Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Feb 15.
Published in final edited form as: J Affect Disord. 2023 Dec 6;347:477–485. doi: 10.1016/j.jad.2023.12.019

Longitudinal Risk of Suicide Outcomes in People with Severe Mental Illness following an Emergency Department Visit and the Effects of Suicide Prevention Treatment

Ana Rabasco 1,2, Sarah Arias 1,2, Madeline B Benz 1,2, Lauren M Weinstock 1, Ivan Miller 1,2, Edwin D Boudreaux 3, Carlos A Camargo Jr 4, Zachary J Kunicki 1, Brandon A Gaudiano 1,2
PMCID: PMC10872614  NIHMSID: NIHMS1950551  PMID: 38065475

Abstract

Background:

Individuals with severe mental illness (SMI), including bipolar disorder (BD) and schizophrenia-spectrum disorders (SSD), are at high risk for suicide. However, suicide research often excludes individuals with SMI. The current research examined differences in suicide outcomes (i.e., suicide attempt or death) for adults with and without BD and SSD diagnoses following an emergency department (ED) visit and investigated the efficacy of the Coping Long Term with Active Suicide Program (CLASP) intervention in reducing suicide outcomes among people with SMI.

Methods:

1,235 adults presenting with recent suicidality were recruited from 8 different EDs across the United States. Using a quasi-experimental, stepped wedge series design, participants were followed for 52-weeks with or without subsequent provision of CLASP.

Results:

Participants in the SSD group and the BD group had significantly shorter time to and higher rate of suicide outcomes than participants with other psychiatric diagnoses in all study phases and in non-CLASP phases, respectively. Participants with BD receiving the CLASP intervention had significantly longer time to suicide outcomes than those not receiving CLASP; these differences were not observed among those with SSD.

Limitations:

Study limitations include self-reported psychiatric diagnosis, exclusion of homeless participants, and small sample size of participants with SSD.

Conclusions:

Participants with SMI were at higher risk for suicide outcomes than participants with other psychiatric diagnoses. CLASP was efficacious among those participants with BD. Psychiatric diagnosis may be a key indicator of prospective suicide risk. More intensive and specialized follow-up mental health treatment may be necessary for those with SSD.

Keywords: Suicide, severe mental illness, bipolar disorder, schizophrenia, intervention, emergency department

Introduction

Suicide is a significant and growing public health concern. In 2021, suicide was the 9th leading cause of death for people ages 10–64 in the United States, with rates increasing by 36% between 2000 and 2021 (Centers for Disease Control and Prevention, 2023). People with severe mental illness (SMI), including those with bipolar disorder (BD) and schizophrenia-spectrum disorders (SSD), are at particularly high risk for suicide (Novick et al., 2010; Palmer et al., 2005; Saha et al., 2007). The lifetime suicide attempt prevalence in individuals with schizophrenia, bipolar I or bipolar II disorder, are 27%, 36%, and 32%, respectively (Novick et al., 2010). These rates are far higher than the 2.7% lifetime suicide attempt prevalence in the general population (Nock et al., 2008). Despite the increased risk of suicide among people with SMI, they are often excluded from suicide research because of concerns about potential challenges with their ability to understand and follow study procedures due to cognitive deficits and management of their psychosis severity and risk (Villa et al., 2020). The exclusion of people with SMI from suicide research limits our understanding of the risk for suicidal behaviors over time and the efficacy of suicide prevention treatments among people with these diagnoses.

Emergency departments (EDs) are a common point of contact for people at high risk for suicide. From 2010–2020, EDs in the United States treated over 500,000 people each year for deliberate self-harm (National Center for Injury Prevention and Control, 2023). In addition, approximately 10% of people who visit EDs each year have experienced suicidality in the two weeks prior to visiting the ED (Claassen & Larkin, 2005; Ilgen et al., 2009). Furthermore, research has found that over 40% of mental health patients who died by suicide had visited the ED in the year prior to their death (Ahmedani et al., 2019; Da Cruz et al., 2011). Research has also shown that individuals with symptoms associated with SMI, such as psychosis, are more likely to have repeated ED visits over a six-month period (Sirotich et al., 2016). A review of studies on suicidality and the ED identified that psychotic disorders were among the top co-occurring psychiatric diagnoses for individuals presenting to the ED for suicidal ideation (Ceniti et al., 2020). Therefore, the ED is a particularly important context in which to conduct suicide risk assessment and intervention (Betz et al., 2016), especially for those with SMI.

Goldman-Mellor et al. (2019) examined the one-year incidence of suicide and other mortality among ED patients in California presenting with nonfatal deliberate self-harm, suicidal ideation, and any other presenting concern. They found that a clinical diagnosis of psychosis or bipolar disorder was associated with increased suicide risk among those presenting with deliberate self-harm. The present research aims to further explore the role of psychiatric diagnosis in suicide risk following an ED visit by expanding the outcome to suicide attempts (not solely death by suicide), along with investigating the efficacy of a suicide-focused intervention among people with SMI.

A number of interventions have been developed to specifically target suicidality, including Dialectical Behavioral Therapy (Linehan et al., 2006), safety planning (Stanley & Brown, 2012), and Coping Long Term with Active Suicide Program (CLASP; Miller et al., 2016, 2022). Multifaceted interventions have been found to be particularly effective in reducing suicidal behaviors (Krysinska et al. 2016; Hofstra et al. 2020; Doupnik et al. 2020) and adjunctive interventions offer increased accessibility and generalizability to “real world” clinical settings, as they are less time and resource intensive. CLASP is a multifaceted, adjunctive intervention designed to target suicidality through a combination of case management, individual psychotherapy, and optional significant other (SO) involvement (Miller et al., 2022). CLASP has been adapted for ED populations, consisting of seven brief telephone calls to the patient and up to four calls to an SO identified by the patient. Calls focus on reducing suicide risk through identifying risk factors, clarifying values and goals, planning for the future, identifying a safety plan, encouraging treatment engagement and adherence, and facilitating patient and SO communication and informal problem solving (Boudreaux et al., 2013; Miller et al., 2017). The parent study of the current research, a multicenter study of adults with recent suicidality visiting the ED, compared rates of suicide attempts over the course of the 52-week study between three sequential phases: 1) treatment as usual, 2) universal screening, and 3) universal and secondary suicide screening, resources and a safety plan provided in the ED, and CLASP. They found that, compared with Phases 1 and 2 of the study, patients in Phase 3 of the study showed a 20% relative risk reduction in suicide attempt risk over one-year follow-up (Miller et al., 2017). Furthermore, CLASP has been found to be cost-effective to provide (Dunlap et al., 2019). However, the effects of CLASP have not been examined among people with SMI. Because people with SMI have unique needs that may impact their ability to engage in CLASP, such as greater social isolation (Bornheimer et al., 2020) and cognitive impairment (Reichenberg et al., 2008), it is essential to specifically investigate their response to the intervention.

