Principal Investigator: Leona L. Eggert
Level of Intervention: Indicated
Target Population: High school students at-risk for suicidal behavior due to risk of school dropout.
References: Randell, Eggert, & Pike (1998)
Theory (Risk & Protective Factors Targeted):
Suicide is a leading cause of death and the frequency of the problem is significant in community samples (11 to 48%). High school dropouts appear to be an at-risk group (40% screen in as at risk for suicide). Lower grades and poor academic orientation are both associated with suicidal ideation. Drug involvement, personal strain, and family strain are often co-occurring risk factors. No single factor can be used to predict suicidal behavior but the interrelationship among suicide-risk behaviors, depression, and anger is important and one of the strongest predictors. Mediators include personal resources (personal control, coping behavior, self-esteem) and family resources (family support, family goals met, family distress). The authors note that the school is considered an appropriate place to conduct a preventive intervention because it provides daily contact with the student and is consistent with the schools role.
Description of Intervention:
Both approaches were designed to build the personal and family strengths, which influence suicide-risk behaviors. The theoretical basis for the programs was the social support literature, social learning theory, and social influence models.
Counselors CARE (C-CARE)
A two-hour computer-assisted, comprehensive assessment of risk and protective factors related to suicide and a brief intervention to designed to provide empathy & support, develop the youths social network connections with adults in school and home, and develop their own personal resources (e.g. positive coping skills and help-seeking behaviors). The program is administered by specially trained, advance practice clinicians at the students school. Total intervention lasts 3.5 to 4 hours.
Coping and Support Training (CAST)
Small group, life skills training provided across 12 sessions that meet twice weekly over a 6-week period in the school. Most groups consisted of 6-7 students and were led by specially trained group leaders. Session content included building group support, helping students problem solve, anger management, strengthening students ability to recognize their own progress, and building self-esteem.
This condition followed the high schools typical procedure for addressing suicidal-risk in a student. A trained interviewer conducted a minimal assessment interview entitled the "Suicidal Ideation and Intent Scales" (Beck, Kovacs, & Weissman, 1979) and then implemented a brief, standardized "social connections" procedure. The students parents/guardian and appropriate school personnel were also contacted.
Total sample consisted of 341 9th through 12th grade students (103 CAST & C-CARE; 117 C-CARE; 121 CONTROL) with an average age between 15 & 16 years. Males and females were equally represented. The sample was 39.9% White, 12.3 African-American, 12.9 Mixed Ethnicity, 12.9% Asian/Pacific Islander, 7% Hispanic/Latino, 2.1% American Indian/Alaska Native, 3.8% Other, and 9.1% Unknown.
Youth were identified as at-risk due to their dropout and suicide potential. A two-stage screening procedure was used. Stage one consisted of examining school records and utilizing school referrals to identify a pool of youth at-risk for dropout. Students in this group had either 1) been a previous dropout, 2) were in the top 25th percentile for absences, behind credits and have a GPA < 2.3 or a drop in GPA to < 0.7, or 3) referred by the school as in jeopardy of school failure or dropout and meeting one of the above criteria. Youth from this screen were randomly selected and invited to participate. Refusal rates similar across cohorts and study groups.
Once students agreed to participate and consent was obtained they participated in a second screen. In stage 2, students completed the High School Questionnaire (HSQ; Eggert, Herting, & Thompson, 1989; 1995) which included the Suicide Risk Screen (SRS; Eggert, Thompson, & Herting, 1994), an assessment of indicators related to suicide-risk. On average 40-50% of the youth in each cohort identified as high-risk of school dropout also screened in as being at suicide-risk. Students classified as "at suicide-risk" based on this measure were randomly assigned to one of the 3 study conditions. Students in the CAST group were more likely to refuse to participate (p<.03) and retention rate for this group was significantly less compared to the other two groups (p<.001). These findings were not surprising to the authors who noted that participation in CAST required greater motivation and time by the participant.
Subjects were randomly assigned to one of the two experimental groups (CAST and C-CARE or C-CARE Only) or a control group.
The C-CARE and CONTROL conditions only lasted 4 weeks and ended prior to the CAST skills training program. An assessment was conducted when these programs ended (Time 2). Participants in these two conditions received an additional set of booster sessions and participated in a second follow-up assessment (Time 3) at 10-weeks after the pre-intervention assessment when the post-intervention assessment for the CAST program was also conducted.
Trend analyses were conducted using MANOVA. Groups were similar on background variables and baseline levels of risk and protective factors except for age. CAST subjects tended to be slightly older.
Post-Test (10 weeks after baseline)
Trend analyses were based on baseline, 4-week, and 10-week follow-up data and included all subjects assigned to each condition (including those who did not participate). Means for all three groups moved towards normative levels suggesting that the changes for all three groups were clinically as well as statistically significant. There were no group differences on suicide-risk behaviors. Significant group differences in levels of depression (p<.01) for the intervention groups compared to the control group were attributed to the C-CARE component. There was a significant decline in anger control problems for all three groups.
Program effects on personal protective factors and family factors were also examined. Group differences were found on all three indicators: self-esteem (p<.001), personal control (p<.02), and problem-solving coping (p<.001). Changes in self-esteem were attributed to both of the intervention conditions, but the changes in personal control appeared to be a function of the CAST intervention. Improvements in problem-solving coping were greatest for CAST participants, followed by those in C-CARE, which were greater than those of participants in the control condition. Increases in coping were evident for the CAST group only at Time 3. All three groups evidenced significant decreases in family distress (p<.000). The authors attributed this to the parental phone call that was part of all 3 conditions. Youth in the two intervention conditions also evidenced significant increases in their ability to meet conventional family goals (p<.02). By the end of the follow-up period it appeared that youth in the CAST condition evidenced significant changes in perceived family support (p<.05) which was attributed to the training CAST youth received on how to seek out support from significant adults.
Strengths & Limitations:
There are very few suicide prevention programs and the C-CARE and CAST programs are two of the only programs that have been evaluated with a randomized clinical trial design. The results of the study provide preliminary evidence for the ability of the two school-based to reduce suicide-risk in youth. Although participants in all three conditions reported significant reductions (statistically and clinically) in suicide-risk behaviors over time, the intervention appeared to contribute to significant reductions in students depression and the CAST program contributed specifically to improvements in self-efficacy (personal control & problem-solving) and perceived family support. The findings should be interpreted cautiously given the short-term nature of the follow-up and the fact that all of the outcome measures were self-report.
The intervention fidelity of the C-CARE and CAST programs was excellent. The CAST program was implemented with a standardized published protocol. Process evaluation of both interventions were conducted by videotaping all sessions and coding for content compliance and leader competency. Student responses and participation were also coded from the videotapes. The Principal Investigator or program supervisor evaluated each of the tapes for reliability and quality control. They also provided supervision and group consultation to the group leaders. The control condition assessments were also videotaped and randomly reviewed by the program supervisor.