The Associations Between Health Risk Behaviours and Suicidal Ideation and Attempts in a Nationally Representative Sample of Young Adolescents

Transcription

1 The Associations Between Health Risk Behaviours and Suicidal Ideation and Attempts in a Nationally Representative Sample of Young Adolescents Tracie O Afifi, MSc 1, Brian J Cox, PhD 2, Laurence Y Katz, MD 3 Objective: To examine associations between health risk behaviours and suicidal ideation and attempts in Canadian adolescents aged 12 to 13 years. Young adolescents think about and attempt suicide. However, most existing research on suicide has been conducted on individuals aged 15 years and older. Method: The present study examined a nationally representative Canadian sample of adolescents aged 12 to 13 years (n = 2090). Health risk behaviours included disruptive (shoplifting, physical fighting, damaging property, fighting with a weapon, carrying a knife, and gambling), sexual (petting below the waist and sexual intercourse), and substance use behaviours (smoking cigarettes, consuming alcohol, marijuana or hash, and glue or solvents). Unadjusted and adjusted (for all significant health risk behaviour and psychiatric symptoms) models were tested. Results: All health risk behaviours were common among male and female adolescents. In unadjusted models, almost all health risk behaviours were associated with suicidal ideation and attempts among adolescent boys. In adjusted models, only damaging property, sexual intercourse, and smoking cigarettes remained statistically associated with suicidal ideation, while smoking cigarettes and using marijuana or hash remained statistically associated with suicide attempts among adolescent boys. All health risk behaviours were statistically associated with suicidal ideation and attempts among female adolescents in unadjusted models. In adjusted models, only carrying a knife remained statistically associated with suicidal ideation, while shoplifting and gambling remained statistically associated with suicide attempts among adolescent girls. Conclusions: Health risk behaviours among young adolescents are associated with suicidal ideation and attempts among young adolescents. Recognizing health risk behaviours among young adolescents may be one means of understanding who among them is at increased risk of suicidality. (Can J Psychiatry 2007;52: ) Information on funding and support and author affiliations appears at the end of the article. Clinical Implications Early intervention, before the adolescent enters an age category when suicidal behaviour dramatically increases, is important. The findings that certain disruptive health risk behaviours among adolescent girls and substance use behaviours among adolescent boys remained significantly associated with suicide attempts in adjusted models in the current research may help to identify young adolescents at risk of self-harm. Among young adolescents, health risk behaviours are associated with suicidal ideation and attempts and, therefore, cannot be ignored. Limitations The data were cross-sectional, which indicates that causal relations cannot be determined. The health risk behaviours included in the current analysis do not represent an exhaustive list. Data were solely based on adolescent self-reports. 666

2 The Associations Between Health Risk Behaviours and Suicidal Ideation and Attempts in a Nationally Representative Sample of Young Adolescents Key Words: adolescent, suicide, health risk behaviours, Canada Adolescent suicidal ideation and attempts are signs of severe emotional distress and are identified risk factors for eventual completed suicide. 1 For individuals aged 10 to 24 years, suicide is the second leading cause of death next only to motor vehicle accidents. 2 Most existing research on suicide has been conducted on individuals aged 15 years and older, although suicide also occurs among children and young adolescents. In 1997, 51 children in Canada aged 14 years or younger committed suicide. 3 Research may focus on older adolescents because the rate of suicide increases tremendously after age 14. For example, the rate of completed youth suicide in Canada in the year 2000 was estimated to be 1.1 per among youth aged 5 to 14 years and 13 per among youth aged 15 to 24 years. 4 An older US study of youth aged 12 to 14 years found the prevalence of suicidal ideation was 8.7% in female youth and 4.0% in male youth, while the prevalence of suicide attempts was 1.5% in female youth and 1.9% in male youth. 