The present research examined suicide risk among adults with SMI (BD or SSD) identified in the ED compared to those with other psychiatric diagnoses (OPD) over the course of a year, along with the effectiveness of the CLASP intervention among those with SSD and BD. We hypothesized that: 1) participants with BD and SSD will have a higher rate of lifetime suicide attempts at baseline compared to participants with OPD; 2) Among participants in Phases 1 and 2 (non-intervention phases), those with BD and SSD will have a shorter time to and higher rate of suicide outcomes over follow-up compared with participants with OPD; 3) Among participants in Phase 3 (CLASP intervention), participants with BD and SSD will have shorter time to and higher rate of suicide outcomes over follow-up compared with participants with OPD; and 4) those with BD and SSD who received CLASP will have longer time to and lower rate of suicide outcomes over follow-up compared to those not receiving CLASP.

Methods

Procedure

This paper is a secondary analysis of the Emergency Department Safety Assessment and Follow-up Evaluation (ED-SAFE) multicenter study (Boudreaux et al., 2013; Miller et al., 2017). The ED-SAFE study, a quasi-experimental, stepped wedge series design (Hemming et al., 2015) tested the effectiveness of universal suicide screening and a suicide prevention intervention in reducing risk for adults recruited from EDs across the United States. Recruitment was conducted at a total of eight EDs, ranging from small community hospitals to large academic centers, across seven states. The study consisted of three sequential phases at each ED site. The first phase was treatment as usual and the second phase was universal screening. This third phase consisted of the following components: 1) universal screening, 2) secondary suicide risk screening, 3) a safety plan and information to patients provided by nursing staff, and 4) CLASP intervention, in which up to seven telephone calls (10–20 minutes) were made to participants, along with optional involvement of their SO (four 10–20 minute phone calls), for 52-weeks following the index ED visit (see Introduction for more detailed description of CLASP). Phone calls were made by 10 CLASP advisors who were trained to fidelity by the developers of CLASP and received weekly supervision.

For all phases, participants completed telephone assessments at weeks 6, 12, 24, 36, and 52 and medical record reviews at 6 and 12 months. This study received institutional review board approval at each site. The National Institute of Mental Health Data and Safety Monitoring Board conducted overall study oversight and monitoring. See Boudreaux et al. (2013) for a full description of the study methods.

Participants

Of the full ED-SAFE sample (N = 1,376), 141 (10%) participants reported no psychiatric diagnosis. The current study only included those participants who reported having a psychiatric diagnosis at baseline (n = 1,235). For this subsample, the mean age was 38 years old (SD = 13) with 57% (n = 709) identifying as female and white (76%, n = 938). Additional demographic characteristics are presented in Table 1.

Table 1.

Demographic Characteristics by Psychiatric Diagnosis (N = 1,235)

BD (n = 456) SSD (n = 149) OPD (n = 630) Group Differences
Variable Mean (SD) or N (%) F/X 2 η 2 /Cramer’s V
Age 38.5 (11.63) 39.38 (1117) 37.02 (14.11) 2.94 0.01
Sex 9.67** 0.09
 Male 185a (41%) 81b (54%) 260a (41%)
 Female 271a (59%) 68b (46%) 370b (59%)
Ethnicity 3.62 0.05
 Latine 51 (11%) 24 (16%) 67 (11%)
 Non-Latine 405 (89%) 125 (84%) 563 (89%)
Race 6.90* 0.08
 White 361a,b (79%) 104b (70%) 500a (79%)
 Non-White 95a,b (21%) 45b (30%) 130a (21%)
Highest Level of Education 44.49*** 0.13
 < High School 86a (19%) 50b (34%) 114a (18%)
 High School 128a,b (28%) 57b (38%) 164a (26%)
 Vocati onal/Techni cal 16a (4%) 5a (3%) 30a (5%)
 Some college 137a (30%) 28b (19%) 191a (30%)
 College graduate 57a (13%) 5b (3%) 93a (15%)
 Any post-graduate 32a (7%) 4a (3%) 38a (6%)
Sexual Orientation 2.97 0.04
 Heterosexual 383 (84%) 129 (87%) 541 (86%)
 Lesbian/Gay 33 (7%) 10 (7%) 39 (6%)
 Bisexual 32 (7%) 6 (4%) 36 (6%)
Marital Status 14.61 0.08
 Never married 220a (48%) 75 a (50%) 311a (49%)
 Married 77a,b (17%) 18b (12%) 140a (22%)
 Widowed 12 a (3%) 6 a (4%) 17a (3%)
 Divorced 113a (25%) 39a (26%) 121a (19%)
 Other 34 a (8%) 11a (7%) 31a (7%)
Employment 28.20*** 0.15
 Full time 67 (15%)a 7 (5%)b 138 (22%)c
 Non-full time 389 (85%)a 142 (95%)b 492 (78%)c
Household 27.20*** 0.15
 Living alone 113a (25%) 65b (44%) 145a (23%)
 Living with others 343a (75%) 84b (56%) 485a (77%)
Insurance Coverage 0.85 0.03
 No insurance 112 (26%) 33 (24%) 161 (28%)
 Insurance 316 (74%) 106 (76%) 425 (73%)
Open in a new tab
*

p < .05,

**

p < .01,

***

p < .001

Note. Each subscript letter denotes categories whose column proportions do not differ significantly from each other at the .05 level; BD = bipolar disorder; SSD = schizophrenia-spectrum disorder; OPD = other psychiatric diagnosis

Inclusion criteria included having attempted suicide or experiencing active suicidal ideation (answering yes to, “Have you actually had any thoughts of killing yourself?”) in the week prior to the ED visit. Exclusion criteria were: 1) being medically or cognitively unable to participate in study procedures, 2) living in a noncommunity setting, 3) being under state custody or pending legal action, 4) being without permanent residence or reliable telephone service, 5) having an insurmountable language barrier, or 6) no self-reported psychiatric diagnosis.

