Suicidal Motivations Among Opioid Overdose Survivors: Replication and Extension

Abstract

Background:

Several studies suggest suicidal motivation may contribute to opioid overdose fatalities in people with opioid use disorder. In this study, we sought to replicate and extend prior findings suggesting that a desire to die is common prior to nonfatal opioid overdose in people with opioid use disorder.

Methods:

Adults receiving inpatient detoxification and stabilization who reported a history of opioid overdose (N=60) completed questions about suicidal cognition prior to their most recent overdose.

Results:

Approximately 45% reported some desire to die prior to their most recent overdose, with 20% reporting they had some intention to die. The correlation between these ratings was of a moderate magnitude (ρ=.58). Almost 40% of the sample perceived no risk of overdose prior to their most recent overdose event, suggesting a significant underestimation of risk in this population.

Conclusions:

Desire to die was common in adults with opioid use disorder prior to nonfatal opioid overdose events, and 1 in 5 people with a history of opioid overdose reported intention to die prior to their most recent opioid overdose. Careful assessment of suicidal cognition in this population may improve prevention of opioid overdose deaths.

1. Introduction

Despite public health prevention efforts to reduce opioid-related mortality, opioid fatalities, including death by suicide, have substantially increased, and spread throughout the United States (Olfson et al., 2019; Rockett et al., 2021), with opioid overdose deaths in the 12-month period ending in April 2021 estimated at >75,000 (Ahmad et al., 2021; Centers for Disease Control, 2021). Suicide risk is elevated among people who misuse opioids (Bhatia et al., 2020; Montiel Ishino et al., 2020), those with opioid use disorder (OUD; Wilcox et al., 2004) and those surviving a non-fatal opioid overdose (Olfson et al., 2018; Olfson et al., 2020). Associations between opioid misuse, OUD, and suicidal thoughts and behaviors are complex (Bohnert and Ilgen, 2019). For example, the intentionality of a fatal drug overdose is difficult to determine in the absence of an explicit communication by the decedent, e.g., an authenticated suicide note (Rockett et al., 2018).

Research on suicidal motivations reported by those who survived opioid overdose may inform mortality prevention in high-risk populations. Prior qualitative studies in OUD concluded low rates of opioid suicide based on dichotomous (“yes/no”) self-reports about intentional opioid overdose self-harm (Darke and Ross, 2001; Heale et al., 2003; Neale, 2000). Recently we reported dimensional ratings of suicidal motivations (desire to die just prior to opioid overdose) in a small study (N=54) of people receiving inpatient OUD treatment (Connery et al., 2019). More than half (58.5%) reported some desire to die prior to their most recent overdose (> 1 on a 0–10 scale), with 36% reporting a strong desire to die (≥7 on a 0–10 scale). A study (Gicquelais et al., 2020) of 274 people with OUD and nonfatal opioid overdose history in residential treatment reported that 51% of patients endorsed passive or active suicidal ideation prior to overdose. These data suggest that suicidal motivations are common prior to nonfatal opioid overdose and occur along a continuum of severity.

The aim of the current study was to replicate these prior findings, assessing both desire to die and intention to die, as distinct suicidal cognitions, prior to participants’ most recent opioid overdose. Both desire to die and intent to die reflect motivational states, but desire to die may reflect contemplation of suicide without any planning, while intention to die reflects suicide planning and/or transition of suicidal cognitions to active suicidal behavior(s). In non-OUD populations, intention to die has greater predictive value, regarding the probability of future suicide attempts (Nock et al., 2018), than suicidal motivations without intention to die.

2. Methods

This cross-sectional study was conducted as part of a larger ongoing study characterizing adults undergoing inpatient treatment for substance use disorders (SUDs).

2.1. Participants

We recruited 326 treatment-seeking adults (June 2017-March 2020) from an inpatient withdrawal management SUD unit (average length of stay 4–5 days) at an academically affiliated psychiatric hospital in the northeastern United States. All participants were asked whether they had a history of opioid overdose; the 60 participants (52.6% of people currently in treatment for OUD and 7.1% of people seeking treatment for other SUDs) who reported a history of opioid overdose were included in this analysis. Exclusion criteria included a psychiatric, medical, or cognitive condition that would impair the participant’s ability to complete study procedures; no participants were excluded for this reason.

2.2. Procedures

Participants provided written informed consent, followed by self-report measures completed on tablet computers, lasting approximately 30-minutes. Participants were not compensated for participating. All procedures were approved by the local Institutional Review Board.

