Abstract
Objectives. We compared sexual-minority adolescents living in rural communities with their peers in urban areas in British Columbia, exploring differences in emotional health, victimization experiences, sexual behaviors, and substance use.
Methods. We analyzed a population-based sample of self-identified lesbian, gay, or bisexual respondents from the British Columbia Adolescent Health Survey of 2003 (weighted n = 6905). We tested rural–urban differences separately by gender with the χ2 test and logistic regressions.
Results. We found many similarities and several differences. Rural sexual-minority adolescent boys were more likely than were their urban peers to report suicidal behaviors and pregnancy involvement. Rural sexual-minority adolescents, especially girls, were more likely to report various types of substance use. Rural status was associated with a lower risk of dating violence and higher risk of early sexual debut for sexual-minority girls and a higher risk of dating violence and lower risk of early sexual debut for sexual-minority boys.
Conclusions. Location should be a demographic consideration in monitoring the health of sexual-minority adolescents. Lesbian, gay, and bisexual adolescents in rural communities may need additional support and services as they navigate adolescence.
Adolescence is marked by many developmental tasks, including the unfolding of sexual identity. Adolescents who are lesbian, gay, or bisexual (LGB) face the additional challenge of being members of a sexual minority in a heterosexually dominant world. This may be an especially complex task in a rural environment, where heterosexism may be more pronounced1 and supportive resources limited.2 Scant research has compared the experiences of LGB adolescents in rural and urban areas.
In the past decade, a growing number of population-based studies of LGB adolescents addressed the methodological criticisms of earlier studies that relied on urban samples, convenience samples, and samples of adolescents who may have been of legal adult age.2–5 Population-based research on LGB adolescents has demonstrated that they face health disparities when compared with heterosexual adolescents. For example, LGB adolescents are at increased risk of being stigmatized and victimized.3,6–11 LGB adolescents also experience higher rates of emotional distress and suicidality,3,6,8,11–16 substance use,3,6,8,11,17 and risky sex,6,8 which may put them at increased risk for HIV infection18,19 and pregnancy.9
Rural communities have several characteristics that may affect the experiences of LGB adolescents. Social isolation may be greater in rural regions, because the chance to identify with an LGB peer group may be limited or nonexistent.2,20,21 Adolescents and their families may also lack access to resources for information and support.2,20 Furthermore, more conservative attitudes in general, and negative attitudes or misconceptions about nonheterosexual orientations specifically, and less anonymity in small communities may make it harder for adolescents to openly express same-gender affections2 and risk public disclosure.22 Thus, LGB adolescents in rural areas may face greater disparities and challenges compared with their urban counterparts.
Results of research on rural–urban differences in the general population have been mixed. For example, a study of multiple data sets by the US National Center on Addiction and Substance Abuse found substance-use rates among rural adolescents to be generally higher.23 Likewise, a study of more than 2000 students from urban, suburban, and rural schools in upstate New York found that rural adolescents were twice as likely as their suburban and urban counterparts to frequently use tobacco, alcohol, and other drugs and to have had sex.24 Conversely, analyses of the 1999 US Youth Risk Behavior Survey found no differences among urban and nonurban adolescents on several substance-use and sexual behavior variables.25 A separate report on health in the United States found that adolescent childbearing rates were lowest in suburban areas.26 Studies of dating violence have also had mixed results, finding rural adolescents at higher risk27 or lower risk.28
Only a few studies have directly compared LGB adolescents living in rural areas with those living in urban areas. Waldo et al. studied LGB rural and urban young people aged 15 to 21 years.29 Urban participants were recruited through metropolitan community centers and rural participants through university student groups in a politically conservative town more than an hour away from a metropolitan center. They found rates of assault were lower for rural young people, but suicidality did not differ between rural and urban participants. In an analysis of population-based data from the US National Longitudinal Study of Adolescent Health, Galliher et al. defined urban, suburban, and rural from interviewers’ observations of the immediate are around the participant's residence.12 They found that rural sexual-minority adolescent girls reported more depressive symptoms, but rural adolescent boys, regardless of attraction status, reported a greater sense of school belonging, greater self-esteem, and fewer depressive symptoms.
