Abstract
Objectives. To investigate the prevalence and co-occurrence of heavy drinking, anxiety, and mood disorders among Canadians who self-identified as gay, lesbian, bisexual, or heterosexual.
Methods. Pooled data from the 2007 to 2012 cycles of the Canadian Community Health Survey (n = 222 548) were used to fit logistic regression models controlling for sociodemographic characteristics.
Results. In adjusted logistic regression models, gay or lesbian respondents had greater odds than heterosexual respondents of reporting anxiety disorders, mood disorders, and anxiety–mood disorders. Bisexual respondents had greater odds of reporting anxiety disorders, mood disorders, anxiety–mood disorders, and heavy drinking. Gay or lesbian and bisexual respondents had greater odds than heterosexuals of reporting co-occurring anxiety or mood disorders and heavy drinking. The highest rates of disorders were observed among bisexual respondents, with nearly quadruple the rates of anxiety, mood, and combined anxiety and mood disorders relative to heterosexuals and approximately twice the rates of gay or lesbian respondents.
Conclusions. Members of sexual minority groups in Canada, in particular those self-identifying as bisexual, experience disproportionate rates of anxiety and mood disorders, heavy drinking, and co-occurring disorders.
Evidence from several countries indicates that lesbian, gay, and bisexual (LGB) people experience poorer mental health than their heterosexual counterparts. For example, studies conducted in the United States and Europe have shown consistently elevated rates of mental disorders, substance use, violence, victimization, self-harm, and suicidality in this population.1–4 From the standpoint of minority stress theory, the leading framework for studying sexual minority health, these observed disparities implicate adverse distal (e.g., prejudice, discrimination) and proximal (e.g., expectations of rejection, hypervigilance) stressors1 that LGB people experience throughout their lives. Such stressors increase the risk of health problems,5 including mental health problems (particularly in adolescence).1,6
Although Canada has made substantial gains in affording the LGB population the legal rights enjoyed by heterosexuals, discrimination and victimization experienced by LGB people remain significant social and public health problems. For example, a recent analysis of police-reported data revealed that two thirds of all reported hate crimes in Canada that targeted people because of their actual or perceived sexual orientation involved violent offenses.7 Analyses of the 2007–2008 Canadian Community Health Survey (CCHS), Canada’s national population health surveillance survey, indicated that 17.1% of LGB respondents reported having a current mood disorder (e.g., depression or bipolar disorder) compared with 6.9% of heterosexual respondents.8
Mood and anxiety disorders frequently co-occur with heavy drinking, defined as consuming 5 or more drinks on a single occasion 12 or more times in a year.9 Although 24.8% of Canadian men and 10.1% of Canadian women reported heavy drinking in 2010,10 no national estimates of heavy drinking or of co-occurring heavy drinking and mental disorders are available for LGB Canadians. Despite this context and the likelihood that being a member of a stigmatized sexual minority group may lead to mental health challenges, few Canadian studies have examined the relationship of sexual orientation with mental health and substance misuse in national samples.8,11
These knowledge gaps are not specific to Canada. Population-based studies are scant; the majority of research has been conducted in the United States, where existing evidence remains sparse.1,4,12 Few studies have examined mental health and alcohol misuse comorbidities.13 Also, because of methodological limitations (e.g., insufficient sample sizes), previous researchers have been limited in their ability to determine precise prevalence estimates and to disaggregate findings for lesbian, gay, and bisexual people (as opposed to treating them as an undifferentiated LGB group).12
Our goal in this study was to address these existing substantive and methodological knowledge gaps. By combining multiple cycles of the CCHS (2007–2012), we were able to investigate the prevalence of anxiety and mood disorders, heavy drinking, and co-occurring anxiety or mood disorders and heavy drinking among Canadians who self-identified as gay or lesbian, bisexual, or heterosexual. This relatively new, pooled approach to combining multiple cycles of a population health survey offers important advantages, including an increased sample size, thereby enabling disaggregated analyses by sexual identity and improved precision of sexual identity–specific prevalence estimates.
METHODS
We derived our data from the 2007 to 2012 cycles of the CCHS, conducted by Statistics Canada. The CCHS is an ongoing national, cross-sectional survey that collects health-related information representative of approximately 98% of Canadian residents 12 years or older from all of the country’s provinces and territories who live in private dwellings.14 The sampling frame excludes individuals living on First Nations Reserves and Crown Lands, institutional residents, full-time members of the Canadian Forces, and residents of extremely remote regions. Since 2007, Statistics Canada has collected CCHS data on an ongoing basis, producing annual and biannual files.15
We obtained data from 3 biannual files (2007–2008, 2009–2010, and 2011–2012) and formed a combined pooled sample (2007–2012). The response rate across cycles ranged from 68.4% to 76.0%. Statistics Canada has reported further details about the CCHS methodology, including measures for detecting and minimizing errors.14
Pooled Data
We used published guidelines to combine the study data.16 Before pooling the data, we merged the cycle-specific external files containing bootstrap weights and imputed income variables (the 2011–2012 cycle contained imputed income data). Data from each biannual file were combined into a single data set. This data set pooling was feasible because the CCHS cycles involved comparable sample designs, survey content, and interview modes. Because the data collected in the CCHS represent the changing characteristics of an evolving population, our pooled data set should be considered a sample of an “average” population.16 The combined sample is not necessarily representative of the population included in any single cycle; rather, it is representative of the combined population, with estimates representing the period 2007 to 2012.