5 A more recent study investigating suicide among young Canadian adolescents (aged 12 to 13 years) found that suicidal ideation and attempts among girls (8.4% and 5.9%, respectively) and boys (4.6% and 2.4%, respectively) were common. 6 It is reasonable to speculate that studying a group before it enters an age cohort with severely increased risk of suicide may provide insight for prevention or intervention measures. Suicidal behaviour among adolescents has been found to cooccur with other health risk behaviours. Previous research on suicidality and health risk behaviours has found that smoking cigarettes, 7,8 drug use, 7 9 gambling, 8 physical fighting, 10 carrying a weapon, 9 and fighting while carrying a weapon 11 were associated with suicidal behaviour in male and female adolescents. Although inconsistent findings do exist in the literature, sex differences have been noted in the relation between suicidality and health risk behaviours. One study found that carrying a weapon and using hallucinogenic drugs was related to suicide attempts among male adolescents but not among female adolescents. 7 Health risk behaviours that have been found to be significantly associated with suicide attempts among female adolescents and not male adolescents include drinking alcohol 8 and physical fighting. 7 However, in Abbreviations used in this article CI NLSCY OR SE confidence interval National Longitudinal Survey of Children and Youth odds ratio standard error another sample of male adolescents, alcohol use was associated with suicidal ideation and suicide attempts. 12 In addition, a longitudinal study found that aggressive behaviour among girls (physical fighting, carrying a weapon, and using a weapon) at age 14 years was associated with suicidal behaviour at age Among boys, it was substance use at age 14 years that was associated with suicidal behaviour at age Although numerous studies have been conducted on suicidality and health risk behaviours, the literature is limited for several reasons. First, many studies have used selective samples that are not representative of the general population. Therefore, findings from such studies cannot be generalized. Further, to date, no studies on health risk behaviours and suicidality have been conducted in a sample of Canadian adolescents. Second, some studies only considered a small scope of health risk behaviours, such as addictive behaviours. It is necessary to investigate a broad spectrum of health risk behaviours to gain a more comprehensive understanding of the correlates of suicidal ideation and attempts among young adolescents. Finally, many studies on health risk behaviours and suicidality have either focused on older adolescents only 7,8,10,11 or combined younger and older adolescents into one group. 9,12 The current research addresses the limitations of previous studies since it is the first investigation to study the relation between numerous health risk behaviours and suicidal ideation and attempts exclusively in a young nationally representative sample of Canadian adolescents. The purpose of the present study was to investigate several health risk behaviours and determine whether associations existed with suicidal ideations and attempts in a representative sample of young adolescents. It was hypothesized that health risk behaviours would be prevalent among male and female adolescents, that health risk behaviours would have positive associations with suicidal ideation and attempts among both male and female adolescents in unadjusted logistic regression models, and that health risk behaviours would have positive associations with suicidal ideation and attempts among both male and female adolescents after adjusting for other significant health risk behaviours, depressive symptom scores, emotional or anxiety symptom scores, and physical aggression scores. Methods Sample The dataset used for this research was Statistics Canada s NLSCY Cycle 2. The data were collected in 1996 and 1997 according to a cluster sampling research design. A parent provided informed consent for his or her child to participate in the survey. The self-administered questionnaire used for this research was confidential and was completed in the privacy of the adolescent s home while an interviewer waited. The The Canadian Journal of Psychiatry, Vol 52, No 10, October

