Characteristics of adolescents aged 10–19 years who died by suicide, and circumstances preceding the suicide, were examined using CDC’s National Violent Death Reporting System (NVDRS) Restricted Access Database. NVDRS is an ongoing, state-based, active surveillance system that collects data on all violent deaths, including suicides. NVDRS data are collected from three sources: death certificates, coroner/medical examiner reports, and law enforcement reports. The sources are linked for each incident to provide a comprehensive picture of the death. Trained data abstractors compile information from these records into standardized variables in the NVDRS web-based system using CDC guidance and definitions. Final manner of death is assigned by the NVDRS data abstractors using information from these sources [25]. NVDRS was selected due to its inclusion of circumstance data on each suicide decedent. This analysis included suicide fatality data reported from participating U.S. states, districts, and territories between 2013 and 2018 (Appendix A) [26]. Data were restricted to 2013–2018 due to key circumstances of interest (i.e., history of suicide thoughts/plans and crisis-related circumstance variables) not collected in NVDRS prior to 2013 and 2018 being the latest year of NVDRS data available.

Descriptive analyses of the number and percentage of suicide deaths by demographic characteristics, method of injury, substance use, and a range of circumstances (i.e., alcohol-related and nonalcohol-related substance misuse and treatment, relationship problems, life stressors, recent crisis, and suicide event history) were conducted. We examined differences in characteristics and circumstances between adolescents with a known mental health condition versus no known mental health condition similar to a prior CDC study [5]. NVDRS defines mental health conditions as disorders and syndromes listed in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition [27], with the exception of problematic alcohol use and other substance use, which are captured separately in NVDRS [25]. Circumstances captured are those identified as contributing to the decedent’s suicide in coroner/medical examiner or law enforcement reports. Decedents could have experienced multiple circumstances. Circumstances are coded as “Yes/Present” or “No/Not Present/Unknown.” “No” and “Unknown” are grouped together because NVDRS cannot determine if the circumstance did not occur or occurred and was not discovered during the investigation by authorities [25]. Chi-squared tests of proportions were performed to determine the statistical difference between the proportion with and without a known mental health condition. Two-tailed p values < .05 were considered statistically significant. Logistic regression models were used to estimate adjusted odds ratios (aORs) and 95% confidence intervals (CIs) controlling for sex, age group, and race/ethnicity of specific circumstances among suicide decedents with and without a known mental health condition. Those with no known mental health condition were used as the reference group. SAS Version 9.4 (SAS Institute Inc.) was used to conduct statistical analyses. The study was exempt by the Institutional Review Board as it constituted the use of secondary analysis with deidentified data.

Between 2013 and 2018, 8,298 suicide deaths among adolescents aged 10–19 years were reported in the participating NVDRS states. Of these, 3,492 were of adolescents with a known mental health condition and 4,806 were of adolescents with no known mental health condition (Table 1). Overall, decedents were predominately male (74.7%), predominately older adolescents between the ages of 15 and 19 years (81.9%; mean age = 16.5, data not shown), and predominately non-Hispanic White (68.0%). Among adolescents with a known mental health condition, males comprised 67.9% of decedents; however, they were less likely to have a known mental health condition relative to females (odds ratio [OR] = 0.54, 95% CI = 0.49–0.60). Younger adolescents (aged 10–14 years) were less likely to have a known mental health condition compared to older adolescents (OR = 0.81, CI = 0.73–0.91). Among racial and ethnic minorities, decedents were predominantly Hispanic (12.4%), followed by non-Hispanic Black (10.4%), non-Hispanic Asian/Pacific Islander (3.5%), and non-Hispanic American Indian/Alaskan Native (2.7%). Other and unknown races constituted 2.9% of the remaining decedents. Percentages constitute the share of deaths reported during the study time period. All racial and ethnic minority groups were less likely to have a known mental health condition, relative to Whites (OR range = 0.41–0.76). Decedents identified as a sexual minority (i.e., Lesbian, Gay, or Bisexual) constituted 2.1% of deaths (only 11% of adolescents had sexual orientation reported); sexual minorities were more likely to have a known mental health condition relative to those identified as straight/heterosexual (OR = 1.76, CI = 1.26–2.46).