We used school-based representative data from British Columbia to explore differences in health and risk between LGB adolescents in rural communities and their urban counterparts. We focused on LGB adolescents only, because previous research has documented the health disparities between LGB and heterosexual adolescents. In contrast to previous studies,12,29 we looked at a broader range of behaviors and used a standardized definition of rurality to categorize location.
We conducted a secondary analysis of LGB participants in the 2003 British Columbia Adolescent Health Survey (BCAHS). This paper-and-pencil, anonymous, voluntary questionnaire contained 140 items assessing physical and mental health, risky behaviors, health-promoting behaviors, and protective factors. The questionnaire was administered to a cluster-stratified random sample of more than 1500 classes in grades 7 to 12 in public schools in British Columbia, stratified geographically by health service delivery area and by grade level. Original data from 30 588 students were weighted to adjust for differential probability of sampling and differential response rates to be representative of the approximately 290 000 students enrolled during the 2002 to 2003 school year. Statistics Canada served as methodology consultant for the survey, designed the sampling frame, and weighted the data. Individual school districts determined whether participation required written parental consent or parental notification with student consent. The overall response rate for the survey was 76%.
Sexual orientation was assessed with an item about romantic attractions used in several studies,9,13,18 with the response options of “100% heterosexual (attracted to persons of the opposite sex),” “mostly heterosexual,” “bisexual (attracted to both males and females),” “mostly homosexual,” “100% homosexual (‘gay/lesbian’; attracted to persons of the same sex),” and “not sure.”
We explored the survey items concerning emotional health, victimization, sexual behavior, and substance use. Emotional health items included worries about family, emotional distress, and suicidality. Victimization items assessed abuse, dating violence, sexual harassment, and discrimination, as well as verbal harassment (being teased), purposeful exclusion, or physical assault experienced at school or on the way to or from school. Sexual health information included ever having sex, number of partners, use of condoms and birth control, and pregnancy involvement. Items about substance use concerned alcohol, marijuana, other drugs, and problems experienced following alcohol or drug use (e.g., got into car accident, damaged property). A detailed description of the BCAHS instrument and methodology is available elsewhere.30
We adopted the definitions of rural and urban from Statistics Canada's census definitions. Urban areas were classified as census metropolitan areas (urban core ≥ 100 000) or census agglomerations (urban core ≥ 10 000). An urban area consisted of the urban core as well as neighboring municipalities from which at least 50% of the labor force commuted to the urban core.31 Rural areas had populations under 10 000, with less than 50% of the labor force commuting to an urban center.31 The rural classification, because it considered commuting patterns as well as population size, reflected the level of social and economic integration between a rural area and urban center. This integration included access that rural residents might have to the urban center's health care, educational, recreational, financial, and cultural services.32
The rural–urban classification was applied to student records by matching records from the 2003 BCAHS and the 2001 Canadian Census through Canada's 6-character alphanumeric postal codes. Because the BCAHS did not collect data about students’ home addresses, we used school addresses to categorize students as rural or urban. Other school-based studies have used the school as a proxy measure of rurality for students.25,27,33–35 This approach seemed reasonable because the school is the primary locale of school-aged adolescents’ lives.
Analyses compared LGB adolescents living in rural areas with their counterparts in urban areas. Because of maturational and gender differences associated with many health and risk behaviors, particularly for LGB adolescents,36 we controlled for age and conducted analyses separately by gender. All variables were categorical and were analyzed by the Pearson χ2 test and logistic regression.
As a result of differences found between rural and urban adolescents in ethnicity, immigrant status, and recent relocation (Table 1), we incorporated ethnicity and moving status as covariates in logistic regression analyses. Because respondents could choose more than 1 ethnicity, this variable was dichotomized into European heritage only versus others (i.e., adolescents of mixed European and non-European heritages and adolescents with only non-European heritages). Immigrant status was not included as a covariate because it is closely linked with ethnicity. Moving status was dichotomized as yes or no.
