in suicide among African-Americans—specifically, young black men.
And yet, the demographic group at highest risk is older white men.
Female anorexics, prostitutes, athletes, and physicians all have ele-
vated suicide rates. A theory that can account for these diverse facts
would be persuasive.
Such a theory would not only advance scientific knowledge, but
deepen the understanding of suicidal behavior among clinicians who
need to assess risk, intervene in crises, and design treatment and pre-
16
What We Know and Don’t Know about Suicide ● 17
vention protocols. It would also help those who have lost a loved one
to suicide, who suffer much misunderstanding.
In this chapter, I describe some of my own clinical work and the
supervision of others’ clinical work with suicidal patients. In the clin-
ical literature, suicide is often described as an “urgent,” “vexing,” or
“pressing” issue, one that preoccupies clinicians. Suicide is an urgent issue—it kills people—but urgency need not entail panic. Suicide
can be understood in ways that resolutely point to clear clinical deci-
sions . . . given, that is, a full explanatory model. My and others’ clini-
cal experiences with suicidal patients will highlight how a compre-
hensive account of suicide would have reduced confusion and panic
and facilitated clinical progress.
This chapter also touches on some of my scientific work on sui-
cide. My research group is one of many that have produced new and
important findings regarding suicide. The chapter will include some
basic scientific findings on suicide produced by my and other re-
search groups—facts that any compelling account of suicide must
explain.
I also summarize existing models of suicide in this chapter—theo-
retical accounts that have been developed to explain some of these
facts. One of the best ways to evaluate a theoretical model is the
number of facts it can explain, and some of these models are more
successful than others, as we shall see. My hope is that this book’s ex-
planation of suicide will save people some of the misunderstandings
my family and I went through, will refine clinicians’ approach to
treating suicidal behavior, and will set a scientific agenda for the
study of suicide. In the process, some interesting questions will be
raised and addressed. For example, should family members tell the
truth about the cause of death when a loved one has died by suicide?
What constitutes a proper definition of suicide itself? How are we to
understand the deaths of those who jumped from the World Trade
Center towers’ upper floors on September 11, of the September 11
18 ● WHY PEOPLE DIE BY SUICIDE
terrorists, and of those in mass suicides in cults? What protects most
women from suicide, and yet, why do some very different subgroups
of women—such as prostitutes and physicians—share similarly high
suicide rates? Why are older, white men the demographic group in
the United States most vulnerable to suicide? Why do suicide rates
decrease in the United States during times of national crisis and de-
crease in a particular city when the city’s professional sports team is
making a championship run? What are the constituent parts of the
genuine desire for death? These and other questions will be raised
and addressed throughout the book.
Notes from the Clinic
My first job after getting my doctorate was as an assistant professor
of psychiatry at the University of Texas Medical Branch at Galveston.
What a blessing this job was in many ways. I saw many psychother-
apy patients and worked with skilled psychiatrists who taught me a
lot about the biological bases of mental disorders. Biology appears to
play some role in why people die by suicide, a fact I will address later
in this book. But they also taught me something more—an attitude
about suicide risk in patients that was neither dismissive nor alarm-
ist. The alarmist position is perhaps the easiest to understand—this
threatening situation and alarms should be sounded. This idea oc-
curs in settings in which staff see relatively few people with serious
mood disorders. In settings where serious mood disorders are com-
mon, people understand that suicidality is just part of the disorder;
the majority of people who experience mood disorders will have
ideas about suicide, and the vast majority will neither attempt sui-
cide nor die by suicide. If 911 were called in each of these cases, a
“cry wolf ” scenario would quickly develop. Alarmists are making a
mistake in conditional probability. Given the existence of a suicidal
What We Know and Don’t Know about Suicide ● 19
thought or behavior, they mistakenly estimate the probability of
death or serious injury by suicide to be higher than it is.
Although alarmists make a mistake, it is not hard to see why they
do. When people have ideas about suicide, it is quite true that risk is
elevated compared to people who do not have suicidal ideas. More-
over, suicide is irreversible, and everything possible should be done
to prevent it. Alarmists overreact, but they are doing so in the safe di-
rection; “better safe than sorry,” they might say.