3 parent was not permitted to view the adolescent s responses and the parent was informed of this before giving permission for the adolescent to participate in the survey. The sample used for the current research consisted of male and female adolescents aged 12 to 13 years. The total sample size was 2090 (1041 boys and 1049 girls). Appropriate statistical weights were applied to the sample. In addition, to account for the complex sampling design, SEs were estimated with the Taylor series linearization method in SUDAAN software. 14 Such a variance estimation technique is recommended for data with a complex sampling design. Measures The 12 health risk behaviours in the study that would be considered harmful, dangerous, or age inappropriate were labelled as health risk behaviours and included disruptive behaviours (shoplifting, physical fighting, damaging others property, fighting with a weapon, carrying a knife, and gambling), sexual behaviours (petting below the waist and having sexual intercourse), and substance use behaviours (smoking cigarettes, drinking alcohol, using marijuana or hash, and sniffing glue or solvents). The frequency with which all health risk behaviours occurred was recorded. Owing to small cell sizes, variables were dichotomized to indicate whether the adolescent did or did not participate in such activities. To further understand the unique impact of health risk behaviours on suicidal ideation and attempts, we used measures of psychopathology as covariates in adjusted logistic regression models. Psychiatric disorders were not assessed in the NLSCY, but measures of psychiatric symptoms were included. Statistics Canada calculated depressive symptoms scores from 12 items asking about things such as lack of appetite, having the blues, having trouble sleeping, having trouble concentrating, and so on. The depressive symptoms scale ranged from a score of 0 to 36, with higher scores corresponding to greater depressive symptoms. Emotional or anxiety symptoms scores were calculated from 8 items asking the respondents about emotions such as feeling fearful, anxious, nervous, worried, and so on. The emotional or anxiety symptoms scale ranged from a score of 0 to 16, with higher scores corresponding to higher levels of emotional problems or anxiety symptoms. Physical aggression scores were calculated from 6 items asking about the respondent s tendency to, for example, react with anger and fighting, physically attack people, and bully other children. The physical aggression scale ranged from a score of 0 to 12, with higher scores corresponding to higher levels of physical aggression. The continuous-level depressive symptoms, emotional or anxiety symptoms, and physical aggression variables were used as covariates in the adjusted logistic regression models. Suicidality was measured with 2 variables indicating ideation and attempts. Participants were asked, In the past 12 months have you seriously considered trying to kill yourself? and In the past 12 months have you tried to kill yourself? The response for the suicidal ideation variable was either yes or no. Responses for the variable measuring attempts included never or none, 1 time, 2 or 3 times, 4 or 5 times, and 6 or more times. Owing to small cell sizes, it was not possible to conduct analyses to compare repeat attempters with one-time attempters. Therefore, this variable was dichotomized to indicate the presence or absence of the suicide attempts. Statistical Analysis The prevalence of all health risk and suicidal behaviour was calculated among male and female adolescents. All reported cell counts (n) in the analysis were based on unweighted data, while the prevalence, SE, ORs, and 95%CIs were based on weighted analyses carried out with SUDAAN software. For the logistic regression analysis, unadjusted logistic regression models were used to determine the associations between each health risk behaviour and suicidal ideation and attempts among male and female adolescents. Next, to understand the unique impact of each health risk behaviour, models were rerun, adjusting for all significant health risk behaviours, depressive symptoms scores, emotional or anxiety symptom scores, and physical aggression scores. Results The prevalences of health risk behaviours for male and female adolescents are presented in Table 1. Suicidal ideation, attempts, and health risk behaviours were common among young male and female adolescents. The results from the logistic regression analyses that examined the relation between health risk behaviours and suicidality among male and female adolescents are presented in Table 2 and Table 3, respectively. The 95%CIs for some ORs were wide, reflecting low-powered models that indicated low-prevalence