2.3. Measures

The Opioid History Questionnaire was a 5-item measure adapted from the Prescription Opiate Analgesic Use History Questionnaire (Weiss et al., 2011) to characterize lifetime history of opioid use in participants with current OUD. The Opioid Overdose Questionnaire was a 15-item measure developed by our group to assess the thoughts surrounding opioid overdose in participants with OUD (Connery et al., 2019). Items on this questionnaire assessed overdose history, number of past overdoses, beliefs about likelihood of overdose, and thoughts about death from overdose. The questionnaire comprised the following items regarding cognition prior to overdose: 1) “Just before your most recent opioid overdose, how likely did you think that you would overdose?” (from 0 “no chance” to 10 “extremely likely”) 2) “Just before your most recent opioid overdose, how strongly did you want to die?” (from 0 “I did not want to die” to 10 “I definitely wanted to die”), and 3) “At the time of your most recent opioid overdose, were you trying to kill yourself?” (from 0 “not at all” to 10 “definitely”).

2.4. Data Analysis

We first evaluated evidence for skewness and outliers in our variables of interest. One participant was excluded from analyses due to inconsistent responding. Participants were given the opportunity to decline to answer questions; all results reflect a minimum of 53 participants. We then characterized our a priori outcomes of interest, self-reported perceived likelihood of overdose, desire to die, and intention to die using descriptive statistics. Spearman’s correlation (ρ) was used to assess the associations between these variables due to non-normal distributions. Analyses were conducted in SPSS Version 20.

3. Results

Sample demographic characteristics are presented Table 1. The mean number of overdoses was 4.5 (SD=6.2, range=1–35). Approximately 48% reported that their most recent overdose was within the past year, with 20% reporting that it occurred within the past month.

Table 1.

Sample Characteristics (N=60)

	mean (SD) or n (%)
Age	37.8 (11.0)
Gender (% female)	23 (39%)
Education
Less than high school	3 (5.1%)
High school or equivalent	22 (37.3%)
Some college	20 (33.9%)
College graduate	14 (23.7%)
Employment
Unemployed	28 (47.5%)
Employed full-time	15 (25.4%)
Student	1 (1.7%)
Retired	2 (3.4%)
Disabled	11 (18.6%)
Homemaker	2 (3.4%)
Marial Status
Never married	37 (62.7%)
Married	9 (15.3%)
Divorced	5 (8.5%)
Separated	2 (3.4%)
Widow/widower	3 (5.1%)
Partnered	3 (5.1%)
Race
White	55 (93.2%)
Black/African American	1 (1.7%)
More than one race	1 (1.7%)
Other Race	2 (3.4%)
Hispanic/Latinx	2 (3.4%)

Prior to the most recent overdose, 57% reported using opioids daily, 17% reported intermittent opioid use, and 26% reported recurrent opioid use following a period of abstinence. To the question, “Do you believe that any events or stressors led to your most recent opioid overdose?” most participants (64%) reported “no.”

Responses to questions about perceived risk of overdose and desire to die prior to the most recent overdose were quite varied (see Figures 1a and 1b). When asked the likelihood of overdose prior to the most recent overdose, 39.3% of participants reported no chance, whereas 14.3% reported a high likelihood (≥7 on a 0–10 scale); the mean score on this variable was 2.4 (SD=3.0). When asked about desire to die, 54.4% of participants reported they did not want to die (0 on a 0–10 scale), 45.6% reported some desire to die (> 0 on a 0–10 scale), and 19.4% reported a strong desire to die (≥7 on a 0–10 scale), with a mean of 2.7 (SD=3.6).

Figure 1.

Self-Reported Cognition Prior to Most Recent Opioid Overdose (N=59)

As an extension of our prior study, we added a question about intention to die, “At the time of your most recent overdose, were you trying to kill yourself?” scored on a scale of 0–10 where 0 = not at all and 10 = definitely (Figure 1c). To this question, 80.4% of respondents reported no intention to kill themselves (0 on a 0–10 scale), 8.9% of respondents reported a mild (1–3/10) intention, 5.3% reported a moderate (4–6/10) intention and 5.3% reported strong intention to die (>7 on a 0–10 scale, with 10 “I was definitely trying to kill myself”); the mean score for this item was 0.9 (SD=2.6). See Figure 1c.

