Risk Factors for Attempted Suicide in Gay and

Bisexual Youth

Gary Remafedi, MD, MPH*; James A. Farrow, MD4; and

Robert W. Deisher, MD*

From the *Adolescent Health Program, Department of Pediatrics, University of Minnesota

Hospital and Clinic, Minneapolis, and jDivision of Adolescent Medicine, Department of
Pediatrics, University of Washington, Seattle

ABSTRACT. Studies of human sexuality have noted high The impact of sexual identity on suicide risk is a
rates of suicidality among homosexual youth, but the
relatively uncharted area of research. To date, most
problem has not been systematically examined. This
surveys of suicide attempters and psychological au-
work was undertaken to identify risk factors for suicide
attempts among bisexual and homosexual male youth. topsies of victims have not examined sexual dimen-
Subjects were 137 gay and bisexual males, 14 through 21 sions beyond gender.3 However, surveys of homo-
years of age, from the upper Midwest and Pacific North- sexual populations have raised questions about su-
west. Forty-one subjects (41/137) reported a suicide at-
icide risk in relation to sexual orientation. An
tempt; and almost half of them described multiple at-
tempts. Twenty-one percent of all attempts resulted in
unusual prevalence of suicide attempts and ideation
medical or psychiatric admissions. Compared with non- among homosexual persons has surfaced repeatedly
attempters, attempters had more feminine gender roles as an incidental finding in studies of human sex-
and adopted a bisexual or homosexual identity at younger uality.48
ages. Attempters were more likely than peers to report Forty percent of 5000 homosexual men and
sexual abuse, drug abuse, and arrests for misconduct. The
findings parallel previous studies’ results and also intro-
women who were surveyed by Jay and Young4 se-
duce novel suicide risk factors related to gender noncon- miously considered or attempted suicide. Bell and
formity and sexual milestones. Pediatrics 1991;87:869- Weinbemg found that 1000 black and white homo-
875; adolescence, suicide, homosexuality, gender role, sub- sexual men were, respectively, 12 and 3 times more
stance abuse, sexual abuse.
likely than heterosexual men to report suicidal idea-
tion or attempts. The homosexual men were more
likely to have made attempts during adolescence
than adulthood. Saghim and Robins6 reported that
According to most recent statistics, more than all suicide attempts in a cohort of homosexual
5000 US adolescents and young adults (aged 15 to adults occurred during adolescence, often in asso-
24) take their own lives each year.’ During the past ciation with a history of childhood gender atypical
25 years, suicide mates for young men quadrupled, behavior or emotional disturbance. In two different
and self-inflicted death became the second leading studies of homosexual and bisexual adolescents,7’8
cause of adolescent mortality. These disturbing one third of boys reported attempts, and repeat
trends have led to an ongoing search for epidemio- attempts were common.
logical, psychological, and sociological risk factors. Two thirds of randomly sampled US psychia-
Despite considerable progress and new information, trists#{176}believed that the self-injurious acts of ho-
the unifying characteristics of young victims are mosexual adolescents were more serious and lethal
still incompletely understood.2 than those of their heterosexual peers. The recent
Report of the Secretary ‘s Task Force on Youth
0 projected that gay adolescents were two
Received for publication May 14, 1990; accepted Jul 30, 1990. to three times more likely than peers to attempt
Reprint requests to (G.R.) Adolescent Health Program, Box 721,
suicide, accounting for as many as 30% of corn-
University of Minnesota Hospital and Clinic, Minneapolis, MN
pleted youth suicides each year. Theoretical risk
55455.
PEDIATRICS (ISSN 0031 4005). Copyright © 1991 by the factors include “coming out” at a young age, gender
American Academy of Pediatrics. atypicality, low self-esteem, substance abuse, run-

Downloaded from www.aappublications.org/news at Indonesia:AAP SponsoredVol.

