Antecedents, Reasons for, and Consequences of Suicide Attempts: Results from a Qualitative Study of 89 Suicide Attempts among Army Soldiers

Abstract

Most studies aimed at understanding suicidal behavior have focused on quantifying the associations between putative risk factors and suicidal behavior in comparative studies of cases and controls. The current study, in comparison, exclusively focused on cases – 89 Army soldiers presenting for hospital care following a suicide attempt – and attempted to reveal the antecedents of, reasons for, and consequences of suicide attempts. This mixed-methods study using qualitative interviews and self-report surveys/interviews revealed that in most cases, the most recent onset of suicidal thoughts began shortly before the suicide attempt and were not disclosed to others, limiting opportunities for intervention via traditional approaches. The primary reason given for attempting suicide was to escape from psychologically aversive conditions after concluding that no other effective strategies or options were available. Participants reported both negative (e.g., self-view, guilt) and positive (e.g., learning new skills, receiving support) consequences of their suicide attempt – and described things they believe would have prevented them from making the attempt. These findings provide new insights into the motivational and contextual factors for suicidal behavior and highlight several novel directions for prevention and intervention efforts.

General Scientific Summary

In this study, researchers conducted in depth interviews with 89 Army soldiers who had just tried to kill themselves. The aim was to increase understanding of why people engage in such behavior. Results revealed that the primary motivation for suicide was to escape from seemingly intolerable psychological states. Most people attempting suicide did not tell someone else about their plans ahead of time. Respondents described things that would have made them change their mind about making a suicide attempt. The results of this study provide valuable information about why people try to kill themselves and how we might better prevent them from doing so.

Suicide is among the most perplexing aspects of human behavior. Over the past 50 years, psychiatric and psychological researchers have increasingly studied the question of why people die by suicide. Most of this research compares cases and controls on baseline factors that predict the subsequent occurrence of suicidal behavior, with 70-80% of these studies evaluating associations involving a relatively small set of potential explanatory factors, such as the presence of mental disorders (Franklin et al., 2017). Despite these efforts, we lack a firm understanding of many fundamental aspects of suicidal behavior (Millner et al., 2020).

Virtually all prior research has relied on testing the hypotheses that researchers have about the factors hypothesized to differ between people who have tried to kill themselves and people who have not. One potential approach to increase our understanding of why people try to kill themselves is to change the lens through which we view the behavior by asking people who made suicide attempts to describe, in their own words why they made that decision, and what they think might have prevented their suicide attempt.

Several prior qualitative studies of this sort have been carried out. These studies focused primarily on the psychological experiences and external events that preceded the suicide attempts. One common finding from these studies is that people who attempt suicide describe experiencing multiple interacting stressors and negative affective states leading them to attempt suicide (Denneson et al., 2020; Kidd, 2004; Montross et al., 2014; O’Brien et al., 2021).

However, prior qualitative studies have been limited in several important ways. First, most of them examined narratives from people who interacted with suicide decedents (e.g., Skopp et al., 2019), those with suicidal ideation (e.g., Denneson, 2015), or those with either suicidal or nonsuicidal self-harm experiences (Crawford, 2009) rather than interviewing those surviving a suicide attempt. Second, prior studies did not report information obtained from respondents about the specific reasons or purposes of the suicide attempts (e.g., to end aversive psychological states, to benefit their families), nor did they ask why they attempted suicide as opposed to another behavior/option, thus requiring the researcher (and reader) to make these important inferences for themselves. Third, most of the few studies that interviewed individuals after their suicide attempt were based on relatively small samples (e.g., 1-20) and reported asking relatively few questions (e.g., 1-5). Taken together, these issues limit the value of prior qualitative studies, leaving important information about the context of suicide attempts uninvestigated.

Here we report on the results of a mixed-methods (qualitative and quantitative) study of people who made a recent suicide attempt. We focused on a group at recently elevated risk for suicide, current soldiers in the US Army. The suicide rate among soldiers has increased dramatically in recent years, and so there is an urgent need to understand why soldiers are dying by suicide at such elevated rates (Nock et al., 2013). Although some risk factors for suicide may be unique to Army soldiers, there is evidence that many of the key risk factors for suicide among soldiers are similar to those in the general population (Nock et al., 2018; Ursano et al., 2015), suggesting that results from such a study may help to understand suicide more generally. This study was carried out as a component of Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS), a large, multi-component, epidemiological and neurobiological study aimed at identifying modifiable risk and protective factors for suicide (Kessler et al., 2013; Naifeh et al., 2019; Ursano et al., 2014).

This study was designed to overcome many of the limitations of prior qualitative studies of suicide. The primary aims of this first analysis of the data are to describe the antecedents, reasons for, and consequences of suicide attempts as described from the perspective of those who recently made a suicide attempt to advance the understanding, prediction, and prevention of suicidal behavior.

Method

Participants

Potential participants included 104 active-duty Army soldiers currently hospitalized following a suicide attempt, who were recruited from five Army hospitals across the US: Walter Reed National Military Medical Center (Washington DC), Fort Bragg (North Carolina), Fort Hood (Texas), Joint Base Lewis-McChord (Washington state), and Fort Stewart (Georgia). Inclusion criteria included: admission to an inpatient unit following a reported suicide attempt, ≥18 years old, medically stable as determined by a physician, and English fluency. Exclusion criteria included: any factor that may have impeded an individual’s ability to understand and appropriately complete the study (e.g., organic brain damage, florid psychosis), and not making a full suicide attempt involving self-injurious behavior with intent to die (n=15). All study methods were approved by the Committee for Use of Human Subjects at Harvard University and the Uniformed Services University of Health Sciences.