Measures

Psychiatric Diagnosis

Participants’ psychiatric diagnoses were assessed using the self-report question: “Have you ever been diagnosed by a doctor or therapist with…?” Response options included: depression, bipolar disorder (“sometimes known as manic-depression”), alcohol use disorder, any drug use disorder, anxiety disorder, attention deficit disorder (ADD/ADHD), eating disorder (“like anorexia or bulimia”), schizophrenia or schizoaffective disorder, or any other psychiatric disorder.

Mental Healthcare Utilization

To assess for mental healthcare utilization, participants were asked whether, over the past six months, they had: 1) visited an outpatient mental health provider, 2) visited the emergency room for psychiatric reasons, or 3) stayed overnight in the hospital for psychiatric reasons.

Suicide Outcomes

Lifetime history of suicide attempts was assessed at baseline with the following question: “At any time in your life, including today, have you made a suicide attempt?”

The primary outcome for all follow-up analyses was a combination of suicide attempts or death by suicide, referred to as “suicide outcomes” moving forward. Suicide attempts over the follow-up period were ascertained through self-report on telephone-based follow-up assessments (“After you completed the last assessment on [date] until now, have you made a suicide attempt?”) and participant chart reviews. Participant death was determined through a combination of SO report and review of state and national vital statistics registries and the National Death Index; only suicide deaths were included in the current study analyses. Documentation of suicide outcomes at any of the follow-up assessment time points counted as a “yes” for the outcome. Time to first suicide attempt or suicide death was also recorded.

Statistical Analysis

First, differences in demographic characteristics, treatment utilization, and lifetime suicide attempts by psychiatric diagnosis (BD, SSD, or OPD) were examined using one-way ANOVA, chi-square, and Kruskal-Wallis H test analyses. Incidence rates by psychiatric group were also calculated for participants in each of the study phases (see supplemental materials). A Kaplan-Meier survival analysis was then conducted among those participants in the subsample in Phases 1 and 2, who did not receive study intervention, to examine differences in time to suicide outcomes by psychiatric diagnosis (BD, SSD, or OPD) over the course of the study period. Because there were differences in sex and age by psychiatric diagnosis within Phases 1 and 2 (see supplemental materials), we conducted Cox proportional hazards models adjusting for those covariates. Other baseline differences in education, employment, and living alone were not included as covariates, due to their being influenced by one’s psychiatric diagnosis and course of illness.

A Kaplan-Meier survival analysis and a Cox proportional hazards analysis were also conducted among just those participants in CLASP, examining differences in time to suicide outcomes between BD, SSD, and OPD groups. Although there were baseline differences by psychiatric diagnosis in education, employment, and living alone, they were not included as covariates due to the likelihood of their being influenced by one’s psychiatric diagnosis and course of illness. No other differences in baseline demographic characteristics by psychiatric diagnostic group were identified (see supplemental materials); therefore, no covariates were included in the Cox proportional hazards analysis.

Finally, a Kaplan-Meier survival analysis was conducted among just those participants with BD, examining differences in time to suicide outcomes between participants in Phases 1 and 2 (no intervention) versus participants in Phase 3 (intervention). To control for potential differences between study phases, we conducted a Cox proportional hazards analysis, adjusting for confounders (age, sex, history of suicidal behaviors, alcohol abuse, substance use, and living alone). These were a subgroup of the covariates included in the primary ED-SAFE trial analysis (Miller et al., 2017), which were chosen based on their relevance to the study question and in light of the reduced sample size of the current study. Kaplan-Meier survival and Cox proportional hazards analyses were repeated just among those participants with SSD. All analyses utilized an intent-to-treat (ITT) sample.

Results

Demographic Descriptives and Baseline Psychiatric Treatment by Psychiatric Diagnosis

The total sample consisted of 1,235 participants. Thirty-six percent (n = 497) of the sample was recruited in Phase 1, 27% (n = 377) in Phase 2, and 37% (n = 502) in Phase 3. Thirty-three percent (n = 456) of the sample had a diagnosis of BD, 11% (n = 149) had a diagnosis of SSD, and 46% (n = 630) were diagnosed with an OPD (e.g., unipolar depression, anxiety disorder, substance use disorder, eating disorder, attention deficit disorder, other). Descriptive statistics by psychiatric diagnosis group are presented in Table 1, along with group differences. There were significant differences in sex, race, level of education, living alone, insurance coverage, and employment by psychiatric diagnosis. There were no significant differences on other demographic variables.

Baseline differences in psychiatric treatment over the six months prior to the ED visit based on psychiatric diagnosis were examined for the full sample (all three phases before study intervention was provided) using chi-square analyses. Participants with BD and SSD had significantly higher psychiatric treatment utilization compared with participants with OPD (see Table 2).

Table 2.

