Victimization, Substance Use, and HIV Risk Behaviors Among Gay/Bisexual/Two-Spirit and Heterosexual American Indian Men in New York City

Abstract

Objectives.secondary aims included describing condom-use attitudes, beliefs about HIV/AIDS in the Indian community, HIV knowledge, HIV status, and preference for and access to HIV prevention services in this population.

Methods. A survey was mailed to all members of an American Indian community organization in New York City.

Results. The 20 men self-identifying as gay, two-spirit, or bisexual (hereafter, “two-spirit”) were more likely to report being victimized and engaging in HIV risk behaviors than the 51 heterosexual respondents, although they reported comparable levels of recent substance use. Overall, victimization was associated with lifetime HIV risk behaviors (even after control for sexual orientation) but not with substance use or unsafe sex in the past 12 months. The percentage of HIV infection was surprisingly high (10% of two-spirit men and 6% of heterosexual men).

Conclusions. Two-spirit men are a vulnerable population whose victimization must be understood within an appropriate historical and political context.

American Indians and Alaska Natives are a people under siege, living within a colonial nation-state in a “fourth world.”1 This “indigenist” perspective foregrounds their historical and political context, including their experience of a succession of traumatic assaults, and is essential to contexualizing their current health status.2^,3

Although preliminary evidence suggests that American Indian men might respond differently than American Indian women to the history of indigenous oppression,4 there is little empirical research on the health status of American Indian men. Data from the only 2 comprehensive reviews of American Indian men’s health indicate that they suffer from disproportionate rates of poor health compared with men in the general population.5 Additionally, life expectancy for American Indian men is significantly shorter than for White men or American Indian women, with early death caused in part by disproportionate rates of homicide, suicide, and motor vehicle accidents.6 Indeed, 5 of the top 10 leading causes of death for American Indian men (i.e., unintentional injuries, chronic liver disease and cirrhosis, diabetes mellitus, suicide, and homicide) are related to voluntary risky behaviors such as alcohol abuse, and might be preventable with appropriate public health intervention.7

A subgroup of American Indian men at increased risk for adverse health outcomes is self-identified gay, bisexual, or “two-spirit” men. Indigenous activists have adopted the term “two-spirit” as a way for gay, bisexual, or transgendered American Indians to name themselves outside the colonizing terms previously imposed upon them (e.g., “berdache”), to reconnect with tribal traditions related to sexuality and gender identity, and to transcend the Eurocentric binary categorizations of heterosexual versus homosexual and male versus female.8 Gay and bisexual American Indian men confront stressors associated with negotiating their multiple oppressed statuses and often must contend with heterosexism in American Indian communities as well as racism in gay communities.9^–11

Victimization Experiences

The rate of violent victimization for American Indian adults is more than 2.5 times that of the overall US population,12 and data from general populations of gay and bisexual men indicate relatively high rates of self-reported childhood sexual abuse and other victimization experiences that generally surpass those reported by heterosexual men.13^,14 In one of the few studies of victimization experiences specifically among two-spirit men, Walters et al.11 reported lifetime prevalences of physical assault (36%) and sexual assault (29%) that were much higher than those reported in non–American Indian gay, lesbian, and bisexual samples. Saewyc et al.15 reported that American Indian gay and bisexual male youth reported a significantly higher prevalence of lifetime sexual abuse and physical abuse than their heterosexual counterparts.

Among men in the general population, both physical and sexual victimization have been linked to substance abuse.16^–19 Preliminary research indicates that these links also exist among gay/bisexual men from the general population20 and American Indian men.21^,22 We could locate no studies of victimization and substance use among two-spirit men.