relations between some health risk behaviours and suicidality. In unadjusted models, all health risk behaviours were positively associated with suicidal ideation among male adolescents, with the exception of shoplifting, petting below the waist, and drinking alcohol. The health risk behaviour with the strongest association with suicidal ideation among male adolescents was sexual intercourse. In the model adjusting for all significant health risk behaviours, depressive symptoms, emotional or anxiety symptoms, and physical aggression scores, only damaging property, sexual intercourse, and smoking cigarettes remained significantly associated with suicidal ideation among male adolescents. Additionally, in unadjusted models, all health risk behaviours among male adolescents were positively associated with suicide attempts, with the exception of sniffing glue or solvents. 668

4 The Associations Between Health Risk Behaviours and Suicidal Ideation and Attempts in a Nationally Representative Sample of Young Adolescents Table 1 Prevalence of suicidal ideation and attempts and health risk behaviours among male and female adolescents Boys (n = 1041) Girls (n = 1049) Behaviours n % (SE) n % (SE) Health risk Ideation (1.01) (1.20) Attempts (0.74) (1.02) Shoplifting (1.89) (1.58) Physical fight (2.47) (1.97) Damaged property (1.75) (1.35) Fought with a weapon (1.08) (0.79) Carried a knife (1.42) (0.87) Gambling (2.45) (2.03) Petting below the waist (2.15) (1.61) intercourse (1.23) (0.72) Cigarettes (1.20) (1.37) Alcohol (2.38) (2.11) Marijuana or hash (1.36) (1.09) Glue or solvents (0.83) (0.82) Note: n are based on unweighted data, while % and SE are based on weighted data. The health risk behaviours with the strongest associations with suicide attempts among male adolescents included sexual intercourse, smoking cigarettes, fighting with a weapon, and using marijuana or hash. In the model adjusting for all covariates, only smoking cigarettes and using marijuana or hash remained significantly associated with suicide attempts among male adolescents. In unadjusted models for female adolescents, all 12 health risk behaviours were positively associated with suicidal ideation and attempts. The health risk behaviours with the strongest associations included fighting with a weapon, carrying a knife, damaging property, and sexual intercourse. In the model adjusting for all covariates, only carrying a knife remained significantly associated with suicidal ideation among female adolescents. In the adjusted suicide attempt model, only shoplifting and gambling remained statistically significant among female adolescents. Discussion The current study is the first of its kind to examine self-reported health risk behaviours and the odds of suicidal ideation and attempts among a young, nationally representative cohort of Canadian adolescents. The results supported the research hypotheses that stated health risk behaviours would be prevalent among male and female adolescents, health risk behaviours would have positive associations with suicidal ideation and attempts among both male and female adolescents in unadjusted logistic regression models, and health risk behaviours would have positive associations with suicidal ideation and attempts among both male and female adolescents after adjusting for other significant health risk behaviours, depressive symptom scores, emotional or anxiety symptom scores, and physical aggression scores. Results from the study are also consistent with past research, which found a relation between several health risk behaviours and suicidality among adolescents. 7 9,11 13 In the unadjusted models, 8 health risk behaviours were robustly associated with suicidal ideation and attempts for both male and female adolescents. These health risk behaviours included being in physical fights, damaging property, fighting with a weapon, carrying a knife, gambling, sexual intercourse, smoking cigarettes, and using marijuana or hash. The Canadian Journal of Psychiatry, Vol 52, No 10, October

