Professor Richard Ramsay
Richard Ramsay: To Home Page
Email to Richard RamsayEmail to Pierre Tremblay
University of Calgary Home Page



To Home of: Increasing Awareness of Gay, Lesbian, Bisexual, Transgender, Two Spirit, Queer... Suicide Issues
T0 Index Page for: Changing Construction of Western Male Homosexuality: Suicide Problems
Faculty of Social Work Home Page


 
To Search For Anything At This Site!


An updated interim paper prepared in support of a Poster Presentation with the same name at the 2004 CASP (Canadian Association For Suicide Prevention) Conference held in Edmonton, Alberta. Original Paper. Authors: Pierre Tremblay & Richard Ramsay



Male Youth Suicidality: Worsening over Time, Greater Risk for Homosexually Oriented Males


The male youth (15 - 24 years) suicide problem increased 3-fold from 1950 to 1990 (Table 3). The greatest increase (5-fold) was for adolescent males in the 15 - 19 age group. Between 1950 and 1990, males accounted for 88.5 percent of the increase in suicide deaths in the adolescent age group and 95.5 percent in the 20 – 24 age group. The increase in youth suicides in this period was almost exclusively a male problem. Overall males accounted for 92.3 percent of the increase in youth suicides. Some speculations have been made to explain the worsening male adolescent suicide problem. Gibson (1989) suggested that homosexually oriented adolescents may account for about one-third of these deaths, but "some experts rejected the conclusions as being drawn from biased samples" (Remafedi, 1999: 1291; Remafedi 1999a).
 
 

Table 3 - American Youth Suicide Rates: 1950 to 1990
Increasing Youth Suicides: A 90% Male Problem
Age Range 
(Years)
Suicide Rate 1950 [% Male]
Suicide Rate 
1990 [% Male]
Suicide Rate
Increase [% Male]
15 to 19
M - 3.5
F - 1.8
[66%]
M - 18.1
F - 3.70
[83%]
M: 14.6 ( 5.2X)
F: 1.90 (2.1X)
[88.5%]
20 to 24
M - 9.3
F - 3.3
[73.8%]
M - 25.7
F - 4.10
[86.2%]
M: 16.4 (2.8X)
F: 0.8 (1.2X)
[95.3%]
15 to 24
M - 6.5
F - 2.6
[71.4%]
M - 22.0
F - 3.9
[84.9%]
M: 15.5 (3.4X)
F: 1.3 (1.5X)
[92.3%]
M = Males, F = Females, X = Suicide Rate Increase Multiple.
Suicide Rates: Deaths per 100,000 individuals per year.
Data Source: CDC, 1994, Table 1, p. 3.

Tremblay (1995) proposed that up to 50 percent of male adolescent youth suicide may involve homosexually oriented males. This proposal contrasted with the results from two questionable methodology post-mortem studies (Rich et al., 1986; Shaffer et al., 1995) that are commonly cited to suggest that homosexually oriented adolescents are not at greater risk for suicide than their heterosexual counterparts (Muehrer, 1995; Moscicki, 1995). On the basis of reported suicide risk indicators and a secondary analysis of a community random sample that included sexual orientation data, Bagley and Tremblay (1997a) speculated that more than half of male suicide deaths may involve homosexually oriented males. Given the controversy associated with homosexual orientation and suicide and that male youth, especially adolescents, accounted for the vast majority of additional youth suicide since 1950, and the existence of considerable information on adolescent male suicidality and sexual orientation, a review of the information is warranted.

In mainstream suicidology, the traditional assumption that sexual orientation is not significantly related to male youth suicide rates or other suicide problems has created a general lack of awareness that homosexually oriented males may be greatly overrepresented in worsening youth suicide problems (Gibson, 1989; Remafedi, 1994; Savin-Williams, 1994; Tremblay, 1995). This assumption has persisted in spite of the Bell and Weinberg (1978) study that reported white predominantly homosexual males, by the age of 20 years, were 14 times more at risk for a "first time" suicide attempt than their heterosexual counterparts. Bagley and Tremblay (1997) replicated this multiplier factor (to the average age of 22.7 years) for a large random sample of 18- to 27-year-old Calgary males. A further analysis of the Cochran and Mays (2000) "suicide attempt" data for males 17 to 29 years in the National Health and Nutrition Examination Survey III (NHANES) produced Odds Ratios of 12.1 and 14.4 (Appendix A). These are very similar results to those in the two previously noted studies (OR range for the three studies: 12-15, Note 6). Given the mean age of 37 years for the males studied in 1969 by Bell and Weinberg (1978), the suicidality results for homosexual males best represent the situation existing in the early 1950s. At that time, "first time" suicide attempt incidence to the age of 17 years for white predominantly homosexual males was 4.9% percent: 28/575 (Bell and Weinberg, 1978; 453).

The recent studies that have reported on the suicidality of homosexually oriented male youth have used varied methodologies, thus creating interpretation problems and making it difficult to be absolute about the numbers, but a general trend is apparent from the study results. The lifetime suicide attempt incidence for homosexually oriented adolescent males has increased over the years (about 6 times) given the Bell and Weinberg data (4.9% "first time" suicide attempt incidence by the age of 17), compared to the results in more recent North American studies. A variety of volunteer community sample studies of gay and bisexual males have reported lifetime suicide attempt incidences averaging 30 percent (Bagley and Tremblay 1997a: 197, data summary for 12 North American Studies; Table 4, data for 6 studies). A range of 20 to 42 percent was found in ten published American studies (Remafedi et al., 1999). A recent study of 192 gay and bisexual males (14 to 21 years) associated with gay community organisations and university groups in the United States, Canada, and New Zealand produced results near the 30 percent lifetime incidence. Thirty-four percent of the males studied had attempted suicide at least once, and 19.7 percent had attempted suicide more than once (D'Augelli et al., 2001). These previously disputed suicidality (not suicide death) results have been replicated in school based random sampling YRBS (Youth Risk Behaviour Surveys, or similar surveys), and two studies specifically report suicidality results for male adolescents based on sexual orientation.
 