Study Sample
The initial survey samples were restricted to people who participated in the 3 biannual cycles (2007–2008, n = 131 959; 2009–2010, n = 124 870; 2011–2012, n = 125 645). Subsequently, these samples were further restricted to individuals who had valid responses on the sexual identity item. People who responded “don’t know,” did not answer the question, or refused to answer the question were excluded from the analyses described here but were included in supplemental analyses. Because the sexual identity question was asked only of respondents aged 18 to 59 years, the study sample was restricted to people in this age range.
These restrictions yielded cycle samples of 79 957 in 2007–2008, 72 554 in 2009–2010, and 70 037 in 2011–2012, leading to a final pooled sample of 222 548. Approximately 4% of pooled respondents had missing data on the sexual identity item; missing data on all other study variables were minimal (less than 2.5%).
Study Variables
Our primary explanatory variable, self-reported sexual identity, was based on the item
Do you consider yourself to be: heterosexual (sexual relations with people of the opposite sex)? Homosexual, that is lesbian or gay (sexual relations with people of your own sex)? Bisexual (sexual relations with people of both sexes)?
The mood disorder variable (yes or no) was based on the following item:
I’d like to ask about certain long-term health conditions which you may have. We are interested in “long-term conditions” which are expected to last or have already lasted 6 months or more and that have been diagnosed by a health professional. Do you have a mood disorder such as depression, bipolar disorder, mania, or dysthymia?
The interviewers were instructed to include manic depression if it was mentioned by respondents.
The anxiety disorder variable (yes or no) was based on the item “Do you have an anxiety disorder such as a phobia, obsessive–compulsive disorder, or a panic disorder?” (diagnosed by a health professional and with a duration of 6 months or more). The anxiety–mood disorder variable (yes or no) was used to indicate whether a respondent reported having both an anxiety disorder and a mood disorder.
The heavy drinking variable was based on the item “How often in the past 12 months have you had 5 or more drinks on one occasion?” Consistent with the World Health Organization and Health Canada definitions of heavy drinking,17 responses were recoded to define heavy drinking as consuming 5 or more drinks on a single occasion 12 or more times over the preceding year (i.e., a yes response indicates once a month or more often). Respondents who reported not drinking at all, drinking 5 or more drinks less often than once a month, or never consuming 5 or more drinks on a single occasion were recoded as responding no to the question. The co-occurring anxiety or mood disorder and heavy drinking variable (yes or no) indicates whether a respondent reported an anxiety or a mood disorder and heavy drinking.
We selected known and potential confounders from a literature review focusing on associations between sexual identity, mental health, and substance misuse.2,18 Study covariates included sex, age, educational attainment, household income, marital status, racial minority status, and residence (i.e., region of Canada). Racial minority status was determined through a question asking respondents to indicate their cultural or racial background. Those who self-identified as White were coded “not a racial minority.” All others, including those who self-identified as Aboriginal or First Nations, were coded as belonging to a racial minority group.
Data Analysis
All analyses were conducted with Stata version 13 (StataCorp LP, College Station, TX). We used svyset procedures and Statistics Canada’s guidelines to apply design and bootstrap weights in an effort to produce unbiased estimates with variances adjusted for the sampling method.19 Analyses were initially performed on each biannual data cycle and then on the pooled sample. We conducted the χ2 test to examine bivariable relationships between sexual identity, the study outcomes, and the confounders. Associations between sexual identity and the study outcomes were examined via unadjusted and adjusted logistic regression models and odds ratios (ORs) with 95% confidence intervals. We also obtained weighted prevalence rates stratified by sex and tested sex–sexual identity interaction terms for significance in all models.
With respect to assumptions of time equality across the survey cycles, the use of a pooled approach in fitting models can be justified according to a model-design-based view.20 Consequently, we examined differences in the variable distributions across the cycles, included a survey cycle–time effect in the models, and tested for statistical significance.
RESULTS
During the period 2007 to 2012, 97.8% of CCHS respondents self-identified as heterosexual; 1.3% and 1.0% identified as gay or lesbian and bisexual, respectively. Differences in self-identification between cycles were minimal and not statistically significant. Table 1 presents information on the sociodemographic characteristics of the respondents. Whereas nearly equal proportions of heterosexual respondents were men and women, more than 62% of those who identified as gay or lesbian were men, and more than 69% of those identifying as bisexual were women.