Baseline Treatment Utilization and Lifetime Suicide Attempts by Psychiatric Diagnosis (N = 1,235)

BD (n = 456) SSD (n = 149) OPD (n = 630) Group Differences
Variable N (%) X 2 Cramer’s V
Outpatient Psychiatric Treatment 19.65*** 0.13
 Yes 307a (67%) 107a (72%) 356b (57%)
 No 149a (33%) 42a (28%) 274b (44%)
ED Psychiatric Visit 33.75*** 0.17
 Yes 177a (39%) 77b (52%) 178c (28%)
 No 279a (61%) 72b (48%) 452c (72%)
Inpatient Psychiatric Treatment 28.78*** 0.15
 Yes 134a (29%) 54a (36%) 117b (19%)
 No 322a (71%) 95a (64%) 513b (81%)
Lifetime Suicide Attempt 28.17*** 0.15
 Yes 353a (77%) 128a (86%) 423b (67%)
 No 103a (23%) 21a (14%) 207b (33%)
Open in a new tab
*

p < .05,

**

p < .01,

***

p < .001

Note. Each subscript letter denotes categories whose column proportions do not differ significantly from each other at the .05 level; Assessment period for all treatment utilization variables is past six months; BD = bipolar disorder; SSD = schizophrenia-spectrum disorder; OPD = other psychiatric diagnosis

Baseline Suicide Attempt History by Psychiatric Diagnosis

Baseline differences in lifetime suicide attempt frequency (including at index ED visit) between psychiatric diagnostic groups were examined for the full sample (all three phases before study intervention; see Table 2). A Shapiro-Wilks test showed that lifetime suicide attempt frequency was not normally distributed for psychiatric diagnosis groups (p < .001, Cramer’s V = 0.15). Therefore, a Kruskal-Wallis H test was used to determine if there were differences in suicide attempt frequency between the three diagnostic category groups (BD, SSD, OPD). Median lifetime suicide attempt frequency was significantly different between groups, H(2) = 59.00, p < .001. Pairwise comparisons revealed significant differences in suicide attempt frequency between the OPD diagnosis group (Mdn = 2.00) and BD group (Mdn = 3.00) (p < .001), between the OPD group and the SSD group (Mdn = 4.00) (p < .001), and between the BD group and SSD group (p = .002).

Suicide Outcomes by Psychiatric Diagnostic Group Across Study Follow-up for Participants in Phases 1 and 2

For the following analyses, only those participants in Phases 1 and 2 (n = 787) were included, as they did not receive the study intervention and no outcome differences were found between Phases 1 and 2 in the original study (Miller et al., 2017). No significant differences in outpatient mental health treatment utilization by psychiatric diagnosis over the course of the study period were identified (p = .78). Over the course of the follow-up period, 19% (n = 75) OPD participants, 28% (n = 28) of SSD participants, and 29% (n = 86) of BD participants had a suicide outcome. The suicide outcome incidence rates per 100 person-years were 51.2 for OPD participants, 93.5 for BD participants, and 87.1 for SSD participants, reflecting the average number of suicide outcome events if 100 participants were observed for one year (see supplemental materials).

Kaplan-Meier survival analyses were conducted to compare whether there were significant differences in the amount of time to suicide outcomes during the follow-up period between diagnostic groups. Presence of suicide outcomes over the 52-week study period were entered as the “event” and weeks to first suicide outcome was entered as the “time.” The BD group had a mean time to suicide outcome of 40.15 weeks (95% CI, 37.91 to 42.39), which was similar to the SSD group mean time of 40.76 weeks (95% CI, 36.95 to 44.57). The OPD group had a longer time to suicide outcome of 44.07 weeks (95% CI, 42.36 to 45.77). A log rank test showed that the survival distributions for the three groups were statistically significantly different, χ2(2) = 9.63, p = .008. Pairwise log rank comparisons showed statistically significant differences in survival distributions for the BD versus OPD groups, χ2(1) = 8.53, p = .003, and for the SSD versus OPD groups, χ2(1) = 4.25, p = .04, with both BD and SSD groups having significantly shorter time to suicide outcome compared to the OPD group. The survival distributions for the BD and SSD groups were not statistically significantly different, χ2(1) = 0.00, p = .98 (see Figure 1).

Figure 1.

Figure 1.

Open in a new tab

Rate of survival (i.e., lack of suicide outcome) over the course of the study period by psychiatric diagnosis for participants within control (Phases 1 and 2) conditions

Note. BD = bipolar disorder; SSD = schizophrenia-spectrum disorder; OPD = other psychiatric diagnosis

To include age and sex as covariates, a Cox regression was conducted, with psychiatric diagnostic group as the predictor (BD, SSD, OPD), and suicide attempt or death as the outcome. Participants with BD had significantly higher risk of suicide outcomes over the study period than participants in the OPD group (hazard ratio, 1.61, 95% CI, 1.18–2.20, p = .003), as did participants in the SSD group (hazard ratio, 1.64, 95% CI, 1.06–2.55, p = .03). There was no significant difference in suicide outcomes between the BD and SSD groups (hazard ratio, 1.02, 95% CI, 0.67–1.57, p = .92).

Suicide Outcomes by Psychiatric Diagnosis Group Across Study Follow-up for Participants in CLASP

For the following analyses, only those participants in the CLASP intervention (Phase 3; n = 448) were included. Over the CLASP follow-up period, 16% (n = 39) OPD participants, 31% (n = 15) of SSD participants, and 20% (n = 31) of BD participants had a suicide outcome. The suicide incidence rates per 100 person-years were 51.6 for OPD participants, 56.5 for BD participants, and 99.2 for SSD participants (see supplemental material). We then examined differences in CLASP session attendance rates, inclusion of an SO in the intervention, or outpatient mental health treatment utilization between psychiatric diagnostic groups over the 52-week study period and found no differences (all p’s > .20).

Using the methods described in the previous section, Kaplan-Meier survival analyses were conducted comparing the amount of time to suicide outcomes during the follow-up period between diagnostic groups. A log rank test showed that the survival distributions for the three groups were not statistically significantly different (only approaching significance), χ2(2) = 5.34, p = .07. However, pairwise log rank comparisons showed significant differences between groups. Participants in the SSD group had a significantly shorter mean time to suicide outcome (40.89 weeks; 95% CI, 35.77 to 46.01) compared to participants in the OPD group (45.80 weeks; 95% CI, 43.85 to 47.75), χ2(1) = 5.52, p = .02, but not the BD group (44.33 weeks; 95% CI, 41.62 to 47.04), χ2(1) = 2.58, p = .11. There was no significant difference in time to suicide outcome between participants in the BD and OPD groups, χ2(1) = 0.56, p = .45 (see Figure 2).

Figure 2.

Figure 2.