Potential participants were identified within 72 hours of hospital admission via attending physician referral, morning staff meetings/review of admission lists, or review of patient registries. Soldiers who met criteria were approached by an on-site research assistant. Following informed consent, participants completed a telephone-based qualitative interview (M=32 mins; SD=12 mins; range=8-80 mins) with an off-site study team member located at Harvard University. The on-site research assistants and clinical care providers (e.g., nurses, doctors) performed the participant risk assessment and monitoring before and after the interviews.

Study Procedures

Qualitative interview.

We developed a qualitative interview that probed the antecedents (e.g., internal and external experiences/events), reasons for, and consequences of suicide attempts. The interview questions were generated by seeking the guidance of clinicians and clinical researchers with decades of experience specializing in the assessment and treatment of suicidal patients, (e.g., Drs. Aaron T. Beck and Marsha Linehan). The resulting interview consists of 11 open-ended questions (see Table 1). The interviews with suicide attempt survivors included these open-ended questions, followed by person-specific probes to clarify and gain additional information (e.g., Merton, Fisk, & Kendall, 1990). All interviews were audio-recorded (with permission) and transcribed verbatim by trained transcribers.

Table 1.

Qualitative Interview Questions

1. Please tell me about the period leading up to your suicide attempt. That is, what was happening in your life, what were you doing, what were you thinking, and what were you feeling during this period?

2. Was there a particular moment when the thought of suicide came into your mind? What was happening when that occurred? What made you think of suicide?

3. Was there a specific moment when you decided to act on your thoughts and make a suicide attempt? What was happening when that occurred? What made you decide to attempt to kill yourself? After you made that decision, how long was it until you made the attempt?

4. Did you tell anyone that you were feeling distressed (or whatever word they used to describe how they were feeling)? Did you tell anyone that you were feeling suicidal? Who did you tell? What was the response?

5. Do you think something could have occurred that would have made you change your mind and not make the attempt? What?

6. Why did you try to kill yourself? (Ask follow-up questions as needed to get at: What were you trying to accomplish by making a suicide attempt? What was the purpose of your attempt?)

7. Why did you try to kill yourself, rather than use some other strategy to achieve this end? What else did you consider? What else did you try? Were any of those things helpful? Why not? How did you come to think about using suicide as a means to achieve this end?

8. What did you think would happen to you after you made the attempt? What actually happened?

9. Did your suicide attempt have an impact on your relationships, your problems, the way you have been treated, your work situation? If so, how? What was the most negative effect? The most positive effect?

10. Looking back, how do you feel about having made the suicide attempt?

11. Why in your opinion is suicide a problem in the Army right now? What is your take on why service members are making suicide attempts?

Transcription.

Audio recordings were transcribed using F4 transcription software. A transcription manual was created to ensure that transcriptions were accurate, complete, and consistent across interviews. Transcribers were trained undergraduates and volunteer post-baccalaureate research assistants. Each transcription was reviewed twice by independent reviewers to ensure accuracy and completeness. After all transcriptions were finalized, interview audio recordings were destroyed to ensure participants’ confidentiality.

Qualitative Coding.

Interview transcripts were human-coded (i.e., with no automated/ natural language processing tools) and analyzed using a method that integrated two distinct qualitative analysis approaches: (a) directed content analysis (Hsieh & Shannon, 2005), a deductive, theory-based approach aiming to corroborate and extend existing constructs/understandings of qualia (i.e., subjective experiences); and (b) grounded theory (Glasser & Strauss, 1967), an inductive, atheoretical approach positing that understanding arises as, or is an emergent property of, being grounded in the qualia (not existing notions). Concretely, qualitative coding consisted of three phases. First, a taxonomy of codes (termed coding scheme or ‘codebook’) was developed and implemented via the qualitative analysis computer program MAXQDA. Second, coders were trained and tested on their knowledge and application of this codebook and program. Third, coders classified/characterized the meaning of interview content (instantiated in words, phrases, sentences, and paragraphs) by applying codes to it. Codebook development, coder training, and interview transcription methods are described in detail in the Supplemental Text.

Sociodemographic and Clinical Assessment.

All participants also completed self-reported surveys assessing their sociodemographic. Participant demographic data were compared against the Department of Defense Suicide Event Report (DoDSER; Smolenski et al., 2013) to evaluate the generality of the results to the rest of the US Army. We also generated current (past 30-day) DSM-IV diagnoses for 12 of the most common disorders. All participants completed the Composite International Diagnostic Interview screening scales (CIDI-SC)(Kessler, Calabrese, et al., 2013) and a modified version of the PTSD Checklist (PCL)(Weathers, Litz, Herman, Huska, & Keane, 1993). We assessed for the presence of both internalizing disorders (major depression, bipolar, generalized anxiety, panic, and PTSD) externalizing disorders (ADHD, intermittent explosive disorder, and substance use disorder). All disorders were assessed without DSM-IV diagnostic hierarchy or organic exclusion rules. The CIDI-SC and PCL have good concordance with independent clinical diagnoses (Kessler, Santiago, et al., 2013).

Analysis Plan

This study aimed to provide a better understanding of the antecedents, reasons for, and consequences of suicide attempts through the perspectives and own words of those who have attempted to kill themselves. We summarized the data derived from the methods above in two ways to achieve this. First, we provide a quantitative summary of the major themes discussed by the participants in response to each interview question in the form of frequency counts, rank-orderings, and other statistics describing the codes that were applied to classify interview content (c.f., Creswell, 2003; Onwuegbuzie & Teddlie, 2002). Second, alongside the quantitative results, we provide sample qualitative/textual data in the form of direct quotes from participants that exemplify particular codes. Third, we used a mixed-methods approach to compare the results of our qualitative interviews with those from our clinically assessed quantitative data.