5 Table 2 The association of health risk behaviours and suicidal ideation and attempts among adolescent males (n = 1041) Behaviours OR (95%CI) Adjusted OR (95%CI) Ideation Shoplifting 2.43 ( ) na Physical fight 5.33 ( ) a 2.05 ( ) Damaging property 5.31 ( ) a 6.10 ( ) a Fighting with a weapon 6.20 ( ) a 0.75 ( ) Carrying a knife 4.02 ( ) a 0.41 ( ) Gambling 3.13 ( ) a 0.95 ( ) Petting below the waist 2.34 ( ) na intercourse ( ) a 6.59 ( ) a Attempts Cigarettes 6.42 ( ) a 3.88 ( ) a Alcohol 2.49 ( ) na Marijuana or hash 4.35 ( ) a 1.45 ( ) Glue or solvents 5.10 ( ) a 2.60 ( ) Shoplifting ( ) a 1.53 ( ) Physical fight 4.55 ( ) a 1.32 ( ) Damaging property 8.58 ( ) a 1.82 ( ) Fighting with a weapon ( ) a 1.12 ( ) Carrying a knife 4.90 ( ) a 0.16 ( ) Gambling 3.17 ( ) a 1.95 ( ) Petting below the waist ( ) a 5.18 ( ) intercourse ( ) a 1.04 ( ) Cigarettes ( ) a 6.08 ( ) a Alcohol 5.21 ( ) a 1.43 ( ) Marijuana or hash ( ) a 7.54 ( ) a Glue or solvents 3.48 ( ) na ORs are not adjusted for any covariates and a separate logistic regression model has been conducted for each health risk behaviour. Adjusted ORs are adjusted for all significant health risk behaviours, depressive symptom scores, emotional and anxious symptom scores, and physical aggression scores entered simultaneously. a Significant ORs na = not applicable 670

6 The Associations Between Health Risk Behaviours and Suicidal Ideation and Attempts in a Nationally Representative Sample of Young Adolescents Table 3 The association of health risk behaviours and suicidal ideation and attempts among adolescent girls (n = 1049) Behaviours OR (95%CI) Adjusted OR (95%CI) Ideation Shoplifting 4.58 ( ) a 2.19 ( ) Physical fight 3.45 ( ) a 1.24 ( ) Damaging property 8.11 ( ) a 2.40 ( ) Fighting with a weapon ( ) a 1.79 ( ) Carrying a knife ( ) a 4.27 ( ) a Gambling 2.27 ( ) a 2.16 ( ) Petting below the waist 4.63 ( ) a 1.76 ( ) intercourse 6.82 ( ) a 2.13 ( ) Attempts Cigarettes 5.47 ( ) a 1.27 ( ) Alcohol 3.30 ( ) a 1.01 ( ) Marijuana or hash 4.77 ( ) a 0.66 ( ) Glue or solvents 4.30 ( ) a 0.38 ( ) Shoplifting 7.66 ( ) a 5.53 ( ) a Physical fight 3.62 ( ) a 0.93 ( ) Damaging property 9.19 ( ) a 1.67 ( ) Fighting with a weapon ( ) a 3.77 ( ) Carrying a knife ( ) a 3.94 ( ) Gambling 3.00 ( ) a 4.79 ( ) a Petting below the waist 4.89 ( ) a 1.66 ( ) intercourse 8.71 ( ) a 3.59 ( ) Cigarettes 6.34 ( ) a 1.23 ( ) Alcohol 3.16 ( ) a 1.02 ( ) Marijuana or hash 5.50 ( ) a 0.75 ( ) Glue or solvents 5.82 ( ) a 0.81 ( ) ORs are not adjusted for any covariates and a separate logistic regression model has been conducted for each health risk behaviour. Adjusted ORs are adjusted for all significant health risk behaviours, depressive symptom scores, emotional and anxious symptom scores, and physical aggression scores entered simultaneously. a Significant ORs The Canadian Journal of Psychiatry, Vol 52, No 10, October

7 These robust findings may help to identify health risk behaviours that are associated with at-risk self-harm behaviour for both male and female adolescents. More specifically, the identification of such behaviours among boys and girls may help to identify individuals at heightened risk of suicidal ideation and attempts and could provide insight for targeting at-risk individuals for intervention purposes. Further, because all physical aggression health risk behaviours were robustly associated with suicidal ideation and attempts among male and female adolescents in all unadjusted models, interventions that target these behaviours may be useful. Although several health risk behaviours were consistently associated with suicidal ideation and attempts among male and female adolescents, sex differences were noted in unadjusted models. First, sexual intercourse had the strongest association with suicidal ideation among male adolescents. Although this relation was also significant among female adolescents, the strength of the association was not as great, nor was it found to be the health risk behaviour with the largest effect size. The health risk behaviours with the largest effect size among girls were the disruptive behaviours, including damaging property, fighting with a weapon, and carrying a knife. These health risk behaviours were also significantly associated with suicidal ideation and attempts among boys, but with the exception of fighting with a weapon, these disruptive behaviours did not carry the largest odds of suicidality among boys. It seems that even though adolescent girls may not engage in disruptive health risk behaviours at rates as high as adolescent boys, when they did exhibit these disruptive health risk behaviours the association with suicidal ideation and attempts was substantial. Conducting the current analysis separately for male and female adolescents furthers our understanding of sex differences that exist concerning health risk behaviours and suicidality and highlights the fact that the relation between certain health risk behaviours and suicidal ideation and attempts might function differently