The threat of separation from a parent theoretically increases the risk of adolescent suicide attempts. The present study evaluated this and other hypothesized risk factors in a sample of adolescent suicide attempters and nonsuicidal controls, using the Psychiatric Consultation Checklist (Lyon, 1987). Stepwise logistic regression was used to predict group membership. It was found that threat of separation from a parental figure, insomnia, neglect, substance abuse, suicidal ideation, and failing grades were the strongest predictors of suicide attempt. Ten predictor variables correctly identified 97% of suicide attempters and 86% of nonattempters. Unexpected findings included high levels of truancy, threatening others, and separation from a parent before the age of 12 among nonattempters.
Suicide is one of the leading causes of death among Americans aged 15 to 24 years (Centers for Disease Control, 1998c; Guyer, MacDorman, Martin, Peters, & Strobino, 1998). Among those aged 15 to 19, suicide accounted for 13% of deaths in 1997 (Guyer, MacDorman, Martin, Peters, & Strobino, 1998). Although African American youths have historically had lower suicide rates than have whites, from 1980 to 1995 the suicide rate for African Americans aged 10 to 19 increased from 2.1 to 4.5 per 100,000 population. Among African Americans aged 10 to 14, the suicide rate increased the most (233%), while increasing 126% for African Americans aged 15 to 19. The largest rate of increase occurred for African Americans who live in the South (214%). By gender, the largest increase in suicide occurred among African American males (Centers for Disease Control, 1998a).
Those who make initial suicide attempts are at greater risk for subsequent suicide, with repeated attempts increasing in lethality over time (Eyman & Smith, 1986; Robbins & Alessi, 1985). Significant relationships have been found between major depression, suicidal ideation, health problems, and suicide attempts (Garrison, Jackson, Addy, et al., 1991; Alcohol, Drug Abuse, and Mental Health Administration, 1989).
The Centers for Disease Control (1998b) reported that, in a 12-month period, 20.5% of students in grades 9 through 12 seriously considered attempting suicide, 7.7% had attempted suicide one or more times, and 2.6% sought medical care after a suicide attempt. African American high school students (7.3%) were as likely as white students (7.3%) to attempt suicide.
One factor hypothesized to contribute to suicide attempts among adolescents is disturbance in the parent-child relationship (Sabbath, 1969). For example, a parent may have an unconscious or conscious wish to be rid of a child, who is regarded as "expendable." It has been postulated that people treat themselves as important others have treated them; thus, it may be that an adolescent who is no longer tolerated by a parent will engage in suicidal behavior. The significantly higher incidence of suicide attempts among rejected and abused young children (Brent, 1987), physically abused grade school children (Robbins & Alessi, 1985), sexually abused children (Adams-Tucker, 1982), and abused and/or neglected adolescents (Khan, 1987) provides support for this position. Further, Green (1978) found that self-destructive behavior in children was often precipitated by parental beatings or occurred in response to actual or threatened separation from key parental figures (see also Asarnow, 1992, regarding the association of suicidal ideation and attempts with family stress).
Studies have examined psychopathology, family functioning, and cognitive style in urban adolescents with suicide attempts (Summerville, Kaslow, Abbate, & Cronan, 1994); depression and distress among gay and lesbian African Americans (Cochran & Mays, 1994); depression and suicidal ideation in African American high school students (Lester & Anderson, 1992); psychopathology in African American female high school students with suicide attempts (Summerville, Abbate, Siegel, et al., 1992); and homicide and suicide among African American males aged 15 to 24 (Greenberg & Schneider, 1992). Other variables associated with suicide attempts among adolescents include: aggressive or hostile behavior (Brent, Perper, Goldstein, Kolko, Allan, Allman, & Zelenak, 1988; Cohen-Sandier, Berman, & King, 1982; Joffe, Offord, & Boyle, 1988; Shaffer, Garland, Gould, Fisher, & Trautman, 1988); separation from either parent during critical periods of development (Cohen-Sandier, Berman, & King, 1982); and modeling of the suicidal behavi or of others (Diekstra, 1985; Shafi, Carrigan, Whittinghill, & Derrick, 1985). Finally, alcohol and substance abuse seem to be playing an increasing role in adolescent suicide attempts (Vega, Gil, & Warheit, 1993).