PEDIATRICS on August
87 No.19, 2019
6 June 1 991 869
ning away, involvement in prostitution, and other percent of participants (41/137) reported at least
psychosocial morbidities. However, such predictors one suicide attempt.
have not been studied empirically. This work was
undertaken to identify risk factors for suicide at- METHODS
tempts among gay and bisexual youth, with a
broader goal to advance understanding of self-in- All participants completed a structured interview
regarding demography, education, home environ-
flicted deaths among adolescents.
ment, sexuality, and psychosocial history (adapted
from a previous study of adolescent homo-
SUBJECTS
sexuality).” Suicide attempts were defined as delib-
Subjects were 137 males, between 14 and 21 years emate acts intended to cause death. Suicide attempts
of age, who identified themselves as gay (88%) or were assessed in interview, using Weisman and
bisexual (12%). They were recruited during a 1- Worden’s Risk Rescue Rating Scale.’2 When sub-
year period of time (1988) through advertisements jects reported multiple attempts, only the latest
in gay publications (30%) and bars (5%), social three were used.
support groups for gay and lesbian youth (20%) and According to Weisman and Womden,12 the overall
university students (15%), a youth drop-in center seriousness of a suicide attempt is “a balance of
(19%), and referral from peers (11%). None were calculated factors related to the degree of immevems-
referred from mental heath treatment facilities. ible damage and the resources that facilitate or
Interviews took place in Minneapolis or Seattle. hinder rescue.” Weisman and Womden’s method
The project was advertised as a study of health rates attempts by degrees of risk and rescuability
issues for gay and bisexual male youth, 21 years of and provides a composite ratio of risk and rescue
age or younger. factors. “Risk points” correspond to the method of
Participants resided in the states of Minnesota injury, subsequent physical impairment, and the
(67%), Washington (31%), South Dakota (1%), or type of medical care that was administered. “Rescue
Wisconsin (1 %). The ethnic/racial composition of points” are determined by the victim’s location,
the group was 82% white, 13% African-American, request for help, the rescuer’s identity, and the
4% Hispanic, and 1% Asian. Other demographic duration of time until discovery. The total risk and
characteristics are summarized in Table 1. Thirty rescue points are converted to respective scores,
which correspond to intervals on Likert-type scales
TABLE 1. Demographic Characteristics (Means and ranging from “least” to “most” risky or mescuable.
Percent Frequency Distribution [%F]) of Suicide At- When the victims sought help for themselves, their
tempters and Nonattempters attempts automatically receive a “most rescuable”
Variable Attempters Nonattempters score. The composite risk-rescue ratio is computed
(n=41) (n=96) by dividing the risk score by the sum of risk and
Mean age, y (SD) 19.25 (1.63) 19.83 (1.63) rescue factors.
Mean grade level (SD) 11.58 (1.86) 12.42 (1.98)
All participants also completed (in interview) the
Race, %F
White 76 85 Scale for Suicide Ideation,’3 which measures cur-
Black, Native Amen- 24 15 rent suicidal intent by scaling various dimensions
can, Hispanic, of self-destructive thoughts and wishes. Four other
Asian written instruments were administered. The Bern
Residence, %F
Sex Role Inventory’4 was used to rate masculinity
Urban 66 71
Suburban/rural 34 29 and femininity and to classify sex-mole as feminine,
Interview location, %F masculine, androgynous (ie, high masculine/high
Minneapolis 59 74 feminine), or undifferentiated (ie, low masculine!
Seattle 42 26
low feminine). Current levels of depression and
Primary financial sup-
hopelessness were respectively measured with the
port, %F
Employment (self) 43 55 Beck Depression Inventory15 and the Hopelessness
Parents 35 34 Scale.’6 Finally, personal attitudes toward horno-
Other 22 11 sexuality were rated by the Modified Attitudes To-
Religion, %F ward Homosexuality Scale.’7
Catholic 20 24
Protestant 24 35 Subjects gave verbal and written consent to study
Other* 29 24 procedures. Participation was voluntary and confi-
None 27 17 dential. Completion of procedures required approx-
* Includes non-Christian faiths or religiosity, without imately 1.5 hours. Subjects were reimbursed for
specific affiliation. participation and were given a list of resources for

