Advertisement

Archives of Sexual Behavior

, Volume 48, Issue 1, pp 89–111 | Cite as

A Systematic Review and Meta-Analysis of Disparities in the Prevalence of Suicide Ideation and Attempt Among Bisexual Populations

  • Travis SalwayEmail author
  • Lori E. Ross
  • Charles P. Fehr
  • Joseph Burley
  • Shayan Asadi
  • Blake Hawkins
  • Lesley A. Tarasoff
Special Section: Bisexual Health

Abstract

Sexual minorities are at increased risk of suicide; however, it is unclear whether there are within-sexual minority differences in risk across specific sexual identities—notably between bisexual and lesbian/gay subgroups. We therefore conducted a systematic review and meta-analysis to quantify associations between bisexual identity and self-reported suicide ideation and attempt and the moderation of these associations by gender/sex, age, sampling strategy, and measurement of sexuality. Abstracts and full texts were independently screened by two reviewers, resulting in a total of 46 studies that met inclusion criteria and reported 12-month or lifetime prevalence estimates for suicide ideation or attempt. A consistent gradient was observed across all four outcomes, whereby bisexual respondents reported the highest proportion of suicide ideation or attempt, lesbian/gay respondents the next highest proportion, and heterosexual respondents the lowest proportion. Random-effects meta-analysis comparing bisexual individuals with lesbian/gay individuals yielded odds ratios (ORs) ranging between 1.22–1.52 across the four outcomes examined. Between-study variability in ORs was large. Thirty-one percent of heterogeneity was explained by sample type (e.g., probability vs. non-probability) and 17% by gender/sex. ORs were consistently larger for women (range: 1.48–1.95, all statistically significant at p < .05) than for men (range: 1.00–1.48, all p > .05), suggesting that gender/sex moderates the association between bisexual identity and suicide risk. Within-sexual minority differences in suicide risk may be attributed to structural and interpersonal experiences of monosexism, bisexual erasure and invisibility, or lack of bisexual-affirming social support, each of which may be experienced differently across gender/sex identities.

Keywords

Suicide Bisexual Sexual orientation Mental health Meta-analysis 

Introduction

Empirical evidence of inequity in suicide-related outcomes among sexual minorities (i.e., bisexual, lesbian, and gay people) relative to heterosexual comparators is robust. Four systematic reviews and meta-analyses—combining data from over 75 primary research studies from several North American and Western European countries—collectively indicate that sexual minorities are two to four times more likely than heterosexuals to have attempted suicide (Hottes, Bogaert, Rhodes, Brennan, & Gesink, 2016; King et al., 2008; Marshal et al., 2011; Plöderl & Tremblay, 2015). This large body of epidemiological between-group research (i.e., comparing sexual minorities to heterosexuals) has led to a growing consensus about the need for more intensive research and interventions focused on preventing sexual minority suicide (Haas et al., 2011; Plöderl et al., 2013).

Do Bisexual Populations Experience Higher Risk of Suicide than do Lesbian and Gay Populations?

Sexual minorities are not a monolithic population (Weiss, 2003); analyses of lesbian, gay, bisexual, and transgender (LGBT) persons in aggregate obscure potentially important differences in suicide risk across specific sexual minority identities (Bostwick & Hequembourg, 2013; Salway et al., 2017). Although calls to stratify analyses of sexual minority populations by sexual identity have been successful in drawing attention to the importance of within-sexual minority population differences (Bostwick & Hequembourg, 2013), many studies continue the practice of reporting sexual minority results in aggregate, often because of limited sample sizes (Institute of Medicine, 2011). More work is needed to disaggregate quantitative data for bisexual groups within the body of sexual minority health research (Ross et al., 2017).

While all four of the above-cited reviews focused primary analysis and interpretation on sexual minorities in aggregate, subgroup analyses signal an elevated risk of suicide in bisexual populations as compared with lesbian and gay populations. First, a meta-analysis of adolescents found that bisexual status moderated the observed association between sexual minority status and suicidality (odds ratio [OR] = 4.92, 95% confidence interval [CI] 2.82, 8.59 for bisexual adolescents, versus OR = 1.87, 95% CI 1.35, 2.58 for non-bisexual sexual minority adolescents) (Marshal et al., 2011). A second meta-analysis, of data from adults, included bisexual identity in mixed effects meta-regression analysis and found that bisexual adults had a significantly higher lifetime prevalence of suicide attempt than lesbian and gay adults (average absolute difference: + 7%) (Hottes et al., 2016a). A third review of suicide attempts by sexual orientation found that “the majority of studies [reviewed] reported the largest effects for bisexual individuals, compared to homosexuals” (Plöderl & Tremblay, 2015). While two of these studies examined a moderating effect of bisexuality (Hottes et al., 2016a; Marshal et al., 2011), no meta-analysis to date has quantified the burden of suicide ideation or attempts specifically among bisexual people.

One additional systematic review of suicide-related outcomes—without quantitative meta-analysis—focused specifically on bisexual populations and found that bisexual groups were more likely to report suicide ideation and attempt than heterosexual groups in 13 of 15 studies examined (Pompili et al., 2014). Results comparing bisexual groups to lesbian and gay groups, however, were mixed, with 5 studies showing increased rates of suicide ideation or attempt in bisexual groups relative to lesbian and gay groups, 1 showing decreased rates, and 13 showing statistically equivalent or inconclusive differences in rates. This review did not meta-analyze differences in suicide-related outcomes by sexual identity and was unable to offer any interpretation for the heterogeneity in effect estimates from these 19 studies.

Several hypotheses may explain the higher rate of suicide ideation or attempt among bisexual groups, relative to lesbian and gay groups, observed in some of these studies. First, bisexual people may experience additional forms of sexual orientation-based stigma above and beyond those experienced by lesbian and gay people. For example, bisexual people encounter various forms of biphobia—i.e., rejection, prejudice, and discrimination on the basis of social attitudes and beliefs about bisexuality—and monosexism—the limits and barriers that accompany a structural assumption of monosexuality (sexual orientation to a single gender)—in their everyday lives (Israel & Mohr, 2004; Ross, Dobinson, & Eady, 2010). A robust and growing literature on how sexual minority stress “gets under the skin” offers a plausible pathway between bisexual status and adverse mental health outcomes (Hatzenbuehler, 2009; Meyer, 2003). In short, minority stress models predict that various forms of biphobia and monosexism will accumulate across the life span of bisexual individuals, ultimately creating emotional and cognitive dysfunction that may manifest as depression, anxiety, or in some cases, suicide ideation or attempt.

A second possible mechanism for higher rates of suicide-related outcomes in bisexual people is the pervasive erasure and invisibility of bisexual people in society. Yoshino (2001) described bisexual invisibility as the result of an unconscious “epistemic contract” between gay and straight people, who have mutual interests in maintaining the primacy of monosexual assumptions and practices including “stable,” binary sexual orientation and gender categories. This phenomenon may contribute to an internalized sense of invisibility among bisexual persons, which in turn may lead to what Joiner (2007) has called “thwarted belongingness”—one of the components of suicidal desire in the interpersonal-psychological theory of suicide (with analogues of social alienation and marginalization existing in other prominent theories of suicide).

Third, a structural deficit in bisexual-affirmative social and healthcare support is a by-product of the social processes of biphobia, monosexism, and bisexual invisibility and erasure. For example, in Canada, 29% of bisexual women, as compared with 20% of lesbian women, and 18% of bisexual men, as compared with 14% of gay men, report an unmet healthcare need in the past 12 months (Tjepkema, 2008). This lack of support may also directly cause or contribute to well-established risk factors for suicide among those who would otherwise seek and receive help; social isolation is indeed one of the strongest and most consistently associated predictors of suicide ideation and attempt—in both heterosexual and sexual minority populations (Nock et al., 2008; Salway et al., 2017; Van Orden et al., 2010). Furthermore, not having access to supportive and sexual minority-affirming healthcare may increase the duration of suicide ideation or the repetition of suicide attempts.

Quantifying comparative estimates of suicide-related outcomes between bisexual and lesbian and gay groups is important as it can provide an understanding of the degree to which sexual identity-specific experiences are driving rates of suicide, as well as the opportunity to tailor suicide prevention approaches within sexual minority communities. Such prevention work may in fact look very different in bisexual communities than in gay or lesbian communities. For example, sexual minority mental health researchers have proposed sexual minority-affirming cognitive-behavioral therapy; however, these approaches have been developed using samples predominantly constituted by gay men (Millar, Wang, & Pachankis, 2016; Pachankis, Hatzenbuehler, Rendina, Safren, & Parsons, 2015), and there is a strong rationale to expect that minority stress-related experiences may differ between gay and bisexual people, as discussed above (Persson & Pfaus, 2015). Quantifying the difference in rates of suicide-related outcomes between bisexual and lesbian and gay groups can help motivate and justify bisexual-specific research for suicide interventions such as these. Effectively delivering suicide prevention interventions to bisexual people will likely require service provision outside of traditional LGBT venues and a critical re-evaluation of the well-intentioned practice of discussing LGBT health promotion and disease prevention as a unified and undifferentiated task (Haas et al., 2011; Weiss, 2003).

What Factors Moderate the Association Between Bisexuality and Suicide-Related Outcomes?

Meta-analyses that are stratified by social subgroups within the bisexual population (e.g., age, gender/sex) and by methodological aspects related to how bisexual individuals are sampled and measured can further help as a first step toward identifying bisexual-specific causal pathways to suicide related to these particular social categories. Indeed, the heterogeneous results in comparative estimates of suicide ideation and attempt between bisexual and lesbian or gay groups highlighted in the Pompili et al. (2014) review may be explained by some of these measurable aspects of within-bisexual population diversity. Below, we briefly review each of these hypothesized factors and consider how they may act as moderators of within-sexual minority disparities in suicide-related outcomes and thus signal potential points of intervention.

Gender/Sex

Gender fundamentally patterns experiences of socialization, exposure to stigma and stressors, and access to resources (Herek, Norton, Allen, & Sims, 2010; Katz-Wise & Hyde, 2012), and modifies experiences of sexual identity development and expression (Savin-Williams & Diamond, 2000). Meanwhile, sexual identity affects perceived and performed gender roles (Diamond & Butterworth, 2008). The intersecting effects of gender/sex and sexual identities are notably not always predictable (Katz-Wise, 2015; Veenstra, 2011). In this paper, we use the term gender/sex, as discussed by van Anders (2015), given that a gender or sex construct is infrequently defined when quantitative sexual minority population health studies report differences between men/male and women/female gender/sex groups, and given that few general population probability surveys include separate measures for gender and sex (Ross et al., 2017).

With regard to suicide, a persistent “gender paradox” is observed in most countries whereby rates of suicide ideation and attempt are higher in women, while rates of suicide mortality are higher in men (Rhodes et al., 2014). Differences in suicide-related outcomes between bisexual men and women have not been widely investigated, though one systematic review and meta-analysis found evidence of effect modification by gender when comparing rates of depression and anxiety between bisexual and lesbian/gay populations; that is, both bisexual men and women had higher rates of depression and anxiety than gay men and lesbian comparators, but the difference between bisexual and lesbian/gay groups was more pronounced among women (Ross et al., 2017). Given that depression is a well-established risk factor for suicide (Nock et al., 2008), we may expect a similar gender/sex-related pattern in meta-analysis of suicide ideation and attempt.