The alarmist problem is easy to notice in training clinics. Most of
the pages I receive on my beeper are from therapists at the training
clinic I direct who are worried that they should do more for a patient
with suicidal symptoms. When I return the call, I ask a series of ques-
tions to see if the therapist is meeting the standard of care. In our
clinic, meeting the standard of care is routine. And so I will then say,
“Well, you’ve done everything I would’ve done; I wonder, what else is
it that you think you’re supposed to do?” The answer is often, “I’m
not sure, I just have this feeling that there’s something else I should
do.” Then I’ll say, “Well, there’s not; but don’t lose that feeling, be-
cause it will ensure that you regularly do what’s best for patients;
also, though, don’t let that feeling get out of hand, because it can
burn you out, plus, ultimately these choices are not up to us, they’re
up to patients.” Make no mistake, the standard of care is impor-
tant—at times even life-saving—and therapists are expected to meet
it rigorously, including involuntary hospitalization of the patient if
needed. But beyond that, responsibility for life choices resides with
patients. Therapists who see this are likely to enjoy their work more,
to not be distracted by one patient when dealing with another, and,
importantly, to enjoy their nonwork time as well.
The alarmist attitude is understandable but, especially if exagger-
ated, mistaken. Those who take a dismissive approach make a mis-
take in the opposite direction. They become blasé about suicidal be-
havior, often attributing it to manipulation or gesturing on the part
20 ● WHY PEOPLE DIE BY SUICIDE
of the po
tentially suicidal person. This problem is acute when it
comes to the often misunderstood borderline personality disorder,
which is characterized by a long-standing pattern of out-of-control
emotions, interpersonal storminess, feelings of emptiness, and im-
pulsive behaviors, including impulses toward self-injury. Some clini-
cians take a dismissive attitude toward patients with this disorder be-
cause they believe that these patients merely “gesture” suicide. In
other words, they engage in suicidal behaviors, such as cutting them-
selves, but do not really intend to kill themselves; instead, they only
intend to provoke or manipulate others. I wish this were true, but it
is not—approximately 10 percent of patients with this disorder end
up dying from their suicidal gestures (comparable to the rate for pa-
tients with mood disorders). The following quotation illustrates this
misunderstanding:
The borderline patient is a therapist’s nightmare . . . because border-
lines never really get better. The best you can do is help them coast,
without getting sucked into their pathology . . . They’re the chroni-
cally depressed, the determinedly addictive, the compulsively di-
vorced, living from one emotional disaster to the next. Bed hoppers,
stomach pumpers, freeway jumpers, and sad-eyed bench-sitters with
arms stitched up like footballs and psychic wounds that can never be
sutured . . . Borderlines go from therapist to therapist, hoping to find
a magic bullet for the crushing feelings of emptiness.1
This characterization is demonstrably false. Patients with border-
line personality disorder do get better. A persuasive study found that 34.5 percent of a sample of borderline patients met the criteria for
remission at two years, 49.4 percent at four years, 68.6 percent at six
years, and 73.5 percent over the entire follow-up. Only around 6 per-
cent of those who remitted then experienced a recurrence.2
The dismissive attitude is dangerous for another reason. A main
thesis of this book is that those who die by suicide work up to the act.
What We Know and Don’t Know about Suicide ● 21
They do this in various ways—for instance, previous suicide at-
tempts—and all of these various ways have the effect of insulating
people from danger signals. They get used to the pain and fear asso-
ciated with self-harm, and thus gradually lose natural inhibitions
against it. Clinicians’ dismissive attitudes have the potential to model
a blasé attitude about self-harm. If clinicians blithely get used to sui-
cidal behavior, their patients may vicariously do so as well.
The psychiatrists at my first job balanced the alarmist and dismiss-
ive positions very well. They clearly understood the danger and hor-
ror; in fact, most of them had had a patient who had died by suicide.
They knew the standards of care for suicide risk assessment and the
treatment of suicidal behavior, and they followed them faithfully. But
they understood the limits of their interventions, they understood
people’s ultimate autonomy, including their freedom to occasion
their own death if they really were committed to doing so. My im-
pression was that these psychiatrists did their job well during the day,
and slept well at night.
Consider for example the case of Gayle (a false name). In retro-
spect, I understand Gayle’s situation clearly, but when I was seeing
her, I was uneasy. She was the sort of patient who seemed potentially
self-destructive. Indeed, she often fantasized about death by suicide,
envisioning a particularly graphic means—severing her hand with a
machete and bleeding to death (people have died in just this way, in-
cidentally). She even owned a machete. This would be enough to
concern any clinician, and I was no exception. I recommended that
Gayle be hospitalized, so that she would remain safe while treatments
for her substantial depression were started.