870 SUICIDE Downloaded

AND from www.aappublications.org/newsYOUTH
HOMOSEXUAL/BISEXUAL at Indonesia:AAP Sponsored on August 19, 2019
social support and mental health and medical care. stance abuse (15%). In one case, a suicide attempt
Those who indicated active suicidal intent or other followed notification of a partner’s human immu-
acute problems were immediately referred for care, nodeficiency virus seropositivity.
according to a protocol approved by the University Fifty-four percent of all suicide attempts (37/68)
of Minnesota Human Subjects Committee. Study received risk scores in the “moderate to high” le-
procedures were administered by a physician (Mm- thality mange. The remaining attempts were of the
neapolis) and a nurse (Seattle). “low risk” type. In one third of all cases (23/68),
recovery was unassisted, and rescue scores could
STATISTICAL ANALYSIS not be assigned by Weisman and Wonden’s criteria.
These ranged from trivial injuries not needing
Response frequencies and means were computed
treatment to severe insults with fortuitous recov-
for the suicide attempter and the nonattemptem
emies. A rescue was initiated by the victim (24%) or
comparison groups. Independent samples x2 and t
by another person (76%) in the remaining 45 cases.
tests were used to compare responses from the two
Fifty-eight percent (26/45) of these cases received
groups. To limit the chance occurrence of signifi-
scores in the “moderate to least” rescuable mange.
cant findings, a .01 level of statistical significance
In other words, the predicted likelihood of rescue
was chosen for univariate analyses. Multiple logis-
was moderate to low, despite the actual occurrence
tic regression analysis was used to identify variables
of an intervention.
that were independently predictive of suicide at-
Table 2 depicts the distribution of risk and rescue
tempts at a .05 level of significance.
scores for the 45 attempts for which both scores
were available. Almost half of them (45%) were
RESULTS
associated with moderate risk and moderate likeli-
hood of rescue. Composite risk-rescue scores also
Description of Suicide Attempts
were computed for these attempts. Thirty-six per-
Thirty percent of subjects (41/137) reported at cent (18/45) received scores at or above Weisman
least one suicide attempt, and almost half of the and Worden’s empirically derived mean (40) for
attempters (18/41) reported more than one at- adult attempters hospitalized in medical or psychi-
tempt. A total of 68 suicide attempts were described atmic units.
and mated by Weisman and Womden’s method. The
mean age at the time of suicide attempts was 15.5 Univanate Analyses
years. Ingestion ofprescniption and/or nonprescnip- The following is a summary of the similarities
tion drugs and self-laceration accounted for 80% of
and differences among the comparison groups. Fur-
attempts. The remainder involved hangings, carbon
them detail is provided in Table 3. There were no
monoxide poisonings, jumping, firearms, and auto- statistically significant differences (P < .01) be-
motive crashes. Twenty-one percent of the suicide
tween attempters and nonattempters with regard
attempts (14/68) resulted in a medical or psychi- to age, educational level, race, religion, residence,
atnic hospitalization. Almost 3 out of 4 attempts or source of financial support. The majority of all
(50/68) did not receive any medical attention.
participants (107/137, 78%) had received some type
These included some attempts with high potential
of professional mental health came. However, mean
lethality, including two intentional car accidents,
scores for current depression, hopelessness, and
two carbon monoxide asphyxiations, and ingestions suicide intent (as measured by the respective in-
of large quantities of antidepmessants, acetamino-
struments) were uniformly low for attempters and
phen, aspirin, or intravenous heroin.
nonattemptems. Also, no statistically significant (P
In 44% of cases, subjects attributed suicide at- < .01) differences were noted in the reported occur-
tempts to “family problems,” including conflict rence of depression, attempted or completed sui-
with family members and parents’ marital discord,
cide, and psychiatric hospitalization among family
divorce, or alcoholism. One third of attempts were
related to personal or interpersonal turmoil regard-
TABLE 2. Number (and Percent Frequency Distnibu-
ing homosexuality. Almost one third of subjects
tion) of Suicide Attempts (n = 45), Classified by Likeli-
made their first suicide attempt in the same year hood of Risk and Rescue
that they identified themselves as bisexual or ho-
Likelihood Risk of Morbidity/Mortality
mosexual. Overall, three fourths of all first attempts of Rescue
temporally followed self-labeling. Other common Low Moderate High

precipitants were depression (30%), conflict with Least 1(2) 0 0

peers (22%), problems in a romantic relationship Moderate 5 (11) 20 (45) 1 (2)