Results

Characteristics of the Sample and of their Suicide Attempts

An initial review of the details of the suicide attempts revealed that 15 of the 104 participants referred to the study team did not engage in behavior that meets the definition of a suicide attempt (i.e., self-injurious behavior with some intent to die). More specifically, seven reported making an interrupted suicide attempt (i.e., someone stopped them before they attempted suicide), and six made an aborted suicide attempt (i.e., they stopped themselves behavior any injury occurred). Another participant led someone to believe they would attempt suicide but did not, and one engaged in potentially lethal behavior without suicidal intent. This resulted in a final analytic sample of 89 participants.

The majority of participants were male, white, under 30 years old, and had less than a college degree (see Table 2). The sociodemographic characteristics of the sample generally matched that of the population of US Army soldiers who attempted suicide at the time of the study (Smolenski et al., 2013), suggesting this study includes a representative sample of participants. Ninety-one percent of participants met criteria for at least one mental disorder (Table 2), with major depression, generalized anxiety, and post-traumatic stress disorder the most frequently endorsed.

Table 2.

Sociodemographic and Diagnostic Characteristics of Suicide Attempters, and Comparison with the US Army

	Suicide Attempt Cases Completing Interviews (n = 89)		Suicide Attempts in US Army (n = 365)a
	N	%	N	%
Gender
Female	20	22.5	82	22.5
Male	69	77.5	283	77.5
Race/ethnicity
White	40	44.9	229	62.7
Black	23	25.8	70	19.2
Hispanic	13	14.6	–	–
Other	7	7.9	66	18.1
Multi-Racial	6	6.7	–	–
Current Age
≤ 24 years	28	31.4	180	49.3
25 – 29 years	26	29.2	105	28.8
30 – 34 years	18	20.2	35	9.6
≥ 35 years	17	19.1	45	12.3
Education
< High schoolb	8	9.0	32	8.7
High school diploma	30	33.7	140	38.4
Some collegec	41	46.1	107	29.3
≥ 4-year college degree	10	11.2	21	5.8
Unknown	–	–	65	17.8
Marital status
Never married	21	23.6	131	35.9
Currently married	36	40.5	160	43.8
Previously married	23	25.9	46	12.6
Missing	9	10.1	28	7.7
Current (30-day) DSM Disorders Internalizing Disorders
Major Depressive Disorder	58	65.2
Bipolar Disorder	29	32.6
Generalized Anxiety Disorder	50	56.2
Panic Disorder	26	29.2
Posttraumatic Stress Disorder	47	52.8
Externalizing Disorders
Substance Use Disorder	27	30.3
Intermittent Explosive Disorder	34	38.2
Attention-Deficit/Hyperactivity Disorder	29	32.6
Any Mental Disorder
Yes	81	91.0
No	8	9.0

^aArmy suicide attempt data are from the 2012 DoDSER (Smolenski et al., 2013).

^bIncludes GED.

^cIncludes post-high school education without a diploma, post-high school education with a technical certification, and 2-year Associate’s Degree.

The majority of the 89 suicide attempts occurred within the week before the interview (74.1%) and virtually all (96.4%) occurred within the month before. Most suicide attempts occurred in the interviewee’s home or barracks (74.1%), their car (9.0%), or a range of other areas (e.g., work, outside). The primary methods employed were drug overdose (67.4%), cutting with a sharp object (12.4%), hanging (6.7%), and carbon monoxide poisoning (4.5%). Most participants reported high (33.7%) or moderate (49.5%) intent to die, whereas a minority (15.7%) reported low intent to die (1.1%=unclear). Relatedly, most participants (61.8%) reported that they believed they would die from their attempt, 24.7% thought the result would be sleep, and 18.0% were unsure.

Nearly all (90%) of these soldiers reported that they began thinking about suicide continuously (without a break of seven or more days) within the year before their attempt, 78.6% within the past month, 43.9% within the past week, 31.4% the day before, and 27% the day of the attempt (see Table S1 for more detail). Note, this was not necessarily their first time thinking about suicide in their lifetime, but the onset of continuous thoughts leading up to this attempt. Most soldiers (60.7%) had disclosed to someone else that they were experiencing distress; however, most (58.4%) did not disclose their suicidal thoughts to others prior to their attempt; 34.8% did, and disclosure was unclear in 6.8% of cases. Discovery of the suicide attempt occurred via being found by others (43.8%), telling others after the attempt (31.5%), passive communication to someone during the attempt (15.7%), or being observed by others during the attempt (7.9%).

Antecedents of Suicidal Thoughts and Attempts

Participants’ descriptions of the factors that occurred just before their suicide attempt that they believe contributed to their thinking about suicide and deciding to kill themselves are summarized in Table 3. The factors described by soldiers were similar for suicidal thoughts (Table 3a) and the decision to attempt (Table 3b), primarily relating to negative internal/psychological experiences and stressful life events. In terms of psychiatric symptoms, emotional factors (e.g., depression, exhaustion, anxiety, agitation) were most common, followed by cognitive (e.g., hopelessness, rumination) and behavioral factors (e.g., sleep difficulties, alcohol problems). Results from the diagnostic assessment provided convergent validity such that rates of internalizing disorders were highest among those reporting emotional factors (55.7%-63.3%), slightly lower among those reporting cognitive factors (39.2%-49.4%), and lower still among those reporting behavioral factors (16.5%-20.3%). In terms of stressful life events, the most commonly described were those involving social or relationship problems (e.g., argument with partner/friend, separation from family), work stressors (e.g., workload increases, unsupportive supervisor), romantic/relationship issues (e.g., divorce, separation), financial concerns, and concerns about the military system. A range of other factors endorsed less frequently (by <10% of respondents) are presented in Table 3.

Table 3a.