 

Table 4 - G(L)B Youth Lifetime "Suicide Attempt" DATA
Published Studies (American)
STUDIES1
(N)1
MALES
MEAN
AGE
(years)
ATTEMPTED
SUICIDE
Percent
MULTIPLE 
ATTEMPTS
Percent
Roesler & Deisher, 
1972
60 GB Males
20
31.0%
11.5%
Remafedi, 1987
29 GB Males
18.3
31.0%
6.2%
Schneider et al., 
1989
108 GB Males
20.6
20.0%
9.0%
Remafedi et al., 
1991
137 GB Males
19.6
30.0%
13.2%
Rotheram Borus, 1994
138 GB Males
+/- 17
(14-19)
39.0%
20.3%
Grossman & 
Kerner, 1998
90 GLB
Youth 2
+/- 18
(14-21)
30.0%
20.1%
Means
(6 Studies)
--
--
30.2%
13.4%
1. A variety of community-based volunteer samples.
2. Sample of gay (G) / bisexual (B) / lesbian (L) individuals with suicide attempt prevalence given for entire group.

On the basis of a 1987 Minnesota survey of Grade 7 to 12 students, Remafedi et al. (1998) report a lifetime suicide attempt incidence of 28.1 percent by the average age of about 15 years for males identified as gay or bisexual, compared to 4.2 percent for heterosexual identified males. Using the same sample data, Saewyc et al. (1998) report that older gay and bisexual adolescent males (15- to 19-years-old) have a lifetime suicide attempt incidence of about 33% by the average age of about 17 years. Garofalo et al. (1999) do not specifically give the suicide attempt rate (based on a 12-month period) for males identified as gay, bisexual, and "not sure" about their sexual orientation (GBN), but an estimate of 33 percent was determined on the basis of the information given (Note 7). Other studies (some not yet published, but the data is available) report similar results on average for homosexually oriented adolescents not differentiated on the basis of sex (Table 5).
 
 

Table 5 - Increasing Risk for the More Serious Suicide Problems: 
Youth Risk Behaviour Survey Results Based on Self-Identification 
as Gay / Lesbian / Bisexual, and/or Same-Sex Sexual Behaviour
Survey ®
/ Suicide
Behaviour Category
Mass. 95 1
GLBN* (Gay,
Lesbian, Bi- 
sexual, Not 
Sure) vs.
Non-GBLN*
Mass. 97 
GLB and/ or Homo-sex 
Active vs. 
Heterosexual
-Identified
Seattle 95 GLB-
Identified vs. Hetero-
sexual
dentified
Vermont 951
Males &
Females:
Homo-sex active
vs. Hetero-sex
active.
Mass. 931
Males &
Females:
Homo-sex active
vs. Hetero-sex
active.
Suicide 
Seriously
Considered
46.4% vs.
24.6%
**RR: 1.9
54% vs.
22%
RR: 2.5
34.2% vs.
16.7%
RR: 2.1
59.2% vs. 
37.1%
RR: 1.6
41.7% vs.
28.6%
RR: 1.5
Suicide
Planned
40.5% vs.
17.5%
RR: 2.3
41% vs.
18%
RR: 2.3
31.1% vs.
15.7% 
RR: 2.0
52.8% vs.
28.7%
RR: 1.8
29.7% vs.
24.5%
RR: 1.2
Attempted
Suicide
36.0% vs.
8.9%
RR: 4.0
37% vs. 
8%
RR: 4.6
20.6% vs.
6.7%
RR: 3.1
40.7% vs. 
15.2%
RR: 2.7
27.5% vs.
13.4%
RR: 2.1
Multiple
Attempts
N/A
N/A
N/A
N/A
N/A
Suicide
Attempt(s)
& Medical Care
N/A
19% vs.
3%
RR: 6.3
9.4% vs. 
2.2% 
RR: 4.3
26.5% vs. 
6.2% 
RR: 4.3
20% vs. 
4.7% 
RR: 4.2
1. The superscript indicates that the studies with some results were published in peer-reviewed journals, but relevant data / results cited may not have been given in these papers. The information references for all the cited surveys are given in the Bibliography as "Massachusetts Youth Risk Behaviour Survey (1995)," "Massachusetts Youth Risk Behaviour Survey (1997)," etc.

*GLB = Gay, lesbian, or bisexual self-labelling and/or Identification; N = "Not Sure of one's sexual orientation." N/A = Not Available.

** "RR" is the Risk Ratio: "More at risk" factor for the indicated "homosexual orientation" category compared to other adolescents. The Chi Squared statistical significance for the cited incidences was given in data sources as p < .05.

Suicide Attempt Incidences (12 month period), Massachusetts YRBS (GLB and/or Homo-Sex Active): 1999 (29%, RR = 6.0), 2001 (31%, RR = 5.4), 2003 (32%, RR = 4.6). 

Youth Risk Behaviour Survey results have repeatedly indicated that homosexually oriented adolescent males and females, based on self-labelling and/or reporting same-sex sexual activity, are generally more at risk (compared to heterosexual youth) for increasingly serious suicide behaviours: (1) considered suicide, (2) planning a suicide, (3) attempting suicide, and (4) attempting suicide that resulted in receiving medical attention (Table 5). It is likely that the most life-threatening suicide attempts, or suicide attempts associated with a definite intent to die, would be those that resulted in adolescents receiving medical attention, but this may not always apply, at least for some gay and bisexual males (Remafedi et al. 1991). Some individuals who attempt suicide in very serious ways survive their attempt but do not seek medical care.