Correlations between these factors were in the range of a medium effect size (p<.001): ρ=.47 for likelihood of overdose and desire to die, ρ=.52 for likelihood of overdose and intention to die, and ρ=.58 for desire to die and intention to die. As would be predicted, the 11 participants who reported a non-zero response to the suicidal intent question also reported some desire to die: 3 reported a low desire to die (1–3), 3 reported a moderate desire to die (4–6), 1 reported a high desire to die (7–9), and 4 reported an extremely high desire to die (10). Among those reporting no perceived likelihood of overdose (0), 1 reported some intention to die (4), and 6 reported some desire to die (range 2–10). Conversely, among the 34 participants who reported some likelihood of overdose (i.e., > 1/10), 14 reported neither desire nor intent to die (0).

4. Discussion

This study affirmed and extended previous findings that approximately half of treatment-seeking adults with OUD and history of opioid overdose endorse some desire to die prior to overdose, and one in five answers positively the question, “Were you trying to kill yourself?” on a rating scale. High prevalence of both suicidal desire without planning or suicidal behavior and more intentional suicidal behaviors (also discussed as “passive” versus “active” suicidal ideation) have been consistently reported for OUD as well as for other SUD populations and are associated with worse SUD treatment outcomes and elevated self-harm risk (Breet et al., 2018; Kelly et al., 2020).

Participant ratings of suicidal motivation and intention suggest that these are dimensional, rather than binary, factors, with the range from none to high represented. Future research is needed to understand the implications of these dimensional ratings (e.g., the prognostic significance of binary vs. dimensional indicators).

Reports of intention to die and desire to die were moderately correlated, yet not redundant, suggesting that the assessment of intention to die alone may fail to detect potentially clinically relevant suicidal thinking. Furthermore, one individual reported intent to die despite no perceived risk of overdose death, and 14 of 34 participants reporting some risk of overdose reported neither intent nor desire to die. Our findings highlight the need to better understand (1) suicide risk that may be associated with passive suicidal ideation (desire to die) in OUD (Gicquelais et al., 2020); (2) suicide risk related to “tempting fate,” defined as risk-taking behaviors associated with the wish to be dead, which in community populations is a predictor of suicide attempt (Nock et al., 2018); and (3) fatality risk related to engagement in known potentially deadly behaviors even in the absence of identified desire or intent to die.

These findings support a need for suicide prevention interventions, specifically tailored to opioid-using populations, implemented alongside other evidence-based interventions (medications for OUD, naloxone training and dissemination, and harm reduction interventions) to amplify efforts to reduce opioid overdose deaths. Both a psychological autopsy study (Abiragi et al., 2020) and a Danish national record-linkage study (Hesse et al., 2020) suggest that individuals with OUD and co-occurring mental health disorders are at greatest risk for intentional opioid overdose. A national data claims study (Busch et al., 2020) reported that OUD patients with nonfatal overdose were unlikely to receive either psychosocial treatment or FDA-approved medications to treat OUD, despite evidence that receipt of buprenorphine or methadone treatment following opioid overdose reduces both all-cause mortality and opioid-related mortality (LaRochelle et al., 2018), and methadone maintenance significantly reduces suicidal behaviors (Molero et al., 2018). While it may be feasible and useful to screen OUD populations to stratify risk for intentional and unintentional overdose (Copeland et al., 2020; Stover et al., 2020), the prevalence of suicide risk factors and the dimensionality of suicidal motivations and intentions reported in this study and prior studies would favor the integration of universal suicide prevention screening and education, and personalized safety planning (including attention to underestimation of overdose risk), into care for all OUD patients. This may be particularly important for times of crises like the SARS-CoV-2 pandemic, during which emergency department visits surged for both suicide attempts and drug overdoses (Holland et al., 2021).

Study limitations include cross-sectional assessment among predominantly white OUD treatment-seekers, retrospective self-report bias, and absence of validated diagnostic assessments for co-occurring mental health disorders, which are significantly associated with OUD suicide risk and suicidal motivations (Bohnert et al., 2017). This limits our ability to extend correlational analysis to examine mediating factors, such as polysubstance use; however, the replication of our original report (Connery et al., 2019) and extension of results to include dimensional ratings of suicide intention in nonfatal opioid overdose contributes to the emerging evidence base and provides a hopeful opportunity to incorporate existing, well-defined suicide prevention interventions to stem the rising burden of U.S. drug overdose deaths. Future research on the fluctuation of suicidal motivations and intentions over time in OUD populations, as well as studies of predictors of suicidal thinking and behavior, may clarify links between suicidal cognition and opioid overdoses.

Highlights.

• 60 adults with a history of overdose answered questions about most recent overdose.

• Desire to die was reported by 45% of participants prior to the most recent overdose.

• Intention to die prior to overdose was reported by 20% of participants.