strengthen individuals’ internal braking systems, helping them to choose not to act upon suicidal impulses and urges in stressful situations.

 

DBT

If mental illness is unrelated to the reduction of suicidal behaviors in suicide-focused treatments, what is the “something else” that these treatments are acting upon? Based on the research reviewed in the previous chapters, I would argue that the “something” involves decision-making processes related to self-regulation. Specifically, DBT and CBT-SP seem to reduce the probability of suicidal behaviors because they strengthen individuals’ internal braking systems, helping them to choose not to act upon suicidal impulses and urges in stressful situations.