The YRBS results listed in Table 5 include homosexually oriented females. Including male and female data together tends to reduce major differences existing between males as indicated from the Garofalo et al. (1999) and Remafedi et al. (1998) study results. In the two studies, statistical significance was eliminated between homosexual and heterosexual female suicide attempters in the multivariate analysis, but great differences for suicide attempt incidences were reported for homosexually oriented males compared to their heterosexual counterparts. In the Garofalo et al. (1999) study, LBN females are 2.0-times more likely to report a suicide attempt than heterosexual females, while the factor is 6.5-times for GBN males compared to heterosexual males. In the Remafedi et al. (1998) study, the higher risk factor is 1.4 times for LB females and 7 times for GB males compared to their heterosexual counterparts. Therefore, it is suspected and predicted that an analysis of only the male data in other YRBS studies would likely show more serious suicidal behaviours than their heterosexual counterparts, compared to results generated from the combined male and female data (Table 5).

The effects of separating males and females in the statistical analyses of YRBS data is illustrated in the suicidality results (4 variables) for the Oregon 1999 Youth Risk Behaviour Survey that are based on assumptions by others that one is homosexual, and not on self-identification as homosexual or bisexual. That is, a "Yes" answer to the following harassment question asked in the Oregon survey:

"During the past 12 months, have you ever been harassed at school (or on the way to and from school) because someone thought you were gay, lesbian or bisexual?" (Yes / No)
implies that an adolescent is not only being identified or labelled as "gay, lesbian or bisexual by others," but that they have been harassed and abused to varying degrees because this assumption was made by individuals who perceive homosexuality very negatively and believe it should be stigmatised. As noted by both Goffman (1963) and Becker (1963) in their studies of stigmatised individuals, or those deemed to be "outsiders" by the more acceptable members of society (the "normal" ones), negative results may be expected from the stigmatisation and related abuses. For adolescents reporting anti-gay harassment (some are likely homosexual oriented but most would not have revealed this fact to others), an increasing relative risk for more serious suicide behaviours is observed with males and females together (Table 6, column 1). The same observation is true when males and females are separated (columns 2 and 3). The "suicide attempt" Risk Ratios for males (column 3) when compared to in the corresponding Risk Ratios for females (column 2), and for male and females together (column 1) illustrates the importance of separating male and female data for analysis. Given that females, as a rule, report significantly more suicide attempts than males, and also more suicide attempts in association with anti-gay harassment, their contribution to the risk ratios when males and females are analysed together tends to produce results that are more like "female" results. The "suicide attempt" Risk Ratios for anti-gay harassed males are noticeably higher, compared to other males, than are the equivalent Risk Ratios for females, or for males and females analysed together.
 
 
Table 6 - Oregon 1999 Youth Risk Behaviour Survey Results: 
Harassment Based on One's Presumed Homosexual Orientation
Increasing Risk for the More Serious Suicide Problems
Group ®
/ Suicide
Behaviour Category
Males &
Females:
All Ethnic
Groups
N = 22,921
Females:
All
Ethnic
Groups
n = 11,656
Males:
All
Ethnic
Groups
n = 11,265
Males:
White
(One
Ethnicity)
n = 9,109
Males
Of Colour:
(One
Ethnicity)
n = 1,166
Males: 
2 or More
Ethnicities Reported
n = 382
Harassment
Incidence
6.8%
1557 / 22,858
5.7%
664 / 11,630
8.0%
893 / 11,228
7.6%
691 / 9,081
8.0%
133 / 1,661
13.4%
51 / 381
Suicide 
Seriously
Considered
34.0% v 14.7%
2.1<2.3<2.5
0.133
43.4% v 19.4%
2.0<2.2<2.4
0.137
27.0% v 9.7%
2.4<2.8<3.1
0.149
25.9% v 9.4%
2.4<2.8<3.2
0.142
28.8% v 10.7%
2.0<2.7<3.6
0.150
36.5% v 15.0%
1.6<2.4<3.8
0.193
Attempted
Suicide
16.7% v 5.7%
2.6<2.9<3.3
0.114
22.1% v 8.4%
2.2<2.6<3.1
0.111
12.4% v 2.7%
3.7<4.6<5.7
0.144
10.6% v 2.4%
3.4<4.4<5.8
0.128
15.5% v 4.0%
2.4<3.9<6.4
0.146
30.4% v 5.1%
3.1<5.9<11.3
0.306
Multiple
Suicide Attempts
9.2% v 2.3%
3.4<4.0<4.9
0.107
12.8% v 3.4%
3.0<3.8<4.8
0.113
6.6% v 1.1%
4.4<6.0<8.3
0.122
5.1% v 1.0%
3.3<4.9<7.2
0.093
7.8% v 1.4%
2.6<5.6<12.3
0.128
21.7% v 1.6%
4.8<16.6<37.9
0.336
Suicide
Attempt(s) & Medical Care
5.5% v 1.4%
3.1<4.0<5.1
0.082
7.8% v 2.0%
2.8<3.6<4.7
0.091
3.6% v 0.7%
3.2<5.0<7.6
0.081
3.9% v 0.7%
3.3<5.2<8.3
0.085
3.4% v 0.5%
2.1<7.5<26.2
0.-098
4.3% v 1.6%*
.54<2.7<13.4
0.066
Suicide
Attempt(s): Multiple and/or Medical C.
11.6% v 2.9%
3.3<3.9<4.6
0.119
16.0% v 4.3%
3.0<3.7<4.5
0.125
8.3% v 1.5%
4.3<5.7<7.5
0.133
6.8% v 1.4%
3.5<5.0<7.0
0.109
9.4% v 1.9%
2.5<4.9<9.8
0.131
26.1% v 2.6%
4.4<10.2<23.6
0.343
1. Statistical results based on one's sex (male and females), but not on one's ethnicity, are calculated using a weighted values normalised to the original total counts for males and females. The weighting values were supplied with the data set obtained from the Oregon Health Department. The weighting adjusts values on the basis of age, sex, grade, and ethnicity for the State of Oregon.
2. Results generated for males on the basis of ethnic groupings are calculated using weighted values normalised to the original total male counts. These values were calculated by the author as a proportion of the above noted weights supplied by the Oregon Health Department.
3. Students were asked to classify themselves in one or more the following five 'race' / 'ethnic' categories: White, American Indian / Alaska Native (Aboriginal), Black / African American, Latino / Hispanic, Asian, and Hawaiian / Pacific Islander. There were 303 students (1.3%) who did respond to this question.
4. Data Reporting: Row Variable Incidences (in percent) for the group reporting anti-gay/lesbian harassment versus (v) Row Variable Incidence for the ones not harassed in this way. Risk Ratios (in bold and larger font: 2.1<2.3<2.5) are given with a 95% Confidence Interval: from 2.1 to 2.5 in the example given. Risk Ratio = the factor relationship between the two incidences given as a percent value. Definition using 2X2 Table language: "top row against column totals."
5. The Chi Square statistical significance for the reported differences in each table cell and the Correlation (0.113) are p < .000 (mostly p< .0000), with one exception indicated by an asterisk. For males reporting more than one ethnicity, the difference is not significant (p = .214) for reports of receiving medical attention in association with a suicide attempt, compared to other multi-ethnic males. This does not mean, however, that they are less at risk for this outcome than males in other ethnic categories. The opposite applies.