TABLE 1—
Sociodemographic Characteristics of Heterosexual, Gay or Lesbian, and Bisexual Respondents: Canadian Community Health Survey, 2007–2012
| Characteristic | Total, % | Heterosexual (97.8%), % | Gay/Lesbian (1.3%), % | Bisexual (1.0%), % |
| Sexa | ||||
| Female | 50.1 | 50.1 | 37.7 | 69.2 |
| Male | 49.9 | 49.9 | 62.3 | 30.8 |
| Age, ya | ||||
| 18–29 | 27.5 | 27.2 | 26.4 | 51.3 |
| 30–39 | 22.5 | 22.5 | 20.5 | 18.5 |
| 40–49 | 25.7 | 25.7 | 29.4 | 18.1 |
| 50–59 | 24.4 | 24.5 | 23.6 | 12.1 |
| Educational attainmenta | ||||
| Some secondary school | 9.3 | 9.4 | 4.9 | 13.4 |
| Secondary school | 17.6 | 17.7 | 13.0 | 20.7 |
| Some postsecondary | 8.7 | 8.6 | 9.1 | 16.2 |
| Postsecondary | 64.3 | 64.3 | 73.0 | 49.7 |
| Household income, $a | ||||
| 0–39 999 | 21.5 | 21.3 | 23.1 | 38.9 |
| 40 000–59 999 | 16.9 | 16.9 | 18.0 | 18.6 |
| 60 000–99 999 | 29.6 | 29.7 | 27.0 | 25.0 |
| ≥ 100 000 | 32.0 | 32.1 | 31.8 | 17.5 |
| Racial minoritya | ||||
| Yes | 21.9 | 22.0 | 13.9 | 21.2 |
| No | 78.1 | 78.0 | 86.1 | 78.8 |
| Marital statusa | ||||
| Single/widowed/divorced | 37.4 | 36.8 | 60.9 | 67.2 |
| Married/common law union | 62.6 | 63.2 | 39.1 | 32.8 |
| Regiona | ||||
| Atlantic | 6.9 | 6.9 | 6.7 | 6.0 |
| Quebec | 23.2 | 23.1 | 31.1 | 19.8 |
| Prairies | 17.5 | 17.6 | 11.4 | 17.2 |
| British Columbia | 13.2 | 13.2 | 14.1 | 19.7 |
| Northern Territories | 0.3 | 0.3 | 0.2 | 0.4 |
| Ontario | 39.0 | 39.0 | 36.5 | 36.9 |
| Area of residencea | ||||
| Rural | 17.0 | 17.1 | 9.7 | 11.8 |
| Urban | 83.0 | 82.9 | 90.3 | 88.2 |
Note. Data are weighted. The sample size was n = 222 548.
Significant at P < .001 based on χ2 test of association between covariate and sexual identity (heterosexual, gay or lesbian, bisexual).
Rates of Anxiety and Mood Disorders and Heavy Drinking
Table 2 displays the prevalence rates of the study outcomes. All confounders were significantly associated with the study outcomes. Across all study outcomes, higher prevalence rates were observed among sexual minority (vs heterosexual) respondents, with rates substantially higher among bisexual respondents than among respondents of any other sexual identity.
TABLE 2—
Prevalence Rates of Anxiety, Mood, and Anxiety–Mood Disorders; Heavy Drinking; and Co-Occurring Anxiety or Mood Disorders and Heavy Drinking; by Sexual Identity and Sociodemographic Characteristics: Canadian Community Health Survey, 2007–2012
| Variable | Anxiety Disorder, % (95% CI) | Mood Disorder, % (95% CI) | Anxiety–Mood Disorder, % (95% CI) | Heavy Drinking, % (95% CI) | Co-Occurring Anxiety or Mood Disorder and Heavy Drinking, % (95% CI) |
| Sexual identity | |||||
| Heterosexual | 5.9 (5.8, 6.1) | 7.0 (6.8, 7.2) | 2.7 (2.6, 2.8) | 22.6 (22.3, 22.9) | 2.2 (2.1, 2.3) |
| Gay/lesbian | 11.4 (9.9, 13.2) | 14.5 (12.5, 16.7) | 6.6 (5.5, 7.8) | 27.9 (25.2, 30.8) | 5.2 (4.1, 6.5) |
| Bisexual | 20.7 (18.3, 23.4) | 24.8 (22.0, 27.7) | 13.5 (11.4, 15.8) | 30.4 (27.2, 33.8) | 10.0 (7.9, 12.4) |
| Sexa | |||||
| Female | 8.0 (7.7, 8.2) | 9.6 (9.3, 9.9) | 3.8 (3.7, 4.0) | 13.6 (13.3, 13.9) | 2.3 (2.2, 2.4) |
| Male | 4.5 (4.3, 4.7) | 5.2 (5.0, 5.4) | 2.0 (1.9, 2.1) | 31.3 (30.8, 31.8) | 2.4 (2.2, 2.5) |
| Age, y | |||||
| 18–29 | 6.3 (6.0, 6.7) | 5.9 (5.6, 6.3) | 2.7 (2.5, 2.9) | 34.8 (34.2, 35.5) | 3.3 (3.1, 3.6) |
| 30–39 | 5.9 (5.6, 6.2) | 7.0 (6.7, 7.4) | 2.8 (2.6, 3.0) | 20.2 (19.6, 20.8) | 2.1 (1.9, 2.3) |
| 40–49 | 6.5 (6.2, 6.9) | 8.0 (7.6, 8.4) | 3.1 (2.9, 3.4) | 17.9 (17.3, 18.4) | 2.0 (1.9, 2.2) |
| 50–59 | 6.1 (5.8, 6.5) | 8.7 (8.4, 9.1) | 3.0 (2.8, 3.2) | 15.4 (14.8, 15.9) | 1.7 (1.5, 1.9) |
