Oral presentation at the 8th annual Canadian Conference on HIV/AIDS Research, in May 1999, in Victoria, BC.

Social Determinants of Suicide Attempts in a Cohort of 
Young Men Who Have Sex With Men

Michael R. Botnick, Katherine V. Heath, Peter G.A. Cornelisse, Steffanie A. Strathdee, Stephen L. Martindale, Martin T. Schechter, Michael V. O'Shaughnessy, Robert S. Hogg.

British Columbia Centre for Excellence in HIV/AIDS, University of British Columbia, St. Paul's Hospital, Vancouver, Canada.

Objective: To investigate socio-demographic, psychosocial and behavioural characteristics which may be readily assessed in an attempt to identify those characteristics associated with increased risk of attempted suicide among young MSM.

Methods: MSM between the ages of 18 and 30 completed questionnaires detailing demographic information, sexual behaviours, history of forced and paid sex, comfort with sexual orientation, use of illicit drugs, and previously validated measures of depression, social support, alcohol dependency and self-esteem. Participants were also asked whether they had ever attempted to commit suicide. Comparisons between those who had and had not attempted suicide were carried out using contingency table analysis. Step-wise logistic regression was used to identify predictors of suicide attempt after adjustment for multiple explanatory variables.

Results: Of the 345 participants, 19.4% had attempted suicide. Those who had attempted suicide were significantly less likely to have completed high school (p=0.006) and more likely to have annual incomes below $10,000 CA (p=0.002). As regards social-sexual characteristics, forced sex in childhood and adolescence and having been paid for sex were more commonly reported by attempters than non-attempters (all p<0.01). Suicide attempters had higher depression scores and lower levels of social support and self esteem (all p<0.01). The use of nitrite inhalants (poppers) and alcohol addiction were also elevated among suicide attempters (p=0.002 and p=0.001 respectively). In multivariate analysis the use of poppers (AOR=2.37; 95%CI 1.30,4.33), low social support (AOR=2.19; 95%CI 1.18,4.09) and low/moderate self-esteem (AOR=3.73; 95%CI 2.03,6.86), were associated with elevated risk of attempted suicide.

Conclusion: Previously validated tools used to measure self esteem and social support may of value for health care providers and counsellors in identifying young gay and bisexual men at elevated risk for attempted suicide.

Slide 1:
TITLE & ACKNOWLEDGEMENTS

Before I begin, I would like to thank the many people who were co-investigators on this project. I especially would like to thank Kate Heath.

I am pleased to have the opportunity to come before you and speak briefly on some of the social determinants of suicide attempts in a cohort of young men.

In and age where schoolyard violence claims dozens each year, gang related activities claim hundreds, and AIDS claims thousands, it is especially tragic to note that death by one's own hand claims any at all.

Slide 2:
OBJECTIVES

• To compare socio-demographic, psychosocial and behavioural characteristics of young gay and bisexual men who have attempted suicide with those who have not
• To identify factors that may be associated with elevated risk of suicide attempts
  

Last summer, we set out to investigate some sociodemographic, psychosocial, and behavioral characteristics which may be readily accessed and assessed by professionals who work with youth in order to assist in identifying those young people who are at an increased risk of attempted suicide. Fortunately, the Vanguard database contains a wealth of information that permitted us to tap into a number of potential determinants.

Slide 3:
OTHER STUDIES

• Berman & Jobes 1995
• Lipshitz 1995
• Health Canada 1994
• Gibson 1989
• Garafalo et al 1998
• Bagley & Tremblay 1996
• D'Augelli & Hershberger 1993
• Ensel & Woelfel 1986
• Hammelman 1993
• Jay & Young 1979
• Hickson et al 1994
• Remafedi et al 1991
• Rosenberg 1989,1965
• Strathdee et al (in press)
• . . . and many others
  

We became intrigued by this social problem after reading a number of other studies that led us to question the degree to which young gay and bisexual men and men who have sex with men exhibit suicide rates higher than the general population. The estimates range from a high of 10 times more risk according to Bagley and Tremblay, to a low of 2 to 3 times more risk as reported by Gibson, as well as DuRant & Garofalo.