To date, published North American studies on suicide behaviours associated with homosexuality have only reported results on the basis of gay / lesbian / homosexual / bisexual / heterosexual self-identification (or self-labelling) and/or on the basis of individuals reporting same-sex sexual behaviour, and/or having related sexually with same-sex individuals. However, it is not only these youth who happen to be at risk for reasons associated with society's homonegativity as indicated by the results listed in Table 6. Adolescent males "assumed" to be gay or bisexual and are reporting harassment for this reason, whether or not they are identified as gay, lesbian, or bisexual, are generally at increasing risk for the more serious suicide behaviours. As a rule, their relative risk for suicide behaviours increases from considering suicide to attempting suicide, and the relative risk is even higher for reports of "multiple suicide attempts" (an indication of ongoing suicide problems) and/or a suicide attempt(s) associated with having received medical care (an indication of a life-threatening suicide attempt). Variations and exceptions to the rule, however, may be expected on the basis of ethnicity. This is indicated (Table 6, columns 3 to 6) for white males, males of colour in one of the identified 'ethnic' / 'race' categories, and for the males associating themselves with two or more of the 'ethnic' / 'race' categories

The likelihood of anti-gay related harassment, abuses, violence, and criminal acts being implicated in negative mental health outcomes has been reported in several studies of homosexually oriented youths and adults (Hunter, 1990; Garnets et al., 1992; Proctor and Groze, 1994; Meyer, 1995; Hershberger and D'Augelli, 1995; Pilkington and D'Augelli, 1995; Herek et al, 1997; Herek et al., 1999; D'Augelli and Grossman, 2001). One study exploring the possible association between victimisation and suicidality in a sample of gay and lesbian youth did not report a direct association between the two variables (Hershberger and D'Augelli, 1995). However this does not necessarily mean that victimisation is not related to suicidality as noted by McDaniel et al. (2001). One British study of 119 gay, lesbian, and bisexual adults (16 – 66 years; mean = 29 years) explored the subjects' history of being bullied and the types of abuses they had experienced. With respect to suicidality, this retrospective study reported:

"In total, 53 per cent of the sample said that they had contemplated self-harm as a result of being bullied. Of that number over three quarters (40 per cent of the total sample) indicated that they had attempted self-harm/suicide on at least one occasion and three quarters of the participants had attempted on more than one occasion" (Rivers, 2001: 39-40).
Rivers (43) noted that the incidence of physical attacks in his study (31 percent reported being "hit or kicked") was similar to the 29 percent incidence of "direct-physical bullying" reported by Pilkington and D'Augelli (1995) for North American GLB youth. These estimates are also similar to the 34% of GLB identified youth (110/324) in the 1995 Seattle YRBS who reported being the "target of offensive comments or attacks" related to their presumed homosexual orientation at school or on the way to school. Such abuse and victimisation is common and its near epidemic status was reported in the 2001 Human Rights Watch study, Hatred in the Hallways: Violence and Discrimination Against Lesbian, Gay, Bisexual and Transgender Students in U.S. Schools. As a rule, however, when anti-gay bullying occurs, targeted homosexually oriented adolescents had not told others about their same-sex sexual desires, but many adolescents are very alert to the possibility that some peers are homosexual. As one British gay adolescent reported, "even if you are not out, teenagers have a sixth sense to spot lesbian and gay men. They just home in on them and take the piss. So you tend to get a lot of hassle and a lot of disruption in lessons" (Crowley et al., 2000). As it will be discussed later, one fact almost always omitted from the discourses related to the victimisation of gay/bisexual identified male youth and associated problems is the role that homophobia related abuses play in the current social construction of masculinity. It is true that a significant number of gay / bisexual identified males are suspected to be homosexual and are abused accordingly. But, for every one of these victimised males, there may be four heterosexual identified adolescents who are also suspected of being homosexual and being abused that may be experiencing serious suicide problems.