Most 10 (22) 8 (18) 0
(19%), and dysphoria associated with personal sub-

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on August 19, 2019

ARTICLES 871
 TABLE 3. Psychosocial Characteristics (Mean and Percent Frequency Distribution
%F]) of Suicide Attempters and Nonattempters
Variables Attempters Nonattempters df P Value*
(n=41) (n=96)

Sexual milestones: mean ages,

y (SD)
Earliest bi/homosexual attrac- 9.27 (2.89) 10.66 (3.84) 130 .050
tions
Bi/homosexual self-labeling 13.74 (3.22) 15.44 (2.61) 132 .002
First homosexual experience 14.38 (2.48) 16.13 (2.54) 129 .000
First heterosexual experience 14.61 (2.56) 16.10 (1.87) 57 .030
“Coming out” to others 15.62 (2.92) 16.95 (2.22) 128 .005
Psychosocial stressors, %F
Parents’ marital status
Unmarried 7 6
Divorced 61 42 3 .08
Married 27 50
Mother knows son’s sexuality 87 81 1 .44
Supportive maternal response 32 38 4 .62
Father knows son’s sexuality 65 56 1 .45
Supportive paternal response 23 15 4 .58
Friendship loss 42 28 1 .18
Peer suicide 39 27 1 .24
Discrimination 61 46 1 .15
Violence 39 38 1 1.00
Sexual abuse 61 29 1 .0008
Psychosocial problems, %F
Running away 49 36 1 .25
Arrest 51 28 1 .01
Prostitution 29 17 1 .20
Ethanol ever use 78 83 1 .62
Illicit drug use 85 63 1 .01
Personal mental health, %F
Any mental health services 88 74 1 .12
Chemical dependency treatment 22 6 1 .01
No current suicide intent (item 68 84 1 .06
from Scale for Suicide Idea-
tion)
No suicide plan (item from Scale 93 98 1 .32
for Suicide Ideation)
Family mental health,t %F
Depression 43 40 1 .92
Attempted/complete suicide 38 22 1 .12
Psychiatric hospitalization 44 45 1 .12
Psychometric indices: mean scores
(SD)
Beck Depression Inventory 10.78 (8.60) 7.68 (7.20) 135 .03
Scale for Suicide Ideation 6.90 (6.40) 4.61 (4.50) 135 .04
Hopelessness Scale 4.37 (0.50) 2.90 (0.40) 135 .10
Attitudes toward homosexuality 56.24 (14.5) 56.35 (0.19) 133 .97
Bern-masculinity raw score 4.75 (0.66) 5.04 (0.79) 135 .04
Bem-femininity raw score 5.08 (0.66) 4.91 (0.61) 135 .15
Bern classification, %F
Masculine 7.3 26.0
Feminine 36.6 17.7
Androgynous 26.8 32.3 3 .01
Undifferentiated 29.3 24.0
* P values correspond to x2 tests for nominal variables and t tests for continuous variables.
t Includes first- to third-degree relatives.