A major obstacle in the prevention of suicide and suicide attempts has been the inability to predict who is most likely to engage in these behaviors. For example, in a prospective study of 4,800 psychiatric patients, efforts to predict subsequent suicides or suicide attempts were unsuccessful (Pokorny, 1983). However, screening for risk factors in a clinical setting remains the best approach.
At this time, there is no brief, empirically based screening instrument to assess adolescent suicidal risk factors. Instruments based on adult samples cannot be assumed to be useful with adolescents, because adolescence is a unique developmental phase. Thus, the purpose of the present study was twofold: first, to study the specificity and sensitivity of the brief (10 minute) Psychiatric Consultation Checklist (Lyon, 1987) as a screening instrument for adolescents at risk for suicide attempts; second, to examine ten purported risk factors relevant to adolescent suicide attempts (using stepwise logistic regression).
Fifty-one adolescent suicide attempters, between the ages of 12 and 18, were evaluated at an inner-city pediatric hospital over a two-year period. They represented the most serious cases, having sought medical attention following an unsuccessful suicide attempt. The control group consisted of 124 "walk-in" patients who used the adolescent clinic for primary care (those with a history of suicide attempts were excluded from the study). The white adolescents primarily lived in the suburbs and all of the African American youth resided in the inner city.
A decision was made to drop the small group of white adolescents and to examine data only for the African American suicide attempters (n = 38) and controls (n = 76), matched 1:2 on age and gender (with one mismatch for gender). Ages ranged from 12 to 17 years, with a mean of 14.7 for suicide attempters and 14.9 for controls. Over four-fifths of the adolescents in both groups were females (82.2% and 81.6%, respectively), and all were from lower socioeconomic backgrounds.
The Psychiatric Consultation Checklist (see Figure 1) was used to collect data, decreasing the chances of underreporting and increasing the reliability of already established procedures. Checklist items are based on a literature review, theory, and a pilot study involving a review of one-year medical records (Lyon, 1987).
Within 24 hours of a suicide attempt, the checklist was completed by a mental health professional who had been trained (by the first author) in its administration. Data collected from the suicide attempters were verified through medical records and interviews with parental figures. There were incomplete data for two suicide attempters and two controls. Their data were not used in the regression model.
Two-way chi-square analyses or Fisher's exact tests, as appropriate, were used to evaluate the bivariate significance of each categorical variable and the response variable, suicide attempt. Univariate tests were two-tailed, with significance set at the .05 alpha level. The Bonferroni correction for multiple comparisons was applied. Student's t was used to compare ages.
Stepwise logistic regression analysis was employed to develop a model for predicting suicide attempts using the hypothesized predictor variables. Forward variable selection was used, entering stepwise the variable with the lowestp value. This stepwise process continued until all variables with p values less than .10 were included. A previously entered variable was deleted if its p value increased above .15. Two-and three-way interaction terms based on hypothesized relationships were tested. This technique yielded a formula that contained only the best predictors of suicide attempt and weighted each variable according to its relative importance. Relative odds ratios and 95% confidence intervals were calculated. Sensitivity and specificity of the model were determined, with sensitivity maximized (i.e., a cutoff was chosen that correctly identified at least 97% of the suicide attempters without losing model efficiency). Maximum likelihood ratio (MLR) methods were used. Bio-Medical Data Package (BMDP) statistica l software (Dixon, Brown, Engelman, & Jeunrich, 1990) was employed for all analyses.
Exploratory Univariate Analysis
Table 1 summarizes the univariate tests (Pearson chi-square). Suicide attempters (73.7%) were significantly more likely than were controls (32.9%) to have ever experienced the threat of separation from a parental figure. A significantly greater percentage of suicide attempters, as compared with controls, reported depressed mood (81.6% vs. 36.8%), suicidal ideation (64.9% vs. 22.4%), insomnia (48.6% vs. 11.8%), psychomotor retardation (32.4% vs. 3.9%), and excessive guilt (44.4% vs. 22.4%). There were no significant differences between groups in regard to other symptoms of major depressive disorder: diminished interest in usual activities, weight gain or loss when not on a diet, fatigue, and difficulty concentrating. Among controls who were depressed, 73.7% were aged 16 to 18 and 26.3% were under age 16, while among suicide attempters it was the younger adolescents who were more likely to be depressed (62.5%) as compared with the older adolescents (37.5%). However, in a 3-way log-linear analysis, this interaction was not statistically significant.