Many factors associated with society's homonegativity are harmful to homosexually oriented youth. Bagley and Tremblay (1997) reported related suicidality outcomes by the end of adolescence or early adulthood from a secondary analysis of a random sample of 750 males ranging in age from 18- to 27-years-old. Their results are presented in two categories of self-harm behaviours based on previous epidemiological studies: "self-harm" (often called suicide attempts by adolescents, but without the intent to die, and the behaviour is usually not life threatening), and "suicide attempts" representing a definite attempt at self-killing (Note 7). Compared to heterosexual males, homosexually oriented young adult males were 3 times more likely to report "self-harm." This result is similar to the reported differences, based on sexual orientation, for "suicide attempts" reported in YRBS studies (Table 5). For strictly defined "suicide attempts," however, the authors reported that young adult homosexually oriented males were 14 times more at risk than their heterosexual counterparts. The magnitude of risk also increases four to five times from their relative risk for having engaged in "self-harm" activities, to their relative risk for a "suicide attempt" (Note 8).

Homosexually oriented adolescent males are at high risk for serious suicide attempts is also reported by Remafedi et al. (1991) from their sample of 137 gay and bisexual male youth ranging in ages from 14 – 21 years. They found "fifty four percent of all suicide attempts (37/68) received risk scores in the 'moderate to high' lethality range." In the 45 attempted suicide cases where the rescuer was not the victim (76% of cases), 58% "received scores in the 'moderate to least' rescuable range. In other words, the predicted likelihood of rescue was moderate to low despite the actual occurrence of an intervention" (p. 871). Unfortunately, since 1991, most studies reporting on the suicidality of homosexually oriented youth have only reported "suicide attempts" (with a related analysis) without investigating their degree of lethality. Nor were "suicide attempts" rigorously evaluated to separate the ones best classified as "self-harm" from the ones best classified as deliberate attempts to take one's life (Note 6). The Remafedi et al. (1991) and YRBS data (Table 5) also suggests that 25 to 50 percent of suicide attempts by gay or bisexual identified adolescent males result in medical interventions, thus indicating a greater likelihood for suicide by these males.

The high proportion of homosexually oriented male adolescents who attempt suicide places them at higher risk for an eventual suicide either as adolescents or later in life (Lewinsohn et al., 1993; Kotila and Lonnqvist, 1989; Sellar et al., 1990; Shafii et al., 1985; Otto, 1972). In this respect, it is reported that the "best single predictor of death by suicide is probably a previous suicide attempt" (Garland and Ziegler, 1993: 172). Community samples of homosexually oriented youth have also produced high rates for repeat suicide attempters, six study samples producing a 13.4 percent incidence for gay and bisexual male reporting multiple attempts (Table 4), thus indicating that suicide is an ongoing problem for many of these youth. In the 1996 Omega AIDS Cohort Study of 629 MSM (men who have sex with men; mean age = 32 years) men in Montreal, 36.1 percent had attempted suicide, and 14.7 percent reported multiple attempts (Otis, 2000: 16). Unfortunately, published studies have not been located on the risk for an eventual suicide by adolescents with a history of repeated suicide attempts. It is possible that such individuals, especially males, are at higher risk for committing suicide than males with a history of only one suicide attempt.

The lifetime incidence of "suicide attempts" for homosexually oriented adolescent males has increased about 6-fold since 1950, from about 5 to 30 percent. Tremblay and Ramsay (2002) carried out an exploration of this increase. They recognised that the Bell and Weinberg. sample of White American predominantly homosexual males taken in 1969 had an average age of 36 years and could therefore produce estimates for males attempting suicide to the age of 20 years over a time period ranging from about 1920 to 1970. The results, if an increase in suicidality existed for these males over time (roughly equivalent to age categories of 10 years), would also be statistically significant given that about 54 suicide attempters were available for analysis in a sample of 575 males. Table 7 indicates that the first time suicide attempt incidence for White American predominantly white homosexual males had increased significantly since the 1920's to the 1960s: from 2.3 percent to 18.1 percent. By the early 1990s, their first time attempted suicide incidence reached about 30 percent, this being an estimated mean for the studies available.
 
 

Table 7 - Suicidality of White Predominantly Homosexual Males 
From the 1920s to the 1990s
Bell & Weinberg (1978): 1969 Sample Data 1
Age Groups ®
/ Categories
> 46 Years
n = 132
37-46 Years
n = 146
27-36 Years
n = 142
16-26 Years
n = 155
Studies 5
Table 4
First Attempt
by Age 20a2, 3
Date for Attempt4
2.3%
(3/132)
1920-42
5.5%
(8/146)
1936-52
10.6%
(15/142)
1946-62
18.1%
(28/155)
1956-69
30.0%
(Mean)
1963-96
Percentage Point
Increase
?
3.3%
5.1%
7.5%
12.1%
1. Analysis of Bell and Weinberg (1978) study sample data by Tremblay and Ramsay (2002). Comparison with the control sample of 284 predominantly heterosexual males used in this study was not possible given that only two of these males had attempted suicide by the age of 20 years.
2. The mean suicide attempt incidence to the age of 20 years for sample of 575 predominantly homosexual males: 9.4% (54/575) representing a time period from about the 1920s to 1970. This mean would represent the suicide attempt incidence existing for these males around 1950. The above suicide attempt incidence distribution, however, would place the 9.4% mean at about 1959-60.
3. The Chi Squared analysis for attempting suicide by predominantly homosexual males divided in 4 age categories in the Bell and Weinberg (1978) study sample: 24.4 (df = 3), p < .000
4. The time range given for reported attempts are approximated, the assumption being that most attempts would have occurred between the ages of 13 and 20 years. A time range is also estimated for more recent "suicide attempt" mean for the more recent community based samples5.
5. A general estimate for community samples of gay and bisexual male youth, including the six studies listed in Table 4, and other studies available. This mean would represent the situation existing around 1990 given that most study samples date from about 1984 to 1996.

