Predicting future suicidal events in adolescents using the Concise Health Risk Tracking Self-Report (CHRT-SR)

Highlights

• The CHRT-SR predicts suicide attempts with at least 80% sensitivity and 50% specificity in suicidal adolescents.

• Findings are consistent with previous reports that the CHRT-SR predicted suicidal events in adults with Bipolar Disorder.

• This is one of the first rating scales to predict future suicide attempts with acceptable sensitivity or specificity.

Abstract

Background

Several self-report rating scales have been developed to assess suicidal ideation, yet most have limited utility in predicting future suicide attempts. This is particularly critical in adolescence, where suicide is the second leading cause of death. This study evaluated the Concise Health Risk Tracking Self-Report (CHRT-SR) as a prospective predictor of suicide attempts and events in high-risk adolescents enrolled in a suicide-prevention intensive outpatient program (IOP).

Methods

Data were collected by retrospective chart review of adolescents treated in IOP for youth with severe suicidality. At baseline, youth completed the 14-item CHRT-SR (CHRT-SR₁₄), which assesses suicide risk based on 3 subscales: Propensity, Impulsivity, and Suicidal Thoughts. Two outcomes were assessed: actual suicide attempts and suicidal events (suicide attempt, inpatient hospitalization, or emergency department visit) during the IOP.

Results

Of the 251 adolescents who completed the baseline CHRT-SR₁₄, 26 had a suicidal event during IOP (mean time in IOP: 5.4 ± 2.3 weeks), of whom 14 had an actual suicide attempt. Youth with any suicidal event had higher scores than those without an event on the CHRT-SR₁₄ Total (p = .005), Propensity (p = .008), and Suicidal Thoughts (p = .001) scales at baseline. Youth who made a suicide attempt had significantly higher scores than those without an event for the Total Score, Propensity, and Suicidal Thoughts subscales. CHRT-SR₁₄ Total Score of 28 had a sensitivity of 85.7% and specificity of 56.5% in predicting suicide attempts. A score of 22 predicted suicidal events, with a sensitivity of 80.8% and specificity of 40.9%. CHRT-SR₇ Total Score of 12 predicted suicide attempts, with a sensitivity of 85.7% and specificity of 53.4%.

Conclusions

The CHRT-SR₁₄ self-report predicts suicide attempts and events with at least 80% sensitivity and acceptable specificity in adolescents at high-risk for suicide.

Introduction

Adolescent suicidal behavior is common, with around one million youth attempting suicide each year in the United States (Nock et al., 2013). Almost 5000 adolescents in the U.S. die by suicide annually; suicide has risen to the 2nd leading cause of death among 15-24-year-olds (Centers for Disease Control, 2016; Curtin et al., 2016). Risk factors for suicide and suicide attempts include a combination of genetic vulnerability, psychiatric disorders, life adversity, and familial, societal, and cultural factors (Hawton, Saunders and O'Connor, 2012). Given increased risk for suicide in youth with mental illness, particularly depression, the U.S. Preventive Services Task Force recommends that all primary care doctors, including pediatricians and family physicians, should routinely screen adolescents for depression, and provide recommendations or referrals for treatment for depression if needed (Siu and Force, 2016).

Regardless of age, identifying individuals who will make a suicide attempt remains a challenge. Runeson and colleagues conducted a systematic review of 21 studies (20 in psychiatric settings and one in primary care) that examined fifteen rating scales to evaluate whether these rating scales could predict suicide attempts. Most studies included only adults, although several also included adolescents. They set a clinically acceptable threshold of at least 80% sensitivity with at least a 50% specificity for evaluation of the scale. No measure met this threshold (Runeson et al., 2017). Madan and colleagues also report on the sensitivity and specificity of several measures, including the Columbia Suicide Severity Rating Scale (C-SSRS), Patient Health Questionnaire (PHQ-9), Beck Hopelessness Scale, and Suicide Cognitions Scale in 1075 adults admitted to a psychiatric hospital. Similar to Runeson and colleagues’ findings, none of the suicide measures met the threshold for reliability in predicting future suicide attempts (Madan et al., 2016). Thus, prediction of future suicidal behavior remains problematic across the lifespan, but is particularly germane in adolescents where suicide is a leading cause of death.

The Concise Health Risk Tracking© Self Report (CHRT-SR), a self-report that focuses on the past week, was designed to assess the severity of current suicide risk. The CHRT-SR assesses both suicidal ideation and other current clinical factors associated with suicide attempts. The original scale included 12 items (CHRT-SR₁₂) to assess helplessness, pessimism, perceived lack of social support, despair, and suicidal thoughts (Ostacher et al., 2015; Sanchez et al., 2018; Trivedi et al., 2011). Two additional items to assess impulsivity were later added (CHRT-SR₁₄) (Mayes et al., 2018; Reilly-Harrington et al., 2016; Sanchez et al., 2018). The CHRT-SR₁₂ has established reliability and validity in adults (Ostacher et al., 2015; Trivedi et al., 2011; Villegas et al., 2017), and the CHRT-SR₁₄ has established reliability and validity in both adults and adolescents (Mayes et al., 2018; Reilly-Harrington et al., 2016). A shortened 7-item version (CHRT-SR₇) was also validated in depressed and bipolar adult samples (Ostacher et al., 2015; Reilly-Harrington et al., 2016; Trivedi et al., 2011; Villegas et al., 2017). The scale has excellent internal reliability (Cronbach's coefficient alpha from .77 to .90) in depressed and bipolar adult populations (Ostacher et al., 2015; Reilly-Harrington et al., 2016; Trivedi et al., 2011; Villegas et al., 2017) and in suicidal adolescents (Cronbach's coefficient alpha of .91) (Mayes et al., 2018). The CHRT-SR₁₄ total score predicted suicide-related serious adverse events in 482 adults with bipolar disorder treated in a 6-month randomized clinical trial comparing lithium and quetiapine (Reilly-Harrington et al., 2016).

This report evaluated whether the CHRT-SR₁₄ could predict future suicide events and attempts in a sample of adolescents determined to be at high risk for an attempt who were treated in a suicide-prevention intensive outpatient program (IOP). We addressed the following questions: 1) Does CHRT-SR₁₄ predict suicidal events (i.e. attempts, hospitalizations, emergency department visits, etc.) and attempts? 2) What scores provide acceptable sensitivity and specificity? We also report on the performance of the previously published shorter versions, the CHRT₇ and the CHRT SR₁₂, which we extracted from the 14-item version used in this study.