Age

The relationships between age and sexual identity development, and age and suicide-related outcomes, are likewise complex yet fundamental to research on suicide and sexual orientation. During adolescence, many sexual minorities will begin to develop and engage with their own sexuality, while also experiencing acute and overt stigma (e.g., bullying and harassment from peers) (Saewyc, 2011; Savin-Williams & Cohen, 2007). Adolescence is also the age when suicide ideation and attempts are most frequent (Nock et al., 2008). Thus, it is perhaps not surprising that some researchers have found the greatest disparities in suicide ideation by sexual orientation during adolescent years (Russell & Toomey, 2012). Relative to adolescence, young adulthood is associated with some reduction in exposure to sexual stigma (Ferlatte, Hottes, Trussler, & Marchand, 2013)—whether through age-related changes in peer behavior, changes in physical location, or changes in social circles. For these reasons, we hypothesize that any disparity in suicide risk between bisexual and lesbian/gay populations may be highest during the adolescent period.

Sample Type

Recruiting sexual minorities—regardless of specific sexual identities—is a perennial challenge for sexual minority health researchers (Binson et al., 2007; Meyer & Wilson, 2009). In the absence of a defined sampling frame, researchers predominantly rely on non-probability sampling via LGBT community venues to recruit study participants (hereafter termed “LGBT non-probability” samples) (Brennan, Bauer, Bradley, & Tran, 2017). These LGBT non-probability samples, however, estimate a substantially higher burden of lifetime suicide attempt (20, 95% CI 18, 22%) than general population probability surveys do for LGBT respondents (11, 95% CI 8, 15%) (Hottes et al., 2016a). Differences in burden of disease estimates by sample type may be further modified when examining within-sexual minority disparities (e.g., comparing bisexual groups to lesbian or gay groups). First, many bisexual persons may not identify with a broader LGBT community nor frequent LGBT-branded venues; by consequence, probability samples often include proportionately more bisexual respondents than do LGBT non-probability samples (Kuyper, Fernee, & Keuzenkamp, 2016; Prah et al., 2016; Weiss, 2003). Second, those bisexual persons who would be recruited through LGBT non-probability sampling differ from bisexual respondents to general population surveys with regard to demographics, social characteristics, aspects of social support, and mental health concerns—including risk of suicide (Hartman, 2011; Hottes et al., 2016b; Persson & Pfaus, 2015). In sum, sampling bias must be accounted for in order to achieve accurate estimates of the burden of suicide ideation and attempts among bisexual people. On the basis of previous research showing larger aggregate sexual minority/heterosexual disparities in suicide attempts in non-probability samples than in general population probability samples (Hottes et al., 2016a), we may hypothesize that non-probability samples will similarly show larger differences in within-sexual minority disparities than will probability samples.

Measurement

Three distinct—albeit overlapping—constructs are commonly used to identify sexual minorities: (1) sexual identity, i.e., an adopted label used to signify membership in a particular social group; (2) sexual behavior, i.e., classification by the gender(s)/sex(es) of persons with whom one has sex (typically measured in a binary fashion); and (3) sexual attraction, i.e., classified by the gender(s)/sex(es) of persons to whom one experiences sexual attraction (Eliason & Schope, 2007; Horowitz & Newcomb, 2001; Sell, 2007). These three constructs are frequently conflated by researchers aiming to make inferences about sexual minorities in the context of imperfect population health survey measures (e.g., federally funded surveys may include sexual behavior but not identity or vice versa); however, there is evidence that these distinct constructs may have distinct associations with particular health outcomes (Bauer & Brennan, 2013; Institute of Medicine, 2011). In particular, behavioral measures of bisexuality are associated with a bias caused by the inherent requirement that those classified as behaviorally bisexual must have had at least two sex partners in the recall period (e.g., past year), while those classified as behaviorally and “exclusively” heterosexual or homosexual may have had only one partner; the effect of this bias is a tendency to over-estimate differences in health outcomes between bisexual groups and gay/lesbian comparators (Bauer & Brennan, 2013). We thus similarly anticipate a larger disparity in suicide-related outcomes between bisexual and gay/lesbian groups when measured by behavior as compared with identity-based measures.

Summary and Study Objective

Evidence for a disparity in suicide risk between sexual minorities (in aggregate) and heterosexuals is robust; however, it remains unclear whether there are within-sexual minority differences in risk across specific sexual identities—notably between bisexual and lesbian/gay subgroups. Heterogeneity in effect measures comparing these subgroups may be explained by between-sample differences in gender/sex, age, and study methods. We therefore conducted a systematic review and meta-analysis with the following objectives: (1) to quantify associations between bisexual identity and self-reported suicide ideation and attempt; and (2) to examine moderation of these associations by gender/sex, age, sampling strategy, and measurement of sexuality.

Method

Literature Search and Selection

We report systematic review methods below, following the items recommended in the PRISMA consensus statement (where applicable) (Moher, Liberati, Tetzlaff, & Altman, 2009). First, we searched Medline, PsychINFO, CINAHL, LGBT Life, and Scopus, from 1995 to December 15, 2016, combining the following subject headings (SH) and keywords: bisexuality (SH) OR bisexual* (title or abstract), AND suicide (SH) OR “suicide, attempted” (SH) OR “suicidal ideation” (SH) OR “suicid*” (title or abstract). We chose 1995 as the earliest date for our search because of significant changes in sampling methodologies after this date—notably the availability of probability samples that include bisexual-identified respondents (Kaestle & Ivory, 2012)—and because a previous systematic review (without meta-analysis) on this topic failed to identify any bisexual-specific studies with suicide-related outcomes prior to 1995 (Pompili et al., 2014). Searches were simultaneously performed for other mental health outcomes of interest, including depression, anxiety, and substance use, thus enabling identification of additional articles that report suicide-related outcomes as part of a broader mental health investigation but do not list suicide in the title or abstract. Results for depression and anxiety outcomes are summarized in a separate publication (Ross et al., 2017). We also manually reviewed the references of previously published systematic reviews on suicide-related outcomes among sexual minorities to identify articles that were absent from database searches (Hottes et al., 2016a; King et al., 2008; Marshal et al., 2011; Plöderl & Tremblay, 2015; Pompili et al., 2014).

Inclusion Criteria

Two co-authors independently reviewed each abstract and full text for inclusion. Disagreements were resolved by the senior author. English, French, or Spanish language primary research studies published in peer-reviewed journals were included, provided they reported one of the following four outcomes of interest. These outcomes were selected because of the relatively high frequency of use across studies: lifetime suicide ideation, suicide ideation in last 12 months, lifetime suicide attempt, or suicide attempt in last 12 months. We excluded studies that reported: suicide ideation or attempt using a different recall period, given the lack of comparability with the 12-month recall period; other suicide-related communications or behaviors (e.g., suicide plan), given the relative infrequency of these outcomes; suicide scales, given inconsistency in constructs and scales used across identified studies; and self-harm, given that this is conceptually distinct from suicide ideation and behavior (Silverman, Berman, Sanddal, O’Carroll, & Joiner, 2007). We also excluded studies that failed to report bisexual-specific results, and those that recruited participants from specialized settings that may be associated with a higher burden of suicide ideation and attempts (e.g., mental health clinics or HIV care clinics).

Data Extraction

The following data were independently extracted by two co-authors, with discrepancies checked and resolved by the first author. The number of events (suicide ideation and/or attempt) and sample size (denominator) were separately recorded for bisexual, lesbian, gay, and heterosexual subgroups, as available for each study. Other sexual identities (e.g., queer, pansexual, mostly heterosexual) were not consistently measured or reported and were therefore excluded from the present analysis. The number of events was estimated for studies that only reported proportions by multiplying the proportion of respondents reporting suicide ideation or attempt by the subgroup denominator. Where results were stratified by gender/sex, age, sample type, or country, separate data were extracted.

Sample-level characteristics were also recorded, in order to be analyzed as possible moderators of outcomes. Gender/sex was categorized as male, female, or gender/sex-pooled (where not reported separately by gender/sex). Transgender identities were rarely reported, and where reported were not further disaggregated by bisexual (or other) sexual identity. Adolescent samples were defined as those exclusively < 21 years of age, young adult samples as those exclusively < 30 years of age—which in some cases may have also included some participants < 21 years of age—and adult samples as those including anyone ≥ 30 years of age. Sample type was coded into a three-level categorical variable. LGBT non-probability samples were defined as venue and/or event-based (including online) samples obtained through LGBT venues or media. General population probability samples were defined as randomly selected samples based on defined sampling frames. Network-based samples and other (non-LGBT) non-probability samples were combined in a third “other” category, given the relatively small number of studies identified using these methods. Snowball and other network-based samples are similar to other non-probability methods (i.e., those that recruit through LGBT venues) in their use of non-random sampling but also distinct—particularly for the purpose of sampling bisexual individuals—in that they reach into particular social networks and thereby may sample individuals who are unlikely to show up at LGBT-branded community venues (Meyer & Wilson, 2009). For this reason, we kept these network-based samples in a separate category. Sexuality measurements were coded on the basis of identity-, behavior-, or attraction-based definitions. Region, median year of data collection, and mode of data collection (interviewer-collected or self-administered) were also recorded.

Reports of duplicate datasets (e.g., population-based surveys) were only extracted once. In these cases, we selected the report that included subgroup analysis, or else the most recent publication. Traditional risk of bias (quality assessment) scales are not well-suited to meta-analyses of sexual minority status and associated burden of disease; for example, representativeness of sample (Wells et al., 2013) is unknowable because there is no defined sampling frame of sexual minorities. We therefore instead extracted the most relevant methodological variables—sample type, sexuality definition, and mode of data collection—and evaluated the effect of these variables on meta-analytic results, as recommended by the Meta-analysis of Observational Studies in Epidemiology (M.O.O.S.E.) guidelines (Stroup et al., 2000).

Analysis

Meta-analysis was conducted in two steps. First, DerSimonian–Laird random-effects meta-analysis was used to estimate pooled proportions for each of the sexual identity subgroups—bisexual, lesbian/gay, and heterosexual—and ORs comparing bisexual to heterosexual, and bisexual to lesbian/gay groups, for each of the four outcomes (DerSimonian & Laird, 1986). Proportions were double-arcsine-transformed to constrain CIs between 0 and 1 (Freeman & Tukey, 1950). Because one of our research objectives was to explore heterogeneity, we analyzed samples at the level of subgroup, i.e., kept estimates separate when reported by gender/sex, age, sample type, or country. Crude (unadjusted) ORs were calculated using raw counts of events if available; otherwise, using crude ORs reported by the authors. Adjusted ORs were not used, given the inconsistency in covariates included in multivariable models across studies.

Second, heterogeneity in estimates comparing bisexual groups to lesbian/gay groups was explored using two methods. We generated stratified pooled proportions and odds ratios for every gender/sex, age, sample type, and sexuality definition stratum with > 1 study. Then, we used meta-regression to assess moderating effects of all four hypothesized sample-level moderators (gender/sex, age, sample type, and sexuality definition) on ORs comparing bisexual to lesbian/gay groups, while controlling for other sample-level covariates of potential importance (year of survey, region, and mode of interview) (Thompson & Higgins, 2002). The Knapp and Hartrung adjustment was applied to all models to reduce the type I error rate in the context of few studies in some moderator subgroups and large heterogeneity (Viechtbauer, 2010). Effect estimates were evaluated based on both magnitude of effect and statistical significance (CIs excluding 1 for ORs, and CIs excluding 0 for β). Because some strata had relatively few studies for some of the outcomes—and therefore higher possibility of type I error—we emphasize cross-strata differences that were consistently observed across all four outcomes in our interpretation of results. Heterogeneity was expected a priori and measured using τ2, a measure of between-study variance, and Cochran’s Q-test (p < .05 considered statistically significant); the degree to which each moderator helped to explain heterogeneity was assessed using R2, calculated as (τ null 2 modelτ model with moderator 2 /τ null model 2 ) (Borenstein, Hedges, Higgins, & Rothstein, 2009). Publication bias was assessed using funnel plots and the Egger test for asymmetry (Egger, Davey Smith, Schneider, & Minder, 1997). Analysis was conducted using the metafor package in R, version 3.3.3 (Viechtbauer, 2010). This systematic review protocol was registered in the PROSPERO database: CRD42016053991.