Open in a new tab

Rate of survival (i.e., lack of suicide outcome) over the course of the study period by psychiatric diagnosis for participants within CLASP (Phase 3) condition

Note. BD = bipolar disorder; SSD = schizophrenia-spectrum disorder; OPD = other psychiatric diagnosis

A Cox regression showed the same pattern of results. Participants with SSD had a significantly higher risk of suicide outcomes over the study period than participants in the OPD group (hazard ratio, 1.95, 95% CI, 1.06–3.59, p = .03). There was no significant difference in suicide outcomes between the BD and SSD groups (hazard ratio, 1.16, 95% CI, 0.72–1.87, p = .54) or the BD and OPD groups (hazard ratio, 1.68, 95% CI, 0.90–3.13, p = .11).

CLASP Suicide Outcome Follow-Up Comparisons among Participants with BD and SSD

First, we identified that there were no significant differences in outpatient mental health treatment utilization over the course of the study period by study phase (Phases 1 and 2 did not involve active intervention and were therefore combined) among participants with BD or SSD (all p’s > .36).

Then, a set of analyses were conducted among all participants with BD to compare differences in suicide outcomes between participants in Phases 1 and 2 (N = 298) and participants in Phase 3 (CLASP; N = 156) over the course of the study period. A Kaplan-Meier survival analysis showed that participants in the Phase 3 group had a mean time to suicide outcome of 44.33 weeks (95% CI, 41.62 to 47.04). This was longer than the Phase 1 and 2 groups, which had a mean time to suicide outcome of 40.15 weeks (95% CI, 37.91 to 42.39). A log rank test showed a statistically significant difference in the survival distributions for the two groups, χ2(1) = 4.39, p = .04 (see Figure 3). To include covariates (age, sex, history of suicidal behaviors, alcohol abuse, substance use, and living alone) in the survival analysis, a Cox regression was conducted, with study phase as the predictor (Phases 1 and 2 vs. Phase 3) and suicide attempt or death as the outcome. The pattern of results matched the Kaplan-Meier analysis; those participants in Phases 1 and 2 had higher risk of suicide outcomes compared with participants in Phase 3 over the course of the study period (hazard ratio, 1.71, 95% CI, 1.06 – 2.78, p = .03).

Figure 3.

Figure 3.

Open in a new tab

Rate of survival (i.e., lack of suicide outcomes) over the course of the study period among participants with bipolar disorder

Note. Phases 1 and 2 = control conditions; Phase 3 = CLASP intervention condition

Analyses were repeated just among those participants with SSD. Participants in the Phase 3 group (N = 49) had a mean time to suicide outcomes of 40.89 weeks (95% CI, 35.77 to 46.01). This was about the same as the Phases 1 and 2 group (N = 100), which had a mean time to suicide outcomes of 40.76 weeks (95% CI, 36.95 to 44.57). A log rank test showed that the survival distributions for the two groups were not statistically significantly different, χ2(1) = 0.03, p = .86 (see Figure 4). A Cox regression was conducted using the methods described above for the BD group. Like the Kaplan-Meier analysis results, there were no significant differences in risk of suicide outcomes between the Phases 1 and 2 group and Phase 3 group over the course of the study period (hazard ratio, 0.98, 95% CI, 0.49 – 2.00, p = .96).

Figure 4.

Figure 4.

Open in a new tab

Rate of survival (i.e., lack of suicide outcomes) over the course of the study period among participants with a schizophrenia-spectrum disorder

Note. Phases 1 and 2 = control conditions; Phase 3 = CLASP intervention condition

Discussion

Suicide is a significant public health problem that disproportionately impacts people with SMI, including BD and SSD (Lu et al., 2020; Novick et al., 2010). Yet, people with BD and SSD are often excluded from suicide research (Villa et al., 2020), limiting our understanding of the course of suicidal behaviors among these individuals over time, along with the efficacy of existing suicide interventions for these populations. The months following discharge from the ED are a time of heightened risk for death by suicide for those presenting with mental health concerns (Olfson et al., 2021), especially for people with a clinical diagnosis of psychosis or bipolar disorder reporting self-harm (Goldman-Mellor et al., 2019). The current study extends previous research by examining differences in suicide attempts and suicide mortality by psychiatric diagnosis following discharge from the ED. The ED is also a particularly useful setting to connect individuals at high risk for suicide with suicide prevention interventions (Betz et al., 2016). Suicide interventions developed for adults presenting to the ED with suicidality, such as CLASP, have been shown to be feasible, acceptable, and effective in reducing suicidal behaviors (Miller et al., 2017; Stanley et al., 2016); however, there is a lack of research on their efficacy for individuals with SMI. To address these gaps in the literature, the present research analyzed multicenter data from ED patients at high risk for suicide, comparing suicide outcomes over the course of the 52-week study between people with SMI and other psychiatric diagnoses, both in the control and CLASP groups, along with the effects of the CLASP intervention on suicide outcomes among those with BD and SSD diagnoses.

Baseline and Control Condition Suicide Outcomes

The present research found that participants with BD and SSD diagnoses had higher rates of lifetime suicide attempts than those with OPDs. In addition, over the course of the study period, those participants in the non-intervention phases with BD and SSD diagnoses had shorter time to suicide outcomes and higher rates of suicide outcomes over one-year post-ED visit than participants with OPDs. These findings align with previous research showing that people with BD and SSD are at particularly high risk for suicide (Lu et al., 2020; Novick et al., 2010). They also extend previous research by showing that, compared to people with OPDs, people with BD and SSD presenting to the ED with suicidality have higher rates of suicide attempts, in addition to death by suicide (Goldman-Mellor et al., 2019).

These results suggest that psychiatric diagnosis may serve as an important indicator of future suicide risk. There is a particular need for additional, accurate markers of suicide risk, as one review found insufficient evidence for current risk assessment tools in reducing risk of suicide among patients presenting to the ED (Stewart & Lees-Deutsch, 2022). Furthermore, recent research has found that patients with psychotic symptoms are less likely to endorse typical suicide risk factors, such as depressive symptoms or suicidal intent, compared to those without psychosis (Bornheimer et al., 2023), underscoring that people with SMI may require additional types of suicide risk assessment. The potential use of psychiatric diagnosis as an indicator of prospective suicide risk has significant clinical implications. For example, clinicians in the ED setting could use psychiatric diagnosis as a brief method to indicate the need to connect certain patients, especially those with SMI, to additional mental health treatments and resources, even in the absence of those patients presenting with imminent suicide risk.