Of great importance are the YRBS results that show a "suicide attempt" average about 30 percent for a 12-month period: 32.3% for the 5 studies listed in Table 5. Given that these rates coincide with lifetime estimates from community samples (Table 4), this similarity may reflect, at least in part, the possibility that adolescents in school define a "suicide attempt" in a more liberal way. A minor injury may be interpreted to be a suicide attempt as suggested by Moscicki et al. (1989):

"'Suicide attempt' could thus have been interpreted by the respondent as an act of self-harm without the intention of dying, a genuine (and failed) attempt to end one's life, or as something else (perhaps an 'accident' had been explained to them by a physician that had it as an unconscious suicide attempt). Suicide ideation, similarly, may mean different things different people" (p. 122)
Despite this caveat, it may be concluded that the suicidality results from the random sampling YRBS studies do suggest that the suicide problems of homosexually oriented adolescent males are more serious than what has been proposed in worst case scenarios. This conclusion is based on the greater representation of gay and homosexuality oriented adolescents in YRBS samples compared to their representation in volunteer community samples and the definition of "gay" or "bisexual" on the basis of self-labelling or identification and/or being homo-sex active.

For many years, the representation of homosexually oriented youth in youth suicide problems has been a debated issue, especially with respect to their representation in suicide statistics. Mainstream suicidologists have often criticised suicidality results from gay community based samples given their assumed biased / non-random nature (Moscicki, 1995; Muehrer, 1995). These criticism were also generally used to counter efforts to have homosexuality issues addressed in youth suicide education and prevention programs where indifference to homosexuality was common, and the same indifference has ruled in mainstream youth suicidality research (Remafedi, 1994; Savin-Williams, 1994; Tremblay, 1995). Tremblay (1995) suggested that a mean spirited objective motivated the production of the 1995 Suicide and Life Threatening Behaviour special issues on "Sexual Orientation" edited by Muehrer, Moscicki, and Potter (1995) who are from the National Institute of Mental Health (NIMH) and the Centers for Disease Control (CDC).

Tremblay (1995) felt that the document was not created by individuals who wanted to see homosexuality issues addressed in mainstream youth suicidality education and prevention programs, and that mainstream youth suicidality researchers were not being encouraged to begin soliciting sexual orientation information in their research. By 1999-2001, this outcome was the rule in mainstream youth suicidality research and it was an ongoing research behaviour that the CDC was modelling as a part of its boldly emphasised "Leadership Role." The CDC's role is "to ensure the availability of accurate and current information on health risk behaviours among young people, [and to provide] funding and technical support to states and major cities to conduct a Youth Risk Behaviour Survey" (CDC, 1999). But, the CDC has continued to avoid being a good role model in youth research by not soliciting "sexual orientation" information from adolescents. The CDC policy in this respect might be seen to be similar to the U.S. Military policy approved by the U.S. Government: "Don't ask. Don't tell." This CDC policy is being maintained despite the conclusion of Sell and Becker (2001) in the first issue of the American Journal of Public Health devoted to lesbian, gay, bisexual and transgender issues:

"…YRBSS [Youth Risk Behaviour Survey Surveillance] data from the localities that measure sexual orientation show that whatever the dimension of sexual orientation measured (sexual orientation identity, sexual behaviour, or sexual attraction), sexual minority youths have higher rates of suicide attempts, victimisation in school violence, drug and alcohol abuse, early onset of sexual behaviour, eating disorders, and teenage pregnancy that other youths. As surveys of adolescents in Connecticut, Wisconsin, Oregon, and Seattle have demonstrated, even a measure of sexual orientation as indirect as perceived homosexual sexual orientation [harassment based on the assumed homosexual orientation of an individual] elicits disturbing correlations with deleterious health outcomes" (p. 878).
The CDC's official YRBS questionnaire has a history of not soliciting "sexual orientation" information (CDC, 2001), meaning that the CDC apparently does not believe that soliciting such information would be of value in better understanding adolescent problems, in spite of the wealth of information indicating otherwise. Nonetheless, the official YRBS questionnaire does solicit information about "sexual intercourse," generally assumed to mean "penis-vagina" sex given the heterosexual nature of Question #63: "The last time you had sexual intercourse, what one method did you or your partner use to prevent pregnancy?" Therefore, to the year 2001, the CDC has continued to ignore "homosexual orientation" issues in youth suicide problems and its indifference will likely have negative results. When any "at risk" group of adolescents is ignored, their concerns are also ignored, and the ones responsible for this indifference are essentially working to maintain a maximum casualty status in this sector of the adolescent population.