Suicide attempters (30.6%), as compared with controls (5.3%), were significantly more likely to have a parent with a psychiatric history (excluding depression), [[chi].sup.2](1, n = 114) = 13.24, p = .0003. Suicide attempters (27%) were also significantly more likely to have had a history of neglect as compared with controls (4%), [[chi].sup.2](1, n = 114) = 12.805, p = .0003. In all cases of neglect among suicide attempters, the adolescent had been removed from the home, at least temporarily. Analysis of neglect by age and suicide attempt revealed that neglected younger (under age 16) children (n = 6) were all suicide attempters, [[chi].sup.2](1, n = 112) = 12.86, p = .0003. However, this interaction was left out of the final regression model because of the small cell sizes.
Contrary to expectations, controls (46.1%) were more likely to report threatening others than were suicide attempters (24.3%). Controls (60.5%) were also more likely to be truant than were suicide attempters (39.5%).
Although there was a large difference in reported substance abuse between suicide attempters (15.8%) and controls (1.3%), this difference was not statistically significant when the critical alpha level was corrected for the number of comparisons made. There were no significant differences between groups in regard to modeling of suicidal behavior by parents, siblings, other relatives, or friends (examined individually). There was also no difference in family abuse (physical or sexual) between controls and suicide attempters.
Logistic Regression Analysis
Stepwise logistic regression using the maximum likelihood ratio revealed the individual contribution of checklist items in determining group classification (see Table 2). Consistent with expectations, the strongest predictors of suicide attempt were insomnia, threatened separation from parental figure, alcohol/drug abuse, neglect, academic problems (poor grades), and suicidal ideation. Relative odds ratios indicated that suicide attempters, compared with controls, were 231 times more likely to have abused drugs or alcohol, over 59 times more likely to have insomnia, nearly 41 times more likely to have been neglected, more than 20 times more likely to have experienced the threat of separation from a parental figure, almost 7 times more likely to have academic problems, and over 5 times more likely to have a history of suicidal ideation.
A number of variables proved to be negative predictors. According to relative odds ratios, separation from mother for more than two weeks before the age of 12 was approximately 5 times less likely among suicide attempters, threatening others with physical harm was almost 5 times less likely among suicide attempters, and school truancy was over 51 times less likely among suicide attempters, compared with controls.
The Psychiatric Consultation Checklist (Lyon, 1987) proved to be highly sensitive and specific. With a probability cutoff of .142, 97% of suicide attempters (true positives) and 86% of nonattempters (true negatives) were correctly classified (see Figure 2).
The results clarify the nature of the acute parent-child conflict associated with a suicide attempt. Parents of adolescent suicide attempters threatened their children with separation at a rate significantly higher than that of other parents living in the same impoverished community. This finding is consistent with the hypothesis that there is a wish to be rid of the child (Sabbath, 1969). It also supports the findings of Lyon (1987), who reported that seven days prior to a suicide attempt, one-third of the adolescents had received an overt threat of separation.
Attachment theory (Bowlby, 1980) suggests that, more than parents' efforts to control their children or simple rejection, threats of separation may signal a major failure in attachment. Threats of separation may cause fear of abandonment, which may in turn precipitate a suicide attempt in an effort to secure attachment. Given that younger adolescents are jeopardized more than are older adolescents by threats of separation, it is not surprising that, in the present study, those under age 16 were more likely to be depressed and to make suicide attempts.
There appeared to be a failure in the ability of some parental figures to nurture, consistent with the high incidence of parental psychiatric illness and neglect. That all of the neglected younger adolescents (under age 16) were suicide attempters further suggests that this seemingly maladaptive behavior may have survival value for youths too young to care for themselves, in terms of engaging the parent in their care and/or in alerting the community to their needs. Further research is needed to determine whether suicidal behavior can be understood in terms of insecure attachment.
Neglect rather than physical or sexual abuse distinguished suicide attempters from nonattempters. This is consistent with an earlier study that found that the greater the degree of neglect, the greater the probability of a serious mental disorder in the parent (Gilbert, 1989). In the present study, there was a higher incidence of psychiatric problems (excluding depression) among the parents of suicide attempters when compared with nonattempters. Research to examine the hypothesis that neglect implies a failure in attachment, in a way that abuse does not, is needed.