Results

Study Selection

A total of 1074 full texts were screened for eligibility, of which 46 met the inclusion criteria (Fig. 1; Supplementary materials). The most common reason for exclusion was not reporting bisexual-specific data (57% of excluded studies). There were a total of 66 samples or subsamples across these studies, from which data were extracted. The majority of samples were derived from general population surveys and used identity-based definitions of sexuality (Table 1). No single outcome was consistently reported across more than half of the studies, and each outcome was reported in between 15 and 31 samples.
Fig. 1

Flowchart of studies screened and included in systematic review and meta-analysis of suicide-related outcomes in bisexual populations

Table 1

Characteristics of samples included in systematic review and meta-analysis of suicide-related outcomes in bisexual populations, N = 66 subgroups/samples, derived from 46 publications

Characteristic

N (%), or median (range)

Gender/sex

 Male

27 (41%)

 Female

23 (35%)

 Gender-/sex-pooled

16 (24%)

Age

 Adolescents

18 (27%)

 Young adults

13 (20%)

 Adults

35 (53%)

Sample type

 LGBT non-probability

16 (24%)

 General population probability

43 (65%)

 Other

7 (11%)

Sexuality definition

 Identity

52 (79%)

 Behavior

7 (11%)

 Attraction

7 (11%)

Median year of survey

2007 (1992–2014)

Region

 United States of America

32 (48%)

 Canada

9 (14%)

 Australia and New Zealand

8 (12%)

 Western and Northern Europe

13 (20%)

 Japan and South Korea

4 (6%)

Mode of data collection

 Interviewer

20 (30%)

 Self-administered

46 (70%)

Outcomes

 Suicide ideation, lifetime

15 (23%)

 Suicide ideation, last 12 months

27 (41%)

 Suicide attempt, lifetime

31 (47%)

 Suicide attempt, last 12 months

25 (38%)

Pooled Estimates

Pooled estimates were used to address our first objective of estimating differences in suicide-related outcomes between bisexual, heterosexual, and gay/lesbian groups; these estimates included all studies that reported each of the respective outcomes. A consistent gradient was observed across all four outcomes, whereby bisexual respondents reported the highest proportion of suicide ideation or attempt, lesbian/gay respondents the next highest proportion, and heterosexual respondents the lowest proportion (Table 2; Supplementary Figures 1–12). Proportion CIs overlapped between bisexual and lesbian/gay groups but were mutually exclusive between bisexual and heterosexual groups.
Table 2

Pooled proportion of bisexual, lesbian/gay, and heterosexual study respondents reporting suicide ideation and attempt (95% confidence intervals)

 

Suicide ideation, lifetime

Suicide ideation, last 12 months

Suicide attempt, lifetime

Suicide attempt, last 12 months

Bisexual

0.38 (0.33, 0.44)

0.21 (0.16, 0.26)

0.18 (0.15, 0.20)

0.16 (0.10, 0.22)

Lesbian/gay

0.34 (0.25, 0.43)

0.16 (0.11, 0.21)

0.15 (0.13, 0.17)

0.11 (0.08, 0.15)

Heterosexual

0.11 (0.10, 0.13)

0.07 (0.05, 0.09)

0.04 (0.04, 0.05)

0.06 (0.04, 0.09)

See supplemental figures S1–S12 for forest plots

Pooled ORs comparing suicide ideation and attempt between bisexual and heterosexual groups ranged 3.96–4.81, and all ORs were statistically significant. More specifically, the OR for lifetime suicide ideation was 4.31 (95% CI 3.23, 5.74) (Fig. 2a), the OR for 12-month suicide ideation was 3.96 (95% CI 3.17, 4.95) (Fig. 2b), the OR for lifetime suicide attempt was 4.44 (95% CI 3.52, 5.60) (Fig. 2c), and the OR for 12-month suicide attempt was 4.81 (95% CI 3.68, 6.29) (Fig. 2d). Moreover, every individual sample estimate was positive (OR > 1), and the majority of these estimates (58/64, 91%) were statistically significant. Cochran’s Q-test was significant for all four outcomes, indicating substantial between-study heterogeneity (p < .001 for all tests).
Fig. 2

Forest plots of odds ratio of a lifetime suicide ideation, b last 12-months suicide ideation, c lifetime suicide attempt, and d last 12-months suicide attempt, comparing bisexual respondents to heterosexual respondents

Pooled ORs comparing suicide ideation and attempt between bisexual and gay/lesbian groups ranged 1.22–1.52 and were statistically significant for suicide ideation in last 12 months, lifetime suicide attempt, and suicide attempt in the last 12 months, but not lifetime suicide ideation. More specifically, the OR for lifetime suicide ideation was 1.22 (95% CI 0.99, 1.51) (Fig. 3a), the OR for 12-month suicide ideation was 1.52 (95% CI 1.16, 2.00) (Fig. 3b), the OR for lifetime suicide attempt was 1.25 (95% CI 1.02, 1.54) (Fig. 3c), and the OR for 12-month suicide attempt was 1.52 (95% CI 1.20, 1.93) (Fig. 3d). The direction of individual effect estimates was mixed, though most (65/92, 71%) were positive, of which 51% (33/65) were statistically significant. Cochran’s Q-test was significant for all four outcomes, indicating substantial between-study heterogeneity (p < .001 for all tests). Funnel plots of pooled ORs comparing bisexual and gay/lesbian groups appeared symmetrical, and the Egger test was nonsignificant (p > .05) for all four outcomes (Supplementary Figures 13–16).
Fig. 3

Forest plots of odds ratio of a lifetime suicide ideation, b last 12-months suicide ideation, c lifetime suicide attempt, and d last 12-months suicide attempt, comparing bisexual respondents to lesbian/gay respondents

Stratified Estimates

Stratified estimates and meta-regression were used to address our second objective of examining factors that may explain between-sample variation in ORs examined in the results above. Stratified estimates focused on within-sexual minority comparisons, i.e., comparing suicide-related outcomes in bisexual groups relative to lesbian and gay counterparts (Tables 3, 4). Stratified pooled ORs for all four outcomes were greater in magnitude among women than among men, and were consistently statistically significant for women but not significant for men. More specifically, for lifetime suicide ideation, the OR was 1.76 among women and 1.07 among men; for 12-month suicide ideation, the OR was 1.95 among women and 1.27 among men; for lifetime suicide attempt, the OR was 1.48 among women and 1.00 among men; and for 12-month suicide attempt, the OR was 1.51 among women and 1.48 among men. The ORs for lifetime suicide ideation and attempts were greater among young adults (lifetime ideation: OR = 1.17; lifetime attempts: OR = 1.60) and adults (lifetime ideation: OR = 1.36; lifetime attempts: OR = 1.29) than among adolescents (lifetime ideation: OR = 0.91; lifetime attempts: OR = 0.73); however, a “U”-shaped trend was apparent for recent (12-month) recall periods for both outcomes, with ORs greatest among adolescents (OR = 1.74 and 1.76 for 12-month ideation and attempts, respectively) and adults (OR = 1.72 and 1.47, respectively), and lower among young adults (OR = 1.25 and 1.01, respectively). With regard to sample type, ORs were greatest in magnitude (and consistently statistically significant) for general population probability surveys (ORs range 1.50–1.69). By comparison, ORs for LGBT non-probability samples ranged 0.73–1.44, and were nonsignificant for all but one outcome (12-month suicide attempt). There were insufficient data to stratify lifetime measures by sexuality definitions other than identity. For the 12-month recall periods, ORs were positive for all three definitions but greater in magnitude for behavior and attraction than for identity.
Table 3

Pooled estimates of proportions and odds ratios of suicide ideation in bisexual respondents as compared with lesbian/gay respondents, stratified by gender/sex, age, sample type, and sexuality definition

 

Suicide ideation, lifetime

Suicide ideation, last 12 months

Proportion (95% CI), bisexual respondents

Proportion (95% CI), lesbian/gay respondents

OR (95% CI), bisexual compared with lesbian/gay respondents

N

Proportion (95% CI), bisexual respondents

Proportion (95% CI), lesbian/gay respondents

OR (95% CI), bisexual compared with lesbian/gay respondents

N

Gender/sex

 Male

0.34 (0.29, 0.40)

0.33 (0.19, 0.49)

1.07 (0.73, 1.57)

6

0.16 (0.11, 0.22)

0.14 (0.08, 0.20)

1.27 (0.88, 1.84)

10

 Female

0.42 (0.37, 0.48)

0.32 (0.26, 0.38)

1.76 (1.46, 2.12)

5

0.20 (0.13, 0.27)

0.12 (0.05, 0.23)

1.95 (1.08, 3.53)

10

Age

 Adolescents

0.34 (0.14, 0.57)

0.35 (0.14, 0.59)

0.91 (0.76, 1.08)

4

0.35 (0.27, 0.45)

0.24 (0.16, 0.34)

1.74 (1.19, 2.55)

6

 Young adults

0.42 (0.13, 0.75)

0.39 (0.12, 0.71)

1.17 (0.95, 1.44)

2

0.17 (0.13, 0.22)

0.14 (0.10, 0.19)

1.25 (0.96, 1.64)

9

 Adults

0.39 (0.34, 0.44)

0.32 (0.21, 0.44)

1.36 (1.01, 1.82)

9

0.17 (0.11, 0.24)

0.12 (0.06, 0.20)

1.72 (0.98, 3.03)

10

Sample type

 LGBT non-probability

0.39 (0.22, 0.57)

0.45 (0.31, 0.59)

0.73 (0.47, 1.11)

3

0.24 (0.13, 0.37)

0.22 (0.16, 0.29)

0.99 (0.71, 1.36)

6

 General population probability

0.39 (0.32, 0.45)

0.30 (0.23, 0.37)

1.50 (1.29, 1.74)

9

0.21 (0.14, 0.28)

0.14 (0.10, 0.19)

1.69 (1.22, 2.35)

18

 Other

0.36 (0.26, 0.47)

0.39 (0.31, 0.46)

0.99 (0.68, 1.44)

3

0.20 (0.13, 0.28)

NSD

NSD

Sexuality definition

 Identity

0.38 (0.32, 0.44)

0.33 (0.24, 0.43)

1.22 (0.97, 1.54)

13

0.22 (0.16, 0.28)

0.17 (0.12, 0.23)

1.43 (1.05, 1.96)

20

 Behavior

NSD

NSD

NSD

0.11 (0.07, 0.15)

0.08 (0.02, 0.18)

1.46 (0.70, 3.03)

2

 Attraction

NSD

NSD

NSD

0.26 (0.17, 0.36)

0.14 (0.09, 0.20)

2.47 (1.09, 5.59)

3

Statistically significant values are given in bold

OR Odds ratio, N number of studies included in subgroup analysis comparing bisexual with lesbian/gay respondents, NSD not sufficient data (< 2 studies)

Table 4

Pooled estimates of proportions and odds ratios of suicide attempt in bisexual respondents as compared with lesbian/gay respondents, stratified by gender/sex, age, sample type, and sexuality definition

 

Suicide attempt, lifetime

Suicide attempt, last 12 months

Proportion (95% CI), bisexual respondents

Proportion (95% CI), lesbian/gay respondents

OR (95% CI), bisexual compared with lesbian/gay respondents

N

Proportion (95% CI), bisexual respondents

Proportion (95% CI), lesbian/gay respondents

OR (95% CI), bisexual compared with lesbian/gay respondents

N

Gender/sex

 Male

0.12 (0.10, 0.15)

0.13 (0.10, 0.15)

1.00 (0.78, 1.28)

14

0.11 (0.04, 0.22)

0.07 (0.03, 0.13)

1.48 (0.93, 2.35)

6

 Female

0.23 (0.20, 0.26)

0.17 (0.13, 0.22)

1.48 (1.05, 2.10)