HIV Sexual Risk Behaviors

Findings from general population studies of men who have sex with men suggest unsafe sex persists despite HIV prevention efforts.23 The scant data available on condom use among American Indians indicate considerable risk behavior.24^–26 For example, in a study of 100 American Indians in New York City, 73% of those who reported sexual activity in the past 6 months had engaged in unprotected vaginal or anal sex.27 Research on sexual behavior and sexual orientation among American Indian populations is limited to a few studies (e.g., Saewyc et al.15^,28; Warren et al.29), which indicate that American Indian youth have a higher prevalence of self-reported gay, lesbian, bisexual, and “unsure” sexual identities than non–American Indian youth and that these youth report earlier onset of heterosexual intercourse and less frequent use of contraceptives compared with their heterosexual American Indian peers.

Predictors of Sexual Risk Behaviors: Victimization and Substance Use

Although we could find no relevant references for American Indian men, recent studies among general populations of men30 and men who have sex with men31^,32 suggest that there is a link between childhood sexual abuse and adult HIV risk behaviors.

Only recently has research begun to investigate the co-occurrence of substance use and risky sex among American Indians.21^,27^,33 In one study of American Indian/Alaska Native drug users, 50% reported drinking until drunk and engaging in unprotected sexual intercourse while in a blackout state.34 Other preliminary research indicates that urban American Indian/Alaska Native drug users are at greater risk for HIV infection than their reservation counterparts because of sex work (i.e., having sex in exchange for money, drugs, or other favors) and unprotected sex.35^,36 Other studies have confirmed the association of injection drug use and HIV risk behaviors among American Indians.37^,38

Present Study

In the present study, we surveyed American Indians in New York City, which has the largest urban American Indian population in the United States.39 On the basis of the literature delineating their dually oppressed status, we expected that two-spirit men would report higher rates of victimization, substance use, and HIV-risk behaviors than would heterosexual men. Overall, we predicted that victimization would be associated with substance use and sexual risk behaviors in the past 12 months and that both victimization and substance abuse would correlate with lifetime sexual risk behaviors. Additionally, we queried the men regarding their HIV knowledge, attitudes toward condom use and HIV/AIDS in the Indian community, access to and preference for HIV prevention services, and their HIV status.

METHODS

Respondents

The final sample consisted of 71 American Indian men who self-identified as heterosexual (72%), gay or two-spirited (24%), or bisexual (4%; see Simoni et al.38 for findings with respect to the 165 women in this study). The 20 men in the latter 2 categories were combined to form a “two-spirit” subgroup. There were no major sociodemographic differences between the two-spirit and heterosexual men. Overall, the men ranged in age from 22 to 75 years (mean = 42.97; SD= 11.71). In terms of education, 90% had a high-school degree or general equivalency degree, and 42% completed at least a bachelor’s degree. Monthly household income from all sources was as follows: $501 to $1000 (21%), $1001 to $1500 (24%), $1501 to $5000 (28%), and more than $5000 (27%). Sixty-five percent reported a steady partner; 51% were currently living with a steady partner; and the mean length of the partnership was 11.69 years (SD= 10.26).

Measures

We used investigator-developed scales to assess victimization, lifetime HIV risk behaviors, recent (i.e., in past 12 months) substance use, recent (i.e., in past 12 months) sexual behaviors, sexual partner risk factors, HIV status and other HIV-related variables, and HIV prevention and program planning issues (specific items are included in the Results and tables). Standardized scales were used to assess other variables. Specifically, attitudes toward condom use were measured with 13 items scored on a Likert-type scale from 1 (strongly disagree) to 4 (strongly agree) adapted from Helweg-Larsen and Collins.39 Knowledge about HIV was assessed with 11 yes/no/do-not-know items (S. Kalichman, PhD, written communication, June 1998; the Cronbach α in the present sample was 0.63). To measure attitudes toward HIV/AIDS in the Indian community, respondents rated 13 attitudinal items adapted from Myers et al.40 on a Likert-type scale from 1 (strongly disagree) to 4 (strongly agree).