Percent of Sample Endorsing Antecedents for Suicidal Thoughts Relevant to their Recent Suicide Attempt

		Quotes
Psychiatric Symptoms
Any Emotional Symptom Endorsed	60.7%	R: I have, uh, I guess diagnosed OCD (Obsessive-Compulsive Disorder). Started cleaning, getting a little agitated, angry, ended up started throwing things, um, kind of went into an angry fit.
Any Cognitive Symptom Endorsed	40.4%	R: Um, yea (yes), uh, tired of being depressed and sad and tired of like thinking about those same images and alcohol’s impulsiveness, I guess (chuckles). 00:10:52-8
Any Behavioral Symptom Endorsed	20.2%	R: Um, and just basically I was crying for about 2 weeks. Nobody paid any attention, I stopped talking to people, I stopped smiling. 00:03:15-6
External Major Life Events
Any Work Stressor Endorsed	47.2%	R: There was a lot of us that had lives outside of work. And they didn’t let us live our lives outside of work, they wanted us to do extra, extra stuff and then come back early in the morning to like feed 500 soldiers and they were causing stress on us, we didn’t get no sleep or nothing, and I was like I couldn’t do this anymore. It was like a suicidal thought like a few weeks prior to me actually getting the chance to handle the suicidal thought in my head…Yea, it was like the first phase of it. It was, it was going on, it was like it was something that I thought about and I was like, I wasn’t really serious about it but it was something that I was thinking about. And I was like, “I don’t know, I don’t know if I should do it or not.”
Any Social Support Stressor	46.1%	R: Oh, family issues. I’m separated from my family um (pause). My wife is, um, in [State A], and I’m in [State B]…And, my motheriin-law has cancer and won’t have it removed, and is kind of in early stages of dementia…So my wife is stressed out every day, not wantin’ (wanting) to go to her own house. And sent me a missed text talking about, um, “do you know any good divorce lawyers?”… And another one that it said, “sometimes I wish I was dead.”
Any Romantic Relationship Stressor	25.8%	R: On [date of suicide attempt] I found out that that [my partner] was cheating, and uh the way I found out was via text message, pictures, and a voicemail.
Any Financial Stressor	18.0%	R: The military, and me, being separated from the military, I did not know what I’d be able to do, because, uh, they were going to try to chapter me out and you know, uh, I thought that, all my benefits were gonna (going to) (sigh) get taken away from me…You know, I was gonna (going to) fall and relapse back into I guess, the street life, I guess you could say.
Any Stressful Experience within Larger Military System	11.2%	R: Um just uh (3s pause) just a lot of like me and my first sergeant, um or should I just say me and my second in command aren’t really getting along. Lately they’ve been trying to, you know, chapter me and stuff like that out of the Army for um pretty much just minor infractions and stuff like that. Things that usually, you know, you’ve seen, or I’ve seen, other people would just get away with uh which is kind of irritating.
Any Death of Other	7.9%	R: Um, I had a lot of depression and anxiety. Uh, ah, I bottled my emotions and I really hadn’t grieved over like my best friend dying from my last deployment.
Any Crime/Legal Stressor	6.7%	R: I was falsely accused of sexual assault.
Any Deployment Issue (Back-to-Back Deployment, Stressful Reintegration)	4.5%	R: It’s home and my husband. I just feel like I’ve started, I just felt like I lived in, uh, Iraq or in Afghanistan [more than in the United States]…would be in the States for about 4 months, and then I went on another deployment, and it was the same routine every year.
Any Health Stressor	4.5%	R: And, he, uh, he basically accused me of not wanting to be a Chief anymore. And I informed him that, no, that was not the case, it was that I was physically unable to continue doing this and I, I um, like I continued to push myself and I basically got to a breaking point because I kept pushing myself, kept pushing myself, kept doing as best as I could but my pain was really, um, wearing me down quite a bit.

Participants also provided information about whether “something could have occurred that would have made [them] change [their] mind and not make the attempt.” Most (58.4%) indicated that something could have changed their mind. The factors most often described were if others knew about and helped manage their distress (27%), if the aversive conditions they were experiencing were relieved (27%), as well as a range of other factors presented in Supplemental Table 2. Table S2 presents sample quotes from soldiers that provide more vivid explanations of these factors in their own words.

Reasons for/Function of Suicide Attempts

When asked “why did you try to kill yourself?”, most (88.8%) participants endorsed doing so to escape an aversive psychological state (intrapersonal negative reinforcement), as presented with example quotes in Table 4. There was a high rate of DSM internalizing (89.9%) and/or an externalizing (58.2%) disorders among those endorsing an escape motivation for their suicide attempt. The next most commonly endorsed reason was to escape an aversive social situation (interpersonal negative reinforcement; 21.4%). Those endorsing such a motivation for their suicide attempt reported a higher number of 12-month social support stressors (M=1.3; SD=1.6) in their self-report survey than those not endorsing a social motivation (M=0.6; SD=1.2), though this difference was only marginally significant (t[87]=1.91, p=.06). Other reasons endorsed by multiple soldiers, albeit less frequently, included ‘to die’ (12.4%) and to achieve some other interpersonal outcome (11.2%; e.g., benefit their family financially).

Table 4a.