| Educational attainment | |||||
| Some secondary school | 10.3 (9.8, 10.9) | 11.6 (10.9, 12.3) | 5.7 (5.3, 6.1) | 23.8 (22.9, 24.8) | 3.8 (3.4, 4.3) |
| Secondary school | 6.3 (6.0, 6.6) | 7.5 (7.1, 7.9) | 2.9 (2.7, 3.2) | 24.6 (24.0, 25.3) | 2.5 (2.3, 2.7) |
| Some postsecondary | 8.1 (7.5, 8.8) | 8.9 (8.3, 9.6) | 4.2 (3.7, 4.7) | 29.4 (28.4, 30.5) | 3.3 (2.9, 3.7) |
| Postsecondary | 5.3 (5.1, 5.5) | 6.5 (6.3, 6.8) | 2.3 (2.2, 2.4) | 20.7 (20.3, 21.1) | 2.0 (1.8, 2.1) |
| Household income, $ | |||||
| 0–39 999 | 10.0 (9.7, 10.4) | 12.3 (11.9, 12.7) | 5.7 (5.4, 6.0) | 20.4 (19.9, 21.0) | 3.6 (3.4, 3.9) |
| 40 000–59 999 | 6.5 (6.1, 6.9) | 7.7 (7.3, 8.1) | 3.0 (2.7, 3.3) | 21.0 (20.4, 21.6) | 2.5 (2.3, 2.8) |
| 60 000–99 999 | 5.1 (4.9, 5.4) | 6.1 (5.9, 6.5) | 2.2 (2.1, 2.4) | 22.2 (21.6, 22.7) | 1.9 (1.7, 2.1) |
| ≥ 100 000 | 4.3 (4.1, 4.6) | 4.8 (4.5, 5.1) | 1.5 (1.4, 1.7) | 24.9 (24.3, 25.4) | 1.7 (1.6, 1.9) |
| Racial minority | |||||
| No | 6.8 (6.7, 7.0) | 8.0 (7.8, 8.3) | 3.2 (3.1, 3.3) | 25.2 (24.9, 25.5) | 2.6 (2.5, 2.7) |
| Yes | 4.1 (3.8, 4.4) | 5.1 (4.8, 5.5) | 1.9 (1.7, 2.2) | 12.7 (12.1, 13.3) | 1.3 (1.2, 1.5) |
| Marital status | |||||
| Married/common law union | 5.0 (4.9, 5.2) | 5.9 (5.7, 6.1) | 2.1 (2.0, 2.3) | 17.9 (17.6, 18.3) | 1.5 (1.4, 1.6) |
| Single/widowed/divorced | 8.2 (7.9, 8.5) | 9.9 (9.6, 10.2) | 4.2 (4.0, 4.4) | 29.9 (29.4, 30.4) | 3.7 (3.5, 3.9) |
| Region | |||||
| Ontario | 6.4 (6.1, 6.7) | 7.8 (7.4, 8.1) | 3.2 (3.0, 3.4) | 20.9 (20.4, 21.4) | 2.3 (2.1, 2.4) |
| Atlantic | 7.9 (7.5, 8.4) | 8.5 (8.0, 9.0) | 3.8 (3.4, 4.1) | 28.0 (27.2, 28.7) | 3.3 (3.0, 3.6) |
| Quebec | 5.9 (5.6, 6.2) | 5.6 (5.3, 6.0) | 2.1 (2.0, 2.3) | 23.0 (22.4, 23.7) | 2.1 (1.9, 2.3) |
| Prairies | 5.7 (5.4, 6.0) | 7.8 (7.4, 8.2) | 2.8 (2.5, 3.0) | 24.4 (23.7, 25.1) | 2.5 (2.3, 2.7) |
| British Columbia | 6.2 (5.7, 6.6) | 8.4 (7.9, 9.0) | 3.1 (2.8, 3.5) | 20.3 (19.6, 21.1) | 2.3 (2.1, 2.6) |
| Northern Territories | 5.2 (4.5, 6.0) | 7.1 (6.4, 7.9) | 2.5 (2.0, 3.0) | 32.1 (30.5, 33.6) | 3.5 (3.0, 4.1) |
| Area of residencea | |||||
| Rural | 5.9 (5.6, 6.2) | 6.9 (6.6, 7.3) | 2.6 (2.4, 2.8) | 24.5 (23.9, 25.1) | 2.2 (2.0, 2.4) |
| Urban | 6.3 (6.1, 6.5) | 7.5 (7.3, 7.7) | 3.0 (2.9, 3.1) | 22.0 (21.7, 22.3) | 2.4 (2.3, 2.5) |
Note. CI = confidence interval. Data are weighted. All associations between sociodemographic characteristics and study outcomes were estimated with χ2 tests of association. Other than the exceptions noted, all corresponding P values were significant at the .05 level. The sample size was n = 222 548.
Nonsignificant association with co-occurring disorders.
In the 2007 to 2012 period, anxiety disorders were reported by 5.9% of heterosexual respondents, 11.4% of gay or lesbian respondents, and 20.7% of bisexual respondents. Mood disorders were reported by 7.0% of heterosexuals, 14.5% of gay or lesbian respondents, and 24.8% of bisexual respondents. Overall, 2.7% of heterosexuals, 6.6% of gay or lesbian respondents, and 13.5% of bisexuals reported combined anxiety–mood disorders. Heavy drinking was reported by 22.6% of heterosexual respondents, 27.9% of gay or lesbian respondents, and 30.4% of bisexual respondents. A total of 2.3% of heterosexuals, 5.2% of gay or lesbian respondents, and 10.0% of bisexuals reported co-occurring anxiety or mood disorders and heavy drinking. This pattern was observed for each of the separate CCHS cycles, with small cycle-to-cycle differences (Table A, available as a supplement to the online version of this article at http://www.ajph.org).