Procedure

The current survey was part of a larger project to assess HIV risk behaviors and prevention needs in the American Indian community in New York City. A survey packet containing an anonymous 6-page questionnaire, a $1 appreciation gift, a ticket for a $500 lottery gift, and a postage-paid return envelope was mailed to all 748 members of an American Indian community organization in December 1998. Membership requires documentation of enrollment or eligibility for enrollment in a state or federally recognized American Indian or Alaska Native tribe or a written statement of support from tribal leaders. Although there are no requirements regarding sobriety to join the association, the center denies access to the premises to any individual obviously intoxicated or high. A postcard reminder was sent 2 weeks after the survey.

Forty surveys were returned because of errors in name or address. Approximately one third of the remaining surveys were returned, which is comparable to response rates typically observed in unsolicited mail surveys. Considering that the survey was mailed during the holiday season, targeted an oppressed population that is justifiably distrustful of researchers, concerned sensitive subject matter, and guaranteed no monetary remuneration, the response rate was considered acceptable for preliminary analyses.

Data Analysis

Analyses proceeded in 4 stages. First, we computed descriptive statistics, including frequencies for all variables. Second, t tests, χ², and multiple analyses of variance (MANOVA) were used to examine differences between two-spirit and heterosexual men for all variables; results are reported for respondents overall where no differences were found. Third, we conducted bivariate analyses to examine the association of victimization with sociodemographics, recent substance use, and recent sexual risk behaviors. Finally, we examined correlates of lifetime HIV risk behaviors, first with bivariate analyses and then in a multiple regression analysis to determine independent predictors.

RESULTS

Descriptive Data on Two-Spirit and Heterosexual Men

Victimization.

Overall, 20% of respondents reported experiencing at least 1 type of victimization in their lifetime (i.e., physical or sexual assault or abuse by a partner or non-partner), with significant differences between two-spirit and heterosexual men (Table 1 ▶).

TABLE 1—

Victimization Among Two-Spirit and Heterosexual American Indian Men

Type of Victimization	Two-Spirits (n = 20), %	Heterosexuals (n = 51), %	χ2
Physical abuse or assault by a spouse/sexual partner	10	2	2.23
Physical abuse or assault by someone other than a spouse/sexual partner	45	6	15.30***
Sexual abuse or assault by a spouse/sexual partner	10	0	5.25*
Sexual abuse or assault by someone other than a spouse/sexual partner	45	2	21.99***
Any 1 of these 4 categories	55	6	21.89***
Mean total number of victimization categories	1.10	0.10	t = −3.49**

* P < .05; **P < .01; ***P < .001.

Lifetime HIV risk behaviors.

Two-spirit men also reported greater exposure to lifetime HIV risk behaviors than heterosexual men (Table 2 ▶). There were no differences between the 2 groups with respect to risky behaviors during their most recent sexual experience: Overall, 24% of respondents reported that they used alcohol or drugs before or during their most recent sexual encounter, and 68% reported that they did not use a condom during their most recent sexual encounter.

TABLE 2—

Lifetime HIV Risk Behaviors Among Two-Spirit and Heterosexual American Indian Men

HIV Risk Behavior	Two-Spirits (n = 20), %	Heterosexuals (n = 51), %	χ2
Oral sex without a condom	100	75	6.24*
Anal intercourse without a condom	85	24	22.47***
Sex with a stranger	85	49	7.70**
More than 2 sexual partners in a month	80	51	5.01*
Unsafe sex while drunk or high	60	46	1.12
Had a sexually transmitted disease	50	24	4.71*
Vaginal intercourse without a condom	37	86	17.11***
Had sex with an HIV+ person	35	6	10.07**
Had sex with an injection drug user	30	8	5.83*
Traded sex for money, drugs, or favors	20	2	7.14**
Injected nonprescription drugs	10	10	0.00
Shared dirty needles when injecting drugs	0	6	1.23
Mean total number of HIV risk behavior categories	5.90	3.84	t = −0.31**

* P < .05; **P < .01; ***P < .001.