13

0.16 (0.04, 0.33)

0.10 (0.04, 0.19)

1.51 (1.14, 2.00)

5

Age

 Adolescents

0.18 (0.12, 0.27)

0.22 (0.10, 0.38)

0.73 (0.47, 1.13)

3

0.28 (0.23, 0.33)

0.19 (0.16, 0.22)

1.76 (1.26, 2.47)

13

 Young adults

0.18 (0.11, 0.25)

0.15 (0.10, 0.21)

1.60 (1.18, 2.17)

5

0.04 (0.03, 0.06)

0.04 (0.03, 0.06)

1.01 (0.63, 1.61)

3

 Adults

0.18 (0.15, 0.21)

0.14 (0.12, 0.16)

1.29 (1.01, 1.65)

23

0.03 (0.02, 0.05)

0.03 (0.02, 0.04)

1.47 (1.22, 1.77)

5

Sample type

 LGBT non-probability

0.20 (0.16, 0.25)

0.20 (0.17, 0.24)

1.06 (0.86, 1.31)

10

0.03 (0.02, 0.06)

0.03 (0.02, 0.05)

1.44 (1.20, 1.73)

6

 General population probability

0.16 (0.13, 0.20)

0.10 (0.09, 0.12)

1.59 (1.15, 2.21)

16

0.24 (0.17, 0.31)

0.16 (0.11, 0.20)

1.67 (1.23, 2.27)

15

 Other

0.17 (0.12, 0.22)

0.19 (0.11, 0.29)

0.76 (0.53, 1.09)

5

NSD

NSD

NSD

Sexuality definition

 Identity

0.17 (0.15, 0.20)

0.15 (0.13, 0.17)

1.22 (0.98, 1.52)

29

0.10 (0.05, 0.17)

0.09 (0.05, 0.13)

1.23 (1.09, 1.40)

12

 Behavior

NSD

NSD

NSD

 

0.28 (0.12, 0.48)

0.16 (0.08, 0.25)

2.02 (1.07, 3.79)

5

 Attraction

NSD

NSD

NSD

 

0.21 (0.06, 0.42)

0.12 (0.01, 0.30)

2.00 (1.22, 3.29)

4

Statistically significant values are given in bold

OR Odds ratio, N number of studies included in subgroup analysis comparing bisexual with lesbian/gay respondents, NSD not sufficient data (< 2 studies)

Meta-Regression

The number of samples by moderator subgroups was only sufficient to perform meta-regression for lifetime suicide attempt. All 29 of these subgroups measured sexuality using an identity-based definition; therefore, we were unable to include sexuality definition as a covariate in meta-regression. The only moderator that was associated with OR comparing lifetime suicide attempt between bisexual and lesbian/gay groups at p < .05 was gender/sex (Table 5). Sample type explained 31%, gender/sex 17%, and mode of data collection 8% of the between-estimate heterogeneity; other moderators did not result in lower τ2 than that of the null model. Collectively, 47% of the between-study variance was explained by the model.
Table 5

Multivariable meta-regression model of associations between study and sample characteristics and log odds ratios comparing lifetime suicide attempt between bisexual and lesbian/gay respondents

Sample characteristics (number of samples)

Univariate β (95% CI)

R 2

Multivariable β (95% CI)

Gender/sex

 Male (14)

Referent

17%

Referent

 Female (13)

0.41 (− 0.06, 0.87)

 

0.67 (0.16, 1.18)

Age

 Adolescents (2)

Referent

0%

Referent

 Young adults (3)

− 0.01 (− 1.42, 1.40)

 

− 0.57 (− 2.56, 1.42)

 Adults (22)

0.29 (− 0.71, 1.28)

 

− 0.75 (− 2.08, 0.57)

Sample type

 LGBT non-probability (8)

Referent

31%

Referent

 General population probability (14)

0.40 (− 0.08, 0.89)

 

0.40 (− 0.31, 1.12)

 Other (5)

− 0.29 (− 1.00, 0.43)

 

− 0.81 (− 1.92, 0.31)

Median year of survey (continuous)

− 0.01 (− 0.06, 0.04)

0%

− 0.01 (− 0.06, 0.05)

Region

 United States of America (11)

Referent

0%

Referent

 Canada (3)

0.06 (− 1.31, 1.44)

 

0.10 (− 1.38, 1.59)

 Australia and New Zealand (4)

− 0.24 (− 1.18, 0.69)

 

− 0.35 (− 1.57, 0.88)

 Western and Northern Europe (8)

− 0.17 (− 0.78, 0.43)

 

0.22 (−0.54, 0.99)

 Japan and South Korea (1)

− 0.14 (− 1.38, 1.10)

 

0.62 (− 0.68, 1.92)

Mode of data collection

   

 Interviewer (10)

Referent

8%

Referent

 Self-administered (17)

− 0.23 (− 0.74, 0.27)

 

− 0.41 (− 1.19, 0.36)

Statistically significant value is given in bold

R2 for multivariable model = 47%

Discussion

This was the first study to meta-analyze comparative estimates (ORs) in suicide ideation and attempt between bisexual and lesbian/gay populations. Our results indicate that bisexual populations experience an increased risk of suicide, relative to other sexual minority identities; however, between-study variability in ORs was large, suggesting that moderating variables should be accounted for when interpreting within-sexual minority risk of suicide. We specifically examined four moderators—gender/sex, age, sample type, and sexuality definition—and found evidence of effect modification for each. These moderating effects are further interpreted below, drawing from interpretations of individual included studies that exemplify the general trends observed.

Gender/Sex

Gender/sex was the largest and most consistent moderator in our analysis. Moreover, gender/sex was also the only covariate that was statistically significant in regression models, though this finding should be interpreted with caution, given the relatively small number of samples in levels of some other covariates, as shown in Tables 3 and 4 (e.g., only 3 samples measured lifetime suicide attempts among adolescents). While a gender/sex paradox is well established in the general suicide literature (Rhodes et al., 2014), this paradox does not explain the moderating effect of gender/sex on associations between bisexuality and suicide-related outcomes. In summary, we observed a greater disparity in suicide risk within sexual minority populations (i.e., in comparison with lesbian/gay populations) for bisexual women (OR range 1.48–1.95) than for bisexual men (OR range 1.00–1.48). Assuming this moderating effect is real, it likely reflects an important interaction between experiences of gender/sex and sexuality. To our knowledge, this finding is novel. Nuanced interpretation will require more in-depth studies of mental health status that allow for stratification by gender/sex and sexuality (Persson & Pfaus, 2015; Schick & Dodge, 2012). In the absence of such data, we offer tentative and speculative hypotheses for this finding below.

First, reflecting on the relative lack of within-LGBT community support available to bisexual individuals (Persson & Pfaus, 2015; Weiss, 2003), there may be reasons why bisexual women experience this lack of community support—or even within-LGBT community stigma and prejudice—differently, or in greater magnitude, than do bisexual men. In providing a “History of Biphobia,” Weiss (2003) noted that following the “gender divide” in the gay and lesbian movement, bisexual women activists’ movements occurred in immediate response to a lesbian feminist movement (a historical evolution within the intersecting systems of privilege and oppression). Weiss (2003) goes on to note that “tensions between lesbian and bisexual women are understood as much more problematic than tensions between gay and bisexual men, caused by the politics of lesbian separatism” (p. 44). That is, bisexual women’s involvement with men was, by some, seen as a betrayal to the lesbian separatist movement, in ways that do not have a parallel for bisexual and gay men. In keeping with this historical evolution of within-sexual minority community politics, bisexual women therefore may continue to experience greater LGBT community exclusion related to their bisexual identity, than do bisexual men. To our knowledge, research investigating potential differences in LGBT community support for bisexual women and bisexual men has not been conducted to test this hypothesis.

Second, patterns and effects of partnership status may differ between bisexual women and bisexual men (Dyar, Feinstein, & London, 2014; Molina et al., 2015). While both bisexual women and bisexual men are more likely than lesbian/gay comparators to be partnered with someone of a different gender (Herek et al., 2010), the intersecting effects of sexism and gender-related male privileges may result in greater stressors and fewer benefits conferred to different-gender-partnered bisexual women than to different-gender-partnered bisexual men. This would be consistent with some literature showing that heterosexual men appear to derive greater mental health-related benefits from marriage than do heterosexual women (in part due to the role of women’s care work as benefiting men while possibly conferring mental health risk to the women providing it) (Gove, 1972; Thomeer, Umberson, & Pudrovska, 2013). There is some limited data to support this hypothesis. In one recent study of gay and bisexual men, bisexual men who were partnered with women had significantly lower risk of recent suicide attempt than those who were partnered with men (AOR 0.22, 95% CI 0.07, 0.68) (Ferlatte et al., 2017). In order to fully understand the differential effects of partnership by gender/sex and sexuality, future research should include ample data to inspect the three-/four-way interactions of gender/sex, partnership (including gender of partners), and sexual identity—i.e., including bisexual women partnered with women, bisexual women partnered with men, bisexual men partnered with women, bisexual men partnered with women, and other partnerships that defy binary and monogamist assumptions.

Related to differences in partnership, some studies suggest that bisexual women report higher rates of trauma, particularly intimate partner violence, than do lesbian women—a difference not apparent among men, though research in the latter group is sparse (Hequembourg, Livingston, & Parks, 2013; Hughes et al., 2010; Walters, Chen, & Breiding, 2013). Experiences of trauma, especially in the forms of physical or sexual abuse, are in turn strongly associated with suicide ideation and attempt (O’Connor & Nock, 2014). As such, history of trauma may be an additional potential contributor to gender/sex differences in suicidality among bisexual people and should be considered in future research on this topic.

Third, given gender-/sex-related differences in sexual identity formation and concealment/disclosure across the life course (Brown, 2002; Galupo, Davis, Grynkiewicz, & Mitchell, 2014; Herek et al., 2010), the observed difference in suicide risk may in fact reflect the intersections of gender/sex, sexuality, and age. For example, gender-/sex-stratified analyses of the Canadian Community Health Survey reveal large differences in mean age between bisexual and lesbian/gay groups for women but not for men: 34.3 years (95% CI 32.6, 36.0) for bisexual women versus 40.0 years (38.0, 41.9) for lesbian women; 39.3 years (36.5, 42.1) for bisexual men versus 39.9 years (38.4, 41.4) for gay men (Brennan, Ross, Dobinson, Veldhuizen, & Steele, 2010; Steele, Ross, Dobinson, Veldhuizen, & Tinmouth, 2009). Younger age is consistently associated with greater risk of suicide ideation and attempt—in heterosexual and sexual minority samples alike (Nock et al., 2008; Russell & Toomey, 2012; Safer, 1997). Though we adjusted for age in meta-regression analysis, the age categories available for adjustment in our dataset (adolescent, young adult, adult) may not have been granular enough to account for important within-sexual minority differences, resulting in residual confounding.

Age

No consistent age-related pattern was observed in lifetime measures of suicide ideation or attempt. This lack of age-related pattern in lifetime measures may be explained by a dilution of effects when using a lifetime recall period—by definition, the lifetime experience of suicide ideation and attempt should be expected to increase cumulatively across the life span. For temporally proximal measures of suicide (i.e., last 12-month recall), however, a “U”-shaped age pattern was observed, whereby bisexual adolescents experienced the greatest disparity in suicide risk relative to lesbian and gay adolescents (OR = 1.74 for ideation, OR = 1.76 for attempt), followed closely by bisexual adults (OR = 1.72 for ideation, OR = 1.47 for attempt), with ORs closer to the null among bisexual young adults. This trend could reflect distinct age and life course-specific effects. Given the higher proportion of respondents who identify as bisexual in surveys of adolescents, as compared with surveys of adults (Denny et al., 2016; Hatzenbuehler, 2011), samples of bisexual adolescents may be more likely to include individuals who are still questioning their sexuality, which in turn is associated with psychological distress (D’Augelli, Hershberger, & Pilkington, 2001; Meyer, Teylan, & Schwartz, 2015; Wichstrom & Hegna, 2003). By contrast, the subsequent period of young adulthood (i.e., twenties) may afford some protection from the stressors of adolescence, as some bisexual individuals begin to find social connection and support in communities of bisexual and bisexual-ally peers. Lastly, the relatively higher rates of suicide ideation and attempts during adult years may reflect the cumulative effects of a chronic minority stress process—including the accumulated exposure to biphobia and monosexism.