Reasons for/Function of Suicide Attempt (Not Mutually Exclusive)

Why did you try to kill yourself?		Quotes
Intrapersonal Negative Reinforcement	88.8%	(1) R: To escape from the, the pain and ah…emotional pain and the, the symptoms I was going through, that I had been going through for 3 years, that I felt weren’t going away. (2) R To run away from my problems…To, to, to, to get, to escape…And, and it, was it also to not be tormented …by the negative emotions.
Social Negative Reinforcement	21.4%	(1) R: Just getting rid of my problems. The problem being, what if I got this baby sick and the other problem being I got the husband, the fiancé, mad…Wanted to get rid of the problem (2) R: At least I wouldn’t have to go through the agonizing. Everyday torture…I was thinkin’ (thinking) just not having to face that man.
To Die, Specifically	12.4%	(1) R: Um, after I took the pills, then, I wanted to die. I wasn’t (was not) trying to uh, I guess reach out for anything…Yea (yes), I wanted to die that day. (2) R: Uh, (pause) I actually wanted to die.
Other-Directed Social Outcome	11.2%	(1) R: Um, I was trying to, um, get my wife the money, the 400,000 dollars…I was gonna (going to), you know, my wife was gonna (going to) have 400,000 dollars so that she could take care of the kids. (2) R: Well, knowing my family would be better off with some type of financial, financial, uh, future, versus me just, you know…
Social Positive Reinforcement	9.0%	(1) R: Um, well at the time, uh, I was just thinkin’ (thinking) that maybe, maybe this will show the wife that, you know, the whole, the way our marriage is and everything, it’s not working…And I couldn’t live that way anymore. And I could never get her to actually talk to me or open up…So, during that time I was taking those pills and stuff, I figured, you know, if she, if she really wants anything to do with me, she’ll, you know, show up or come by or something. (2) R: Um, I mean I also, If I lived through it I wanted to be able to be taken seriously too.
Intrapersonal Positive Reinforcement	1.1%	(1) R: I guess to feel more alive (2s pause) cause (because) I don’t feel alive…And so that’s why I thought when I take it, or when I try and kill myself that I’m actually gonna (going to) be alive. Afterwards.
Not Enough Information	1.1%	(1) R: Um with this one, I couldn’t tell you.

Respondents also were asked “why did you try to kill yourself, rather than use some other strategy to achieve this end?” The most commonly endorsed reason was feeling trapped with no other options (46.1%) and similarly that all other options had been tried and failed (28.1%). Soldiers also reported an inability to implement other (non-suicidal) strategies but failed due to intrapersonal (27%; e.g., emotional exhaustion) or interpersonal barriers (13.5%; e.g., not knowing/trusting people who could have helped).

Consequences of Suicide Attempts

Participants were asked “looking back, how do you feel about having made the suicide attempt?” The most commonly reported feelings were negative views of oneself (34.8%), regret (33.7%), and guilt (16.9%) about having made the attempt (Table 5). Only 11.2% of soldiers were disappointed that they survived, and only one respondent (1.1%) reported some intent to make another suicide attempt in the future.

Table 5.

Percent of Sample Endorsing Consequences for having Made a Recent Suicide Attempt (Not Mutually Exclusive)

		Quotes
Negative Views of Self	34.8%	R: Um, I’m very disappointed in myself, um, very ashamed, //um//…I know I let down a lot of people.
Regret It	33.7%	R: I mean, uh, at the time I thought it was the right decision, but now looking back I realize that it wasn’t the right decision. Because, um, (I/A 4 secs)… I mean I realize now it wasn’t a, a smart decision based on all of the life, and everybody else it would have impacted. 00:38:26-9
Helpful or Effective in Long-Term	27.0%	R: I mean, in all honesty, looking back on it, it, it was very stupid. Um, but at the same time, it, uh, it helped…I mean it’s weird to say that it helped because, but it did, I mean it, it put a whole new perspective on life, on my relationship with my wife, what needs to really be done. …So that’s why, in a way, I mean it was stupid, I shouldn’t’ve (should not have) done it, but doing it, you know, I got clarity and…where, how, how bad the relationship really is…And I’ve actually gotten, you know, (2 s pause), um the answers of what I need to do // to fix it
Guilt Due to Impact on Others	16.9%	R: That fact that uh, some of my soldiers saw it…And know that I allowed my emotions to get the better of me. The fact that my wife knows about it and I had to put my family through that. The, the. the fact that I had to put people through that, was the worst part. 00:32:50-9
Not Sure/Don’t Know	13.5%
Disappointed Attempt was Not Successful	11.2%	R: You know um I feel like uh (5 s pause) sometimes I wish that it would have just ended then…A couple of times being in here and knowing that I can no longer avoid and hide in the dark and everything is out now. 01:01:59-2
Feels Fine	5.6%	R: Um, I’m still sort of indifferent to it. Um, I, like I said, I’m, I’ve been running for a while now. I’ve just felt numb to a lot of things…And right now, it’s one of the things I’ve been certainly apathetic towards, like, I mean, it hasn’t even bothered me, really. Honestly, I still even sleep well at night right now. 00:36:32-1
Not Enough Information	4.5%
Intends to Reattempt in the Future	1.1%	R: It’s like any other suicide attempt. I don’t feel about anything except like, I didn’t die…It’s like, and if it happens again I’ll just do the same thing, like either I die.. If I do it I can uh, I can complete the suicide or it’s just another attempt. You know? It’s one of those, it’s gotten that bad where like I don’t feel anything 00:58:05-9

When asked more specifically about the negative and positive impact that their attempt may have had on their lives, soldiers reported both types of consequences (Tables S3 & S4). Negative consequences were most often related to effects on their relationships (33.7%) and their career/reputation (16.9%). Positive effects focused primarily on receiving treatment and learning new skills while in the hospital (57.3%), improved relationships (25.8%), realizing that others cared about them more than they thought (20.2%), and an increased desire to live (19.1%). Reporting specifically on the military response to their suicide attempt, 37.1% of soldiers reported that the response was supportive or helpful, only 5.6% reported that it was unsupportive or unhelpful, and the supportiveness of the response was unclear in 57.3% of cases.