Recent substance use.

Contrary to our hypotheses, a series of t tests revealed no significant differences between two-spirit and heterosexual men in use of any substance in the past 12 months. We found high levels of sobriety overall (Table 3 ▶).

TABLE 3—

Frequency of Substance Use (%) in Past 12 Months Among 71 American Indian Men

Substance	Never	About Monthly	About Weekly	About Daily
Any alcohol at all	43	41	7	9
Six or more drinks containing alcohol on 1 occasion	63	29	9	1
Alcohol to the point of getting drunk	77	13	8	2
Amphetamines	96	3	2	0
Any other illegal drug	87	6	6	1
Any injected illegal drug	96	2	3	0

Note. Results do not differ by sexual orientation.

Recent sexual behaviors.

Analyses of data on sexual behaviors in the past 12 months indicated the majority (82%) of the men reported having had sex. The sexually active two-spirit men (n = 17) reported 0 to 100 male partners (mean = 16.29; SD= 27.64); 2 (12%) reported female partners as well. The sexually active heterosexual men (n = 38) reported from 1 to 5 female partners (mean = 0.41; SD= 1.46) and no male partners. Condom use was reportedly inconsistent, with no statistically significant differences between two-spirit and heterosexual men. Specifically, respective percentages of those never and always using condoms were 70% and 11% for oral sex (n = 46), 50% and 23% for vaginal sex (n = 40), and 36% and 27% for anal sex (n = 22). Among sexually active respondents overall, 72% reported that they had engaged in sex without a condom at least once.

There were no significant differences between two-spirit and heterosexual men with respect to substance-related recent sexual risk behaviors. Specifically, respondents who were sexually active in the past 12 months indicated they had used alcohol or other drugs just before or during sex never (60%), less than half the time (19%), about half the time (5%), more than half the time (7%), and always (7%). Their ability to request or use a condom had been affected by their own or their partner’s alcohol or other drug use never (84%), less than half the time (5%), about half the time (2%), more than half the time (5%), and always (2%).

Sexual partner risk factors.

As shown in Table 4 ▶, two-spirit men (n = 8) were more likely than heterosexual men (n = 29) to attribute 6 of the 8 risk factors to partners with whom they had anal or vaginal sex without a condom in the past 12 months.

TABLE 4—

Sexual Partner Risk Factors Among American Indian Men Who Had Anal or Vaginal Sex Without a Condom in the Past 12 Months

Sexual Partner Risk Factor	Two-Spirits (n = 8). %	Heterosexuals (n = 29), %	χ2
Has had sex with a gay or bisexual man	88	4	25.36***
Was a casual (not steady or regular) partner	75	14	11.43**
Was someone participant did not know well	63	11	9.65**
Was having sex with other partners	71	4	18.15***
Has traded sex for money, drugs, or favors	50	0	15.75***
Has HIV/AIDS	29	0	8.48*
Has injected illegal drugs	14	0	4.12
Has a sexually transmitted disease	0	0	. . .
Mean total number of sexual partner risk factor categories	3.75	0.32	t = −7.08***

* P < .05; **P < .01; ***P < .001 according to Fisher exact test.

Condom attitudes.

A MANOVA revealed no significant differences between two-spirit and heterosexual men with respect to attitudes toward condom use.

HIV knowledge.

Respondents correctly answered from 4 to all 11 of the HIV knowledge items (mean = 9.52; SD = 1.64), with no differences between two-spirit and heterosexual men.

HIV status and other HIV-related variables.