Death by hanging/strangulation/suffocation was the most common method of suicide overall (45.6%), followed by firearms (41.1%); however, a higher prevalence of suicide by firearm (47.6%) was reported for those without a known mental health condition compared to those with a known mental health condition (32.3%). Those with a known mental health condition were more likely to die by suicide involving hanging/strangulation/suffocation, poisoning, or motor vehicle/transport relative to those with no known mental health condition (aOR range = 1.33–2.30). The presence of alcohol and other substances among decedents was examined; of decedents tested for alcohol, 14.9% tested positive, and of decedents tested for marijuana, 30.9% tested positive. Of substances tested, decedents with a known mental health condition were more likely to be tested across all substances (aOR range = 1.32–2.04). Among those tested, those with a known mental health condition were less likely to test positive for alcohol (aOR = 0.72, CI = 0.61–0.86) but more likely to test positive for all other substances (aOR range = 1.36–9.57) compared to decedents with no known mental health condition (exceptions are cocaine, marijuana, and opioids which showed no significant differences between groups). Of decedents with a known mental health condition, 40.9% tested positive for antidepressants.

Of the 8,298 total suicide deaths reported, 7,414 (89.3%) included additional data on circumstances reported as a contributing factor to the suicide (Table 2). As mental health condition is one of the circumstance variables, information on circumstances were available for 100% of decedents with a known mental health condition and 81.6% of decedents with no known mental health condition. Of adolescents with a known mental health condition, the most common diagnoses were depression (71.5%), anxiety disorder (18.6%), attention deficit disorder/attention deficit hyperactivity disorder (14.6%), and bipolar disorder (9.9%). In addition, a total of 2,502 adolescents (33.7%) with circumstances known were identified as having a current depressed mood (feeling down or sad as per reports of informants; not a clinical diagnosis of depression) prior to the suicide. Roughly 29.2% of adolescents were in current treatment for either a mental health or substance misuse condition, while 39.4% reported a history of treatment for a mental health or substance misuse condition.

Adolescents with a known mental health condition were more likely to have a problematic alcohol or other substance misuse problem relative to those with no known mental health condition (aOR = 1.67, CI = 1.34–2.09 and aOR = 1.78, CI = 1.54–2.05, respectively). Adolescents with a known mental health condition were less likely to have family relationship problems (aOR = 0.81, CI = 0.73–0.91) but more likely to have other nonintimate relationship problems (aOR = 1.27, CI = 1.05–1.53) relative to those with no known mental health condition. Adolescents with a known mental health condition were also more likely to have the death of a loved one or suicide of a family member or friend noted as a contributing factor relative to those with no known mental health condition (aOR = 1.41, CI = 1.13–1.77 and aOR = 1.53, CI = 1.21–1.93, respectively). Notably, there were 1,535 adolescents (20.7%) noted to have an argument or conflict prior to suicide and there was no significant difference in this contributing circumstance between those with and without a known mental health condition. Adolescents with a known mental health condition were more likely to have a history of abuse as a child as a contributing circumstance relative to those with no known mental health condition (aOR = 3.03, CI = 2.35–3.90). There were no significant differences between those with and without a known mental health condition for the circumstances of a recent criminal legal problem or any crisis experienced within 2 weeks of death. This included a crisis related to family relationship problems, school crisis, criminal/legal problem crisis, and crisis related to suicide or death of a family member or friend.

Regarding suicide history, 2,614 (35.3%) adolescents had a history of suicidal thoughts and 1,600 (21.6%) had a history of attempts. Relative to the fatal suicide attempt, 2,768 (37.3%) decedents left a suicide note and 1,921 (25.9%) disclosed suicidal intent. Adolescents with a known mental health condition were more likely to disclose intent and have a history of ideation and attempts reported (aOR range = 1.60–5.64) relative to those with no known mental health condition.

This study examined circumstances contributing to suicide among adolescents with and without known mental health conditions. Nearly 58% of adolescents whose suicide deaths were reported in NVDRS between 2013 and 2018 had no known mental health condition. While decedents with known mental health conditions were more likely to experience problematic alcohol misuse, substance misuse, nonintimate relationship problems, and school problems, there were no significant differences between these groups for the contributing factors of arguments or conflicts, criminal or legal problems, or any crisis occurring within the previous two weeks of death. This finding suggests some adolescents may have difficulty coping with conflict and crisis regardless of mental health condition. Experiencing crises may be overwhelming and leave adolescents vulnerable to risk of suicide. Thus, suicide prevention efforts which recognize life stressors and other crises experienced among adolescents irrespective of diagnosed mental health conditions may be instrumental in reducing the risk of suicide.