872 SUICIDE Downloaded from www.aappublications.org/news

AND HOMOSEXUAL/BISEXUAL at Indonesia:AAP Sponsored on August 19, 2019
YOUTH
members. Approximately one third of subjects in gender mole, criminal apprehension, and sexual
both groups were acquainted with a peer who had abuse were included. All five variables were ob-
committed suicide. tamed in 96% (132/137) of cases.
Differences in gender role and sexual orientation Based on the multiple logistic regression analysis,
development were prominent. Based on the Bern ever use of illicit drugs, femine gender role, and age
classification, attempters were more likely than at the time of bisexual or homosexual self-labeling
nonattempters to be feminine or undifferentiated, were independently associated with suicide at-
and less likely to be masculine or androgynous (P tempts at the .05 level of statistical significance
= .01). Compared with nonattempters, attempters (Table 4). Feminine gender roles and illicit drug
described themselves as bisexual or homosexual (P use each were associated with greaten than a three-
= .002) and shamed this with other persons (P = fold risk of attempted suicide. According to the
.005) at younger ages. They also engaged in homo- model, the likelihood of an attempt diminished with
sexual (P = .0001) and heterosexual activity (P = advancing age at the time of bisexual or homosexual
.03) to the point of orgasm at younger ages than self-labeling. With each year’s delay in self-identi-
their peers, although differences in the mean ages fication, the odds of a suicide attempt declined by
of first heterosexual activity were not statistically more than 80%.
significant. As indicated by Table 3, bisexual or
homosexual self-identification generally preceded DISCUSSION
first sexual experiences with either gender.
Approximately one third of gay and bisexual
Suicide attempters were more likely than nonat-
youth in this study reported at least one intentional
tempters (P = .0008) to report sexual abuse, broadly
self-destructive act, and almost half of them me-
defined as being forced, pressured, or tricked to
peatedly attempted suicide. The gravity of some
have sex. Detailed descriptions of incidents were
attempts is reflected in the mate of subsequent hos-
not obtained. Twenty-nine percent of attempters
pitalization (21%), the lethality of methods (54%,
and 17% of nonattempters accepted money for sex
moderate to high risk), and the victims’ inaccessi-
on at least one occasion. Both groups reported
bility to rescue (62%, moderate to least nescuable).
similar personal attitudes toward homosexuality,
The findings support psychiatrists’ concerns about
reactions from parents and friends, and experiences
the severity of suicide attempts among homosexual
with discrimination and violence.
youth.9
Compared with peers, a larger proportion of at-
tempters reported illicit drug use (85% vs 63%, P In this sample, bisexuality on homosexuality pen
se was not associated with self-destructive acts.
= .01) and arrest for criminal activities (51% vs
Most of the subjects did not attempt on plan suicide.
28%, P = .01). Twenty-two percent of attempters
However, from the perspective of many attempters,
(vs 6% of nonattempters, P = .01) had undergone
sexual concerns were circumstantially on tempo-
chemical dependency treatment. Theft (40%) and
rally related to self-harm. One third of all suicide
possession of illicit substances (20%) accounted for
attempts were attributed to personal or interpen-
the majority of reported arrests. The remaining
sonal turmoil about homosexuality. One third of
violations included assault, vandalism, disorderly
first attempts occurred in the same year that sub-
conduct, disturbing the peace, trespassing, prosti-
jects identified their bisexuality on homosexuality,
tution, and other sex offenses.
and most other attempts happened soon thereafter.
The apparent connection between sexual mile-
Multivariate Analyses
stones and attempts may be a clue to the appnopni-
The univaniate analyses revealed a total of nine ate timing of suicide prevention efforts.
variables that were associated with suicide attempts Based on the univaniate analyses, suicide at-
at the .01 level of statistical significance. Multiple tempts were not explained by experiences with dis-
logistic regression analysis was undertaken to iden- cnimination, violence, loss of friendship, on current
tify which were predictive of attempts, when con- personal attitudes toward homosexuality. Unlike
trolling for the other variables. Of the nine van- other reports,’8 this study did not find a significant
ables, five were entered into the analysis based on association between attempts and running away
their conceptual inclusivity and completeness of from home. However, gender nonconformity and
ascertainment in this study. From the four items precocious psychosexual development were predic-
related to sexual milestones, age at the time of tive of self-harm. Compared with peers, suicide
bisexual or homosexual self-labeling was selected. attempters recognized homosexual attractions and
From the two items regarding substance abuse, ever told other persons at younger ages. First sexual
use of illegal drugs was chosen. Finally, feminine experiences with males and females also occurred

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on August 19, 2019

ARTICLES 873
 TABLE 4.__Summary Table:_RegressionAnalysis of Suicide Attempts
Variable Odds Ratio 95% Regression SE P Value
Confidence Interval Coefficient
Constant . . . . . . 0.24 1.24 ...

