Content uploaded by Jane Simoni
Author content
All content in this area was uploaded by Jane Simoni
Content may be subject to copyright.
1
PATTERNS AND PREDICTORS OF HIV RISK AMONG URBAN AMERICAN
INDIANS
Karina L. Walters, Ph.D., Jane M. Simoni, Ph.D.,
and Curtis Harris, B.A.
Abstract: A preliminary survey of HIV risk and service
preferences among American Indians residing in the New
York metropolitan area included 68 women and 32 men (M
age=35.8 years). Overall, the sample was knowledgeable
about the mechanisms of HIV transmission, and 58%
reported having taken an HIV test. However, of the 63%
who reported sexual activity in the last six months, 73%
reported engaging in vaginal or anal sex without a condom
with at least 1 partner, and 52% used condoms none of the
time during vaginal and anal sex. Almost half (43%)
reported alcohol or other drug (AOD) use for nonceremonial
purposes in the last six months. Alarmingly, 44% reported
lifetime trauma, including domestic violence (20%) and
physical (29%) or sexual (26%) assault by a family member
or stranger. Bivariate and multivariate analyses indicated
trauma and drug use were factors that may place
respondents at risk for sexual transmission of HIV. Trauma
variables were better predictors of HIV risk behaviors than
social cognitive variables providing preliminary support for the
use of a postcolonial framework in American Indian HIV
studies.
The Relocation Act (Public Law 959) of the early 1950s instigated a
mass migration of American Indians (AIs) from reservation and rural settings
to large cities across the United States (U.S.). The Act constituted a modern
day attempt by the federal government to deculturate and assimilate AIs. As
a result, today more than 60% of AIs from federally recognized tribes live in
cities (Bureau of the Census, 1991), with urban communities constituting
microcosms of national tribal representation. Migration of AIs from rural to
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
2 VOLUME 9, NUMBER 2
urban settings has led to increases in health problems among this population,
including susceptibility to the HIV epidemic.
Only a handful of studies provide any relevant data on urban AI
health-related concerns. One from Washington (Grossman, Krieger,
Sugarman, & Forquera, 1994) indicated urban AIs were much less healthy
than European Americans. For example, risk factors for poor birth outcomes
were significantly higher for AIs than for European Americans and resembled
the rates among African Americans. All communicable diseases studied
were significantly more common among urban AIs than European Americans.
Urban AIs also suffer a poverty rate three times that of any other ethnic
group, and, as in reservation settings, low socio-economic status correlates
with poor health outcomes. These data on the economic vulnerability and ill
health of AIs suggest they may be disproportionately at risk of HIV infection
(Metler, Conway, & Stehr-Green, 1991).
Despite the risk, only a few published studies in refereed journals
offer empirical research regarding AI HIV/AIDS knowledge, attitudes, and
behaviors (Brassard, Smeja, & Valverde, 1996; Hall, Wilder, Bodenroeder, &
Hess, 1990). The available research indicates AIs are as knowledgeable
regarding HIV as the general population (Hall et al., 1990; Meyers, Calzavara,
Cockerill, Marshall, & Bullock, 1993), tending to know less about HIV
transmission and more regarding specific clinical properties of the disease.
One study indicated no overall knowledge difference between AI men and
women; however, younger respondents were more informed than older
respondents about HIV transmission through blood, casual contact, kissing,
and indirect contact (Meyers et al., 1993). In one unpublished study of HIV
sexual risk behaviors among AIs living off reserve in Canada, 83% of the 376
sexually active AIs reported at least one incident of condomless sex in the
previous 12 months (Bullock et al., 1996). Unprotected sex was significantly
more likely for AI men (vs. women) and individuals with a steady (vs. a
casual) partner. Consistent condom use was more likely among those familiar
with AI traditions and among those reporting a history of physical abuse,
whereas inconsistent condom use was positively associated with age and a
history of sexual abuse.