CLASP Suicide Outcomes

Although previous work has described the efficacy of CLASP, an adjunctive suicide intervention, in reducing suicidal behaviors among the overall sample (Miller et al., 2017), it has not examined differences in suicide outcome within the CLASP condition by psychiatric diagnosis nor the specific efficacy of CLASP for people with SMI. Therefore, this study investigated differences in suicide outcome by psychiatric diagnosis specifically within CLASP, finding that those with SSD had significantly shorter time to and higher rate of suicide outcome compared to participants with OPD, while there was no significant difference in suicide outcomes between participants with BD and OPD. This suggests that participants without SSD may have responded to the intervention better than participants with this diagnosis. We further examined the efficacy of CLASP among just those participants with SMI diagnoses. Among 456 participants with BD, those in the CLASP phase had significantly longer time to and lower rate of suicide outcomes compared to non-intervention phases, suggesting that CLASP was efficacious in reducing suicide outcomes for participants with BD. However, among the 149 participants with SSD, there was no significant difference in time to and rate of suicide outcomes between the CLASP and control conditions, aligning with the finding SSD participants had worse suicide outcomes compared to BD and OPD participants when comparing the three diagnostic groups within CLASP.

The inability to verify a significant effect of CLASP among those with SSD may be due to the smaller sample size of SSD participants. It is also possible that the lack of efficacy was because individuals with SSD face unique stressors that influence suicide risk but are not incorporated into existing suicide interventions, such as CLASP. For example, research shown that people with SSD experience more self-stigma compared to those with mood disorders (Oliveira et al., 2015), and self-stigma has been associated with increased suicide risk among people with SSD (Sharaf et al., 2012). These findings highlight the need future research conducted on CLASP with a larger sample size of individuals with SSD, along with the need for future suicide treatment development research that specifically addresses the experiences of people with SSD.

Limitations

The results from this research should be interpreted in light of its limitations. First, psychiatric diagnoses were self-reported by participants and not confirmed through diagnostic interview. Some participants may have misreported their diagnoses or failed to report a diagnosis that they met criteria for. However, descriptive analyses of differences in demographic characteristics by psychiatric diagnosis generally aligned with previous research. For example, in the current study, there were more male participants (Aleman et al., 2003) and people of color (Schwartz & Blankenship, 2014) in the SSD group compared with the BD and OPD groups, supporting the validity of the self-reported diagnoses. It would be beneficial for future research to replicate and confirm these findings using diagnostic interviews. In addition, being without a permanent residence was one of the exclusion criteria for the study. Individuals with SMI have higher rates of homelessness (Folsom et al., 2005); therefore, generalizability of the current study’s findings are somewhat limited, as they may not fully represent ED patients with SMI. It is possible that the suicide rates reported in this study are underrepresenting suicidality among people with SMI, as being unhoused is associated with increased suicide risk (Ayano et al., 2019). To increase generalizability and clinical applicability, future research on the efficacy of suicide interventions for people with SMI could include those participants experiencing homelessness. Finally, the sample of size of participants with SSD was small (n = 149), so there was low power to detect differences in suicide outcomes between participants in the control and CLASP with SSD.

Conclusion

Participants with BD and SSD had significantly higher rates of and shorter time to suicide outcomes (i.e., suicide attempts or death) compared with participants with OPDs. In addition, the CLASP intervention was efficacious among those participants with BD but results with patients with SSD were inconclusive. Findings highlight self-reported psychiatric diagnosis as a key indicator of prospective suicide risk for patients screened in the ED. Future research could investigate potential changes to suicide interventions that specifically address the needs of people with SSD, as these findings suggest that they may not benefit from existing suicide interventions. The present research also offers clinical practice implications, such as the importance of providing ED patients with BD or SSD with connections to mental health aftercare as a suicide prevention method. More intensive and specialized follow-up mental health treatment may also be necessary for those individuals with SSD. Implementation of interventions that address the specific clinical and social needs of patients with SMI would advance ongoing efforts to reduce the high rate of suicide among this population.

Supplementary Material

1

Highlights.

  • Those with severe mental illness had highest rate of lifetime suicide attempts

  • During the study, those with severe mental illness had shortest time to suicide

  • During the study, those with severe mental illness had the highest rate of suicide

  • The suicide intervention was efficacious for participants with bipolar disorder

  • Additional research is needed on suicide intervention for those with schizophrenia

Role of Funding Source:

This work was supported by the National Institutes of Mental Health [U01MH088278]. In addition, AR and MB’s efforts were in part supported by the National Institute of Mental Health [T32MH126426].