Sample Type

We found that the within-sexual minority disparity in suicide risk was most pronounced among general population probability surveys, contrary to our a priori hypothesis. In order to interpret this effect, we look to aspects of the study methods themselves. Samples of sexual minorities derived from general population probability surveys differ from LGBT non-probability samples in several regards. First, general population surveys randomly select participants within a sampling frame that is typically defined by the geographic distribution of the total population (Meyer & Wilson, 2009). If sexual minorities follow the same geographic distribution—a doubtful assumption-given research suggesting many sexual minorities will relocate to avoid exposure to sexual minority stigma (Lewis, 2014)—then these samples could be said to be more representative than LGBT non-probability samples. Regardless, general population probability surveys likely recruit more sexual minority participants who are not attached to some form of LGBT community, due to the reliance of LGBT non-probability samples on LGBT community venues or websites (Kuyper et al., 2016). Given the long history of bisexual erasure in LGBT communities (Weiss, 2003), many bisexual individuals (and perhaps particularly those with different-gender partners) may not access LGBT community recruitment locations or may opt to not participate in LGBT-branded surveys, presuming that the study will not represent their experiences or identities. In light of the hypothesis that higher rates of suicide-related outcomes among bisexual people may reflect a lack of bisexual-affirming social support (Ross et al., 2017), bisexual respondents to probability surveys, who are less attached to an “umbrella” LGBT community, thus may exhibit greater risk of suicide due to a lower level of social support.

Sexuality Definition

Finally, we saw a modifying effect of the sexuality definition used for the 12-month recall outcomes (the only period for which sufficient data were available to analyze). For both suicide ideation and attempt, the OR comparing bisexual to lesbian/gay respondents was larger when measured using behavior or attraction than when using identity-based measures, though these differences were not statistically significant. We note that our findings are consistent with at least two studies that performed within-sample comparisons using identity, behavior, and attraction-based measures, and similarly found the greatest disparities in mental health outcomes between bisexual and other sexuality groups when using past 12-month sexual behaviors (Bostwick, Boyd, Hughes, & McCabe, 2010; Plöderl, Kralovec, & Fartacek, 2010). With regard to behavior specifically, other research has demonstrated that reliance upon behavioral measures may introduce bias, as discussed in the Introduction; for this reason, behavioral measures are likely poor proxies for studying mental health outcomes among bisexual populations (Bauer & Brennan, 2013). With regard to attraction, these measures tend to be used more frequently in surveys of adolescents (Saewyc et al., 2004); thus, the association between bisexual attraction and the OR of interest may more accurately reflect the association observed for age. In our meta-analysis, three of the five studies that relied upon measures of sexual attraction were studies of adolescents. For all of these reasons, we regard the results for analyses based on identity-based measures to be the most reliable, and those based on behavior or attraction to be at least partially biased or confounded by number of partners (in the case of behavioral measures) or age (in the case of attraction measures).

Limitations

Our results are limited primarily by matters relating to study selection and measurement. First, we regret that none of the studies included provide data on bisexual-identified transgender respondents. Given the high rates of suicide attempts in transgender populations (Bauer, Pyne, Francino, & Hammond, 2013), and the evidence we have presented here for elevated rates of suicide attempts in bisexual populations, future research examining experiences at the intersection of transgender identity and bisexuality is critical.

While our approach was systematic and our review strategy was comprehensive, we acknowledge that we may have missed some relevant studies—notably those that have not been published in peer-reviewed journals and those that report bisexual-specific data but do not include bisexual keywords in the title or abstract. In checking our search findings against those of systematic reviews using broader (“LGBT” or “sexual minority”) search strategies (Hottes et al., 2016a; King et al., 2008; Marshal et al., 2011; Plöderl & Tremblay, 2015), we found only one additional study that was not identified in our initial search. On this basis, we believe a bisexual-focused search strategy is appropriate given our objective but nonetheless acknowledge this potential limitation. Unpublished studies were not included in our review; however, this did not appear to induce bias in estimated ORs, based on funnel plots and the Egger test for asymmetry (Supplementary Figures 13–16).

We did not impose any geographic restrictions in our search because we wanted to identify as many relevant studies as possible. We ultimately were able to include two studies from “non-Western” societies: Japan and South Korea. Notwithstanding these countries’ high rates of suicide—notably as high or higher than those of Western counterparts (World Health Organization, 2008)—they have distinct social and cultural contexts that may make the samples non-comparable to other samples from North America and Western Europe. We addressed this potential limitation in two ways. First, we included region as a covariate in meta-regression (Table 5), though we did not find any significant effect of the Japan/South Korea regions. Second, we repeated our analyses excluding studies from Japan and South Korea and found no differences in the pooled proportion of bisexual respondents reporting each of the four outcomes (data not shown).

Further bias in study selection may result from the fact that we were unable to meta-analyze other suicide-related outcomes (such as continuous scales) and that none of the four outcomes we did analyze was reported in more than half of the studies. We attempted to mitigate this bias by repeating analyses across all four outcomes, where possible, and emphasizing results that were consistently observed across multiple outcomes. Finally, we were unable to contact study authors to obtain crude prevalence rates and ORs when not reported, which may have resulted in more complete data for analysis.

With regard to measurement, our primary variables of interest—bisexual status and suicide behavior—both remain stigmatized in most societies, and as such are inherently difficult to measure. Specifically, both tend to be underreported—in one study, 60% of bisexual men indicated they were unwilling to disclose their sexuality to government interviewers (Ferlatte, Hottes, Trussler, & Marchand, 2017); as many as 25% of those with a history of suicide attempt may refuse to disclose this information (Millner, Lee, & Nock, 2015; Plöderl, Kralovec, Yazdi, & Fartacek, 2011). Future analyses should apply misclassification correction techniques to understand the degree to which these forms of bias may explain observed results (Liu, Gustafson, Cherry, & Burstyn, 2009). We relied upon suicide ideation and attempts in our review because suicide mortality data in sexual minority populations is sparse (Plöderl et al., 2013). Suicide ideation and attempts, however, are imperfect proxies for suicide deaths (Klonsky, May, & Saffer, 2016), and the patterns we observed for ideation and attempts may not hold when examining suicide mortality. In light of the “gender paradox” (i.e., generally higher rates of suicide ideation and attempt among women but higher rates of suicide mortality among men), we in particular caution against extending our finding of greater within-sexual minority population disparities for women than for men to the outcome of suicide deaths. For all of these reasons, future research investigating rates of suicide mortality by specific sexual and gender identities is imperative.

With regard to age, given distinct age-related patterns in suicide-related outcomes (Nock et al., 2008), ideally, we would have used finer and mutually exclusive age categories for our meta-analysis; however, the available data did not allow such precise categorization. In addition, while we were able to analyze six covariates hypothesized to moderate the comparative effects we studied, there are many other sources of heterogeneity that we were unable to explore, as revealed by the R2 (47%) in our multivariable regression model. Such moderators may include socioeconomic status, race and ethnicity, indigeneity, and finer measures of age and geography—all of which have been associated with suicide-related outcomes in other studies (Frohlich, Ross, & Richmond, 2006; Milner, Spittal, Pirkis, & LaMontagne, 2013; Nock et al., 2008; Wray, Colen, & Pescosolido, 2011). Finally, there were too few studies reporting other measures of plurisexual identity (e.g., pansexual, queer, fluid) to analyze these groups in order to detect differences in suicide risk among identities within the “plurisexual umbrella” (Galupo et al., 2014; Mitchell, Davis, & Galupo, 2014).

Implications

Suicide prevention is possible (Zalsman et al., 2016), but effective prevention requires a comprehensive understanding of which social groups are affected and why. The present systematic review and meta-analysis is a modest first step toward future research than can derive more precise understandings of the causes of suicide among bisexual populations. We suggest the following steps be taken, first with regard to research on the topic of bisexual suicide, then with regard to policy and practice.

Given the hypotheses we outlined in the Introduction to explain suicide-related disparities between bisexual and lesbian/gay people—i.e., bisexual stigma, erasure, and lack of social support—we recommend future mediation studies to test the degree to which each of these possible pathways explains the disparity in suicide ideation and attempt among bisexual groups. Studies of Joiner’s (2007) interpersonal-psychological theory of suicide and Meyer’s (2003) minority stress theory among sexual minorities have yielded conflicting results about which constructs within each of these theories are most relevant for suicide in sexual minority populations; for example, thwarted belongingness versus perceived burden in Joiner’s theory (Baams, Grossman, & Russell, 2015; Plöderl et al., 2014; Silva, Chu, Monahan, & Joiner, 2015) and enacted/externalized versus internalized stigma in Meyer’s theory (McLaren, 2016; Plöderl et al., 2014; Salway et al., 2017). Future studies should explore these theories with bisexual samples and use bisexual-specific measures of aspects of minority stress (Ross et al., 2010). Such analyses would help to specify what kinds of interventions are most needed in order to prevent suicide among bisexual people.

Our review detected the highest rates of all four measures of suicide ideation and attempt among bisexual females, and the highest rates of recent suicide ideation and attempts among bisexual adolescents, as shown in Tables 3 and 4. Future studies should specifically focus on these subgroups to understand the particular causes of suicide among bisexual women and adolescents. Moreover, we suggest that research at this “three-way” intersection of bisexuality, gender/sex, and age, may be particularly important. While some probability surveys reveal a younger average age of bisexual women respondents as compared with lesbian women respondents (Steele et al., 2009), at least one study has found that bisexual women come out at later stages of life, on average, than do lesbian women (Rust, 1993), highlighting the complexity of the potential interactions between sexuality, gender/sex, and age. More generally, our fruitful analysis of a small number of socially relevant moderators suggests that intersectionality may be a useful organizing framework for future meta-analytic studies of the sexual minority and mental health literature (Collins, 1989; Crenshaw, 1991); such studies should aim to incorporate additional social axes, most notably race and socioeconomic position (Ghabrial & Ross, 2018). As with intersections between sexuality, gender/sex, and age, the effects of intersections with race and socioeconomic position may not be predicted based on the epidemiologic trends observed for any one of these social categories in isolation (Veenstra, 2011). For instance, while Black Americans as a total population experience lower rates of suicide than do White Americans, some studies among sexual minority populations have demonstrated the opposite trend at this intersection of race and sexuality (Lytle, Blosnich, & Kamen, 2016; Meyer et al., 2015).

We furthermore recommend two methodological advancements to aid future research concerning bisexuality and suicide. First, given that we detected the greatest differences in suicide risk for bisexual people in probability samples, we recommend that future stratified probability samples examine approaches for over-sampling bisexual respondents (and perhaps all sexual minority identity-based subgroups), to enable greater power for future within-group analyses. Second, we recommend the use of qualitative methods specific to the topic of suicide. If qualitative research on suicide in sexual minority populations is sparse (Dorais, 2004; Fenaughty & Harré, 2003; Hottes, 2016), qualitative research on suicide in bisexual populations is elusive. Qualitative methods, however, may be the optimal approach for in-depth investigation of population-specific causes and triggers of suicide (Hjelmeland, 2016) and may be particularly useful in examining suicide at the intersection of bisexuality, gender/sex, and age (Bowleg, 2008).