In contrast to earlier studies (Brent et al., 1988; Joffe et al., 1988; Velez & Cohen, 1988), the present investigation found no positive association between conduct disorders and suicide attempts among adolescents. (This supports traditional psychiatric theories, which propose that suicide attempts represent depression and anger turned inward.) Rather, the high rate of threatening others with physical harm among the nonattempters (46.1%) may be adaptive to the stressful inner-city environment. This possibility is supported by a study with a similar sample of female, inner-city, minority, adolescent suicide attempters (Trautman, Rotheram-Borus, Dopkins, & Lewin, 1991); no significant differences were found between suicide attempters and disturbed nonattempters in the incidence of conduct disorders. Perhaps threatening others with harm and other conduct problems protect poor African American adolescents from acting on suicidal impulses. Alternatively, the high homicide rate among poor, inner-city, African Ame rican youths may have a suicidal component, with adolescents' provocative behavior being a manifestation of the wish that someone would kill them.
While not the focus of this research, a major finding was the extensive amount of hardship experienced by the control subjects: 18.4% had been physically or sexually abused, 4% neglected, and 32.9% threatened with parental separation. Half had experienced the loss of a significant person in their lives. These are only some of the problems facing urban African American youths, for whom poverty, violence, poor schools, and living with relatives other than parents are often the norm.
Limitations of This Study
The sample consisted of suicide attempters and primary care patients who did not have a history of suicide attempts. Two very important groups were missing: those who completed suicide and those who attempted suicide but did not seek medical care.
None of the investigated variables technically were "predictors"; rather, they were used to differentiate matched samples of African American suicide attempters and nonattempters. Greater understanding of cause and effect would require a prospective, not a cross-sectional, design.
One limitation of the Psychiatric Consultation Checklist was that sexual orientation was not assessed. Clinical reports suggest that there is a relationship between adolescent suicide attempts and conflict regarding sexual orientation. A question about sexual contact with a same-sex partner can easily be added to the checklist.
The national health objective for the year 2000 is to reduce by 15% the incidence of injurious suicide attempts among adolescents aged 14 to 17 (Public Health Service, 1991). The African American adolescent should not be forgotten in this endeavor.
Health professionals who treat urban adolescents have a responsibility to screen, in a systematic way, for victimization and other risk factors for suicide attempts. The Psychiatric Consultation Checklist (Lyon, 1987) may prove useful in this regard. In the move toward managed care and cost containment, the checklist or a similar instrument may also be helpful in triaging youths who are in need of more intensive interventions.
The themes identified in this study require further investigation from a culturally and developmentally sensitive framework. Finally, adolescents from a variety of clinics should be sampled to confirm the results.
Marilyn Benoit, M.D., Department of Psychiatry, Children's National Medical Center.
Regina M. O'Donnell, Statistical Associate, Children's National Medical Center.
Pamela R. Getson, Ph.D., Senior Statistician, Children's National Medical Center.
Tomas Silber, M.D., Professor of Pediatrics, George Washington University School of Medicine.
Thomas Walsh, M.D., Department of Psychiatry, Children's National Medical Center.
Adams-Tucker, C. (1982). Proximate effects of sexual abuse in childhood: A report on 28 children. American Journal of Psychiatry, 139, 1252-1256.
Alcohol, Drug Abuse, and Mental Health Administration. (1989). Report of the Secretary's Task Force on Youth Suicide (DHHS Publication No. ADM 89-1621-4). Washington, DC: U.S. Government Printing Office.
Asarnow, J. (1992). Suicidal ideation and attempts during middle childhood:
Associations with perceived family stress and depression among child psychiatry inpatients. Journal of Child Psychology, 21, 35-40.
Bowlby, J. (1980). Attachment and loss: Vol. 3. Loss, sadness and depression. New York: Basic Books.
Brent, D. (1987). Correlates of medical lethality of suicide attempts in children and adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 26, 87-89.
Brent, D., Perper, J., Goldstein, C., Kolko, D., Allan, M., Allman, C., & Zelenak, J. (1988). Risk factors for adolescent suicide: A comparison of adolescent suicide victims with suicidal inpatients. Archives of General Psychiatry, 45, 581-588.
Centers for Disease Control. (1998a, March 20). Suicide among Black youths--United States, 1980-1995. Morbidity and Mortality Weekly Report, 47, 106-193.
