Sexual Orientation and Risk Factors for Suicidal Ideation and Suicide Attempts Among Adolescents and Young Adults

Abstract

Same-gender sexual orientation has been repeatedly shown to exert an independent influence on suicidal ideation and suicide attempts, suggesting that risk factors and markers may differ in relative importance between lesbian, gay, and bisexual individuals and others. Analyses of recent data from the National Longitudinal Study of Adolescent Health revealed that lesbian, gay, and bisexual respondents reported higher rates of suicidal ideation and suicide attempts than did heterosexual respondents and that drug use and depression were associated with adverse outcomes among heterosexual respondents but not among lesbian, gay, and bisexual respondents.

Compared with their heterosexual peers, lesbian, gay, and bisexual (LGB) youths report elevated rates of suicidal ideation and attempted suicide.1^–6 Same-gender orientation is independently associated with suicidal behavior,7^–11 raising the possibility that risk markers for suicidal behavior may differ in their relative importance between LGB and non-LGB individuals because of interaction effects between these markers and sexual orientation. If this is true, interventions that generically target risk markers without regard to sexual orientation may be differentially effective12 and therefore do little to address the underlying disparity. We explored potential differences between groups by examining the statistical interaction between LGB status and putative markers of risk in different subgroups.13

METHODS

In an attempt to identify potential interactions between sexual orientation status and risk factors for suicide ideation or suicide attempts among adolescents and young adults, we analyzed data from the National Longitudinal Study of Adolescent Health, a prospective cohort study that followed a nationally representative sample of adolescents into young adulthood. The study’s sampling design is described elsewhere.14^,15 We analyzed wave 3 data, which contained a weighted sample of 14 322 respondents, aged 18 to 26 years, who were surveyed during August 2001 through April 2002.

Sexual orientation was measured with 1 forced-choice self-description. Consistent with (problematic) convention,16^,17 those who described themselves as bisexual, mostly homosexual, or exclusively homosexual were coded as LGB. Respondents who identified as exclusively or mostly heterosexual, or not sexually attracted to either males or females, were coded as non-LGB. Demographic variables included age, gender, race, ethnicity, and residence status.

Suicide attempts and suicidal ideation within the past 12 months were each measured with 1 item,18^,19 which was dichotomized. Problem drinking and problem drug use in the preceding 12 months were coded as present if the respondents answered yes to any of the 6 items in each domain. Depression was measured in the adolescent health survey with 9 items from the Center for Epidemiologic Studies Depression Scale.20 A factor analysis revealed 1 robust factor containing 7 items. Responses were summed across these 7 items.

Statistical analyses were performed with Stata version 8 (Stata Corp, College Station, Tex). Race, gender, and age were covaried in all multivariate analyses. After conducting descriptive analyses, we conducted regressions to examine the relationship between predictors (depression, problem alcohol use, and problem drug use) and outcomes (suicidal ideation and suicide attempts) in the 2 strata (LGB and non-LGB). Next, we used interaction models to test whether the strength of the relationship between predictors and outcomes differed for the 2 groups of respondents. Each model included 1 of the 3 putative predictors along with LGB status and an interaction term for LGB and the putative predictor. The t value associated with the product term was used to determine if the interaction was significant,21 and the exponent was used to assess the effect size. Listwise deletion was used for each model tested.

RESULTS

Table 1 ▶ shows the descriptive and demographic factors for respondents with completed item responses. LGB respondents reported higher rates of suicidal ideation and suicide attempts than did non-LGB respondents after we controlled for race, gender, and age.

TABLE 1—

Sample Demographics and Adjusted Odds Ratios for Suicidal Ideation and Suicide Attempts by Sexual Orientation Among Adolescents and Young Adults: National Longitudinal Study of Adolescent Health, 2001–2002

	Unweighted Total (Weighted %)
	Total Sample, No. (%)	Non-LGB, No. (%)	LGB, No. (%)	Adjusted Wald F a or AOR (95% CI)
Gender				14.37**
Female	7478 (49.1)	7207 (48.7)	271 (60.1)
Male	6711 (50.9)	6535 (51.3)	176 (39.9)
Raceb
White	8955 (74.0)	8648 (73.8)	307 (81.2)	5.62*
African American	2978 (15.8)	2905 (16.0)	73 (9.9)	6.59*
Native American	410 (2.4)	397 (2.3)	13 (2.4)	0.01
Asian	1056 (3.9)	1031 (4.0)	25 (2.3)	2.64
Biracial	589 (3.6)	566 (3.6)	23 (4.0)	0.12
Ethnicity				0.72
Hispanic	2307 (11.7)	2212 (11.6)	95 (13.4)
Non-Hispanic	11 860 (88.3)	11 510 (88.4)	350 (86.6)
Residence at wave 1c				0.91
Urban	7671 (51.8)	7408 (51.5)	263 (55.1)
Nonurban	6405 (48.2)	6 226 (48.5)	179 (44.9)
Suicidal ideationd				2.94 (2.06, 4.19)
Yes	841 (6.6)	762 (6.3)	79 (17.2)
No	12 997 (93.4)	12 638 (93.7)	359 (82.8)
Suicide attempte				2.96 (1.41, 6.21)
Yes	218 (1.7)	198 (1.6)	20 (4.9)
No	13 616 (98.3)	13 199 (98.4)	417 (95.1)

Note. LGB = lesbian, gay, or bisexual; AOR = adjusted odds ratio; CI = confidence interval. Mean age was 22.3 years for both LGB and non-LGB respondents, with standard errors of 0.17 and 0.12 years, respectively.