TABLE 1 Demographic Characteristics of Rural and Urban Lesbian, Gay, and Bisexual Adolescents: British Columbia Adolescent Health Survey, 2003
| Adolescent Boys | Adolescent Girls | |||||
| Rural (n = 199) | Urban (n = 1913) | χ2 (df) | Rural (n = 538) | Urban (n = 4255) | χ2 (df) | |
| Mean age,a y | 15.8 | 15.7 | … | 15.6 | 15.5 | … |
| European ethnicity,b % | 80.9 | 56.7 | 43.50** (1) | 82.2 | 67.7 | 46.97** (1) |
| Aboriginal ethnicity,b % | 17.6 | 8.6 | 17.24** (1) | 20.0 | 13.1 | 19.10** (1) |
| Recent immigrant (lived in Canada for ≤ 5 y), % | 2.5 | 12.6 | 17.92** (1) | 0.0 | 14.6 | 90.40** (1) |
| Living with 2 parents, % | 62.3 | 62.2 | 0.00 (1) | 61.8 | 57.5 | 3.61 (1) |
| Lived in foster/group home in past year, % | 2.5 | 3.7 | 0.75 (1) | 4.1 | 4.0 | 0.03 (1) |
| Ran away from home in past year, % | 17.6 | 13.0 | 3.21 (1) | 31.1 | 28.6 | 1.47 (1) |
| Moved in past year, % | 37.4 | 25.8 | 12.17** (1) | 38.9 | 40.2 | 0.34 (1) |
Note. Data were weighted to provincial enrollment and adjusted for differential probability of sampling and differential response rates.
a For boys, t = −0.79; df = 255 (not significant); equal variances not assumed. For girls, t = −0.88; df = 4791 (not significant).
b Other ethnicity responses were available and participants could choose more than 1 response.
**P < .001.
Because of the number of multiple comparisons, α levels were set to .01 to control for inflated type I errors. Cases with missing data for any relevant survey items in an analysis were excluded from that analysis. All items had fewer than 5% missing responses with the exception of adolescent boys’ reports of pregnancy involvement. Analyses were conducted with SPSS version 15.0 (SPSS Inc, Chicago, IL). Because of the stratified sampling frame of the BCAHS, we conducted extensive analyses to explore possible design effects. We found that LGB adolescents were randomly distributed throughout the population, and thus no cluster effects were detected (there was no clustering when there was only 1 LGB student in a classroom per grade). Therefore, specialized software, such as SUDAAN or SPSS for complex samples, was not used.
The sample consisted of the 2.4% of respondents who self-identified as lesbian, gay, or bisexual (weighted N = 6905; 2112 adolescent boys and 4793 girls). Of these LGB adolescents, 58.3% of boys and 90.3% of girls identified as bisexual. Population-based studies from the United States have also found LGB samples to be predominantly bisexual,8,18,37 including those taken from rural areas.12 In our sample, 10.7% of the students came from rural areas. Seventy-three percent of rural and 69.0% of urban participants were girls. The relatively small number of adolescents in rural areas did not permit categorization of bisexuals separately from gays and lesbians. Thus, the final LGB sample included 199 rural and 1913 urban adolescent boys, and 538 rural and 4255 urban adolescent girls.
Demographic characteristics of LGB respondents are shown in Table 1. Rural and urban adolescents were similar in age, whether they lived with 2 parents, lived in a foster or group home, or ran away from home in the past year. More rural adolescents identified as European or Aboriginal. Rural respondents were less likely than were urban respondents to be recent immigrants. Sexual-minority adolescent boys from rural areas were more likely than were urban boys to have moved in the past year.
Worries about basic family needs were more prevalent among rural LGB participants, and suicidality was more prevalent among rural than among urban sexual-minority adolescent boys. Overall, 49.4% of sexual-minority boys and 54.8% of sexual-minority girls worried about having enough food or money, and 20.7% and 32.5%, respectively, worried about not having someone to take care of them. As shown in Table 2, however, rural LGB respondents were more likely than were their urban peers to report these worries. Rural and urban LGB adolescents were equally likely to be nervous in the past month, but rural LGB respondents were less likely to be stressed. In the past year, 33.3% of sexual-minority boys had considered suicide and 11.2% had attempted suicide; rural sexual-minority boys were more likely than were their urban peers to report these behaviors. We observed no rural–urban differences in suicidal behaviors among sexual-minority girls.