Associations Between Sexual Identity and the Study Outcomes
Table 3 shows adjusted odds ratios for associations between sexual identity and the study outcomes. Stratified odds are reported if the omnibus test of significance for the interaction between sex and sexual identity was statistically significant (P < .05). In all of the unadjusted and adjusted logistic regression models (for the former, estimates are not shown in tables but are reported for comparison purposes), gay or lesbian or bisexual respondents were significantly more likely than heterosexual respondents to report anxiety, mood, and anxiety–mood disorders; heavy drinking; and co-occurring anxiety or mood disorders and heavy drinking.
TABLE 3—
Adjusted Odds Ratios for Anxiety, Mood, and Anxiety–Mood Disorders; Heavy Drinking; and Co-Occurring Anxiety or Mood Disorders and Heavy Drinking: Canadian Community Health Survey, 2007–2012
| Sexual Identity | Anxiety Disorder (n = 210 852), AOR (95% CI) | Mood Disorder (n = 210 826), AOR (95% CI) | Anxiety–Mood Disorder (n = 210 724), AOR (95% CI) | Heavy Drinking (n = 210 020), AOR (95% CI) | Co-Occurring Anxiety or Mood Disorder and Heavy Drinking (n = 204 208), AOR (95% CI) |
| Heterosexual (Ref) | 1 | 1 | 1 | 1 | 1 |
| Gay/lesbian | 2.0 (1.7, 2.4) | 2.2 (1.8, 2.6) | 2.5 (2.0, 3.0) | 1.0 (0.9, 1.2) | 2.0 (1.6, 2.6) |
| Male | 2.5 (1.9, 3.1) | 2.9 (2.2, 3.9) | 0.9 (0.8, 1.1) | ||
| Female | 1.5 (1.1, 1.9) | 1.9 (1.4, 2.6) | 1.6 (1.2, 2.0) | ||
| Bisexual | 3.0 (2.5, 3.5) | 3.4 (2.8, 4.0) | 3.8 (3.0, 4.7) | 1.4 (1.2, 1.7) | 3.3 (2.5, 4.3) |
| Male | 1.1 (0.8, 1.4) | ||||
| Female | 1.6 (1.3, 1.9) |
Note. AOR = adjusted odds ratio; CI = confidence interval. Data are weighted. Stratified odds are reported when the omnibus test for the interaction between sex and sexual identity was significant at P < .05. All models adjusted for sex, age, educational attainment, household income, racial minority status, marital status, region, area of residence, and survey cycle.
Relative to heterosexuals, gay or lesbian respondents had twice the unadjusted and adjusted odds of reporting an anxiety disorder. In the stratified model, the difference in the estimates for gay men (adjusted OR = 2.5) and lesbian women (adjusted OR = 1.5) was statistically significant. Bisexual respondents had 4.1-fold higher unadjusted odds than heterosexuals of reporting an anxiety disorder; the adjusted odds ratio was 3.0 and remained significant. In addition, gay or lesbian respondents had 2.2-fold higher unadjusted and adjusted odds than heterosexuals of reporting a mood disorder. Bisexual respondents had 4.4-fold higher unadjusted and 3.4-fold higher adjusted odds of a mood disorder than heterosexuals.
Gay or lesbian respondents had 2.5-fold higher unadjusted and adjusted odds of anxiety–mood disorders than heterosexual respondents. In the stratified models, the estimates for gay men (adjusted OR = 2.9) and lesbian women (adjusted OR = 1.9) indicated a statistically significant interaction between sex and sexual identity. Bisexual respondents had 5.6-fold higher unadjusted and 3.8-fold higher adjusted odds than heterosexuals of anxiety–mood disorders.
Gay or lesbian respondents had 1.3-fold higher unadjusted odds than heterosexual respondents of heavy drinking, but this association was not significant in the adjusted model. In the stratified analyses, odds ratios for gay and heterosexual male respondents were not significantly different, but lesbian women had 1.6-fold (significantly) higher odds than heterosexual women of heavy drinking. Relative to heterosexuals, bisexual respondents had 1.5-fold greater unadjusted odds of heavy drinking, and this difference remained statistically significant and of a similar magnitude in the adjusted model (adjusted OR = 1.4). Although the odds of heavy drinking among bisexual and heterosexual men did not differ significantly, bisexual women had 1.6-fold (significantly) greater odds than heterosexual women. Also, the interaction was statistically significant.
Gay or lesbian respondents had 2.4-fold higher unadjusted and 2.0-fold higher adjusted odds than heterosexual respondents of reporting co-occurring anxiety or mood disorders and heavy drinking. Bisexual respondents had 4.8-fold and 3.3-fold higher odds than heterosexuals in respective unadjusted and adjusted models.
The survey cycle–time effect was included in all of the adjusted models and tested for significance. The coefficient for the 2011–2012 cycle (vs the 2007–2008 reference cycle) was statistically significant in all of the adjusted models and associated with higher odds of reporting each of the study outcomes.
Missing Data on Sexual Identity
To determine whether missing data were randomly distributed, we repeated all of our analyses with 4 sexual identity groups: heterosexual (reference), gay or lesbian, bisexual, and missing (i.e., those who answered “don’t know,” refused to answer the question, or neglected to respond). Relative to heterosexuals, those with missing data were more likely to be male, older, and single; to have lower educational attainment; to live in British Columbia; and to self-identify as a member of a racial minority group.