Discussion

There are four key findings from this study. First, in most cases, the most recent episode of suicidal thoughts began shortly before the suicide attempt and these thoughts were not disclosed to others, limiting opportunities for intervention via traditional approaches. Second, this study provides new information about the antecedents and contexts in which suicidal thinking and decisions emerge, with most suicide attempters describing things that could have changed their minds about trying to kill themselves. Third, the primary reason given for attempting suicide was to escape from psychologically aversive conditions after perceiving that there were no other effective strategies or options available. Fourth, although the most reported consequences of having made a suicide attempt were negative (e.g., negative self-view, guilt), some participants also reported various positive consequences (e.g., learning new skills, receiving support). Each of these findings warrants additional comment.

Most participants indicated that a relatively short period had passed between starting to think about suicide and making a suicide attempt. Approximately 80% reported starting to think about suicide in the month before the attempt, 45% in the week before, 30% on the same day, and 15% within one hour before their attempt. These findings add to prior studies on the nature of suicidal thinking in the days and weeks before suicide attempts (Millner et al., 2017; O’Brien et al., 2021). Specifically, this study provides a more granular description of the timeline leading up to suicide attempts, using interviews conducted much closer in time to the suicide attempt, and from a larger sample than in prior studies (Denneson et al., 2020; Montross et al., 2014).

These results also revealed that although most participants shared with others that they had been experiencing distress, less than half (34.8%) disclosed to someone that they had been considering suicide before they made their attempt. This is lower than the rate of disclosure reported by next-of-kin of Army soldiers who had died by suicide (58.4%; Nock et al., 2017) and informants of civilian suicide decedents more generally (Cavanagh et al., 2003). This may represent genuine differences in the rate of such reporting by those making lethal vs. non-lethal suicide attempts. Alternatively, the higher rates of suicide disclosure reported by next-of-kin in previous research may be due to inadvertent or indirect communication that the individual attempting suicide does not realize they are making but next-of-kin recognize. Prior research indicates that direct communication of suicide intent (e.g., “I’m going to kill myself”) occurs 41-43% of the time, whereas talking about “suicidal ideas” or indicating suicide intent in some way happens more frequently 69-83% (Dorpat & Ripley, 1960, Robins et al., 1959). Taken together, the short amount of time between consistently thinking about suicide to attempting and the relatively low rates of purposeful disclosure of suicidal thinking/intent limit opportunities for intervening with those in high-risk states. Such a short timeline and low rate of disclosure suggest that traditional approaches, such as asking those at risk to voluntarily self-identify and come to clinics to see a mental health provider and even attending weekly or monthly sessions with a healthcare provider, are insufficient to identify and intervene with people at risk. Newly developed methods that include identifying individuals with elevated suicide risk using machine learning applied to electronic health records (Barak-Corren et al., 2020; Kessler et al., 2020; Ursano et al., 2018); more frequent monitoring of those determined to be at elevated risk using digital methods like smartphones, sensors, and social media platforms – or additional public health campaigns to educate next-of-kin about indirect warning signs for suicide and steps to help loved ones find treatment (Jaroszewski et al., 2019); and just-in-time interventions that can be deployed via smartphones, emails, and social media (Coppersmith et al., 2022) may prove especially helpful for bridging these gaps.

This study also provided information about the antecedent events, experiences, and contexts in which suicidal thoughts and the decision to make a suicide attempt emerged – along with suggestions for what might have prevented people from attempting suicide. In terms of antecedents, participants most reported that negative emotional and cognitive states and the piling up of work and social stressors led to their suicidal thoughts and decisions. These qualitative findings were largely supported by accompanying quantitative data in this study – such as the presence of internalizing and externalizing DSM disorders. Study findings are generally in line with those from prior quantitative studies of factors predicting suicidal behavior – but provide more vivid detail about constructs that may be fruitful to pursue in future studies. These include time-varying factors not well-captured by measures of DSM-based disorders, such as emotional exhaustion, agitation, separation from family, and increases in workplace stress.

Another key finding from this study was that the vast majority (89%) of those who attempted suicide explained that they did so to escape from a seemingly unbearable psychological state. This finding aligns with prior results (O’Brien et al., 2021; Bryan et al., 2015) and supports escape theories of suicide (Baumeister, 1990; Beck et al., 1990; Linehan, 1993; Shneidman, 1996), possibly helping to explain why stressful life events and mental disorders – and multimorbidity in particular (Kessler et al., 2020; Nock & Kessler, 2006; Nock et al., 2008; Ursano et al., 2018) – are so strongly linked with suicidal behavior. That is, mental disorders likely predict suicidal behavior at least in part because they consist of and co-occur with aversive psychological states from which a person wants to escape, and suicide provides a mechanism through which to do so. Participants reported other interpersonal and intrapersonal motivations far less frequently, mirroring the pattern of motivations observed among those engaging in nonsuicidal self-injury (Nock & Prinstein 2004; Nock, 2010). The presence of motivation to escape from aversive states among those engaging in suicidal behavior, nonsuicidal self-injury, and substance use may explain why these behaviors are often comorbid.

But why do some people engage in suicidal behavior, rather than other behavioral strategies, to escape aversive states? Another key finding from this study was that people reported deciding on suicide most often because they felt trapped in their situation with no other options, or they had tried other options and failed. These findings are consistent with prior qualitative results (O’Brien et al., 2021) and work revealing differences in decision-making (Clark, 2010; Dombrovski & Hallquist, 2017; Jollant et al., 2005, 2010; Millner et al., 2019) and problem-solving (Levenson and Neuringer, 1971; Linehan et al., 1987; Pollock & Williams, 2004; Schotte & Clum, 1982, 1987) among those who experience suicidal thoughts and make suicide attempts, highlighting the need for additional research in these relatively under-studied areas. For instance, work is needed to better understand how, when, and for whom differences in decision-making biases, deficits, and styles may increase risk of onset for suicidal thoughts or behaviors. This information may help us understand how to modify decision-making in ways that decrease risk of suicidal outcomes.