In spite of the ongoing CDC modelling of indifference to homosexuality issues in youth problems, results from relatively rare school-based studies that have solicited "sexual orientation" were being peer reviewed and published:

It was only from the Garofalo et al. (1999) study that sufficient information was made available to permit an estimate of the proportion of male suicide attempters who are homosexually oriented. They reported a previous 12-month incidence of 33 versus 5.1 percent attempted suicide rate for GBN males (gay, bisexual, and "Not Sure" about their sexual orientation) that made up 4.7 percent of the male population. From this information it was then possible to estimate that about one quarter (24.3%) of male suicide attempters are in the GBN category (Note 7). Therefore, 4.7 percent of male students account for about 25% of the adolescent male suicide attempt problem. This study is based on an expanded version of the pencil-and-paper YRBS questionnaire (CDC, 2001). Recent knowledge indicates that pencil-and-paper methodologies may result in underestimates of the numbers of homosexually oriented male youth by factor of up to 3 or 4 (Bagley and Tremblay, 1998; Turner et al., 1998). This kind of underestimation factor suggests the possibility that had all homosexually oriented male adolescents been identified by Garofalo et al. (1999), they would account for more than 25 percent of adolescent male suicide attempters. The corrected figure could easily be 50 percent or more, and such a figure would be consonant with the Bagley and Tremblay (1997) result. Young adult males in the homosexual / bisexual categories (defined on the basis of self-identification or self-labelling and/or being currently homosexually active) accounted for 62.5% of young adult male suicide attempters (lifetime incidence).

Bagley and Tremblay (1997), however, did not include all males who could have been determined to be homosexual in minor ways. This expanded definition of sexual minority was investigated by Savin-Williams (2001) in a study sample of male university students. Little has been published about homosexually oriented male youth that completed high school with grades high enough to meet university entrance requirements. Some difference could be expected given that more representative studies report high levels of problems for these youth in high school, including a risk for dropping out of school. Savin-Williams (2001) studied a sample of 114 males university students enrolled in introductory human development and sexuality courses. They were asked in a pencil and paper questionnaire to rate themselves with respect to "sexual orientation" on the Kinsey Scale ("0" = Exclusively heterosexual, "1" = Predominantly heterosexual, but slightly homosexual, to "6" = Exclusively homosexual). Suicidality information was also solicited from these males, with an effort made to determine "true suicide attempters" as it was done by Bagley and Tremblay (1997). There were 53 males (46.5%) in the 1 to 6 categories, and 61 (54.5%) were deemed to be "exclusively heterosexual ("0"). Only one "true suicide attempter" was in the "exclusively heterosexual" category, and five (83 percent) of the "real suicide attempters" had reported some degree of homosexuality. Out of the six attempters, three males reported that their suicide attempt resulted in receiving medical attention, and all three were in Savin-Williams' expanded definition of "sexual minority youth": anyone with a self-rating of 1 to 6 on the Kinsey Scale. Unfortunately, because the sample was small, the relative "suicide attempts" and "suicide attempts associated with medical care" results were only close to Chi Squared significance: 3.4, p = 0.063 and 3.5, p = 0.060, respectively (Calculated by Pierre Tremblay). The suicide attempt incidence for these homosexually oriented male university students is 11.3%, compared to an incidence of 1.6% for their 100% heterosexual counterparts.

The results of the aforementioned studies leads to the informed proposition that at least 50 percent of male adolescent suicides may be associated with homosexuality issues. This probability is based on suicide attempt comparisons between homosexual and heterosexual males attempts and the fact that these males are more at risk than their heterosexual counterparts for more serious forms of suicide attempts (Bagley and Tremblay, 1997, 1997a; Remafedi et al., 1991; study results in Table 5). A recent study suggests that this probability may apply to a much greater degree for homosexually oriented male university students (Savin-Williams, 2001). Even this probability may be an underestimate because one group of males have yet to be considered: those who are not homosexually oriented but are subjected to anti-gay harassment that is implicated in male youth suicide attempts. Occasionally, deaths related to homosexuality issues and harassment have been reported in the media in both Canada and the United States.

One of the pathways leading some gay and bisexual adolescent males to seeking an end to their lives is illustrated by Nathan life in England:

"I mean at the age (i.e. 12 years) it absolutely scared the shit out of me… I realised that all those jokes said at school - I was the person, I was the object of the hatred."

"I'm not sure whether I thought I was the only person. No, I never thought I was the only person that was gay. I just, I felt that, I felt very, very, isolated, um, very alone and rather unloved… and so I just, I just took and overdose" (Walker, 2001: 52-53, 56).

It is known that some males, such as Bruce C., Bobby Griffith, and others do commit suicide for reasons expressed by Nicolas and Howard (1998) in their study "Better Dead Than Gay!" This fact was also the subject of a web page titled "Better to be Dead Than Gay" (Tremblay, 1998-2002) and Michel Dorais (2000) book, "Mort ou Fif" [Death or Be A Fag], also emphasises the concept. For a significant number of males who recognise their same-sex sexual desires, the situation has been:

Better Dead Than a Fag


NOTES


Note 6

The suicidality data intake methodology used by Bagley and Tremblay (1997) is known in the field of American suicidology. In fact, Moscicki (1989) deemed Bagley & Ramsay's 1985 distinction between "self-harm" and "suicide attempt" to be "very important and she reported on the methodology. "Respondents were asked in a semi structured interview about sociodemographic characteristics, childhood history, social ties, religiosity, stress, health, 'suicidal behaviours,' and 'suicidal actions': 1. Have you ever felt that life was not worth living?  2. Have you wished that you were dead - for instance, that you could go to sleep and not wake up? 3. Have you ever thought of taking your life, even if you would not really do it? 4. Have you ever reached the point where you seriously considered taking your own life and perhaps made plans how you would go about doing it?  5. Have you ever deliberately harmed yourself, but in a way that stopped short of a real intent to take your life? 6. Have you ever made an intentional attempt to take your life?" The six items are cited from the Appendix in Ramsay and Bagley (1985, p. 165), and the correlates of suicidal behaviour were published in the same year by Bagley and Ramsay (1985). The questions related to self-harm and attempting suicide asked in the Bagley and Tremblay 1997 study via laptop computer were:

"Have you ever deliberately harmed your self, but in a way that stopped short of a real intent to take your life?" (Yes / No)
"Have you ever made an intentional attempt to take your life?" (Yes / No)

"Attempted suicide involves either a clear intention to die, use of a potentially lethal method, or both" (Bagley et al., 1994: 686) and additional information was solicited from suicide attempters to make this determination.