Respondents with missing information had higher rates of anxiety (8.0%), mood (9.8%), and combined anxiety and mood (4.3%) disorders than heterosexuals but lower rates than gay or lesbian and bisexual respondents. Rates of heavy drinking (14.5%) and co-occurring anxiety or mood disorders and heavy drinking (2.2%) were lower among respondents with missing information than among those in the other sexual identity groups. The same pattern was observed for respondents with missing information in sensitivity analyses involving adjusted logistic regression models (Table B, available as a supplement to the online version of this article at http://www.ajph.org).
DISCUSSION
Our study, involving pooled data from the 2007 to 2012 cycles of the CCHS, documents disparities in prevalence rates of self-reported anxiety and mood disorders, heavy drinking, and co-occurring anxiety or mood disorders and heavy drinking among gay or lesbian and bisexual Canadians relative to their heterosexual peers. The substantially greater odds of these outcomes among LGB people, even after controlling for multiple confounders, point to the disproportionate mental health burden experienced by this population in Canada. Particularly concerning is the dramatically higher prevalence among bisexual respondents, with nearly quadruple the rates of anxiety, mood, and combined anxiety and mood disorders relative to heterosexuals and approximately twice the rates of gay or lesbian respondents.
The results also point to the important moderating role of sex, with higher adjusted odds of anxiety and anxiety–mood disorders among gay male respondents and higher adjusted odds of heavy drinking among lesbian and bisexual women. Notably, the relatively large adjustment in odds of the study outcomes after control for confounders among bisexual but not gay or lesbian respondents indicates the potential confluence of multiple social factors in influencing bisexual people’s mental health.
Our results corroborate and extend the findings of international population-based studies showing that gay, lesbian, and bisexual sexual identities are associated with poorer mental health and substance misuse.1–4 Our study is the first in Canada to pool epidemiological data to document prevalence rates and co-occurrence of mental disorders and alcohol misuse among gay or lesbian and bisexual Canadians relative to their heterosexual peers. Accordingly, our findings add to the existing national studies of LGB populations8,11 and enlarge the evidence base regarding health disparities related to sexual identity. In addition, our investigation corroborates other studies indicating that bisexuality confers the greatest odds of adverse health outcomes, including mood and anxiety disorders,11,21 by providing new population-level evidence of the disparities experienced by bisexual Canadians.
Our results implicate the consequences of minority stressors experienced by LGB people.1 According to minority stress theory, members of sexual- and gender-identity (and other) minority groups experience chronic stressors as a result of a number of distal and proximal social conditions,1,5 all of which increase the risks of health problems.5 Consistent with this theory, a burgeoning literature on micro-aggressions (i.e., micro-level forms of discrimination and prejudice) considers how seemingly minor events can be psychologically damaging because of the message of rejection they convey, especially when accumulated over time.5,19,22 These experiences can diminish psychological well-being, resulting in symptoms of depression, anxiety, or substance misuse.6,18
Such experiences are not rare. In a 2013 survey of LGB and transgender Americans, 66% of respondents reported experiencing at least one incident of discrimination or exclusion (e.g., being subjected to slurs or jokes, rejected by a family member, or threatened or physically attacked) because of their sexual orientation or gender identity.23 In addition, as noted, a 2012 analysis of Canadian national data suggested that police-reported hate crimes were more likely to be violent when they were motivated by the victim’s sexual orientation.7 Robust evidence links trauma and victimization to negative outcomes such as depression,24 anxiety,25 and alcohol and tobacco use,26 and there is some evidence of higher rates of traumatic stress in LGB samples.27
Possible explanations for the observed disparities among bisexual people focus on the unique, double stigma these individuals experience from within both the heterosexual and gay or lesbian communities. Pervasive stereotypes about and negative attitudes toward bisexuality (e.g., identity confusion, experimentation, promiscuity)21 present consistent messages to bisexual people about the unintelligibility (i.e., bisexuality as an unknowable entity) and illegitimacy of their identity and are often coupled with a lack of an identifiable supportive community.28 Qualitative research has documented how bisexual women experience a range of micro-aggressions (e.g., targeted hostility, perceived hypersexuality, lack of legitimacy within the gay or lesbian community)29 and how notions of monosexism (the belief that one can be only heterosexual or gay or lesbian) and biphobia (aversion toward bisexuality and bisexual people as a social group) exert broad-reaching mental health effects.30
Limitations and Strengths
Although our study has important strengths, 3 limitations deserve noting. First, our self-reported data may be subject to recall and social desirability biases as well as subjective interpretations of the questions. Although the CCHS asked about diagnosed conditions, their clinical nature (e.g., severity and chronicity) is unknown.
Second, because of the stigma some attach to LGB status, collecting information about sexual identity may be sensitive to the mode of data gathering used.23 A review of multiple surveys estimated that approximately 3.5% of adults self-identify as LGB31; the proportion of LGB Canadians in the CCHS is therefore probably underestimated. This discrepancy may lead to underestimates of true associations between sexual identity and the study outcomes because respondents who self-identified might, on average, be healthier and experience less stress than those who did not self-identify or disclose (i.e., those who remained “closeted”).8 Consequently, our results, although informative, may represent imprecise estimates of existing disparities.