Centers for Disease Control. (1998b, August 14). Youth risk behavior surveillance--United States, 1997. Morbidity and Mortality Weekly Report, 47, 1-89.
Centers for Disease Control. (1998c, November 10). National vital statistics report (Vol. 47). Atlanta: National Center for Health Statistics, U.S. Department of Health and Human Services.
Cochran, S., & Mays, V. (1994). Depressive distress among homosexually active African American men and women. American Journal of Psychiatry, 151, 524-529.
Cohen-Sandler, R., Berman, A., & King, R. (1982). Life stress and symptomatology: Determinants of suicidal behavior in children. Journal of the American Academy of Child Psychiatry, 21, 178-186.
Diekstra, R. (1985). Suicide and suicide attempts in the European economic community: An analysis of the trends with special emphasis upon trends among the young. Suicide and Life Threatening Behavior, 15, 27-42.
Dixon, W., Brown, M., Engelman, L., & Jennrich, R. (1990). BMDP statistical software manual (Vol. 1). Berkeley, CA: University of California Press.
Eyman, J., & Smith, K. (1986, April). Lethality trends in multiple suicide attempts. Paper presented at the 19th Annual Meeting of the American Association of Suicidology, Atlanta, GA.
Garrison, C., Jackson, K., Addy, C., et al. (1991). Suicidal behaviors in young adolescents. American Journal of Epidemiology, 133, 1005-1014.
Gilbert, P. (1989). Human nature and suffering. Hillsdale, NJ: Lawrence Erlbaum Associates.
Green, A. (1978). Self destructive behavior in battered children. American Journal of Psychiatry, 135, 579-582.
Greenberg, M., & Schneider, D. (1992). Blue Thursday? Homicide and suicide among urban 15-24 year old black male Americans. Public Health Reports, 107, 264-268.
Guyer, B., MacDorman, M., Martin, J., Peters, K., & Strobino, D. (1998). Annual summary of vital statistics--1997. Pediatrics, 102, 1333-1349.
Joffe, R., Offord, D., & Boyle, M. (1988). Ontario Child Health Study: Suicidal behavior in youth aged 12-16 years. American Journal of Psychiatry, 145, 1420-1423.
Khan, A. (1987). Heterogeneity of suicidal adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 26, 92-96.
Lester, D., & Anderson, D. (1992). Depression and suicidal ideation in African-American and Hispanic-American high school students. Psychological Reports, 71, 618.
Lyon, M. (1987). Suicide attempts in adolescence. Unpublished master's thesis, The American University, Washington, DC (University Microfilms No. 1332195).
Pokorny, A. (1983). Prediction of suicide in psychiatric patients. Archives of General Psychiatry, 40, 249-257.
Public Health Service. (1991). Healthy people 2000: National promotion and disease prevention objectives--Full report, with commentary (DDHS Publication No. PHS 91-50212). Washington, DC: U.S. Government Printing Office.
Robbins, D., & Alessi, N. (1985). Depressive symptoms and suicidal behavior in adolescents. American Journal of Psychiatry, 142, 588-592.
Sabbath J. (1969). The suicidal adolescent--The expendable child. Journal of the American Academy of Child Psychiatry, 8, 272-289.
Shaffer, D., Garland, S., Gould, M., Fisher, P., & Trautman, P. (1988). Preventing teenage suicide: A critical review. Journal of the American Academy of Child and Adolescent Psychiatry, 27, 675-687.
Shafii, M., Carrigan, S., Whittinghill, J., & Derrick, A. (1985). Psychological autopsy of completed suicide in children and adolescents. American Journal of Psychiatry, 142, 1062-1064.
Summerville, M., Abbate, M., Siegel, A., et al. (1992). Psychopathology in urban female minority adolescents with suicide attempts. Journal of the American Academy of Child and Adolescent Psychiatry, 31, 663-668.
Summerville, M., Kaslow, N., Abbate, M., & Cronan, S. (1994). Psychopathology, family functioning, and cognitive style in urban adolescents with suicide attempts. Journal of Abnormal Child Psychology, 22, 221-235.
Trautman, P., Rotheram-Borus, M., Dopkins, S., & Lewin, N. (1991). Psychiatric diagnoses in minority female adolescent suicide attempters. Journal of the American Academy of Child and Adolescent Psychiatry, 30, 617-622.