Label age 0.82 0.69-0.96 -0.21 0.08 0.01

Drug use 3.63 1.07-12.24 1.29 0.61 0.03
Feminine 3.03 1.10-8.31 1.11 0.50 0.03
Sex abuse 2.23 0.90-5.53 0.81 0.45 0.08
Arrest 1.87 0.74-4.67 0.62 0.46 0.06

at younger ages than peers’. For each year’s delay ency treatment. More than half (21/41) of the at-
in bisexual or homosexual self-labeling, the odds of tempters had been arrested for misconduct. Like
a suicide attempt diminished by 80%. These find- other boys who eventually complete suicide,26’35’36
ings support a previously observed, inverse relation- many subjects made multiple unsuccessful at-
ship between psychosocial problems and the age of tempts. In general, many of the psychosocial pmob-
acquiring a homosexual identity.8 Compared with lems associated with gay suicide attempts (eg, fam-
olden persons, early and middle adolescents may be ily discord, substance abuse, and conflict with the
generally less able to cope with the isolation and law) are familiar correlates of completed youth
stigma of a homosexual identity. suicide. Such problems often complicate “coming
A feminine or undifferentiated gender role may out” at an early age8 and may further contribute to
accentuate a gay adolescent’s sense of “different- the high mate of suicide in this group.
ness” and further exacerbate problems.’9 Severely In contrast to some psychological autopsies of
gender-atypical male children have been observed young victims,26’3”35 this study found low levels of
to experience a “pervasive psychological disturb- active depression, hopelessness, and suicidal intent
20 and an “abnormal amount of depression among attempters. Moreover, the attempters and
and social conflict resulting from peer rejection, nonattempters had similar family histories of
isolation, and ridicule of their feminine 2 depression, suicide, and psychiatric hospitalization.
In general, masculine and androgynous boys and Although chronic or heritable forms of depression
girls alike have better self-esteem22 and lower rates did not account for suicide attempts in this group,
of substance use and psychological distress23 than the results must be interpreted cautiously. Acute
feminine or undifferentiated youth. Among male depression and hopelessness at the time of the
and female college students, masculinity and an- attempt may have since lifted with treatment or
drogyny are associated with positive adjustment to the passage of time, and subjects may not have been
stress24 and minimal fearfulness.25 well-informed of family mental histories.
The univaniate analyses also revealed an associ- The circumstances, prevalence, and severity of
ation between sexual abuse and suicide attempts. suicide attempts in this cohort may not reflect the
However, when controlling for other variables in general population of homosexually oriented boys
the multiple regression analysis, sexual abuse was and girls. Because of the social stigma of homo-
not a statistically significant predictor. The contni- sexuality, a probability sampling of homosexual
bution of sexual abuse to suicide risk may have youth was unfeasible. However, an effort was made
been subsumed by other variables such as drug use to enhance the genenalizability of the findings by
or gender atypicality. For example, gender-atypical recruiting subjects from diverse settings and geo-
boys may be vulnerable to sexual assault, as well as graphical areas. Gay-identified adolescents are a
suicide. Alternatively, abused boys may develop subset of all youth who will eventually disclose a
substance abuse problems that heighten suicide homosexual orientation.37 As our own findings sug-
risk. Also noteworthy, sexual abuse did not appear gest, the experiences of openly gay and bisexual
to have a major impact on sexual identity, because youth may be quite different from those of other
bisexual or homosexual identification usually pre- boys who are confused, hiding, or delayed in iden-
ceded sexual experiences. tifying sexual feelings. Likewise, the risks for sui-
The attempters in this study resembled actual cide among lesbian girls may be quite different from
suicide victims in regard to high levels of family those among boys because of the gender, gender
dysfunction,2 personal substance abuse,34 and role, sexual identity development, and cultural dif-
other antisocial behaviors.27’32’35 Family problems fenences.
were the most frequently cited reason for attempts. Acknowledging these limitations, the unusual
Eighty-five percent of attempters reported illicit prevalence of serious suicide attempts remains a
drug use, and 22% had undergone chemical depend- consistent and disturbing finding in the existing