Third, the sexual identity question was asked only of those 18 to 59 years of age, and about 4% of respondents chose not to answer. Therefore, our findings may not be generalizable to all LGB people in Canada, including those who identify with specific dimensions of same-sex sexuality (e.g., behavior or attraction) but do not identify as LGB. Although complete case analysis is a reasonable approach when data are missing for fewer than 10% of cases,32 our sensitivity analyses showed that respondents with missing information on sexual identity had significantly different sociodemographic characteristics, higher rates of mental disorders, and lower rates of heavy drinking than those without missing information.
The primary strength of our study is our use of data from a large national, population-based survey. Pooling data from multiple cycles of the CCHS allowed us to examine disparities across specific sexual identities and to calculate relatively precise prevalence rates and reasonable confidence intervals while adjusting for multiple confounders.
Conclusions
We have documented prevalence rates of common mental disorders, heavy drinking, and their co-occurrence among people of various sexual identities. At a population level, our results point to the importance of concomitantly addressing treatment (responding to current rates of mental disorders and heavy drinking) and prevention (addressing factors known to affect mental health and heavy drinking among LGB communities). At a clinical level, there are existing guidelines for best practices in treating LGB individuals.33
Further research is needed to develop and test explanatory models of the disparities observed in our study, including examinations of stress, coping, resilience, and other factors associated with the health of LGB people. Our sex-stratified findings highlight the utility of analyses that assess potential multiple effects of sexual identity at its intersection with other advantaged and disadvantaged social positions. It is important to remember that, despite being more likely to report mental disorders than their heterosexual peers, the majority of LGB people do not experience mental health problems. Therefore, research on how LGB people not only experience adversity, but overcome it and demonstrate significant resilience,34,35 offers further value in understanding and addressing mental health and substance misuse disparities related to sexual identity.
ACKNOWLEDGMENTS
B. Pakula is the recipient of a Canadian Institutes of Health Research (CIHR) Doctoral Research Award (Frederick Banting and Charles Best Canada Graduate Scholarship), a CIHR Institute of Gender and Health Skills Development Award, and a Killam Doctoral Scholarship. R. Carpiano contributed to this study while receiving funding from a CIHR New Investigator Award.
HUMAN PARTICIPANT PROTECTION
No protocol approval was needed for this study because the data used are legally accessible to the public and appropriately protected by law.
REFERENCES
- 1.Meyer IH. Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: conceptual issues and research evidence. Psychol Bull. 2003;129(5):674–697. doi: 10.1037/0033-2909.129.5.674. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Cochran SD, Mays VM, Sullivan JG. Prevalence of mental disorders, psychological distress, and mental health services use among lesbian, gay, and bisexual adults in the United States. J Consult Clin Psychol. 2003;71(1):53–61. doi: 10.1037//0022-006x.71.1.53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.King M, Semlyen J, Tai SS et al. A systematic review of mental disorder, suicide, and deliberate self harm in lesbian, gay and bisexual people. BMC Psychiatry. 2008;8:70. doi: 10.1186/1471-244X-8-70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Chakraborty A, McManus S, Brugha TS, Bebbington P, King M. Mental health of the non-heterosexual population of England. Br J Psychiatry. 2011;198(2):143–148. doi: 10.1192/bjp.bp.110.082271. [DOI] [PubMed] [Google Scholar]
- 5.Meyer IH, Frost DM. Minority stress and the health of sexual minorities. In: Patterson CJ, D’Augelli A, editors. Handbook of Psychology and Sexual Orientation. New York, NY: Oxford University Press; 2009. pp. 252–266. [Google Scholar]
- 6.D’Augelli AR, Grossman AH, Starks MT. Childhood gender atypicality, victimization, and PTSD among lesbian, gay, and bisexual youth. J Interpers Violence. 2006;21(11):1462–1482. doi: 10.1177/0886260506293482. [DOI] [PubMed] [Google Scholar]
- 7. Statistics Canada. Police-reported hate crimes, 2012. Available at: http://www.statcan.gc.ca/daily-quotidien/140626/dq140626b-eng.htm. Accessed February 20, 2016.
- 8.Pakula B, Shoveller JA. Sexual orientation and self-reported mood disorder diagnosis among Canadian adults. BMC Public Health. 2013;13(1):209. doi: 10.1186/1471-2458-13-209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Skinner WJW, Grady CPO, Bartha C, Parker C. Concurrent Substance Use and Mental Health Disorders: An Information Guide. Toronto, Ontario, Canada: Centre for Addiction and Mental Health; 2004. [Google Scholar]
- 10. Public Health Agency of Canada. Report on the state of public health in Canada 2012: influencing health—the importance of health and gender. Available at: http://www.phac-aspc.gc.ca/cphorsphc-respcacsp/2012/index-eng.php. Accessed February 20, 2016.
- 11.Tjepkema M. Health care use among gay, lesbian and bisexual Canadians. Health Rep. 2008;19(1):53–64. [PubMed] [Google Scholar]
- 12.Institute of Medicine. The Health of Lesbian, Gay, Bisexual, and Transgender People: Building a Foundation for Better Understanding. Washington, DC: National Academies Press; 2011. [PubMed] [Google Scholar]
- 13.Wang J, Hausermann M, Ajdacic-Gross V, Aggleton P, Weiss MG. High prevalence of mental disorders and comorbidity in the Geneva Gay Men’s Health Study. Soc Psychiatry Psychiatr Epidemiol. 2007;42(5):414–420. doi: 10.1007/s00127-007-0190-3. [DOI] [PubMed] [Google Scholar]
- 14. Statistics Canada. Canadian Community Health Survey: annual component (CCHS). Available at: http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&SDDS=3226. Accessed February 20, 2016.