Alcohol use and abuse have been well documented as critical co-
factors for HIV risk behaviors (O’Hara, Parris, Fichtner, & Oster, 1998; Paul,
Stall, & Davis, 1993). Studies have indicated high rates of alcohol abuse
among AIs, although there is tremendous variation over time, by tribe, and
by reservation (May, 1996). Some studies of reservation-based samples
indicate a pattern of lower problem drinking and higher rates of abstinence
compared to the U.S. general population, although some urban samples
have demonstrated higher rates of drinking (Beltrane & McQueen, 1979;
May, 1996). The comorbid relationship between AOD use and precocious
sexual activity and potential HIV sexual-risk behavior among AI youth has
been well documented (Beauvais, 1992; Conner & Conner, 1992; Walker et
al., 1996). However, the specific mechanisms by which alcohol may act as a
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
PATTERNS AND PREDICTORS OF HIV RISK 3
cofactor for HIV infection are still unclear, and no studies exist that specifically
identify the relationship between alcohol use and HIV sexual risk behaviors
among adult urban AIs.
Anecdotal and empirical evidence suggests that sexual and drug-
related risk behaviors are associated with the legacy of trauma and persistent
destructiveness of colonization that AIs have endured (Tafoya & Delvecchio,
1996; Weaver, 1998; Yellow Horse Braveheart, 1998). Alcohol-related
problems have been associated with delayed trauma related to child sexual
abuse (Flanigan, 1990) as well as other forms of violent crime and trauma
(Stewart, 1996). Exposure to traumatic and abusive childhoods (Bartholow,
Doll, Joy, & Douglas, 1994; Singer, 1995); sexual abuse (Miller & Paone,
1998); lifetime physical or sexual abuse by a spouse or boyfriend (Fischbach
& Herbert, 1997); and rape (Cunningham, Stiffman, Dore, & Earls, 1994)
have been associated as well with HIV sexual risk behavior (Wingood &
DiClemente, 1998). Research among other groups has indicated that HIV
disproportionately affects families already confronted with multiple stressors,
including AOD use, abusive childhoods, and other histories of victimization
(Havens, Mellins, & Pilowski, 1996; Singer 1995). The cumulative effect of
these traumas among AIs has been characterized as a “soul wound,” (Duran,
Duran, Yellow Horse Braveheart, & Yellow Horse-Davis, 1998) which must,
according to the AI postcolonial framework (Duran & Duran, 1995), be
incorporated into any conceptualization of contemporary AI health problems.
Given the potential vulnerabilities and dearth of data on urban AI
needs, the current authors undertook a survey of HIV risk behaviors and
service preferences among AIs. The focus was on New York City (NYC), the
epicenter of the AIDS epidemic and the third largest urban AI community in
the U.S. (46,191 in the Metropolitan Statistical Area for NYC according to the
1990 U.S. Census). Although AIs account for fewer than 1% of all AIDS
cases reported in NYC, the 1997 cumulative AIDS case rate of 238/100,000
reflects the sizable impact the epidemic has had on this group (NYC
Department of Health, 1997). Of the reported AI AIDS cases documented in
the AIDS surveillance data in New York City (
N
= 34), 91% are male. Injection
drug use (IDU; 39%) and sex with other men (35%) were nearly equal
transmission risks among men, while heterosexual transmission (66%)
predominated among AI women. HIV seroprevalence studies conducted in
NYC have not collected data on AI ethnicity (see Stevens & Estrada, 2000,
for an overview of national HIV seroprevalence studies among AIs). The
one exception is a linked serostudy from 1991 to 1994 of voluntary testing
among drug treatment clients in NYC. Although there were much lower
numbers of AIs tested each year than other groups, the percentages of AIs
testing positive in each of the four years (20%, 13%, 4%, and 13%,
respectively) were comparable to those of African Americans (19%, 17%,
13%, and 11%, respectively).
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
4 VOLUME 9, NUMBER 2
In this article, we report findings from a focus group and a pilot
survey. In addition to descriptive as well as program planning information,
preliminary data on predictors (including AOD use and trauma variables) of
lifetime HIV risk behaviors and condom use are provided. We aimed to
identify, in a preliminary manner, individual or contextual factors that relate
to risk behaviors, which is a standard approach utilized for HIV needs
assessment studies and community planning groups (Kelly, 1995; Valdiserri,
Aultman, & Curran, 1995). Because the majority of HIV studies utilize
conceptual frameworks based on social cognitive theory (Bandura, 1986)
and the theory of reasoned action (Fishbein & Ajzen, 1975), we assessed
HIV knowledge, self-efficacy for safer behavior, and perceived personal risk
(Jemmott & Jemmott, 1991; Somlai, Kelly, Wagstaff, & Whitson, 1998) as
potential predictors of HIV risk behaviors. We utilized AI postcolonial theory
(Duran & Duran, 1995) as the fundamental framework for the mulitvariate
exploration of the relationships among trauma, substance use, and HIV risk
behaviors.