874 SUICIDE Downloaded from www.aappublications.org/news

AND HOMOSEXUAL/BISEXUAL at Indonesia:AAP Sponsored on August 19, 2019
YOUTH
reports of young homosexual males. The rate of 11. Remafedi G. Male homosexuality: the adolescent’s perspec-
tive. Pediatrics. 1987;79:326-330
completed suicide among homosexual attempters is 12. Weisman AD, Worden JW. Risk-rescue rating in suicide
unknown, as is the relative contribution of homo- assessment. Arch Gen Psychiatry. 1972;26:553-560
sexual adolescents to total youth suicides. These 13. Beck AT, Kovacs M, Weissman A. Assessment of suicidal
intention: the scale for suicide ideation. J Consult Clin
issues merit further investigation to illuminate the Psychol. 1979;47:343-352
epidemiology of adolescent suicide, the direction of 14. Bem S. The measurement of psychological androgyny. J
suicide prevention programs, and the care of mdi- Consult Clin PsychoL 1974;42:155-162
15. Beck AT, Ward CH, Mendelson M, et al. An inventory for
vidual clients. We hope that this study’s findings measuring depression. Arch Gen Psychiatry. 1961;4:561-571
will lead to other investigations of sexuality and 16. Beck AT, Weissman A, Lester D, et al. The measurement
suicide among representative samples of youth in of pessimism: the hopelessness scale. J Consult Clin Psychol.
1974;42:861-865
schools or other community settings. Ultimately, 17. Price JH. High school students’ attitudes toward homo-
the study of suicide among gay and bisexual youth sexuality. J School Health. 1982;52:469-474
may shed new light on the unifying characteristics 18. Rotheram MJ. Evaluation of imminent danger for suicide
among youth. Am J Orthopsychiatry. 1987;57:102-110
of adolescent victims, unraveling the common 19. Martin AD. Learning to hide: the socialization of the gay
threads of risk which transcend the issue of sexual adolescent. Adolesc Psychiatry. 1982;10:52-65
orientation. 20. Coates 5, Person ES. Extreme boyhood femininity: isolated
behavior or pervasive disorder. J Am Acad Child Psychiatry.
1985;24:702-709
21. Rosen AC, Rekers GA, Friar LR. Theoretical and diagnostic
ACKNOWLEDGMENTS issues in child gender disturbances. J Sex Res. 1977;13:89-
103
This work was supported, in part, by grants from the 22. Lamke LK. The impact of sex-role orientation of self-esteem
Bureau of Maternal and Child Health and Resource in early adolescence. Child Dev. 1982;53:1530-1535
Development (MCJ-000985) and the Grant-in-Aid Pro- 23. Horwitz AV, White HR. Gender role orientations and styles
gram of the Graduate School, University of Minnesota. of pathology among adolescents. J Health Soc Be/wv.
Statistical consultation was provided by Kinley Larntz, 1987;28:158-170
24. Roos PE, Cohen LH. Sex roles and social support as mod-
PhD, Professor of Statistics. Special thanks to Robert
erators of life stress adjustment. J Per Soc Psychol.
Blum, MD, Kevin Cwayna, MD, Barry Garfinkel, MD,
1987;3:576-585
Mary Story, PhD, Kenneth Winters, PhD, and John 25. Dillon KM, Wolf E, Katz H. Sex roles, gender, and fear. J
Yoakam, MDiv, for their consultation and to Dean Psychol. 1985;1 19:355-359
McWilliams, Joy Love, MDiv, and W. S. Foster for their 26. Thompson TR. Childhood and adolescent suicide in Mani-
technical assistance. toba: a demographic study. Can J Psychiatry. 1987;32:264-
269
27. Shaffer D. Suicide in childhood and early adolescence. J
Child Psychol Psychiatry. 1974;15:275-291
REFERENCES
28. Hoberman HM, Garfinkel BD. Completed suicide in chil-
1. National Center for Health Statistics. Vital Statistics of the dren and adolescents. J Am Acad Child Adolesc Psychiatry.
United States, Volume 2: Mortality, Part A. Hyattsville, MD; 1988;27:689-695
1986 29. Peck M. Suicide in late adolescence and young adulthood.
2. Hoberman HM. Completed suicide in children and adoles- In: Hatton CL, Velenti SM, eds. Suicide: Assessment and
cents: a review. Residential Treatment Child Youth. Intervention. Norwalk, CT: Appleton-Century-Crofts;
1989;7:61-88 1984:220-230
3. Remafedi G. Adolescent homosexuality: a challenge to con- 30. Brent DA, Perper JA, Goldstein CE, et al. Risk factors for
temporary
society. JAMA. 1988;258:222-225 adolescent suicide. Arch Gen Psychiatry. 1988;45:581-588
4. Jay K, Young A, eds. The Gay Report: Lesbians and Gay 31. Poteet DJ. Adolescent suicide: a review of 87 cases of corn-
Men Speak Out About Their Sexual Experiences and Life- pleted suicide in Shelby County, Tennessee. Am J Forensic
styles. New York, NY: Simon and Schuster; 1979 Med Pathol. 1987;8:12-17