Footnotes

Declarations of Interest: None

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Ahmedani BK, Westphal J, Autio K, Elsiss F, Peterson EL, Beck A, Waitzfelder BE, Rossom RC, Owen-Smith AA, Lynch F, Lu CY, Frank C, Prabhakar D, Braciszewski JM, Miller-Matero LR, Yeh HH, Hu Y, Doshi R, Waring SC, & Simon GE, 2019. Variation in patterns of health care before suicide: A population case-control study. Prev. Med 127, 105796. 10.1016/j.ypmed.2019.105796 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Aleman A, Kahn RS, & Selten JP, 2003. Sex differences in the risk of schizophrenia: Evidence from meta-analysis. Arch. Gen. Psychiatry 60(6), 565–571. 10.1001/archpsyc.60.6.565 [DOI] [PubMed] [Google Scholar]
  3. Ayano G, Tsegay L, Abraha M, & Yohannes K, 2019. Suicidal ideation and attempt among homeless people: A systematic review and meta-analysis. Psychiatr. Q 90(4), 829–842. 10.1007/s11126-019-09667-8 [DOI] [PubMed] [Google Scholar]
  4. Betz ME, Wintersteen MP, Boudreaux ED, Brown G, Capoccia L, Currier G, Goldstein J, King C, Manton A, Stanley B, Moutier C, & Harkavy-Friedman J, 2016. Reducing suicide risk: Challenges and opportunities in the emergency department. Ann. Emerg. Med 68(6), 758–765. 10.1016/j.annemergmed.2016.05.030 [DOI] [PubMed] [Google Scholar]
  5. Bornheimer LA, Li J, Im V, Taylor M, & Himle JA, 2020. The role of social isolation in the relationships between psychosis and suicidal ideation. Clin. Soc. Work J 48(1), 54–62. 10.1007/s10615-019-00735-x [DOI] [Google Scholar]
  6. Bornheimer LA, Wastler H, Li Verdugo J, Im V, Burke H, Holzworth J, Velasquez E, & Mahmood M, 2023. Depression, suicide intent, and plan among adults presenting in an emergency department after making a suicide attempt: Exploratory differences by psychosis symptom experience. J. Psychiatr. Res 157, 197–201. 10.1016/j.jpsychires.2022.11.033 [DOI] [PubMed] [Google Scholar]
  7. Boudreaux ED, Miller I, Goldstein AB, Sullivan AF, Allen MH, Manton AP, Arias SA, & Camargo CA Jr., 2013. The Emergency Department Safety Assessment and Follow-up Evaluation (ED-SAFE): Method and design considerations. Contemp. Clin. Trials 36(1), 14–24. 10.1016/j.cct.2013.05.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Ceniti AK, Heinecke N, & McInerney SJ, 2020. Examining suicide-related presentations to the emergency department. Arch. Gen. Psychiatry 63, 152–157. 10.1016/j.genhosppsych.2018.09.006 [DOI] [PubMed] [Google Scholar]
  9. Centers for Disease Control and Prevention, 2023. Facts about suicide. https://www.cdc.gov/suicide/facts/index.html (accessed 18 June 2023).
  10. Claassen CA, & Larkin GL, 2005. Occult suicidality in an emergency department population. Br. J. Psychiatry 186, 352–353. 10.1192/bjp.186.4.352 [DOI] [PubMed] [Google Scholar]
  11. Da Cruz D, Pearson A, Saini P, Miles C, While D, Swinson N, Williams A, Shaw J, Appleby L, & Kapur N, 2011. Emergency department contact prior to suicide in mental health patients. Emerg. Med 28(6), 467–471. 10.1136/emj.2009.081869 [DOI] [PubMed] [Google Scholar]
  12. Doupnik SK, Rudd B, Schmutte T, Worsley D, Bowden CF, McCarthy E, Eggan E, Bridge JA, & Marcus SC, 2020. Association of suicide prevention interventions with subsequent suicide attempts, linkage to follow-up care, and depression symptoms for acute care settings: A systematic review and meta-analysis. JAMA Psychiatry. 77(10), 1021–1030. 10.1001/jamapsychiatry.2020.1586 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dunlap LJ, Orme S, Zarkin GA, Arias SA, Miller IW, Camargo CA Jr., Sullivan AF, Allen MH, Goldstein AB, Manton AP, Clark R, & Boudreaux ED, 2019. Screening and intervention for suicide prevention: A cost-effectiveness analysis of the ED-SAFE interventions. Psychiatr. Serv 70(12), 1082–1087. 10.1176/appi.ps.201800445 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Folsom DP, Hawthorne W, Lindamer L, Gilmer T, Bailey A, Golshan S, Garcia P, Unützer J, Hough R, & Jeste DV, 2005. Prevalence and risk factors for homelessness and utilization of mental health services among 10,340 patients with serious mental illness in a large public mental health system. Am. J. Psychiatry 162(2), 370–376. 10.1176/appi.ajp.162.2.370 [DOI] [PubMed] [Google Scholar]
  15. Goldman-Mellor S, Olfson M, Lidon-Moyano C, & Schoenbaum M, 2019. Association of suicide and other mortality with emergency department presentation. JAMA Net. Open 2(12), e1917571–e1917571. 10.1001/jamanetworkopen.2019.17571 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hemming K, Haines TP, Chilton PJ, Girling AJ, & Lilford RJ (2015). The stepped wedge cluster randomised trial: rationale, design, analysis, and reporting. BMJ : British Medical Journal, 350, h391. 10.1136/bmj.h391 [DOI] [PubMed] [Google Scholar]
  17. Hofstra E, van Nieuwenhuizen C, Bakker M, Özgül D, Elfeddali I, de Jong SJ, & van der Feltz-Cornelis CM, 2020. Effectiveness of suicide prevention interventions: A systematic review and meta-analysis. Gen. Hosp. Psychiatry 63, 127–140. 10.1016/j.genhosppsych.2019.04.011 [DOI] [PubMed] [Google Scholar]
  18. Ilgen MA, Walton MA, Cunningham RM, Barry KL, Chermack ST, De Chavez P, & Blow FC, 2009. Recent suicidal ideation among patients in an inner city emergency department. Suicide. Life. Threat. Behav 39(5), 508–517. 10.1521/suli.2009.39.5.508 [DOI] [PubMed] [Google Scholar]
  19. Krysinska K, Batterham PJ, Tye M, Shand F, Calear AL, Cockayne N, & Christensen H, 2015. Best strategies for reducing the suicide rate in Australia. Aust. N. Z. J. Psychiatry 50(2), 115–118. 10.1177/0004867415620024 [DOI] [PubMed] [Google Scholar]