- 15.Thomas S, Tremblay S. Interpreting estimates from the redesigned Canadian Community Health Survey (CCHS). Available at: http://www23.statcan.gc.ca/imdb-bmdi/document/3226_D44_T9_V2-eng.pdf. Accessed February 20, 2016.
- 16.Thomas S, Wannell B. Combining cycles of the Canadian Community Health Survey. Health Rep. 2009;20(1):53–58. [PubMed] [Google Scholar]
- 17. Statistics Canada. Heavy drinking, 2013. Available at: http://www.statcan.gc.ca/pub/82-625-x/2014001/article/14019-eng.htm. Accessed February 20, 2016.
- 18.Hatzenbuehler ML. How does sexual minority stigma “get under the skin”? A psychological mediation framework. Psychol Bull. 2009;135(5):707–730. doi: 10.1037/a0016441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Gagné C, Roberts G, Keown L. Weighted estimation and bootstrap variance estimation for analyzing survey data: how to implement in selected software. Available at: http://www.statcan.gc.ca/pub/12-002-x/2014001/article/11901-eng.htm. Accessed February 20, 2016.
- 20.Roberts G, Binder D. Analyses based on combining similar information from multiple surveys. Available at: https://www.amstat.org/sections/srms/proceedings/y2009/Files/303934.pdf. Accessed February 20, 2016.
- 21.Bostwick WB, Boyd CJ, Hughes TL, McCabe SE. Dimensions of sexual orientation and the prevalence of mood and anxiety disorders in the United States. Am J Public Health. 2010;100(3):468–475. doi: 10.2105/AJPH.2008.152942. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Meyer IH, Ouellette SC, Haile R, McFarlane TA. “We’d be free”: narratives of life without homophobia, racism, or sexism. Sex Res Social Policy. 2011;8(3):204–214. doi: 10.1007/s13178-011-0063-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Taylor P. A Survey of LGBT Americans: Attitudes, Experiences and Values in Changing Times. Washington, DC: Pew Research Center; 2013. [Google Scholar]
- 24.Herek GM, Gillis JR, Cogan JC. Psychological sequelae of hate-crime victimization among lesbian, gay, and bisexual adults. J Consult Clin Psychol. 1999;67(6):945–951. doi: 10.1037//0022-006x.67.6.945. [DOI] [PubMed] [Google Scholar]
- 25.Almeida J, Johnson RM, Corliss HL, Molnar BE, Azrael D. Emotional distress among LGBT youth: the influence of perceived discrimination based on sexual orientation. J Youth Adolesc. 2009;38(7):1001–1014. doi: 10.1007/s10964-009-9397-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Schnurr PP, Green BL. Understanding relationships among trauma, post-traumatic stress disorder, and health outcomes. Adv Mind Body Med. 2004;20(1):18–29. [PubMed] [Google Scholar]
- 27.Alessi EJ, Martin JI, Gyamerah A, Meyer IH. Prejudice-related events and traumatic stress among heterosexuals and lesbians, gay men and bisexuals. J Aggress Maltreat Trauma. 2013;22(5):510–526. [PMC free article] [PubMed] [Google Scholar]
- 28.Bostwick W, Hequembourg AL. Minding the noise: conducting health research among bisexual populations and beyond. J Homosex. 2013;60(4):655–661. doi: 10.1080/00918369.2013.760370. [DOI] [PubMed] [Google Scholar]
- 29.Bostwick W, Hequembourg A. “Just a little hint”: bisexual-specific microaggressions and their connection to epistemic injustices. Cult Health Sex. 2014;16(5):488–503. doi: 10.1080/13691058.2014.889754. [DOI] [PubMed] [Google Scholar]
- 30.Ross LE, Dobinson C, Eady A. Perceived determinants of mental health for bisexual people: a qualitative examination. Am J Public Health. 2010;100(3):496–502. doi: 10.2105/AJPH.2008.156307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Gates GJ. How many people are lesbian, gay, bisexual, and transgender? Available at: https://escholarship.org/uc/item/09h684x2. Accessed February 20, 2016.
- 32.Allison P. Missing Data. Thousand Oaks, CA: Sage Publications; 2002. [Google Scholar]
- 33.American Psychological Association. Guidelines for psychological practice with lesbian, gay, and bisexual clients. Am Psychol. 2012;67(1):10–42. doi: 10.1037/a0024659. [DOI] [PubMed] [Google Scholar]
- 34.Herrick AL, Egan JE, Coulter RWS, Friedman MR, Stall R. Raising sexual minority youths’ health levels by incorporating resiliencies into health promotion efforts. Am J Public Health. 2014;104(2):206–210. doi: 10.2105/AJPH.2013.301546. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Riggle EDB, Whitman JS, Olson A, Rostosky SS, Strong S. The positive aspects of being a lesbian or gay man. Prof Psychol Res Pr. 2008;39(2):210–217. [Google Scholar]