5. Bell A, Weinberg M. Homosexualities: A Study of Diversity 32. Shaffer D. The epidemiology ofteen suicide: an examination
Among Men and Women. New York, NY: Simon and Schus- of risk factors. J Clin Psychiatry. 1988;49:36-41
ter; 1978 33. Slap GB, Vorters DF, Chaudhuri S, Centor RM. Risk factors
6. Saghir MT, Robins E. Male and Female Homosexuality: A for attempted suicide during adolescence. Pediatrics.
Comprehensive Investigation. Baltimore, MD: Williams & 1989;84:762-772
Wilkins; 1973 34. Rich CL, Fowler RC, Young D, Blenkush M. San Diego
7. Roesler T, Deisher RW. Youthful male homosexuality. suicide study: comparison of gay to straight males. Suicide
JAMA. 1972;219:1018-1023 Life Threat Behav. 1986;16:448-457
8. Remafedi G. Adolescent homosexuality: psychosocial and 35. Shaffi N, Carrigan 5, Whittinghill JR, Derric A. Psychology
medical implications. Pediatrics. 1987;79:331-337 autopsy of completed suicide in children and adolescents.
9. Kourany RF. Suicide among homosexual adolescents. J Am J Psychiatry. 1985;142:1061-1064
Homosex. 1987;13:111-117 36. Cosand BJ, Bourque ML, Krauss JF. Suicide among adoles-
10. US Dept of Health and Human Services. Report of the cents in Sacramento County, California, 1950-1979. Adoles-
Secretary ‘a Task Force on Youth Suicide, Volume 3: Preven- cence. 1982;17:917-930
tion and Interventions in Youth Suicide. Rockville, MD; 37. Remafedi G. Adolescent homosexuality. Med Clin North
1989 Am. 1990;74:1169-1179

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on August 19, 2019

ARTICLES 875
 Risk Factors for Attempted Suicide in Gay and Bisexual Youth
Gary Remafedi, James A. Farrow and Robert W. Deisher
Pediatrics 1991;87;869

Updated Information & including high resolution figures, can be found at:
Services http://pediatrics.aappublications.org/content/87/6/869
Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its
entirety can be found online at:
http://www.aappublications.org/site/misc/Permissions.xhtml
Reprints Information about ordering reprints can be found online:
http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on August 19, 2019

 Risk Factors for Attempted Suicide in Gay and Bisexual Youth
Gary Remafedi, James A. Farrow and Robert W. Deisher
Pediatrics 1991;87;869

The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://pediatrics.aappublications.org/content/87/6/869

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it has
been published continuously since 1948. Pediatrics is owned, published, and trademarked by the
American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois, 60007.
Copyright © 1991 by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on August 19, 2019