She refused hospitalization and also refused antidepressant medi-
cines; she would agree only to psychotherapy. An initial question,
then, was whether I should hospitalize her involuntarily. I had the
sense that this would not be best, but I was having trouble putting
my finger on exactly why she did not require hospitalization. After
22 ● WHY PEOPLE DIE BY SUICIDE
consultation with colleagues, I was reminded of some mildly reassur-
ing facts. Gayle was around forty-five years old and had never at-
tempted suicide. She had had plenty of time to have tried it, and yet
had not. This is no guarantee. There are people who at age forty-five
or even sixty-five attempt suicide for the first time and die. Still, the
fact that she had not had previous experience with suicidal behavior
was mildly reassuring. Her gender was another mildly reassuring fac-
tor—women are a lot less likely to die by suicide than are men. Also
somewhat reassuring were her connections to life. There were things
that she was proud of regarding her professional life, and more im-
portant, she was deeply connected to her young son. She spontane-
ously mentioned these things as I questioned her about suicide po-
tential.
Gayle was also the rare person who clearly met criteria for a major
depressive episode but who had an absence of depressed mood. In a
study of young adults my colleagues and I conducted, this pattern
was found to occur in only about 5 percent of those who were in a
depressive episode. Recent work has shown lack of depressed mood
to be a positive prognostic indicator among depressed people; that is,
they tend to get better quicker and to have good outcomes.3
Throughout this book, I will argue that the acquired ability to en-
act lethal self-injury is crucial in serious suicidal behavior. People are
not born with the developed capacity to seriously injure themselves
(although they are born with factors, including certain genes, that
may facilitate the future development of this capacity). In fact, if any-
thing, they are born with the opposite—the knee-jerk tendency to
avoid pain, injury, and death. That is, people have strong tendencies
toward self-preservation; evolution has seen to that. Through an ar-
ray of means described later, some people develop the ability to beat
back this pressing urge toward self-preservation. Once they do, ac-
cording to the theory laid out in this book, they are at high risk for
suicide, but only if certain other conditions apply—namely that they
What We Know and Don’t Know about Suicide ● 23
feel real disconnection from others and that they feel ineffective to
the point of seeing themselves as a burden on others. These factors,
like the acquired ability to enact lethal self-injury, are covered in de-
tail in later chapters of the book.
I now understand clearly why Gayle made me feel uneasy, but also
why she was not at particularly high risk for suicide. She had ac-
quired the ability to enact lethal self-injury. A main way that people
develop this capacity is through previous suicidal behavior. As noted
already, Gayle
had not engaged in such behavior. What I believe led
to her developing this capacity was a long history of severe substance
abuse, which included many painful and provocative experiences
(another way to gradually beat back the instinct to survive). Her sub-
stance abuse had ended; she had been clean for around eight years
when I saw her. But her earlier experiences had left various residues.
This ability in Gayle was manifested by her having a clear and de-
tailed suicide plan, but especially in her sense of calm and her lack of
fear about the plan. These were the things that made me want to hos-
pitalize Gayle. Nevertheless she was not at particularly high risk for
suicide, and the reason involves two other factors that I believe are
required for serious suicidal behavior—thwarted belongingness and
perceived burdensomeness. Gayle had a fairly well-developed circle
of friends and was very connected to her son. There was no evidence
that she felt fundamentally disconnected from others, and plenty of
evidence that her sense of belonging was very much intact. Similarly,
Gayle was a particularly capable woman; for instance, even when de-
pressed, she was the office’s top performer in her professional line of
work. There was no evidence that she felt ineffective, certainly not to
the point that she believed she burdened others.
Her sense of belonging and effectiveness buffered her, but it is im-
portant to note that this could have changed rapidly. People cannot
develop the ability to lethally injure themselves quickly; the experi-
ences that are required take time and repetition. By contrast, people
24 ● WHY PEOPLE DIE BY SUICIDE
can quickly develop views that they do not belong or that they are
particularly ineffective. Thus, in a case like Gayle’s, suicide risk can
quickly escalate. Repeated risk assessment is thus necessary in Gayle’s
case (and is a safe clinical practice anyway).
The case of Sharon (a false name) is interesting by way of contrast.
When questioned about suicide risk, Sharon articulated no plan at
all. When pressed a little on the question, she made statements like,
“I can’t imagine actually trying suicide, it’s just that I have the sense
that I’d be better off dead.” Like Gayle, Sharon had never attempted