^aUsed in place of χ² for weighted data.

^bRespondents who identified with 2 or more races were classified as biracial and excluded from the other racial categories.

^cResidence in an urban or nonurban area at Wave 1 determined by census tract.

^dSuicidal ideation within the past 12 months was measured with 1 item, which was dichotomized. Reference is LGB respondents. Adjusted for race, gender, and age.

^eSuicide attempts within the past 12 months was measured with 1 item, which was dichotomized. Reference is LGB respondents. Adjusted for race, gender, and age.

*P ≤.05; **P = .019; ***P ≤.001.

Stratified analyses (Table 2 ▶) showed that problem drinking, problem drug use, and depression were associated with elevated risk for suicidal ideation and suicide attempts among non-LGB respondents. Problem drinking and depression increased risk for suicidal ideation among LGB respondents; drug use did not. Problem drinking, drug use, and depression were not associated with increased risk for suicide attempts among LGB respondents. The interaction models revealed that suicidal ideation was associated with problem drug use and that suicide attempts were associated with depression among non-LGB respondents (Table 3 ▶).

TABLE 2—

Stratified Analyses of Risk Factors for Suicidal Ideation and Suicide Attempts in the Past Year: National Longitudinal Study of Adolescent Health, 2001–2002

	Non-LGB Respondents, AOR (95% CI)	LGB Respondents, AOR (95% CI)
Suicidal ideation
Problem drinkinga	2.47 (2.05, 2.98)*	2.04 (1.02, 4.08)*
Problem drug usea	3.59 (2.80, 4.59)*	1.28 (0.56, 2.94)
Depressionb	1.25 (1.22, 1.28)*	1.29 (1.16, 1.44)*
Suicide attempts
Problem drinkinga	2.58 (1.78, 3.76)*	1.97 (0.42, 9.12)
Problem drug usea	4.84 (3.09, 7.59)*	1.78 (0.46, 6.91)
Depressionb	1.29 (1.23, 1.35)*	1.11 (0.99, 1.25)

Note. LGB = lesbian, gay, or bisexual; AOR = adjusted odds ratio; CI = confidence interval. Adjusted for age, race, and gender.

^aProblem drinking and problem drug use in the preceding 12 months were coded as present if the respondents answered yes to any of the 6 items in each domain. Dichotomous; reference group = no problem drinking.

^bDepression was measured in the adolescent health survey with 9 items from the Center for Epidemiologic Studies Depression Scale.20 For every 1-unit increase in the score, the odds of suicidal ideation or suicide attempts increased by the amount shown (e.g., a 1-unit increase in the score increased the likelihood of reporting suicidal ideation by 25% among the non-LGB respondents and 29% among the LGB respondents).

*Significant at P ≤.05.

TABLE 3—

Analyses With Interaction Terms of Risk Factors for Suicidal Ideation and Suicide Attempts in the Past Year Among Adolescents and Young Adults: National Longitudinal Study of Adolescent Health, 2001–2002

Non-LGB vs LGB Respondents	Exponent of Interaction Terma (95% CI)
Suicidal Ideation
Problem drinkingb	1.21 (0.59, 2.51)
Problem drug useb	2.80 (1.19, 6.57)*
Depressionc	0.97 (0.86, 1.08)
Suicide attempts
Problem drinkingb	1.31 (0.27, 6.38)
Problem drug useb	2.68 (0.65, 11.45)
Depressionc	1.15 (1.01, 1.31)*

Note. LGB = lesbian, gay, or bisexual; CI = confidence interval. Adjusted for age, race, and gender.

^aExponent of interaction term =OR_non-LGB / OR_LGB.

^bProblem drinking and problem drug use in the preceding 12 months were coded as present if the respondents answered yes to any of the 6 items in each domain. Dichotomous; reference group = no problem drinking.

^cDepression was measured in the adolescent health survey with 9 items from the Center for Epidemiologic Studies Depression Scale.20 For every 1-unit increase in the score, the odds of suicidal ideation or suicide attempts increased by the amount shown (e.g., a 1-unit increase in the score increased the likelihood of reporting suicidal ideation by 25% among the non-LGB respondents and 29% among the LGB respondents).

*Significant at P ≤.05.

DISCUSSION

To our knowledge this is the first study to describe the relative differences in risk markers for suicidal ideation and suicide attempts between LGB persons and their peers. Sexual orientation was found to exert significant interaction effects with risk markers for both suicidal ideation and suicide attempts. Specifically, problem drug use was more strongly associated with suicidal ideation among non-LGB respondents than among LGB respondents. Similarly, the association between depression and suicide attempts was stronger among non-LGB respondents than among LGB respondents.

Consistent with earlier findings,11 we observed higher adjusted rates of suicidal ideation and suicide attempts among LGB adolescents and young adults than among non-LGB resondents. It has been proposed that suicidal ideation or suicide attempts may represent something fundamentally different for LGB and non-LGB youths.22 Alternatively, the consistently elevated risk found in this and in previous studies, which controlled for idiosyncratic variables such as victimization or parental support,23^,24 may be mediated by factors that have not yet been tapped in research.

Despite the limitations inherent to secondary data analyses, our findings point to the need for research targeting suicide-related thoughts and behavior among LGB adolescents and young adults. Addressing depression or problem drug use is not unimportant; rather, more information is needed about the nature and source of distress that is driving suicidal behavior in this population. We tentatively conclude that LGB adolescents and young adults may need a different treatment focus and alternative points of entry to health services. Elucidation of risk markers specific to LGB individuals will be necessary to support the design and evaluation of suicide prevention interventions.