TABLE 2 Adjusted Odds Ratios (AORs) for Health and Risk Behaviors of Rural Lesbian, Gay, and Bisexual Adolescents Compared With Their Urban Peers: British Columbia Adolescent Health Survey, 2003
| Boys | Girls | |
| Survey item | AOR (95% CI) | AOR (95% CI) |
| Emotional health | ||
| Worried about family having enough food or money | 2.32** (1.66, 3.26) | 1.68** (1.37, 2.05) |
| Worried about not having someone to take care of him or her | 2.40** (1.66, 3.47) | 2.04** (1.68, 2.47) |
| Worried about violence in the home | 1.14 (0.75, 1.74) | 1.18 (0.96, 1.46) |
| Worried about substance use in the family | 1.60 (1.12, 2.28) | 0.71* (0.57, 0.89) |
| Felt strain, stress, or pressure quite a bit or extremely so in past month | 0.55** (0.40, 0.77) | 0.74* (0.61, 0.90) |
| Bothered by nerves quite a bit or extremely so in past month | 0.83 (0.57, 1.19) | 0.91 (0.74, 1.11) |
| Felt sad, discouraged, or hopeless quite a bit or extremely so in past month | 1.11 (0.79, 1.56) | 0.75* (0.62, 0.91) |
| Seriously considered suicide in past year | 1.68* (1.22, 2.31) | 0.93 (0.77, 1.13) |
| Attempted suicide in past year | 1.89* (1.27, 2.82) | 1.03 (0.83, 1.26) |
| Victimization | ||
| Ever physically abused | 1.06 (0.72, 1.57) | 0.94 (0.77, 1.15) |
| Ever sexually abused | 1.54 (0.99, 2.39) | 1.05 (0.86, 1.28) |
| Physically hurt by boyfriend or girlfriend in past yeara | 3.01** (1.79, 5.07) | 0.50** (0.34, 0.72) |
| Verbally sexually harassed in past year | 1.23 (0.88, 1.71) | 1.77** (1.39, 2.25) |
| Physically sexually harassed in past year | 1.79** (1.28, 2.50) | 1.02 (0.84, 1.24) |
| Verbally harassed (teased) at school in past year | 0.41** (0.30, 0.57) | 1.11 (0.91, 1.35) |
| Purposefully excluded at school in past year | 0.66 (0.48, 0.91) | 1.24 (1.02, 1.52) |
| Physically assaulted at school in past year | 1.22 (0.84, 1.78) | 0.58** (0.44, 0.78) |
| Discriminated against because of sexual orientation in past year | 0.64* (0.46, 0.88) | 1.15 (0.94, 1.41) |
| Discriminated against because of race in past year | 1.00 (0.67, 1.49) | 0.77 (0.55, 1.09) |
| Discriminated against because of physical appearance in past year | 1.28 (0.91, 1.80) | 1.30 (1.06, 1.58) |
| Sexual behavior | ||
| Ever had sex | 1.13 (0.82, 1.56) | 1.30 (1.06, 1.60) |
| Had sex before age 14 yearsb | 0.39* (0.21, 0.73) | 2.21** (1.63, 2.99) |
| Had ≥ 3 sexual partners in lifeb | 1.43 (0.90, 2.27) | 1.40 (1.08, 1.81) |
| Did not use condom at last sexual intercourseb | 0.83 (0.51, 1.36) | 1.40 (1.08, 1.81) |
| Did not use reliable birth control method at last sexual intercourseb | 1.32 (0.80, 2.20) | 0.91 (0.66, 1.25) |
| Ever pregnant or ever caused pregnancyb | 2.80** (1.65, 4.76) | 0.78 (0.48, 1.27) |
| Substance use | ||
| Ever tried alcohol | 0.81 (0.56, 1.17) | 1.93** (1.44, 2.59) |
| Used alcohol on ≥ 10 days in past month | 1.71 (0.96, 3.03) | 2.07** (1.52, 2.82) |
| Engaged in binge drinking in past month | 2.72** (1.95, 3.79) | 1.39** (1.15, 1.69) |
| Ever tried marijuana | 1.03 (0.74, 1.42) | 2.08** (1.67, 2.61) |
| Used marijuana ≥ 10 times in past month | 1.66 (1.07, 2.57) | 0.91 (0.70, 1.18) |