for young male and female adolescents. It is important to recognize the similarities and differences between male and female adolescents when considering strategies for intervention. Sex differences were especially striking in the models adjusting for all significant health risk behaviours, depressive symptoms scores, emotional or anxiety symptom scores, and physical aggression scores. Even after the covariates in the adjusted model were accounted for, damaging property, sexual intercourse, and smoking cigarettes remained significantly associated with suicidal ideation among males. However, carrying a knife was the only health risk behaviour associated with suicidal ideation among female adolescents in the adjusted model. Differences were also noted in adjusted models investigating suicide attempts. In adjusted models, only substance use health risk behaviours were associated with suicide attempts among male adolescents, namely, smoking cigarettes and using marijuana or hash. For female adolescents, only disruptive health risk behaviours, which included shoplifting and gambling, were associated with suicide attempts. A relation between at-risk gambling and suicide behaviour has also been found in previous research using a sample of young adults, with significant findings among young female adults only. 15 Explaining why sex differences were found is beyond the scope of the current analysis. However, the sex differences do suggest that certain behaviours may alert us to health risk behaviours that are associated with greater odds of suicidal ideation or attempts for male and female adolescents. Findings from the current analysis indicate that disruptive health risk behaviours had a more robust association with suicidality among adolescent girls, while in adjusted models health risk behaviours from all 3 categories remained significantly related to suicidal ideation and attempts among adolescent boys. Many of the relations between health risk behaviours and suicidality were no longer significant after adjusting for all other health risk behaviours and measures of psychiatric symptoms. The nonsignificant findings indicate that the covariates in the model account for the variance and help to explain the relations between health risk behaviours and suicidality among young male and female adolescents. More specifically, it may be useful to consider cooccurring health risk behaviours as well as depressive, emotional or anxiety, and physical aggression symptoms when trying to understand the association between health risk behaviours and suicidal ideation and attempts. It is necessary to consider the possible mechanisms that might underlie the observed associations. Many health risk behaviours are characteristics of unsafe and sometimes violent actions. Individuals may engage in health risk behaviours as a means of acting out against others. Suicide may be similar because it is also an unsafe and often violent act and could be considered violence against oneself. In addition, impulsivity in an adolescent may play a role in the occurrence of both health risk behaviours and suicide attempts. Previous research using a clinical sample of adolescent psychiatric inpatients found that positive correlations existed between suicidal behaviour, aggressive behaviour, and impulsivity. 16 This same research also found that the relation between suicidal behaviour and impulsivity remained statistically significant even after controlling for the impact of aggressive behaviour. It is possible that impulsivity may contribute to the occurrence of both health risk behaviours and suicidality among young adolescents. One theory that ties these ideas together is that these behaviours occur either as a result of emotional dysregulation or as a means of escaping emotional dysregulation. The common thread in this theory is the 672