While a majority of decedents were not reported to have a known mental health condition, nearly 40% of those with a known mental health condition were also reported to not be in current treatment for mental health at the time of death. Comprehensive suicide prevention which includes identifying and remedying barriers to treatment and increasing education of the need for mental health services at the community level is also important in reducing the risk of suicide. This study also found that rates of firearm suicide were higher among those with no known mental health condition compared to other methods. This may indicate perceived lethality of firearms among those experiencing crisis and further supports creating protective environments by reducing access to lethal means among adolescents at risk of suicide.

Studies show that one of the strongest predictors of eventual suicide fatality among adolescents is a previous suicide attempt [21,22]; however, adolescents with no known mental health condition were less likely to have a history of suicide thoughts or attempts or to have disclosed their intent or written a note prior to death. Previous research found that adolescents exhibit domains of impulsivity at different stages of their development that are associated with an increased risk of suicidal ideation and attempt [28], therefore a broader approach to suicide prevention that focuses on strategies for reducing impulsive behaviors and impulsive responses to emotions may be beneficial for adolescents both with and without known mental health conditions [20]. In addition, upstream approaches to suicide prevention that include community strategies to reduce the stigma associated with disclosing thoughts of suicide can promote help seeking behavior among adolescents before suicide attempts are made.

Overall, the most common known circumstances contributing to suicide among all decedents were current depressed mood, mental health/substance misuse treatment (current treatment and ever treated), family and intimate partner relationship problems, arguments or conflicts, school problems, and having a school or family relationship crisis within 2 weeks of death. The Surgeon General’s call to action for suicide prevention recognizes the need for broadening of interventions that not only include individual-focused mental health treatment but also a focus on teaching coping and problem-solving skills, strengthening interpersonal relationships, promoting connectedness, and strengthening community and societal-level family support [19]. The CDC’s suicide prevention technical package also provides guidance for implementing interventions aimed at promoting connectedness and teaching coping and problem-solving skills; these interventions include social-emotional learning programs to help adolescents build skills in emotional regulation and conflict resolution [20]. Strong suicide prevention efforts recognize the multilayered challenges faced by adolescents as they cope with physical, physiological, psychological, social, and emotional changes through growth and development [29]. Recognizing challenges can help guide suicide prevention interventions and enable a smooth transition from adolescence into adulthood.

The results of this study are subject to several limitations. First, this report is based on data from U.S. states and jurisdictions that participated in NVDRS between 2013 and 2018. Data collected did not include all 50 states; however, in 2018, NVDRS expanded to include all 50 states, Puerto Rico, and Washington, DC, to begin collecting data in 2019 [25]. A future replication of the present study with multiple years of information available in the expanded NVDRS database might yield more representative results. Second, in our study, 18.4% of adolescents with no known mental health condition were missing additional data on circumstances preceding the death. Consequently, the frequencies of the particular circumstances examined in our study likely are underestimates. Third, because NVDRS cannot determine if a particular circumstance did not occur or occurred and was not discovered by authorities, persons coded as without known mental health conditions may have had mental health challenges that were unknown, undiagnosed, or not reported during investigation. Fourth, prior CDC survey data show higher rates of suicide ideation and attempt among lesbian, gay, and bisexual identifying adolescents [24]; however, sparse data on the sexual orientation status of decedents in NVDRS, an issue common in mortality data due to difficulties with collecting postmortem information about sexual orientation, hindered additional analyses specific to this population. There are significant steps being taken to systematically enhance and standardize NVDRS data collection to identify sexual orientation at the time of death and enhance coding through more in-depth qualitative analysis of the NVDRS narrative data [30,31]. Fifth, NVDRS relies on partnerships between states and local agencies to obtain the information recorded in the database; the degree of completeness might vary among and within states and depends on the degree of knowledge available from family, friends, and other informants during the death investigation. Finally, given that the study is cross-sectional, it should be noted that findings do not imply causality.