This last finding raises a question as to whether something could have been done to expand the range of options that might have prevented the suicide attempts. A novel finding in this study is that most participants (58%) indicated that something could have changed their mind about making a suicide attempt. The most frequently reported factors were getting relief from some aversive psychological state and getting support from others. These findings argue against the myth that those who want to die by suicide cannot be stopped. Indeed, most of those interviewed indicated that they would not have tried to kill themselves if they had gotten relief or support. However, this should be considered in the context of the earlier finding that most participants did not disclose their suicidal thoughts or intentions to others, perhaps due to fear of social stigma or a more accurate perception following cessation of extreme distress. Taken together, these findings demonstrate the need for methods that identify and intervene with those at risk without relying on the self-identification and help-seeking described above (Barak-Corren et al., 2020; Kessler et al., 2020; Ursano et al., 2018). Additionally, these findings beg for societal changes in the stigma associated with help-seeking around mental illness and suicide specifically.

The final set of key findings from this study involves the participant’s reactions to their recent suicide attempt. The most commonly reported responses to having made a suicide attempt were negative, including viewing oneself negatively, experiencing feelings of regret and/or guilt, and perceiving a negative impact on their relationships and career. Participants revealed some perceived positive consequences that they experienced after the attempt, most often involving receiving mental health treatment and experiencing support from others. These findings may be most useful if communicated to those who are considering suicide. That is, knowing that most people who attempt suicide regret doing so and have negative self-views as a result – and that the positive consequences, like getting help, are possible without making a suicide attempt may dissuade people from attempting suicide and encourage them to instead seek help earlier then they may have otherwise.

These findings should be interpreted in the context of several key limitations. First, the sample was composed of Army soldiers who received hospital treatment and agreed to speak with a researcher about their suicide attempt. These factors could limit the extent to which these results generalize to all people making a suicide attempt. Moreover, all interviews were conducted via telephone, which may have limited the interviewers’ ability to read non-verbal cues that may have informed the formulation of follow-up questions about key topics. On balance, prior work has shown that assessment methods that increase participant anonymity tend to yield higher endorsements of sensitive information such as engagement in suicidal behavior (Deming et al., 2021). Thus, there are both limitations and strengths to our assessment approach that should be considered in future studies in this area. Second, the interviewing and coding were done by members of a team of researchers who study suicide. Despite efforts to use an atheoretical/inductive coding approach, how these researchers conceptualize suicide, and its causes may have influenced both the interviewing (e.g., personalized follow-up questions asked) and the coding (which codes were created and/or applied). The latter can be addressed in future analysis of these data using natural language processing methods to extract themes more objectively within the described experiences of the interviewees and the associations among them. A related concern is that there is undoubtedly overlap among some of the codes and categories presented. For instance, work-related difficulties with a co-worker or supervisor could be considered “work-related” or “social” stressors. Carefully parsing out the presence and impacts of putative causal factors remains a difficult task for psychological science, but one that we should continue to be mindful of and work toward better addressing. Third, this analysis was limited to a general description of the qualitative data provided by participants. Future analyses of these data will present more detailed examinations of key themes that emerged (e.g., social stressors, communications, and relationships; aspects of decision-making processes) and will involve mixed method analyses that synthesize qualitative and quantitative data available for these soldiers in other Army STARRS components. Fourth, and perhaps most important, it is unclear to what extent these results inform us about the determinants of suicide death, as the patients surveyed here all made nonfatal suicide attempts. Of course, in the case of suicide decedents, we can interview only those left behind (e.g., Nock et al., 2017) and not the suicidal person themselves. A potential synergistic path forward is to use results of qualitative studies with suicide attempters to inform the survey questions included in future psychological autopsy studies of suicide decedents in an attempt to assess for the presence of these antecedents and reasons for suicidal behavior among those dying by suicide.

These limitations notwithstanding, this study provides valuable new information about the experiences of those who attempt suicide, as well as information about suicide attempts among active-duty Army soldiers more specifically. These results point toward important next steps for understanding, prediction, and prevention efforts that may help to drive down the devastating loss of life due to suicide.

Table 3b.

Percent of Sample Endorsing Antecedent for Suicidal Decision Relevant to their Recent Suicide Attempt

		Quotes
Psychiatric Symptoms
Any Emotional Symptom Endorsed	46.1%	R: Um, I was in the room by myself and I was real (really) down depressed and um and I jus- I just like snapped, I just didn’t want to be here anymore…And it kinda (kind of) was just like a (pause) I was just, it was just like a mood thing like uh (3s pause) that just triggered it…And I just, I decided then that I was gunna (going to) do it. I didn’t care anymore.
Any Cognitive Symptom Endorsed	47.2%	R: And I’d gone out to the field a couple of times in order to do some exercises and (2s pause) when I was out there and everything that was going on just got my anxiety going and bringing up all my memories and everything that I had from my deployments that caused my PTSD
Any Behavioral Symptom Endorsed	18.0%	R: Um, I pretty much turned on the alcohol as soon as I saw [evidence of girlfriend’s infidelity]…I’d already pretty much made up in my mind that I was going to try to drink myself to death that night.
External Major Life Events
Any Work Stressor Endorsed	23.6%	R: It was when my, um my sergeant called to make sure I was going to go to, ah, an appointment, I had that morning…That kind of knocked me out of the spiral. And after that, I was like, I’m sick of thinking. So, I was just never going to think again. That seemed like a good idea at the time.
Any Social Support Stressor	41.6%	R: Um, I uh, (pause) was just sitting there and, and in my barracks room and uh thinking about the wife’s move back to Alabama to be with her family and figuring if I call somebody they’re not gonna (going to) pick up the phone so…I’m gonna (going to) do this [attempt suicide].
Any Romantic Relationship Stressor	20.2%	R: I was talking to my ex-girlfriend, and ya (you) know we started arguing, and she says “I’m done with this” and hung up. And, at the time I was taking off my ACUs, then I took off my belt. And I had my belt in my hand, and I was like, “I could, I could just do this right now, it would be so easy.” And, I was like, “every single problem that you have will be gone. Nothing, (pause) nothing will ever bother you again.” And I tied the belt up on my closet, yanked on it a few times, to make sure it was good. I put the loop around my neck. And tightened it down. And, ah, (2 s pause) and that was that.
Any Financial Stressor	5.6%	R: …When I found out that, um, my husband had stopped paying his portion of the bills at the house is when I kind of just broke.
Any Stressful Experience within Larger Military System	10.1%	R: Um it, it ya (you) know I’m learning here that a big component of what makes us humans is our self-worth. We get that from ya (you) know the jobs we do and what I identify with is I’m a Soldier. I’m a combat medic ya (you) know and I’ve run this tractor vehicle with this crew. Ya (you) know and that was stripped away… my command who I work for, the complete uncaring and very vindictive and angry at me and take work away from me.
Any Crime/Legal Stressor	4.5%	R: Well, there was, uh, there was a certain, certain circumstance that could have led me to go to jail. And I, I told my executive officer, um, that I would rather, you know, I’d, I’d kill myself before I went to jail, is what I told him.