8 The Associations Between Health Risk Behaviours and Suicidal Ideation and Attempts in a Nationally Representative Sample of Young Adolescents presence of emotional dysregulation as a determinant of the problem behaviours. Impulsivity then could be a product of emotional dysregulation or a cofactor that transacts with emotional dysregulation to produce the problem behaviours, including suicidal behaviour. Future research is needed to further investigate the potential role of impulsivity and to understand other possible mechanisms that underlie the observed associations between health risk behaviours and suicidal behaviour. The results from the current study should be considered in light of several limitations. First, the study was based on a cross-sectional sample, which indicates that causal relations cannot be determined. Therefore, it is incorrect to interpret the findings as health risk behaviours causing suicidal ideation or attempts. It can only be stated that there are significant associations between disruptive, sexual, and substance use behaviours and suicidal ideation and attempts. Second, data in the current analysis were based solely on adolescent self-reports and were not verified by using a secondary source. However, adolescents were assured their responses would remain private and that parents would not be allowed to review their questionnaire. The confidential nature of the survey may have helped the collection of accurate information, such as reporting suicidal ideation and attempts, that might not hav been known to the parents. Third, although the current research investigated numerous health risk behaviours, it should be stated that the health risk behaviours in the current analysis do not represent an exhaustive list. For example, in a study of older adolescents, an association between vomiting to lose weight (a health risk behaviour not included in the NLSCY) and suicide attempts was found. 7 It is important to look at additional health risk behaviours in future research. Another limitation of the NLSCY data was the inability to compare one-time suicide attempters with repeat attempters. It is possible that a difference exists between individuals who have attempted suicide one time and multiple attempters. However, owing to small cell counts, it was not possible to investigate one-time attempters and multiple attempters separately in the current analysis. As well, information was not available on the severity of the suicide attempts, which should also be considered a limitation of the data. Finally, a limitation of the current research was the inability to control for socioeconomic status and ethnicity. Such factors can provide important information for understanding suicide. However, Statistics Canada adheres to strict regulations to ensure that confidentially and privacy are provided for all survey respondents. Therefore, owing to the protection of the respondents identity, we could not include such information in this analysis. Conclusions The present research has important implications for prevention and intervention through the identification of disruptive, sexual, and substance use behaviours that have significant positive associations with suicidal ideation and attempts among young Canadian adolescents. Further research is needed to extend the current findings to identify other indicators that may increase the odds of suicidal thoughts and attempts in young cohorts. The current findings contribute to the emerging literature on the identification of the correlates of suicidal behaviour. Health risk behaviours among young adolescents are associated with suicidal ideation and attempts among young adolescents and cannot be ignored. Support and Funding This research was supported by funds awarded to Ms Afifi by The Social Sciences and Humanities Research Council of Canada, Canada Graduate Scholarship, and funds awarded to Dr Cox by the Canada Research Chairs Program. Acknowledgements Statistics Canada provided the data, but the analyses are the sole responsibility of the authors. The opinions expressed do not represent the views of Statistics Canada. References 1. Harriss L, Hawton K, Zahl D. Value of measuring suicidal intent in the assessment of people attending hospital following self-poisoning or self-injury. Br J Psychiatry. 2005;186: Statistics Canada. Leading cause of death at different ages, Canada, Shelf tables. Ottawa (ON): Minister of Industry; Report No: Catalogue XPB. 3. Statistics Canada. Suicides and suicide rates by sex and age group [Internet]. Ottawa (ON): Statistics Canada; 2005 [cited 2006 Nov 4]. Available from: 4. Bridge JA, Goldstein TR, Brent DA. Adolescent suicide and suicidal behavior. J Child Psychol Psychiatry. 2006;47: Garrison CZ, Jackson KL, Addy CL, et al. Suicidal behaviors in young adolescents. Am J Epidemiol. 1991;133: Fotti S, Katz L, Afifi TO, et al. The Associations between peer and parental relationships and suicidal behaviours in early adolescents. Can J Psychiatry. 2006;51: Bae S, Ye R, Chen S, et al. Risky behaviors and factors associated with suicide attempt in adolescents. Arch Suicide Res. 2005;9: Garnefski N, Jan De Wilde J. Addiction-risk behaviours and suicide attempts in adolescents. J Adolesc. 