  20. Linehan MM, Comtois KA, Murray AM, Brown MZ, Gallop RJ, Heard HL, Korslund KE, Tutek DA, Reynolds SK, & Lindenboim N, 2006. Two-year randomized controlled trial and follow-up of dialectical behavior therapy vs therapy by experts for suicidal behaviors and borderline personality disorder. Arch. Gen. Psychiatry 63(7), 757–766. 10.1001/archpsyc.63.7.757 [DOI] [PubMed] [Google Scholar]
  21. Lu L, Dong M, Zhang L, Zhu X-M, Ungvari GS, Ng CH, Wang G, & Xiang Y-T, 2020. Prevalence of suicide attempts in individuals with schizophrenia: a meta-analysis of observational studies. Epidemiol. Psychiatr. Sci 29, e39. 10.1017/S2045796019000313 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Miller I, Gaudiano B, & Weinstock L, 2016. The Coping Long Term with Active Suicide Program (CLASP): description and pilot data. Suicide. Life. Threat. Behav 46, 752–761. 10.1111/sltb.12247 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Miller IW, Camargo CA Jr., Arias SA, Sullivan AF, Allen MH, Goldstein AB, Manton AP, Espinola JA, Jones R, Hasegawa K, & Boudreaux ED, 2017. Suicide prevention in an emergency department population: The ED-SAFE Study. JAMA Psychiatry. 74(6), 563–570. 10.1001/jamapsychiatry.2017.0678 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Miller IW, Gaudiano BA, & Weinstock LM, 2022. Coping long term with active suicide program (CLASP): A multi-modal intervention for suicide prevention. New York, New York: Oxford University Press. [Google Scholar]
  25. National Center for Injury Prevention and Control, Centers for Disease Control, 2023. WISQARS fatal and nonfatal injury reports. https://wisqars.cdc.gov/reports/?o=NFI&y1=2010&y2=2020&d=0&i=2&m=3000&g=00&me=&s=0&r=&e=&a=ALL&g1=0&g2=199&a1=0&a2=199&r1=INTENT&r2=NONE&r3=NONE&r4=NONE&adv=true (accessed 13 June 2023).
  26. Nock MK, Borges G, Bromet EJ, Alonso J, Angermeyer M, Beautrais A, Bruffaerts R, Chiu WT, de Girolamo G, Gluzman S, de Graaf R, Gureje O, Haro JM, Huang Y, Karam E, Kessler RC, Lepine JP, Levinson D, Medina-Mora ME, Ono Y, Posada-Villa J, & Williams D, 2008. Cross-national prevalence and risk factors for suicidal ideation, plans and attempts. Br. J. Psychiatry 192(2), 98–105. 10.1192/bjp.bp.107.040113 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Novick DM, Swartz HA, & Frank E, 2010. Suicide attempts in bipolar I and bipolar II disorder: a review and meta-analysis of the evidence. Bipolar Disord. 12(1), 1–9. 10.1111/j.1399-5618.2009.00786.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Olfson M, Gao YN, Xie M, Wiesel Cullen S, & Marcus SC, 2021. Suicide risk among adults with mental health emergency department visits with and without suicidal symptoms. J. Clin. Psychiatry 82(6). 10.4088/JCP.20m13833 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Oliveira SEH, Esteves F, & Carvalho H, 2015. Clinical profiles of stigma experiences, self-esteem and social relationships among people with schizophrenia, depressive, and bipolar disorders. Psychiatry Res. 229(1), 167–173. 10.1016/j.psychres.2015.07.047 [DOI] [PubMed] [Google Scholar]
  30. Palmer BA, Pankratz VS, & Bostwick JM, 2005. The lifetime risk of suicide in schizophrenia: A reexamination. Arch. Gen. Psychiatry 62(3), 247–253. 10.1001/archpsyc.62.3.247 [DOI] [PubMed] [Google Scholar]
  31. Reichenberg A, Harvey PD, Bowie CR, Mojtabai R, Rabinowitz J, Heaton RK, & Bromet E, 2008. Neuropsychological function and dysfunction in schizophrenia and psychotic affective disorders. Schizophr. Bull 35(5), 1022–1029. 10.1093/schbul/sbn044 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Saha S, Chant D, & McGrath J, 2007. A systematic review of mortality in schizophrenia: Is the differential mortality gap worsening over time? Arch. Gen. Psychiatry 64(10), 1123–1131. 10.1001/archpsyc.64.10.1123 [DOI] [PubMed] [Google Scholar]
  33. Schwartz RC, & Blankenship DM, 2014. Racial disparities in psychotic disorder diagnosis: A review of empirical literature. World J. Psychiatry 4(4), 133–140. 10.5498/wjp.v4.i4.133 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Sharaf AY, Ossman LH, & Lachine OA, 2012. A cross-sectional study of the relationships between illness insight, internalized stigma, and suicide risk in individuals with schizophrenia. Int. J. Nurs. Stud 49(12), 1512–1520. 10.1016/j.ijnurstu.2012.08.006 [DOI] [PubMed] [Google Scholar]
  35. Sirotich F, Durbin A, & Durbin J, 2016. Examining the need profiles of patients with multiple emergency department visits for mental health reasons: a cross-sectional study. Soc. Psychiatry Psychiatr. Epidemiol 51(5), 777–786. 10.1007/s00127-016-1188-5 [DOI] [PubMed] [Google Scholar]
  36. Stanley B, & Brown GK, 2012. Safety planning intervention: A brief intervention to mitigate suicide risk. Cogn. Behav. Pract 19(2), 256–264. doi: 10.1016/j.cbpra.2011.01.001 [DOI] [Google Scholar]
  37. Stanley B, Chaudhury SR, Chesin M, Pontoski K, Bush AM, Knox KL, & Brown GK, 2016. An emergency department intervention and follow-up to reduce suicide risk in the VA: Acceptability and effectiveness. Psychiatr. Serv 67(6), 680–683. 10.1176/appi.ps.201500082 [DOI] [PubMed] [Google Scholar]
  38. Stewart I, & Lees-Deutsch L, 2022. Risk assessment of self-injurious behavior and suicide presentation in the emergency department: An integrative review. J. Emerg. Nurs 48(1), 57–73. https://doi.org/ 10.1016/j.jen.2021.10.002 [DOI] [PubMed] [Google Scholar]
  39. Villa J, Ehret BC, & Depp CA, 2020. Systematic review of the inclusion of people with psychosis in suicide-specific clinical trials. Crisis. 41(3), 233–236. 10.1027/0227-5910/a000628 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1950551-supplement-1.docx (32.3KB, docx)

RESOURCES