Among adolescent MSM, the lifetime suicide attempt rate appears to have tripled over the last 40 years, from 9.6 percent to 31.3 percent. According to Berman and Jobes, Lipschitz, and Health Canada, this figure is in comparison with 3.2 percent of heterosexual men. Clearly, there is a tenfold problem.

Slide 4:
ELIGIBILITY CRITERIA

• 18 to 30 years old at baseline
• Living in Greater Vancouver area
• Gay, bisexual, and/or MSM
• No previous HIV-positive test result

As with all studies emanating from the Vanguard database, the respondents must be between 18 and 30 years old at baseline, live in the greater Vancouver area, self identify as gay, bisexual or MSM, and at the time of sign-up have not received any previous HIV-positive test results.

Slide 5:
METHODS

• Psychosocial and demographic data were provided at baseline and at follow-up
• Comparisons were carried out using contingency table analysis
• Continuous scales were dichotomised above and below the 75th percentile
  

We followed traditional, conservative statistical methods. For the sake of convenience, since our sample base was uncomfortably slim, we dichotomized continuous scales above and below the 75th percentile.

Slide 6:
EVER ATTEMPTED SUICIDE IN RELATION TO DEMOGRAPHICS (at baseline)

	YES (n=67) n (%)	NO (n=278) n (%)	p-value
Caucasian	4 (81)	236 (85)	0.389
Median Age (IQR)	26 (23-28)	27 (24-29)	0.061
Completed High School	35 (52)	194 (70)	0.006
Annual Income <$10,000	24 (36)	49 (18)	0.002
Canadian Born	58 (88)	226 (82)	0.265

Of the 345 participants we studied, 43.5 percent reported that they had considered suicide and 19.4 percent reported that they had attempted suicide at least once. As you can see from this table, there is no significant difference between attempters ­ labeled "yes" -- and non-attempters ­ labeled "no" -- with respect to age, ethnicity or country of birth. However those who attempted suicide were significantly less likely to have completed high school, and were twice as likely to have a low annual income.

Slide 7:
EVER ATTEMPTED SUICIDE IN RELATION TO SOCIAL/SEXUAL FACTORS

	YES (n=67) n (%)		NO (n=278) n (%)	p-value
Anal Insertive	44 (76)		202 (89)	0.013
Anal Receptive	52 (95)		197 (86)	0.096
Ever paid for sex	21 (31)		40 (15)	0.001
Forced/coerced sex:
< 12 years old	18 (27)		38 (14)	0.009
12-17 years old	16 (24)		21 (8)	0.001
> 18 years old	14 (21)		40 (14)	0.188

Curiously, while the majority of participants engaged in anal insertive and receptive intercourse, those who had attempted suicide were less likely to report having had anal insertive intercourse. Of greater consequence is the fact that a large proportion of participants reported at least one episode of non-consensual sex while they were under the age of 18. Additionally the men who had attempted suicide were twice as likely to have been paid for sex.

Slide 8:
EVER ATTEMPTED SUICIDE IN RELATION TO PSYCHO-SOCIAL VARIABLES

	YES (n=67) n (%)	NO (n=278) n (%)	p-value
CES-D median (IQR)	14 (11-20)	12 (10-14)	< 0.001
IES median (IQR)	61 (53-73)	52 (44-61)	< 0.001
Low Rosenberg self-esteem	26 (39)	197 (72)	0.001
Diagnosed mood disorder	24 (36)	48 (18)	0.001
Use poppers	29 (44)	68 (25)	0.002
C.A.G.E.
Low	28 (42)	163 (59)	0.001
Borderline	8 (12)	48 (18)
Alcoholic	31 (46)	64 (23)
Detox Centre	6 (10)	1 (0.4)	0.001
Residential rehabilitation	4 (7)	2 (0.8)	0.003
Psychiatric ward	11 (18)	5 (2)	0.001

This is an abbreviated table. We have deleted the psychosocial variables whose p-value was greater than 0.005. As you can see, the attempters had higher depression scores, lower levels of social support, and what appears to be a misnomer, (and admittedly I labeled it poorly) had significantly lower self-esteem. The third item down refers to low scores on the self-esteem test, which translates into high self-esteem. Not surprisingly, suicide attempters were twice as likely to have been diagnosed with a mood disorder.