Table 4b.

Reasons for not Trying Something other than Suicide to Achieve the Same Goal? (Not Mutually Exclusive)

Why not something else?		Quotes
Trapped/Feels there are No Other Options	46.1%	(1) R: I just felt trapped and (2s pause) just cornered….Boxed in. And I just broke down and (2s pause) just I felt like if I couldn’t be home, you know, which that’s where I felt like I needed to be then…I didn’t wanna (want to) be alive. (2) R: So I just felt (pause) like I felt useless. I felt like there was no hope left. Like everything in my life was goin’ (going) wrong…So I just felt like I needed to get away from those problems. And like trying to commit suicide, like in my life, was the only way that I could get away from the problems.
Tried Other Strategies but Failed/Exhausted All Other Options	28.1%	(1) R: There might have been options left, but I wasn’t seeing them. I mean I feel like I had tried everything. I had gone to doctors, I’d gone to hospitals. I’d tried my unit and I’d tried my family and I just felt like there was nothing else that I could do…Either sit there and suffer or just end it….That’s what I felt my options were. (2) R: Cause (because) everything I was trying wasn’t (was not) working. I’ve (I have) been on multiple types of med and they won’t (will not) work, especially in a field environment, they were not even getting the edge off. So, the pain was just staying constant the whole time.
Unable to Implement Other Strategies (Internal Barriers)	27.0%	(1) R: Um, I don’t know I was, I guess I was just fed up and tired of everything. (2) R: And I, I tend to hold a lot of stuff in, so: …that’s big, a part of the issue there is that I hold in and hold it in and hold it in and it comes out, it, and when it comes out it comes out big.
Unable to Implement Other Strategies (External Barriers)	13.5%	(1) R: Um there was a master sergeant that um had kind of, I kind of talked to about the, I don’t want to say, like I didn’t tell the master sergeant about the harassment, the sexual harassment in the unit cuz (because) it’s such a tight little, like, she might have known her or…But I confided in the master sergeant that I was having issues and problems and she had made herself available a few times but then you know I don’t really know her… (2) R: …soon as I calm down…What they do? They must have smelt it.(Angry)…Come calling me, cussing me out on my phone, telling me to come to the first sergeant office so they can cuss me out to my face. Everyday! (Angry)…( I/A [3 s] words) soon as I thought the day was going to be OK….Like they knew, they was (they were) tracking it.
“I don’t know.”	4.5%	(1) R: Um, I really couldn’t tell you why.
Easy Fix/Easy Way Out	2.2%	(1) R: It just felt like that was the best cure, you know, it would, I mean it’s a selfish act because, you know, I didn’t think of my kids or my wife, but it would have took (taken) all my problems away, I wouldn’t be here anymore, I wouldn’t have to live through it all. (2) R: An escape…Cause it seemed like an easy fix.

Army STARRS Team

The Army STARRS Team consists of Co-Principal Investigators: Robert J. Ursano, MD (Uniformed Services University) and Murray B. Stein, MD, MPH (University of California San Diego and VA San Diego Healthcare System). Site Principal Investigators: James Wagner, PhD (University of Michigan) and Ronald C. Kessler, PhD (Harvard Medical School). Army scientific consultant /liaison: Kenneth Cox, MD, MPH (Office of the Deputy Under Secretary of the Army). Other team members: Pablo A. Aliaga, MA (Uniformed Services University); David M. Benedek, MD (Uniformed Services University); Laura Campbell-Sills, PhD (University of California San Diego); Carol S. Fullerton, PhD (Uniformed Services University); Nancy Gebler, MA (University of Michigan); Meredith House, BA (University of Michigan); Paul E. Hurwitz, MPH (Uniformed Services University); Sonia Jain, PhD (University of California San Diego); Tzu-Cheg Kao, PhD (Uniformed Services University); Lisa Lewandowski-Romps, PhD (University of Michigan); Alex Luedtke, PhD (University of Washington and Fred Hutchinson Cancer Research Center); Holly Herberman Mash, PhD (Uniformed Services University); James A. Naifeh, PhD (Uniformed Services University); Matthew K. Nock, PhD (Harvard University); Victor Puac-Polanco, MD, DrPH (Harvard Medical School); Nancy A. Sampson, BA (Harvard Medical School); and Alan M. Zaslavsky, PhD (Harvard Medical School).