To date, most American suicidality research has ignored the last two distinctions noted by Moscicki (1989) to be very important, and a study that Moscicki's research team published in the same year had also ignored this distinction (Moscicki et al. 1989). Under the heading of "limitations" for their study, the following caveat was written. "'Suicide attempt' could thus have been interpreted by the respondent as an act of deliberate self-harm without the intention of dying, a genuine (and failed) attempt to end one's life, or as something else (perhaps an 'accident' had been explained to them by a physician that had seen it as an unconscious suicide attempt). Suicide ideation, similarly, may mean different things to different people (p. 122)." This serious nomenclature problem was noted by O'Carroll et. al. (1996), with recommendation being made. After reviewing these recommendations, Bagley and Ramsay (1997) stated: "Our definition of suicidality generally accords with those proposed by O'Carroll et al. (1996)" (p. 7).
 

Note 7

Garofalo et al. (1999) reported that GBN males (gay/bisexual identified, and the ones not sure of their sexual orientation) were 6.5 times more likely than heterosexual males to report a suicide attempt in the past 12 months, but it was apparently not deemed important to report the percentage of suicide attempters in each category. Given, however, that males had a suicide attempt incidence of 6.4% (for a 12-month period) and using an estimate of about 4.7% for GBN males (3.8% GB males + 0.9% "not sure") in calculations, the relative incidence of suicide attempters would be approximately 33% versus 5.1% for GBN males compared to heterosexual males. On the basis of this estimate, another estimate is also possible. About one quarter - 24.3% - of male suicide attempters are in the GBN category. Therefore, 4.7% of the students in the GBN category account for about 25% of the male suicide attempt problem. In this study, for males and females combined for analysis, the 1.9% percent of the sample reporting same-sex activity had a suicide attempt rate (for the past 12 months) of 30.6% versus 9.2% for their heterosexual counterpart. The suicide attempt rate for adolescents identified as GLBN was 31.0% versus 9.1% for heterosexual identified adolescents, .22.7% versus 9.1% for adolescents "not sure" of sexual orientation, and 35.3% versus 9.1% for GLB adolescents (not including "not sure" adolescents). GLBN adolescents are also at significantly greater risk than heterosexual adolescents for receiving a high score on the three "risk behaviours" (p < .001): (1) drug use, (2) sexual behaviour, and (3) violence / victimisation.
 

Note 8

Using the 2 X 2 Table located at - http://home.clara.net/sisa/twoby2.htm and the following data from the Bagley and Tremblay (1997) study - 5 homosexual / bisexual suicide attempters and 77 non-attempters versus 3 heterosexual suicide attempters and 665 non-attempters - produces an OR (Odds Ratio for attempting suicide) of 14.4 (95% CI, 3.4-61.4). The OR for "Self-Harm" (13 homosexual / bisexual males reporting self-harm and 69 not reporting self-harm versus 37 heterosexual males reporting self-harm and 631 not reporting self-harm) produces an OR (Odds Ratio for self-harm) of 3.2 (95% CI, 1.6 - 6.3).  This information is not given in the study. The related Chi Square significance is 22.1, p < .0000 and 12.5, p = .0004, respectively.

The "suicide attempt" OR for the Bagley and Tremblay (1997) study of homosexual and bisexual males versus heterosexual males ranging in age from 18 to 27 years (Mean = 22.7 years)  - 14.4 (95% CI, 3.4-61.4) - is similar to the OR calculated for the Bell and Weinberg (1978) data for male to the age of 20: 14.9 (95% CI, 3.6-61.6) - Predominantly homosexual males versus predominantly heterosexual males: 55 suicide-attempters and 520 nonattempters versus 2 suicide attempters and 282 nonattempters. These ORs are also similar to the estimated ORs calculated from the data given by Cochran and Mays (2000) for males (age range = 17 to 29 years) males reporting having had a least one lifetime same-sex partner  versus the males who reported having only female sex partner(s), or versus the males reporting only having female sexual partners combined with the ones reporting not having any sexual partners. Respectively, the ORs are 12.5 (95% CI, 6.4-24.2) and 13.8 (95% CI, 7.1-26.8). The Odds Ratios were calculated using the 2 X 2 Table located at - http://home.clara.net/sisa/twoby2.htm - but they are only "approximations" given that they were estimated from the data given by Cochran and Mays (2000) and not generated from the actual data set:

Odds Ratio Calculation for males reporting lifetime male sex partner(s) versus males reporting only female sex partner(s): 15 / 32 (attempters / nonattempters) versus 60 / 1595 (attempters / nonattempters),  - 95% CI, Odds Ratio: 6.41<12.46<24.24.

Odds Ratio Calculation for males reporting lifetime male sex partner(s) versus males reporting only female sex partner(s) plus males reporting a celibate status: 15 / 32 (attempters / nonattempters) versus 61 [60 + 1]  / 1796 [1595 + 201] - 95% CI, OR: 7.10<13.8<26.82.

Additional Information related to all above estimated OR values is available at - http://www.youth-suicide.com/gay-bisexual/gbsuicide1.htm  and
http://fsw.ucalgary.ca/ramsay/gay-lesbian-bisexual/2g-gay-suicide-study.htm .
 

Previous Section
Top of Page
Email to Richard RamsayEmail to Pierre Tremblay