With regard to policy and practice, our results echo those of others calling for structural supports to validate and honor diverse sexualities (Hatzenbuehler et al., 2013). Empirical research directly linking sexually affirming laws (e.g., legalizing same-sex marriage) and school environments (e.g., queer-straight alliances and anti-gay and bisexual bullying policies) with reduced mental distress and suicidality is accumulating, leading policymakers to reconsider mandates to reduce sources of stigma and stress for sexual minorities (Hatzenbuehler et al., 2010, 2014; Justice Québec, 2016; Saewyc et al., 2014). In most cases, however, these policies do not specifically address bisexual people, thereby further contributing to erasure of bisexual people (Yoshino, 2001); future policy work addressing sexual minority mental health disparities should specifically explore policy mechanisms that will attend to bisexual-specific issues. In the meantime, more proximal measures of suicide prevention that respond to the specific needs of sexual minority communities are urgently needed. Guidelines have been established for sexual minority-affirmative counseling approaches; however, to date these models do not account for differences in sexual identity or sexuality (e.g., those related to plurisexal or bisexual identities or experiences; (American Psychological Association, 2016; Pachankis et al., 2015). Other suicide prevention interventions that specifically respond to the lives and contexts of bisexual people have yet to be imagined. The results of this review provide the quantitative evidence to bolster innovative and creative approaches to help bisexual people struggling with suicide and prevent subsequent suicide. To be successful, such approaches will need to extend beyond LGBT venues, account for intersecting identities and experiences within bisexual populations (e.g., related to gender/sex and age), and meaningfully engage members of bisexual communities (Wexler & Gone, 2012).

Notes

Compliance with Ethical Standards

Conflict of interest

The authors declare that they have no conflicts of interest.

Ethical approval

This study used secondary data; as such, ethical approval was not required.

Supplementary material

10508_2018_1150_MOESM1_ESM.docx (2 mb)
Supplementary material 1 (DOCX 2010 kb)