1998;21: Anteghini M, Fonseca H, Ireland M, et al. Health risk behaviors and associated risk and protective factors among Brazilian adolescents in Santos, Brazil. J Adolesc Health. 2001;28: Sosin DM, Koepsell TD, Rivara FP, et al. Fighting as a marker for multiple problem behaviors in adolescents. J Adolesc Health. 1995;16: Orpinas PK, Basen-Engquist K, Grunbaum J, et al. The co-morbidity of violence-related behaviors with health-risk behaviours in a population of high school students. J Adolesc Health. 1995;16: Vermeiren R, Schwab-Stone M, Ruchkin VV, et al. Suicidal behavior and violence in male adolescents: a school-based study. J Am Acad Child Adolesc Psychiatry. 2003;42: O Donnell L, Stueve A, Wilson-Simmons R. Aggressive behaviors in early adolescence and subsequent suicidality among urban youths. J Adolesc Health. 2005;37: Shah BV, Barnswell BG, Bieler GS. SUDAAN user s manual: software for analysis of correlated data. Release 6.40 ed. Research Triangle Park (NC): Research Triangle Institute; Feigelman W, Gorman BS, Lesieur H. Examining the relationship between at risk gambling and suicidality in a national representative sample of young adults. Suicide Life Threat Behav. 2006;36: The Canadian Journal of Psychiatry, Vol 52, No 10, October

9 16. Horesh N, Gothelf D, Ofek H, et al. Impulsivity as a correlate of suicidal behavior in adolescent psychiatric inpatients. Crisis. 1999;20:8 14. Manuscript received November 2006, revised, and accepted March PhD Candidate, Department of Community Health Sciences, University of Manitoba, Winnipeg, Manitoba; PhD Candidate, Department of Psychiatry, University of Manitoba, Winnipeg, Manitoba. 2 Professor, Department of Community Health Sciences, University of Manitoba, Winnipeg, Manitoba; Professor, Department of Psychiatry, University of Manitoba, Winnipeg, Manitoba. 3 Associate Professor, Department of Psychiatry, University of Manitoba, Winnipeg, Manitoba. Address for correspondence: Dr LY Katz, PZ 162 PsycHealth Centre, 771 Bannatyne Avenue, Winnipeg, MB R3E 3N4; LKatz@exchange.hsc.mb.ca Résumé : Les associations entre les comportements à risque pour la santé et l idéation et les tentatives de suicide dans un échantillon nationalement représentatif de jeunes adolescents Objectif : Les jeunes adolescents pensent au suicide et tentent de se suicider. Cependant, la majorité de la recherche sur le suicide a été menée auprès de jeunes de 15 ans et plus. Méthode : La présente étude a examiné les associations entre les comportements à risque pour la santé et l idéation et les tentatives de suicide dans un échantillon canadien nationalement représentatif d adolescents de 12 à 13 ans (n = 2 090). Les comportements à risque pour la santé comprenaient les comportements perturbateurs (vol à l étalage, bataille physique, dommages à la propriété, combat avec une arme, port d un couteau, et jeu de hasard), sexuels (caresses sous la taille et rapports sexuels), et utilisation de substances (tabagisme, consommation d alcool, de marijuana ou hasch, et de colles ou solvants). Des modèles non corrigés et corrigés (pour tous les comportements à risque significatif pour la santé et les symptômes psychiatriques) ont été testés. Résultats : Tous les comportements à risque pour la santé étaient communs chez les garçons et les filles. Dans les modèles non corrigés, presque tous les comportements à risque pour la santé étaient associés à des idées et des tentatives de suicide chez les garçons. Dans les modèles corrigés, seulement les dommages à la propriété, les rapports sexuels et le tabagisme demeuraient statistiquement associés à l idéation suicidaire, tandis que le tabagisme et la consommation de marijuana ou hasch demeuraient statistiquement associés aux tentatives de suicide chez les garçons. Tous les comportements à risque pour la santé étaient associés à l idéation et aux tentatives de suicide chez les filles, dans les modèles non corrigés. Dans les modèles corrigés, seul le port d un couteau demeurait statistiquement associé à l idéation suicidaire, tandis que le vol à l étalage et le jeu de hasard demeuraient statistiquement associés aux tentatives de suicide chez les filles. Conclusions : Les comportements à risque pour la santé chez les jeunes adolescents sont associés à l idéation et aux tentatives de suicide chez cette population. Reconnaître les comportements à risque pour la santé chez les jeunes adolescents peut être un moyen de distinguer qui parmi eux est à risque accru de suicidabilité. 674