| Ever tried any other drug (other than alcohol and marijuana) | 2.15** (1.57, 2.96) | 1.37* (1.12, 1.66) |
| Ever used cocaine | 0.95 (0.60, 1.50) | 1.25 (0.98, 1.61) |
| Ever used hallucinogens | 1.25 (0.83, 1.87) | 1.27 (1.02, 1.58) |
| Ever used mushrooms | 2.14** (1.53, 2.99) | 1.64** (1.34, 2.00) |
| Ever used inhalants | 1.63 (1.06, 2.49) | 1.06 (0.78, 1.44) |
| Ever used amphetamines | 0.65 (0.39, 1.10) | 1.43* (1.10, 1.85) |
| Ever used heroin | 1.81 (0.60, 5.44) | 2.77** (1.88, 4.08) |
| Ever used injected drugs | NA | 3.55** (2.30, 5.47) |
| Ever used steroids | 7.65** (3.43, 17.09) | 1.10 (0.57, 2.12) |
| Ever used prescription pills without a doctor's consent | 2.43** (1.69, 3.51) | 1.45** (1.17, 1.80) |
| Experienced problems after alcohol or drug use in past yearc | 1.53 (1.01, 2.30) | 1.18 (0.95, 1.47) |
Note. CI = confidence interval; NA = not applicable because an odds ratio could not be calculated. The reference group was lesbian, gay, and bisexual urban adolescents. Analyses controlled for age, ethnicity, and moving status.
a Of those who had a boyfriend or girlfriend (weighted n = 144 rural and 1018 urban boys, and 408 rural and 3134 urban girls).
b Of those who had ever had sex (weighted n = 86 rural and 783 urban boys, and 330 rural and 2188 urban girls).
c Of those who used alcohol or drugs (weighted n = 133 rural and 1076 urban boys, and 470 rural and 2907 urban girls).
*P < .01; **P < .001.
Victimization experiences were common: more than half of the LGB students reported verbal sexual harassment and teasing at school, and nearly half reported being purposefully excluded, but differences in victimization were inconsistent (Table 2). Rural sexual-minority adolescent boys were more likely than were their urban peers to report dating violence and physical sexual harassment, but rural sexual-minority boys were less likely to be teased at school. By contrast, rural sexual-minority girls were less likely than were their urban peers to report dating violence or physical assault at school but more likely to report verbal sexual harassment. Discrimination was also common: 46.0% of sexual-minority adolescent boys and 29.7% of sexual-minority girls reported discrimination in the past year because of their sexual orientation. Sexual-minority boys in rural communities were less likely to report such discrimination, but this discrimination was equally likely among sexual-minority girls in rural and urban areas.
We found few differences in sexual risk behavior (Table 2). Rural LGB students were as likely as their urban peers to report sexual intercourse, multiple sexual partners, condom use, and birth control use. Among sexually experienced LGB adolescents, rural adolescent boys were only one third as likely as their urban counterparts to have had sex before age 14 years, but rural adolescent girls were twice as likely as their urban peers to have done so. Although 16.8% of sexually active sexual-minority boys reported pregnancy involvement overall, rural boys were nearly 3 times more likely than their urban counterparts to have done so. We found no rural–urban differences in pregnancy among sexual-minority adolescent girls.