Fifty-six percent of respondents reported that their physician knows they are American Indian, with no differences between two-spirit and heterosexual men. Two-spirit men were more likely than heterosexual men to know of an HIV-positive immediate family member (20% vs 0%) (χ² [n=71]=11.69; P<.005), or close friend (85% vs 28%) (χ² [n=71]= 19.43; P<.001). Significantly more two-spirit than heterosexual men had been tested for HIV (90% vs 57%) (χ² [n=71]=7.05; P< .01). However, there was no significant difference in the percentage who reported having HIV/AIDS (yes/no/don’t know) among two-spirit (10%/90%/0%) and heterosexual men (6%/84%/10%). Overall, the HIV-negative men reported the following self-perceived risk of contracting HIV: no risk (30%), low risk (48%), moderate risk (8%), high risk (8%), and extremely high risk (0%; 7% had missing data), with no difference by sexual orientation.

Attitudes toward HIV/AIDS in the American Indian community.

A MANOVA revealed no significant differences by sexual orientation with respect to attitudes regarding HIV/ AIDS in the American Indian community. Both two-spirit and heterosexual men recognized the threat of HIV in the American Indian community, acknowledged the need and expressed support for HIV prevention efforts, and wholeheartedly rejected the exclusion of American Indians living with HIV/ AIDS from the community. Thirty percent agreed or strongly agreed with the item “AIDS is another form of germ warfare on Indian people.”

HIV prevention and program planning.

Comparable percentages of two-spirit and heterosexual men thought HIV education services in the home (36%), HIV education services at an American Indian community setting (75%), instruction in getting their partner to use a condom with them (32%), and instruction on using the female condom (32%) would be helpful to them. Two-spirit men were more likely than heterosexual men to consider free condoms/barriers to be a potentially helpful service (80% vs 49%) (χ² [n = 66] = 3.9; P < .05). Finally, 21% of respondents reported needing or wanting information, support, or treatment related to HIV in the past 12 months, with no difference by sexual orientation. Among these respondents, 100% of the 7 two-spirit men reported that they received what they needed or wanted, compared with only 33% of the 8 heterosexual men (χ² = 7.47; P < .01).

Correlates of Victimization

A series of bivariate analyses revealed no significant association between victimization and any sociodemographic indicator. Contrary to our hypotheses, victimization was not associated with any of the indicators of recent substance use. Although it was not significantly correlated with consistency of recent condom use, victimization was significantly associated with the total number of sexual partner risk factors for respondents who reported any anal or vaginal sex without a condom in the past 12 months (r = 0.76; P < .001). Finally, victimization was associated with the frequency of failure to use a condom because of the influence of alcohol or other drugs in the past 12 months (r = 0.37; P < .01).

Correlates of Lifetime HIV Risk Behaviors

Bivariate analyses indicated the only significant correlates of lifetime HIV risk behaviors were victimization (r = 0.51; P < .001), and being two-spirit versus heterosexual (t₆₉ = −3.02; P < .01). To determine whether victimization and sexual orientation independently predicted lifetime HIV risk behaviors, both variables were simultaneously entered into a regression analysis. The overall model accounted for 27% of the variance in HIV risk behaviors (F_2,68 = 12.65; P < .001), but only victimization remained significant as a predictor (B = 0.46; P < .001).

DISCUSSION

Findings from this preliminary study of 71 American Indian men in New York City confirmed expectations of the two-spirit men’s vulnerability to adverse health outcomes. Specifically, compared with their heterosexual counterparts, two-spirit men reported greater victimization, more lifetime HIV risk behaviors, and riskier sexual partners in the past 12 months. These findings parallel results of differences by sexual orientation among European American men.