References

  1. American Psychological Association. (2016). Guidelines for psychological practice with lesbian, gay, and bisexual clients. http://www.apa.org/pi/lgbt/resources/guidelines.aspx. Accessed 17 July 2016.
  2. Baams, L., Grossman, A. H., & Russell, S. T. (2015). Minority stress and mechanisms of risk for depression and suicidal ideation among lesbian, gay, and bisexual youth. Developmental Psychology, 51(5), 688–696.CrossRefGoogle Scholar
  3. Bauer, G. R., & Brennan, D. J. (2013). The problem with “behavioral bisexuality”: Assessing sexual orientation in survey research. Journal of Bisexuality, 13, 148–165.  https://doi.org/10.1080/15299716.2013.782260.CrossRefGoogle Scholar
  4. Bauer, G. R., Pyne, J., Francino, M. C., & Hammond, R. (2013). Suicidality among trans people in Ontario: Implications for social work and social justice. Service Social, 59, 35–62.CrossRefGoogle Scholar
  5. Binson, D., Blair, J., Huebner, D. M., & Woods, W. J. (2007). Sampling in surveys of lesbian, gay, and bisexual people. In I. H. Meyer & M. E. Northridge (Eds.), The health of sexual minorities (p. 375–418). New York, NY: Springer.  https://doi.org/10.1007/978-0-387-31334-4_15.CrossRefGoogle Scholar
  6. Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Meta-regression. In Introduction to meta-analysis.  https://doi.org/10.1002/9780470743386.ch20.
  7. Bostwick, W. B., Boyd, C. J., Hughes, T. L., & McCabe, S. E. (2010). Dimensions of sexual orientation and the prevalence of mood and anxiety disorders in the United States. American Journal of Public Health, 100(3), 468–475.  https://doi.org/10.2105/AJPH.2008.152942.CrossRefPubMedPubMedCentralGoogle Scholar
  8. Bostwick, W., & Hequembourg, A. L. (2013). Minding the noise: Conducting health research among bisexual populations and beyond. Journal of Homosexuality, 60(4), 655–661.  https://doi.org/10.1080/00918369.2013.760370.CrossRefPubMedGoogle Scholar
  9. Bowleg, L. (2008). When Black + Lesbian + Woman ≠ Black Lesbian Woman: The methodological challenges of qualitative and quantitative intersectionality research. Sex Roles, 59(5–6), 312–325.  https://doi.org/10.1007/s11199-008-9400-z.CrossRefGoogle Scholar
  10. Brennan, D. J., Bauer, G. R., Bradley, K., & Tran, O. V. (2017). Methods used and topics addressed in quantitative health research on gay, bisexual and other men who have sex with men: A systematic review of the literature. Journal of Homosexuality, 64, 1519–1538.  https://doi.org/10.1080/00918369.2016.1247537.CrossRefPubMedGoogle Scholar
  11. Brennan, D. J., Ross, L. E., Dobinson, C., Veldhuizen, S., & Steele, L. S. (2010). Men’s sexual orientation and health in Canada. Canadian Journal of Public Health, 101(3), 255–258.  https://doi.org/10.17269/cjph.101.2361.CrossRefPubMedGoogle Scholar
  12. Brown, T. (2002). A proposed model of bisexual identity development that elaborates on experiential differences of women and men. Journal of Bisexuality, 2(4), 67–91.  https://doi.org/10.1300/J159v02n04_05.CrossRefGoogle Scholar
  13. Collins, P. H. (1989). The social construction of black feminist thought. Signs: Journal of Women in Culture and Society, 14(4), 745–773.  https://doi.org/10.1086/494543.CrossRefGoogle Scholar
  14. Crenshaw, K. (1991). Mapping the margins: Intersectionality, identity politics, and violence against women of color. Stanford Law Review, 43(6), 1241–1299.  https://doi.org/10.2307/1229039.CrossRefGoogle Scholar
  15. D’Augelli, A. R., Hershberger, S. L., & Pilkington, N. W. (2001). Suicidality patterns and sexual orientation-related factors among lesbian, gay, and bisexual youths. Suicide and Life-Threatening Behavior, 31(3), 250–264.CrossRefGoogle Scholar
  16. Denny, S., Lucassen, M. F. G., Stuart, J., Fleming, T., Bullen, P., Peiris-John, R., … Utter, J. (2016). The association between supportive high school environments and depressive symptoms and suicidality among sexual minority students. Journal of Clinical Child and Adolescent Psychology, 45(3), 248–261.CrossRefGoogle Scholar
  17. DerSimonian, R., & Laird, N. (1986). Meta-analysis in clinical trials. Controlled Clinical Trials, 7, 177–188.CrossRefGoogle Scholar
  18. Diamond, L. M., & Butterworth, M. (2008). Questioning gender and sexual identity: Dynamic links over time. Sex Roles, 59(5–6), 365–376.  https://doi.org/10.1007/s11199-008-9425-3.CrossRefGoogle Scholar
  19. Dorais, M. (2004). Dead boys can’t dance: Sexual orientation, masculinity, and suicide. Montreal, QC: McGill-Queens University Press.Google Scholar
  20. Dyar, C., Feinstein, B. A., & London, B. (2014). Dimensions of sexual identity and minority stress among bisexual women: The role of partner gender. Psychology of Sexual Orientation and Gender Diversity, 1, 441–451.  https://doi.org/10.1037/sgd0000063.CrossRefGoogle Scholar
  21. Egger, M., Davey Smith, G., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a simple, graphical test. BMJ, 315, 629–634.CrossRefGoogle Scholar
  22. Eliason, M., & Schope, R. (2007). Shifting sands or solid foundation?: Lesbian, gay, bisexual, and transgender identity formation. In I. H. Meyer & M. E. Northridge (Eds.), The health of sexual minorities (pp. 2–26). New York, NY: Springer.Google Scholar
  23. Fenaughty, J., & Harré, N. (2003). Life on the seesaw: A qualitative study of suicide resiliency factors for young gay men. Journal of Homosexuality, 45, 1–22.  https://doi.org/10.1300/J082v45n01_01.CrossRefPubMedGoogle Scholar
  24. Ferlatte, O., Hottes, T. S., Trussler, T., & Marchand, R. (2013). Evidence of a syndemic among young Canadian gay and bisexual men: Uncovering the associations between anti-gay experiences, psychosocial issues, and HIV risk. AIDS and Behavior, 18, 1256–1263.  https://doi.org/10.1007/s10461-013-0639-1.CrossRefGoogle Scholar
  25. Ferlatte, O., Hottes, T. S., Trussler, T., & Marchand, R. (2017a). Disclosure of sexual orientation by gay and bisexual men in government-administered probability surveys. LGBT Health, 4(1), 68–71.  https://doi.org/10.1089/lgbt.2016.0037.CrossRefPubMedGoogle Scholar
  26. Ferlatte, O., Salway, T., Hankivsky, O., Trussler, T., Oliffe, J., & Marchand, R. (2017b). Recent suicide attempts across multiple social identities among gay and bisexual men: An intersectionality analysis. Journal of Homosexuality.  https://doi.org/10.1080/00918369.2017.1377489.CrossRefPubMedGoogle Scholar
  27. Freeman, M. F., & Tukey, J. W. (1950). Transformations related to the angular and the square root. The Annals of Mathematical Statistics, 21, 607–611.  https://doi.org/10.1016/j.healthpol.2005.12.010.CrossRefGoogle Scholar
  28. Frohlich, K. L., Ross, N., & Richmond, C. (2006). Health disparities in Canada today: Some evidence and a theoretical framework. Health Policy, 79, 132–143.  https://doi.org/10.1016/j.healthpol.2005.12.010.CrossRefPubMedGoogle Scholar
  29. Galupo, M. P., Davis, K. S., Grynkiewicz, A. L., & Mitchell, R. C. (2014). Conceptualization of sexual orientation identity among sexual minorities: Patterns across sexual and gender identity. Journal of Bisexuality, 14(3–4), 433–456.  https://doi.org/10.1080/15299716.2014.933466.CrossRefGoogle Scholar
  30. Ghabrial, M., & Ross, L. E. (2018). Representation and erasure of bisexual people of color in mental health research. Manuscript submitted for publication.Google Scholar
  31. Gove, W. R. (1972). The relationship between sex roles, marital status, and mental illness. Social Forces, 51(1), 34–44.  https://doi.org/10.1093/sf/51.1.34.CrossRefGoogle Scholar
  32. Haas, A. P., Eliason, M., Mays, V. M., Mathy, R. M., Cochran, S. D., D’Augelli, A. R., … Russell, T. (2011). Suicide and suicide risk in lesbian, gay, bisexual, and transgender populations: Review and recommendations. Journal of Homosexuality, 58, 10–51.  https://doi.org/10.1080/00918369.2011.534038.CrossRefPubMedPubMedCentralGoogle Scholar
  33. Hartman, J. E. (2011). Finding a needle in a haystack: Methods for sampling in the bisexual community. Journal of Bisexuality, 11(1), 64–74.  https://doi.org/10.1080/15299716.2011.545306.CrossRefGoogle Scholar
  34. Hatzenbuehler, M. L. (2009). How does sexual minority stigma “get under the skin”? A psychological mediation framework. Psychological Bulletin, 135, 707–730.  https://doi.org/10.1037/a0016441.CrossRefPubMedPubMedCentralGoogle Scholar
  35. Hatzenbuehler, M. L. (2011). The social environment and suicide attempts in lesbian, gay, and bisexual youth. Pediatrics, 127(5), 896–903.  https://doi.org/10.1542/peds.2010-3020.CrossRefPubMedPubMedCentralGoogle Scholar
  36. Hatzenbuehler, M. L., Birkett, M., Van Wagenen, A., & Meyer, I. H. (2014). Protective school climates and reduced risk for suicide ideation in sexual minority youths. American Journal of Public Health, 104(2), 279–286.  https://doi.org/10.2105/AJPH.2013.301508.CrossRefPubMedPubMedCentralGoogle Scholar
  37. Hatzenbuehler, M. L., McLaughlin, K. A., Keyes, K. M., & Hasin, D. S. (2010). The impact of institutional discrimination on psychiatric disorders in lesbian, gay, and bisexual populations: A prospective study. American Journal of Public Health, 100(3), 452–459.  https://doi.org/10.2105/AJPH.2009.168815.CrossRefPubMedPubMedCentralGoogle Scholar
  38. Hatzenbuehler, M. L., Phelan, J. C., & Link, B. G. (2013). Stigma as a fundamental cause of population health inequalities. American Journal of Public Health, 103(5), 813–821.CrossRefGoogle Scholar
  39. Hequembourg, A. L., Livingston, J. A., & Parks, K. A. (2013). Sexual victimization and associated risks among lesbian and bisexual women. Violence Against Women, 19(5), 634–657.  https://doi.org/10.1177/1077801213490557.CrossRefPubMedPubMedCentralGoogle Scholar
  40. Herek, G. M., Norton, A. T., Allen, T. J., & Sims, C. L. (2010). Demographic, psychological, and social characteristics of self-identified lesbian, gay, and bisexual adults in a US probability sample. Sexuality Research and Social Policy, 7, 176–200.  https://doi.org/10.1007/s13178-010-0017-y.CrossRefPubMedGoogle Scholar
  41. Hjelmeland, H. (2016). A critical look at the current suicide research. In J. H. White, I. Marsh, M. J. Kral, & J. Morris (Eds.), Critical suicidology: Transforming suicide research and prevention for the 21st century (pp. 31–55). Vancouver, BC: UBC Press.Google Scholar
  42. Horowitz, J. L., & Newcomb, M. D. (2001). A multidimensional approach to homosexual identity. Journal of Homosexuality, 42, 1–19.  https://doi.org/10.1300/J082v42n02_01.CrossRefPubMedGoogle Scholar
  43. Hottes, T. S. (2016). Exploring pathways from sexual stigma to suicide attempts within the life course narratives of adult gay men. In 16th European symposium on suicide and suicidal behavior. Oviedo, Spain.Google Scholar
  44. Hottes, T. S., Bogaert, L., Rhodes, A. E., Brennan, D. J., & Gesink, D. (2016a). Lifetime prevalence of suicide attempts among sexual minority adults by study sampling strategies: A systematic review and meta-analysis. American Journal of Public Health, 106(5), e1–e12.  https://doi.org/10.2105/AJPH.2016.303088a.CrossRefPubMedPubMedCentralGoogle Scholar
  45. Hottes, T. S., Gesink, D., Ferlatte, O., Brennan, D. J., Rhodes, A. E., Marchand, R., & Trussler, T. (2016b). Concealment of sexual minority identities in interviewer-administered government surveys and its impact on estimates of suicide ideation among bisexual and gay men. Journal of Bisexuality, 16(4), 427–453.  https://doi.org/10.1080/15299716.2016.1225622.CrossRefGoogle Scholar
  46. Hughes, T. L., Szalacha, L. A., Johnson, T. P., Kinnison, K. E., Wilsnack, S. C., & Cho, Y. (2010). Sexual victimization and hazardous drinking among heterosexual and sexual minority women. Addictive Behaviors, 35(12), 1152–1156.  https://doi.org/10.1016/j.addbeh.2010.07.004.CrossRefPubMedPubMedCentralGoogle Scholar
  47. Institute of Medicine. (2011). The health of lesbian, gay, bisexual, and transgender people: Building a foundation for better understanding. Washington, DC: National Academies Press.Google Scholar
  48. Israel, T., & Mohr, J. J. (2004). Attitudes toward bisexual women and men. Journal of Bisexuality, 4(1), 117–134.  https://doi.org/10.1300/J159v04n01_09.CrossRefGoogle Scholar
  49. Joiner, T. (2007). Why people die by suicide (1st ed.). Cambridge, MA: Harvard University Press.Google Scholar
  50. Justice Québec. (2016). Québec policy against homophobia. http://www.justice.gouv.qc.ca/english/ministere/dossiers/homophobie/homophobie-a.htm. Accessed 17 Jan 2018.
  51. Kaestle, C. E., & Ivory, A. H. (2012). A forgotten sexuality: Content analysis of bisexuality in the medical literature over two decades. Journal of Bisexuality, 12(1), 35–48.  https://doi.org/10.1080/15299716.2012.645701.CrossRefGoogle Scholar
  52. Katz-Wise, S. L. (2015). Sexual fluidity in young adult women and men: Associations with sexual orientation and sexual identity development. Psychology & Sexuality, 6(2), 189–208.  https://doi.org/10.1080/19419899.2013.876445.CrossRefGoogle Scholar
  53. Katz-Wise, S. L., & Hyde, J. S. (2012). Victimization experiences of lesbian, gay, and bisexual individuals: A meta-analysis. Journal of Sex Research, 49, 142–167.  https://doi.org/10.1080/00224499.2011.637247.CrossRefPubMedGoogle Scholar
  54. King, M., Semlyen, J., Tai, S. S., Killaspy, H., Osborn, D., Popelyuk, D., & Nazareth, I. (2008). A systematic review of mental disorder, suicide, and deliberate self harm in lesbian, gay and bisexual people. BMC Psychiatry, 8(1), 70.  https://doi.org/10.1186/1471-244X-8-70.CrossRefPubMedPubMedCentralGoogle Scholar
  55. Klonsky, E. D., May, A. M., & Saffer, B. Y. (2016). Suicide, suicide attempts, and suicidal ideation. Annual Review of Clinical Psychology, 12(1), 307–330.  https://doi.org/10.1146/annurev-clinpsy-021815-093204.CrossRefPubMedGoogle Scholar
  56. Kuyper, L., Fernee, H., & Keuzenkamp, S. (2016). A comparative analysis of a community and general sample of lesbian, gay, and bisexual individuals. Archives of Sexual Behavior, 45(3), 683–693.CrossRefGoogle Scholar
  57. Lewis, N. M. (2014). Rupture, resilience, and risk: Relationships between mental health and migration among gay-identified men in North America. Health and Place, 27, 212–219.  https://doi.org/10.1016/j.healthplace.2014.03.002.CrossRefPubMedGoogle Scholar
  58. Liu, J., Gustafson, P., Cherry, N., & Burstyn, I. (2009). Bayesian analysis of a matched case-control study with expert prior information on both the misclassification of exposure and the exposure-disease association. Statistics in Medicine, 28(27), 3411–3423.  https://doi.org/10.1002/sim.3694.CrossRefPubMedGoogle Scholar