Rural sexual-minority adolescent girls were generally more likely to report substance use. Among sexual-minority boys, in categories in which we found differences, rural respondents were more likely to report use. As Table 2 shows, rural and urban sexual-minority adolescent boys did not differ in having tried alcohol or marijuana, but more rural than urban sexual-minority girls had tried alcohol and marijuana. Among all LGB students, 30.7% of boys and 41.4% of girls reported binge drinking (defined as having ≥ 5 drinks in a row within a couple of hours) in the past month, but rural adolescents, especially boys, were more likely than were urban adolescents to do so. Furthermore, 35.1% of sexual-minority adolescent boys and 49.5% of sexual-minority girls had tried drugs other than alcohol and marijuana, but rural and urban respondents differed in their experimentation with other drugs by drug and gender (Table 2). Rural sexual-minority boys were more likely than were their urban peers to have tried mushrooms, steroids, and prescription pills. Rural sexual-minority adolescent girls were more likely than were urban girls to have tried mushrooms, amphetamines, heroin, injected drugs, and prescription pills.
As with studies of rural and urban health disparities among the general adolescent population, our study had mixed results. For some health and risk behaviors, rural LGB adolescents were at greater risk; for others, they were at less risk. In many instances, however, no differences emerged between LGB adolescents from rural and urban communities, suggesting they were at equal risk. Although our findings were inconsistent, they suggest the need to consider rural–urban status in future investigations.
Some differences, such as suicidality, are of greater concern because they have more serious long-term health consequences. Rural sexual-minority adolescent boys, but not girls, were at higher risk for suicidal ideation and attempts, in contrast to a study by Waldo et al. that found no rural–urban differences in suicidality.29 Suicidal behaviors may stem partly from turmoil about sexual orientation38 and associated stigma.39 Sexual-minority adolescent boys in rural communities may have particular difficulty dealing with such issues because of experiences of isolation and limited support.2
Rural LGB adolescents engaged in more serious substance-use behaviors, such as binge drinking, using mushrooms, and taking prescription pills without a doctor's consent. Rural sexual-minority adolescent girls, but not boys, were more likely to have had sex before age 14 years, but rural boys were especially at risk for pregnancy involvement. Substance use and risky sexual behavior may be a means of coping for some LGB adolescents, and these behaviors have been linked to stressors associated with the stigma of being or of being perceived to be LGB22,40 and the coming-out process.41–43 Sexual orientation stress may be heightened and the coming-out process more difficult in rural communities, where resources for sexual minorities and opportunities to associate with other LGB adolescents or adults are limited.2
Reports of discrimination based on sexual orientation were particularly interesting. We expected that the relatively more traditional and conservative attitudes of rural communities would result in higher rates of discrimination, but this was not what we found. In fact, among rural sexual-minority adolescent boys, the opposite was true. Perhaps lack of anonymity in these communities deters people from discriminating. Another, more plausible possibility is that rural LGB adolescents are less likely to be out and therefore may not be a target of sexual orientation discrimination. This may also account for the rates of other forms of victimization, such as those experienced at school. If it is riskier to be out in a less anonymous and less safe rural environment,2 perhaps rural LGB adolescents stay closeted and thus are no more likely, or possibly less likely, to be victimized.
Finally, our results showed that gender interacted in different ways with location. Location was a factor for sexual-minority adolescent boys, but not girls, in suicidality and pregnancy involvement. We found another striking difference in dating violence. Rural sexual-minority boys were 3 times as likely, but rural girls only half as likely, as their urban counterparts to report being victims of dating violence in the past year. These incidences of violence may have been perpetrated by opposite-gender partners, because 42.7% of sexual-minority boys and 59.9% of sexual-minority girls reported having only opposite-gender sexual partners in the past year. However, the gender of a respondent's boyfriend or girlfriend may not have been the same as the gender of his or her sexual partners.
The study's geographic definitions had limitations. The use of a sophisticated, standardized census definition may increase its replicability across Canadian provinces but may be more liberal or conservative than definitions based exclusively on population size; other studies have considered much larger communities rural.14 Also, because of the distances some students travel to school, their home may be in a rural area, but the school they attend may be in an urban area. These students would have been misclassified, which could have attenuated differences. The rural and urban LGB adolescents we described were more accurately LGB adolescents who attended rural and urban schools. However, because students spend a great deal of their lives at school, its environment and people are likely to exert considerable influence on their behaviors. Furthermore, in British Columbia, the majority of people live in urban-influenced areas, and urban adolescents would be unlikely to attend rural schools. Therefore, the communities classified as rural were indeed rural and remote.