Additionally, as well as being more likely to know of an HIV-positive immediate family member and close friend and to have been tested for HIV, two-spirit men were nearly twice as likely as heterosexual respondents to have HIV/AIDS (although this difference was not statistically significant in our relatively small sample). These data are not surprising, given the drastically higher rates of HIV among the general population of men who have sex with men versus men who do not have sex with men. However, HIV-negative two-spirit men did not perceive themselves as being at greater risk of contracting HIV than their heterosexual counterparts, suggesting a false sense of security that may underlie their greater involvement in HIV risk behaviors. Indeed, among men who have sex with men, 5-year increases in AIDS incidence rates were higher for American Indians (53%) than for African Americans (45%) or Latinos (23%).41

Most disturbing about the HIV-related findings was the high percentage of American Indian men living with HIV/AIDS in the sample: 10% for two-spirit men and 6% for heterosexual men. The study methodology precludes our interpretation of these as prevalence rates. However, if they are accurate, they suggest that Centers for Disease Control and Prevention surveillance data are vast underestimates. Accurately determining the number of American Indian men infected with HIV is difficult because of the lack of mandatory HIV reporting and the rampant racial misclassification of American Indians.42^,43

Analyses provided only partial support for our hypothesis that lifetime victimization would be associated with recent substance use and recent unsafe sex, perhaps because of our limited measure of substance use and high rates of abstinence. Recall that the community organization we studied denies access to individuals visibly under the influence of drugs or alcohol, a policy that likely discourages current users from active membership. Aside from sexual orientation, the only other significant correlate of lifetime HIV risk behaviors was victimization.

What might account for these sexual orientation differences and the independent association of victimization and lifetime HIV risk behaviors? The higher levels of victimization among two-spirit men may relate to their sexual and romantic involvement with other men, who may be more likely to perpetrate violence than women. Additionally, two-spirit men are more vulnerable to bias-related victimization because of their sexual minority status.44 Future research should examine the extent to which victimization among this group is bias related.

Previous research has implicated intrapersonal explanations for the association between victimization and risk behaviors. For example, the experience of victimization may create a sense of isolation and loneliness that the survivor attempts to assuage with unsafe sex, or there may be a self-destructive motivation in the pursuit of unsafe sex, especially among men who have internalized the earlier victimizing experience.45 However, in line with both a fourth-world perspective and ecosocial theory,46 victimization and HIV risk behaviors may be related to larger structural factors such as social inequality and discrimination. Future research should include culturally specific items about what motivates sexual risk behaviors among men with histories of victimization and the influence of larger societal and structural factors related to their oppressed status.

A significant minority of the sample (30%) espoused support for a conspiracy theory of AIDS. This finding likely reflects the general mistrust within the American Indian community, whose medical “treatment” by the US government has included the distribution of blankets laden with smallpox virus and involuntary sterilization.47 Efforts targeted at American Indians to prevent HIV, therefore, must be prepared to address high levels of justifiable suspicion and distrust and, ideally, should be developed and delivered in collaboration with community representatives. Indigenist2^,3 and postcolonial33 approaches to change as well as more culturally tailored interventions48^,49 should be considered.

Limitations of the present study include its cross-sectional nature and inability to specify the timing of reported events, making it impossible to say with certainty which came first: victimization, HIV risk behaviors, or self-identification as two-spirit. Furthermore, we assessed victimization with subjective questions, which require respondents themselves to determine whether experiences constitute “abuse.”50 The respondents were all non-randomly recruited members of an American Indian community organization, thereby decreasing external validity. Finally, our sample size of 71 is small, limiting statistical power and suggesting the need to validate our findings in larger samples. Given the dearth of literature on American Indian men and the urgency of their health concerns, however, even a small preliminary study such as the current one merits attention.

Despite these limitations, the study suggests topics to consider in future research targeting American Indian men, including types and prevalence of victimization experiences, substance use as a result of victimization and a precursor to sexual risk behavior, physical and mental health outcomes related to victimization, and the most efficacious interventions for this at-risk population. Most importantly, the study underscores how American Indian men’s health behaviors must be considered within the context of their ongoing experience of victimization as members of an oppressed population in a fourth-world context. A larger consideration of historical and cultural trauma perpetrated against their ancestors and tribal communities according to an “indigenist” perspective is needed to appropriately contextualize American Indian men’s current victimization and to understand its effects on their risk behaviors and adverse health outcomes.51