  59. Lytle, M. C., Blosnich, J. R., & Kamen, C. (2016). The association of multiple identities with self-directed violence and depression among transgender individuals. Suicide and Life-Threatening Behavior, 46, 535–544.  https://doi.org/10.1111/sltb.12234.CrossRefPubMedGoogle Scholar
  60. Marshal, M. P., Dietz, L. J., Friedman, M. S., Stall, R., Smith, H. A., McGinley, J., … Brent, D. A. (2011). Suicidality and depression disparities between sexual minority and heterosexual youth: A meta-analytic review. Journal of Adolescent Health, 49(2), 115–123.  https://doi.org/10.1016/j.jadohealth.2011.02.005.CrossRefPubMedGoogle Scholar
  61. McLaren, S. (2016). The interrelations between internalized homophobia, depressive symptoms, and suicidal ideation among Australian gay men, lesbians, and bisexual women. Journal of Homosexuality, 63(2), 156–168.  https://doi.org/10.1080/00918369.2015.1083779.CrossRefPubMedGoogle Scholar
  62. Meyer, I. H. (2003). Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence. Psychological Bulletin, 129, 674–697.  https://doi.org/10.1037/0033-2909.129.5.674.CrossRefPubMedPubMedCentralGoogle Scholar
  63. Meyer, I. H., Teylan, M., & Schwartz, S. (2015). The role of help-seeking in preventing suicide attempts among lesbians, gay men, and bisexuals. Suicide and Life-Threatening Behavior, 45(1), 25–36.CrossRefGoogle Scholar
  64. Meyer, I. H., & Wilson, P. A. (2009). Sampling lesbian, gay, and bisexual populations. Journal of Counseling Psychology, 129, 674–697.  https://doi.org/10.1037/a0014587.CrossRefGoogle Scholar
  65. Millar, B. M., Wang, K., & Pachankis, J. E. (2016). The moderating role of internalized homonegativity on the efficacy of LGB-affirmative psychotherapy: Results from a randomized controlled trial with young adult gay and bisexual men. Journal of Consulting and Clinical Psychology, 84(7), 565–570.  https://doi.org/10.1037/ccp0000113.CrossRefPubMedPubMedCentralGoogle Scholar
  66. Millner, A. J., Lee, M. D., & Nock, M. K. (2015). Single-item measurement of suicidal behaviors: Validity and consequences of misclassification. PLoS ONE, 10(10), e0141606.  https://doi.org/10.1371/journal.pone.0141606.CrossRefPubMedPubMedCentralGoogle Scholar
  67. Milner, A., Spittal, M. J., Pirkis, J., & LaMontagne, A. D. (2013). Suicide by occupation: Systematic review and meta-analysis. British Journal of Psychiatry, 203, 409–416.  https://doi.org/10.1192/bjp.bp.113.128405.CrossRefPubMedGoogle Scholar
  68. Mitchell, R., Davis, K., & Galupo, M. P. (2014). Comparing perceived experiences of prejudice among self-identified plurisexual individuals. Psychology & Sexuality, 6(3), 245–257.  https://doi.org/10.1080/19419899.2014.940372.CrossRefGoogle Scholar
  69. Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. PLoS Medicine, 6(7), e1000097.CrossRefGoogle Scholar
  70. Molina, Y., Marquez, J. H., Logan, D. E., Leeson, C. J., Balsam, K. F., & Kaysen, D. L. (2015). Current intimate relationship status, depression, and alcohol use among bisexual women: The mediating roles of bisexual-specific minority stressors. Sex Roles, 73, 43–57.CrossRefGoogle Scholar
  71. Nock, M. K., Borges, G., Bromet, E. J., Cha, C. B., Kessler, R. C., & Lee, S. (2008). Suicide and suicidal behavior. Epidemiologic Reviews, 30(1), 133–154.  https://doi.org/10.1093/epirev/mxn002.CrossRefPubMedPubMedCentralGoogle Scholar
  72. O’Connor, R. C., & Nock, M. K. (2014). The psychology of suicidal behaviour. Lancet Psychiatry, 1(1), 73–85.  https://doi.org/10.1016/S2215-0366(14)70222-6.CrossRefPubMedGoogle Scholar
  73. Pachankis, J. E., Hatzenbuehler, M. L., Rendina, H. J., Safren, S. A., & Parsons, J. T. (2015). LGB-affirmative cognitive-behavioral therapy for young adult gay and bisexual men: A randomized controlled trial of a transdiagnostic minority stress approach. Journal of Consulting and Clinical Psychology, 83, 875–889.  https://doi.org/10.1037/ccp0000037.CrossRefPubMedPubMedCentralGoogle Scholar
  74. Persson, T. J., & Pfaus, J. G. (2015). Bisexuality and mental health: Future research directions. Journal of Bisexuality, 15(1), 82–98.  https://doi.org/10.1080/15299716.2014.994694.CrossRefGoogle Scholar
  75. Plöderl, M., Kralovec, K., & Fartacek, R. (2010). The relation between sexual orientation and suicide attempts in Austria. Archives of Sexual Behavior, 39(6), 1403–1414.  https://doi.org/10.1007/s10508-009-9597-0.CrossRefPubMedGoogle Scholar
  76. Plöderl, M., Kralovec, K., Yazdi, K., & Fartacek, R. (2011). A closer look at self-reported suicide attempts: False positives and false negatives. Suicide and Life-Threatening Behavior, 41(1), 1–5.  https://doi.org/10.1111/j.1943-278X.2010.00005.x.CrossRefPubMedGoogle Scholar
  77. Plöderl, M., Sellmeier, M., Fartacek, C., Pichler, E., Fartacek, R., & Kralovec, K. (2014). Explaining the suicide risk of sexual minority individuals by contrasting the minority stress model with suicide models. Archives of Sexual Behavior, 43, 1559–1570.CrossRefGoogle Scholar
  78. Plöderl, M., & Tremblay, P. (2015). Mental health of sexual minorities: A systematic review. International Review of Psychiatry, 27(5), 367–385.CrossRefGoogle Scholar
  79. Plöderl, M., Wagenmakers, E. J., Tremblay, P., Ramsay, R., Kralovec, K., Fartacek, C., & Fartacek, R. (2013). Suicide risk and sexual orientation: a critical review. Archives of Sexual Behavior, 42(5), 715–727.  https://doi.org/10.1007/s10508-012-0056-y.CrossRefPubMedGoogle Scholar
  80. Pompili, M., Lester, D., Forte, A., Seretti, M. E., Erbuto, D., Lamis, D. A., … Girardi, P. (2014). Bisexuality and suicide: A systematic review of the current literature. Journal of Sexual Medicine, 11, 1903–1913.  https://doi.org/10.1111/jsm.12581.CrossRefPubMedGoogle Scholar
  81. Prah, P., Hickson, F., Bonell, C., McDaid, L. M., Johnson, A. M., Wayal, S., … Copas, A. J. (2016). Men who have sex with men in Great Britain: Comparing methods and estimates from probability and convenience sample surveys. Sexually Transmitted Infections, 92(6), 455–463.  https://doi.org/10.1136/sextrans-2015-052389.CrossRefPubMedPubMedCentralGoogle Scholar
  82. Rhodes, A. E., Boyle, M. H., Bridge, J. A., Sinyor, M., Links, P. S., Tonmyr, L., … Szatmari, P. (2014). Antecedents and sex/gender differences in youth suicidal behavior. World Journal of Psychiatry, 4(4), 120–132.  https://doi.org/10.5498/wjp.v4.i4.120.CrossRefPubMedPubMedCentralGoogle Scholar
  83. Ross, L. E., Dobinson, C., & Eady, A. (2010). Perceived determinants of mental health for bisexual people: A qualitative examination. American Journal of Public Health, 100(3), 496–502.  https://doi.org/10.2105/AJPH.2008.156307.CrossRefPubMedPubMedCentralGoogle Scholar
  84. Ross, L. E., Salway, T., Tarasoff, L. A., Hawkins, B. W., MacKay, J. M., & Fehr, C. P. (2017). Prevalence of depression and anxiety among bisexual people compared to gay, lesbian, and heterosexual individuals: A systematic review and meta-analysis. Journal of Sex Research.  https://doi.org/10.1080/00224499.2017.1387755.CrossRefPubMedGoogle Scholar
  85. Russell, S. T., & Toomey, R. B. (2012). Men’s sexual orientation and suicide: Evidence for U.S. adolescent-specific risk. Social Science and Medicine, 74(4), 523–529.  https://doi.org/10.1016/j.socscimed.2010.07.038.CrossRefPubMedGoogle Scholar
  86. Rust, P. C. (1993). “Coming out” in the age of social constructionism. Gender & Society, 7(1), 50–77.  https://doi.org/10.1177/089124393007001004.CrossRefGoogle Scholar
  87. Saewyc, E. M. (2011). Research on adolescent sexual orientation: Development, health disparities, stigma, and resilience. Journal of Research on Adolescence, 21(1), 256–272.  https://doi.org/10.1111/j.1532-7795.2010.00727.x.CrossRefPubMedPubMedCentralGoogle Scholar
  88. Saewyc, E. M., Bauer, G. R., Skay, C. L., Bearinger, L. H., Resnick, M. D., Reis, E., et al. (2004). Measuring sexual orientation in adolescent health surveys: Evaluation of eight school-based surveys. Journal of Adolescent Health, 35(4), 345.e1–345.e15.CrossRefGoogle Scholar
  89. Saewyc, E. M., Konishi, C., Rose, H. A., & Homma, Y. (2014). School-based strategies to reduce suicidal ideation, suicide attempts, and discrimination among sexual minority and heterosexual adolescents in Western Canada. International Journal of Child, Youth and Family Studies, 1, 89–112.CrossRefGoogle Scholar
  90. Safer, D. J. (1997). Adolescent/adult differences in suicidal behavior and outcome. Annals of Clinical Psychiatry, 9(1), 61–66.CrossRefGoogle Scholar
  91. Salway, T., Gesink, D., Ibrahim, S., Ferlatte, O., Rhodes, A. E., Brennan, D. J., … Trussler, T. (2017). Evidence of multiple mediating pathways in associations between constructs of stigma and self-reported suicide attempts in a cross-sectional study of adult gay and bisexual men. Archives of Sexual Behavior.  https://doi.org/10.1007/s10508-017-1019-0.CrossRefPubMedGoogle Scholar
  92. Savin-Williams, R. C., & Cohen, K. M. (2007). Development of same-sex attracted youth. In I. H. Meyer & M. E. Northridge (Eds.), The health of sexual minorities (pp. 27–47). Boston, MA: Springer.  https://doi.org/10.1007/978-0-387-31334-4_2.CrossRefGoogle Scholar
  93. Savin-Williams, R. C., & Diamond, L. M. (2000). Sexual identity trajectories among sexual-minority youths: Gender comparisons. Archives of Sexual Behavior, 29(6), 607–627.  https://doi.org/10.1023/A:1002058505138.CrossRefPubMedGoogle Scholar
  94. Schick, V., & Dodge, B. (2012). Introduction to the special issue: Bisexual health: Unpacking the paradox. Journal of Bisexuality, 12(2), 161–167.  https://doi.org/10.1080/15299716.2012.674849.CrossRefPubMedPubMedCentralGoogle Scholar
  95. Sell, R. L. (2007). Defining and measuring sexual orientation for research. In I. H. Meyer & M. E. Northridge (Eds.), The health of sexual minorities (pp. 355–374). Boston, MA: Springer.  https://doi.org/10.1007/978-0-387-31334-4_14.CrossRefGoogle Scholar
  96. Silva, C., Chu, C., Monahan, K. R., & Joiner, T. E. (2015). Suicide risk among sexual minority college students: A mediated moderation model of sex and perceived burdensomeness. Psychology of Sexual Orientation and Gender Diversity, 2, 22–33.CrossRefGoogle Scholar
  97. Silverman, M. M., Berman, A. L., Sanddal, N. D., O’Carroll, P. W., & Joiner, T. E. (2007). Rebuilding the tower of babel: A revised nomenclature for the study of suicide and suicidal behaviors part 1: Background, rationale, and methodology. Suicide and Life-Threatening Behavior, 37(3), 248–263.CrossRefGoogle Scholar
  98. Steele, L. S., Ross, L. E., Dobinson, C., Veldhuizen, S., & Tinmouth, J. M. (2009). Women’s sexual orientation and health: Results from a Canadian population-based survey. Women and Health, 49(5), 353–367.CrossRefGoogle Scholar
  99. Stroup, D. F., Berlin, J. A., Morton, S. C., Olkin, I., Williamson, G. D., Rennie, D., Thacker, … S. B. (2000). Meta-analysis of observational studies in epidemiology: A proposal for reporting. JAMA, 283, 2008–2012.CrossRefGoogle Scholar
  100. Thomeer, M. B., Umberson, D., & Pudrovska, T. (2013). Marital processes around depression. Society and Mental Health, 3(3), 151–169.  https://doi.org/10.1177/2156869313487224.CrossRefPubMedPubMedCentralGoogle Scholar
  101. Thompson, S. G., & Higgins, J. P. T. (2002). How should meta-regression analyses be undertaken and interpreted? Statistics in Medicine, 21, 1559–1573.CrossRefGoogle Scholar
  102. Tjepkema, M. (2008). Health care use among gay, lesbian and bisexual Canadians. Health Reports, 19(1), 53–64. http://www.statcan.gc.ca/pub/82-003-x/2008001/article/10532-eng.htm.
  103. van Anders, S. M. (2015). Beyond sexual orientation: Integrating gender/sex and diverse sexualities via Sexual Configurations Theory. Archives of Sexual Behavior, 44, 1177–1213.  https://doi.org/10.1007/s10508-015-0490-8.CrossRefPubMedGoogle Scholar
  104. Van Orden, K. A., Witte, T. K., Cukrowicz, K. C., Braithwaite, S. R., Selby, E. A., & Joiner, T. E. (2010). The interpersonal theory of suicide. Psychological Review, 117(2), 575–600.  https://doi.org/10.1037/a0018697.CrossRefPubMedPubMedCentralGoogle Scholar
  105. Veenstra, G. (2011). Race, gender, class, and sexual orientation: intersecting axes of inequality and self-rated health in Canada. International Journal for Equity in Health, 10(1), 3.  https://doi.org/10.1186/1475-9276-10-3.CrossRefPubMedPubMedCentralGoogle Scholar
  106. Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Journal of Statistical Software, 36, 1–48.CrossRefGoogle Scholar
  107. Walters, M., Chen, J., & Breiding, M. (2013). The national intimate partner and sexual violence survey: 2010 Findings on victimization by sexual orientation. https://www.ncjrs.gov/App/Publications/abstract.aspx?ID=263171.
  108. Weiss, J. T. (2003). GL vs. BT. Journal of Bisexuality, 3(3–4), 25–55.  https://doi.org/10.1300/J159v03n03_02.CrossRefGoogle Scholar
  109. Wells, G. A., Shea, B., O’Connell, D., Peterson, J., Welch, V., Losos, M., & Tugwell, P. (2013). The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomized studies in meta-analyses. The Ottawa Hospital Research Institute, 3, 1–4.  https://doi.org/10.2307/632432.CrossRefGoogle Scholar
  110. Wexler, L. M., & Gone, J. P. (2012). Culturally responsive suicide prevention in indigenous communities: Unexamined assumptions and new possibilities. American Journal of Public Health, 102(5), 800–806.  https://doi.org/10.2105/AJPH.2011.300432.CrossRefPubMedPubMedCentralGoogle Scholar
  111. Wichstrom, L., & Hegna, K. (2003). Sexual orientation and suicide attempt: A longitudinal study of the general Norwegian adolescent population. Journal of Abnormal Psychology, 112(1), 144–151.CrossRefGoogle Scholar
  112. World Health Organization. (2008). Suicide and suicide prevention in Asia. http://www.who.int/mental_health/resources/suicide_prevention_asia.pdf. Accessed 17 Jan 2018.
  113. Wray, M., Colen, C., & Pescosolido, B. (2011). The sociology of suicide. Annual Review of Sociology, 37(1), 505–528.  https://doi.org/10.1146/annurev-soc-081309-150058.CrossRefGoogle Scholar
  114. Yoshino, K. (2001). The epistemic contract of bisexual erasure. Stanford Law Review, 52(2), 353–456.  https://doi.org/10.2307/1229482.CrossRefGoogle Scholar
  115. Zalsman, G., Hawton, K., Wasserman, D., van Heeringen, K., Arensman, E., Sarchiapone, M., … Purebl, G. (2016). Suicide prevention strategies revisited: 10-year systematic review. Lancet Psychiatry, 3(7), 646–659.  https://doi.org/10.1016/S2215-0366(16)30030-X.CrossRefPubMedGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2018

Authors and Affiliations

  1. 1.School of Population and Public HealthUniversity of British ColumbiaVancouverCanada
  2. 2.Clinical Prevention ServicesBC Centre for Disease ControlVancouverCanada
  3. 3.Dalla Lana School of Public HealthUniversity of TorontoTorontoCanada

Personalised recommendations