The nature of the sample presented other limitations. The large-scale, provincewide data were school based, missing adolescents who were absent or dropped out; LGB students often figure disproportionately in this group.44 The BCAHS was cross-sectional, and because sexual orientation often unfolds during adolescence, the findings may not be representative of respondents who had not yet identified their orientation by the time of the survey. Furthermore, LGB students were not asked whether they had come out, so we cannot know whether they publicly identified or were perceived as LGB.12 This factor may have affected some results; however, internal identification as lesbian, gay, or bisexual or having same-sex attractions still requires coping with a stigmatized orientation.12 Including those who are out and those who are not may be more representative of all LGB adolescents.9
The LGB students in our sample, like those in other population-based studies, were predominantly bisexual. Although we aggregated the LGB sample, other population-based studies that separated bisexual and gay or lesbian adolescents found limited or inconsistent differences between the groups.10,39 Our data were representative of LGB adolescents in school in British Columbia; other studies are needed to determine whether our results are generalizable to other Canadian regions or to other countries.
This study, to our knowledge, was the first to use population-based Canadian data to compare health and risk behaviors among LGB adolescents in rural and urban areas. Given the wide range of behaviors studied, it is not surprising that results were mixed. The findings suggest the need to consider the rural–urban status of LGB adolescents in future research. Further studies are needed to replicate our findings and explore reasons why rural and urban LGB adolescents differ, perhaps investigating the role of orientation stress and the coming-out process.
Our results may not be unique to LGB adolescents. For example, the relatively higher substance use found among rural LGB adolescents is consistent with previous studies of adolescents.23,24 Factors unrelated to sexual orientation should be examined in attempting to explain rural–urban differences. Rural areas in British Columbia, for example, are disadvantaged in employment rates, median incomes, and access to health care providers.45
Location should be considered in health promotion efforts. At the very least, the similarities indicate that rural adolescents constitute an equally at-risk population, with implications for public health policy.25 Intervention approaches may need to be adapted for rural communities' priorities. For example, rural adolescent boys need mental health interventions to reduce suicidality. Rural communities could offer more avenues for emotional and social support for LGB adolescents, perhaps by fostering informal help networks,2 which could link LGB adolescents with peers and LGB adults. Rural sexual-minority adolescent boys also need efforts directed at them to reduce involvement in pregnancies. Rural communities might consider incorporating LGB-sensitive sexual education in schools, which has been shown to reduce risky sexual behavior and pregnancy among LGB adolescents.46 Rural sexual-minority adolescent girls, in particular, would benefit from interventions to address serious substance use, perhaps by reducing sexual orientation stigma and discrimination and by preventing victimization experiences, because these problems may lead to substance use as a means of coping.47
Both rural and urban LGB adolescents would benefit from antibullying policies and interventions that explicitly target homophobia. Communities might consider LGB support groups in schools. Groups such as gay–straight alliances are underrepresented in rural schools: 11% of LGB adolescents in British Columbia attend rural schools but fewer than 5% of the province's gay–straight alliances are in rural schools.48 These groups have been associated with lower victimization and suicidality.44 LGB adolescents face many challenges, and for some, navigating their adolescence in the relative isolation of a rural community may be particularly difficult.
Acknowledgments
This project was supported by funding from the Institute for Gay and Lesbian Strategic Studies/Williams Institute for Sexual Orientation and the Law, University of California, Los Angeles; the Michael Smith Foundation for Health Research, Vancouver, BC; and the National Institute of Drug Abuse, Bethesda, MD (grant R01 DA1797901).
The analyses in this report were part of a larger study on lesbian, gay, and bisexual adolescents. Some of the findings were reported in a monograph for lay audiences, Not Yet Equal: The Health of Lesbian, Gay, and Bisexual Youth in BC, available from the McCreary Centre Society at http://www.mcs.bc.ca.
The authors thank the McCreary Centre Society for access to the BC Adolescent Health Survey 2003.
Human Participant Protection
The University of British Columbia behavioral research ethics board reviewed and approved the project.