The use of alcohol and poppers is also a notable variable.

Slide 9:
INDEPENDENT PREDICTORS OF ATTEMPTED SUICIDE

	Adjusted Odds Ratio	95% CI
Low social support	2.19	1.18-4.09
Low/moderate self-esteem	3.73	2.03-6.86
Use of poppers*	2.37	1.30-4.33
* In previous year

This table shows the final multivariate logistic regression model. After adjustment for multiple explanatory variables, the use of poppers, social support scores below the 75th percentile, and low or moderate self-esteem remained independently associated with elevated risk of attempted suicide.

Slide 10:
SUMMARY & CONCLUSIONS

• Our study confirms an alarmingly high rate of suicide attempts in young gay and bisexual men
• Characteristics of many young suicide attempters include the use of popper (inhalants) and/or alcohol, low social support and self-esteem, mood disorders and rehabilitation attempts, and relative poverty

One of the difficulties in comparing our data with that of others, is the inconsistency among sample bases. Some studies examined gay and lesbian high school students, some studied the general population, and extrapolated from there.

We can however generalize, in order to build a profile of the social problem. Garofalo et al (1998) found that more than one-third of all lesbian and gay high school students had made a suicide attempt in the 12 months prior to his study. We confirm that in our study, 44 percent had considered suicide, and 19 percent had attempted suicide before the age of 30. One third of our total sample did not complete high school. This figure is in agreement with other studies that suggest that a large proportion of homosexual youth drop out of secondary school due to discomfort and fear. In our cohort, those who attempted suicide also tended to be poorer than non-attempters, which is compatible with lower education levels and life chances.

As earlier noted, a large proportion of our suicide attempters reported non-consensual sex. Garofalo (1998) and Hickson et al (1994) both confirm this phenomenon -- 27.6 percent of the Hickson study reported non-consensual sex.

Substance use among gay, bisexual and MSM youth is very high -- 59 percent used alcohol before the age of 13. Alcohol and drug use appear to be more common among the men in our study who had attempted suicide, with 30 percent of attempters having use poppers (which is a cheap high) in the previous year and 43 percent having, or having had alcohol-related problems.

While our study did not directly examine high school and coming out experiences, many of the studies we referenced suggest that there is a very strong correlation between knowing one is gay at an early age, resisting homophobia and heterosexism, and the development of risk behaviors.

Slide 11:
IMPLICATIONS

• Gay positive recognition is the means of access into the psychological morass that is created by homophobia and heteronormativity
• In this manner, perhaps the next generation of gay and bisexual youth will not have to face the added stressors of 'coming out', 'being out' and 'being queer'

Gay and bisexual youth and adolescents live in a confused and hostile environment. Youth confronting the universal stresses of physical, emotional and sexual maturation, within the context of isolation and social rejection occasioned by non-normative sexuality face additional hazards. The complexities of putative sexual deviance and societal disapproval, and real or perceived social rejection can be emotionally damaging and may result in curtailing education, low self-esteem, depression and substance abuse which may in turn lead to an increased risk of suicide.

Our findings may be of value for social workers, school counselors, therapists, street workers and others who interact with and engaged in counseling gay and bisexual youth in an advisory or preventative context, and could be used to identify young men at elevated risk for self harm.

Slide 12:
LIMITATIONS

• Self-defined notion of suicide attempt
• Self-reported behaviours
• Voluntary nature of sample base
• Age restriction (18-30 years at baseline)
• Lack of data from non-returning participants

Thank you very much for your attention. I would be pleased to answer any questions or entertain any comments you may have.