Vega, W., Gil, A., & Warheit, G. (1993). The relationship of drug use to suicide ideation and attempts among African American, Hispanic, and white non-Hispanic male adolescents. Suicide and Life Threatening Behavior, 23, 110-119.
Velez, C., & Cohen, P. (1988). Suicidal behavior and ideation in a community sample of children: Maternal and youth reports. Journal of the American Academy of Child and Adolescent Psychiatry, 27, 349-356.
Univariate Analyses of Hypothesized Risk Factors Suicide Matched p value Attempters Controls (n=38) (n=76) Frequency (%) Frequency (%) Threat of Separation 28 (73.7) 25 (32.9) .0001 Depressed Mood 31 (81.6) 28 (36.8) .0001 Suicidal Ideation 24 (64.9) 17 (22.4) .0001 Insomnia 18 (48.6) 9 (11.8) .0001 Motor Retardation 12 (32.4) 3 (3.9) .0001 Parent Psych. Hist. 11 (30.6) 4 (5.3) .0003 Negleco 10 (27.0) 3 (4.0) .0003 Alcohol/Drug Abuse 6 (15.8) 1 (1.3) .0024 Academic Problems 23 (63.9) 29 (38.2) .0108 Excessive Guilt 16 (44.4) 17 (22.4) .0167 Threatens Others 9 (24.3) 35 (46.1) .0262 Truancy 15 (39.5) 46 (60.5) .0336 Abuse 13 (34.2) 15 (19.7) .0906 Separated [greater than]2 Weeks 23 (60.5) 50 (65.8) .5809 Death of Parent 8 (21.1) 11 (14.5) .3743 Loss of Sig. Other 16 (47.1) 38 (50.0) .7755 Suicidality/Parent 1 (2.6) 3 (4.1) .6918 Suicidality/Relative 1 (2.6) 3 (3.9) .7189 Suicidality/Sibling 2 (5.3) 1 (1.4) .2247 Suicidality/Friend 3 (8.1) 16 (21.1) .0842 Parental Substance Abuse 14 (37.8) 18 (23.7) .1171 Parental Depression 8 (21.6) 16 (21.1) .9447 Expulsion 0 (0.0) 3 (3.9) .2145 Suspension 11 (28.9) 34 (44.7) .1040 Runaway 9 (24.3) 14 (18.4) .4645 Diminished Interest 15 (41.7) 30 (39.5) .2005 Weight Loss/Gain 11 (29.7) 28 (36.8) .4555 Fatigue 18 (48.6) 28 (36.8) .2306 Difficulty Concentrating 15 (41.7) 30 (39.5) .8250 Shoplifting 12 (31.6) 29 (38.2) .4902 Pregnancy (females only) 9 (25.7) 22 (36.1) .2965 Note. Bonferroni correction for 32 comparisons requires p [less than] .0016. Stepwise Logistic Regression - Maximum Likelihood Ratio Risk Factor Coefficient [a] Odds 95% Confidence Ratio Interval Insomnia 4.079 59.1 5.32 - 657 Threat of Separation 3.006 20.2 2.78 - 147 Alcohol/Drug Abuse 5.440 231.0 5.04 -10500 Neglect 3.705 40.6 2.77 - 596 Academic Problems 1.931 6.9 1.20 - 39.8 Suicidal Ideation 1.693 5.44 1.03 - 28.7 Truancy -3.938 51.3 [c] 5.35 - 490.2 [d] Age -3.045 20.2 [c] 2.78 - 147.0 [d] Threatens Others -1.601 4.95 [c] 0.73 - 33.78 [d] Separated [greater than]2 weeks -1.690 5.43 [c] 0.83 - 35.71 [d] Risk Factor p value [b] Insomnia .0001 Threat of Separation .0005 Alcohol/Drug Abuse .0006 Neglect .0007 Academic Problems .0170 Suicidal Ideation .0339 Truancy .0001 Age .0018 Threatens Others .0756 Separated [greater than]2 weeks .0505
(a.)Negative coefficients indicate risk factors operated in a direction opposite to expectations.
(b.)p value (associated with chi-square statistic) for each variable, controlling for effects of all other variables in model.
(c.)Interpreted as number of times less likely to be a risk factor (i.e., suicide attempters were 20.2 times less likely to be age 16 or older).
(d.)Confidence intervals were inverted for clarity.…