Method
Participants
The sample consisted of 100 AIs—68 women and 32 men. They
ranged in age from 18 to 75 years (
M
=35.8;
SD
=12.6) and were fairly well
educated, with 88% having earned a high school diploma/GED and 32% at
least a bachelor’s degree. Monthly income surpassed $2,400 for 47% of the
sample. Over the previous 12 months, 55% were employed full-time, 25%
part-time, and 20% were unemployed. The sample was predominately
heterosexual (91%). Although 55% reported a steady intimate partner, only
16% were legally married (note that “valid” percentages are reported
throughout the paper—i.e., participants with missing data on the particular
item were excluded).
Self-reported individual blood quantum across all tribes varied as
follows: 1-25% (20%), 26-50% (43%), 51-75% (18%), 76-99% (9%), and
full-blooded (10%); 53% of participants were enrolled in an AI tribe. Two
percent had attended an AI boarding school. Only 9% had lived on an AI
reservation or tribal lands within the last year. One-third of the sample had
been adopted (4% by AI families and 28% by non-Indian families). Most
(86%) had learned English as their first language.
Measures
Questionnaire items assessed HIV sexual and drug risk behaviors,
attitudinal items, trauma, barriers to condom use, female condom attitudes,
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
PATTERNS AND PREDICTORS OF HIV RISK 5
and program planning issues. Items validated in other studies of HIV risk
(e.g., Kelly, 1995; Somlai et al., 1998) were used when possible.
Lifetime HIV Risk Behaviors
From a checklist of 15 HIV risk behaviors (see Table 1 for the actual
items), respondents indicated those in which they had engaged at any time
in their life. As indicated in Table 1, the risk factors were collapsed into
three risk categories. Three items referring to IDU either by itself or in
association with sexual behavior were used to calculate an IDU-sex risk
variable. The two items referring to sexual behavior while drunk or high
were used to calculate the high-sex variable. Seven sexual risk behavior
items were combined to form an indicator of sex risk (omitting condomless
vaginal, anal, and forced sex because of their conceptual overlap with trauma
and steady partner). As the sum scores for these three variables were not
normally distributed, dichotomous indices were created such that respondents
having engaged in any of the relevant items were assigned a “1” and others
a “0”.
Sexual Behaviors
Respondents indicated whether they had sex with any man or woman
in the past 6 months (we defined “sex” for the respondents as physical
contact that goes beyond hugging and kissing but does not necessarily include
intercourse). Those who answered affirmatively indicated the number of
their male and female sexual partners during that period as well as the
number of partners with whom they had engaged in vaginal, anal, and oral
sex with and without a condom (e.g., “your penis in her vagina—without a
condom”). Consistency of condom use during oral sex and vaginal/anal sex
was assessed with two additional items scored from
none of the time
(1) to
every time
(5).
Substance Use
Respondents reporting AOD use outside of ceremonial or religious
settings in the last 6 months described how frequently they had used alcohol
to the point of being drunk, cocaine powder, crack cocaine, marijuana,
inhalants, amphetamines, heroin, ecstasy, hallucinogens, sedatives, and IDs
as
not at all
,
a few times
, or
fairly often
. The IDUs indicated if they had used
only clean needles. Additionally, respondents indicated with two yes/no items
whether their condom use had been affected by their own (or their partner’s)
AOD use in the past 6 months.
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
6 VOLUME 9, NUMBER 2
Attitudinal Factors
Three measures were used to assess cognitive and attitudinal
constructs according to social cognitive theories. Participants’
knowledge
of
HIV transmission was assessed with 23 true or false items (Kelly, St.
Lawrence, Hood, & Brasfield, 1989). Internal consistency (Cronbach’s alpha)
in the present sample was .69. A scale of nine items was used to assess
self-efficacy
for safer drug and sexual behaviors (Smith, McGraw, Costa, &
McKinlay, 1996). Respondents were asked how sure they were that they
could, for example, talk about safe sex with a sexual partner or buy condoms.
Each item was scored from
not at all sure
(1) to
extremely sure
(5). Internal
consistency in the present sample was .86. Finally, one item asked
respondents to estimate their
perceived risk
of contracting HIV/AIDS from
no risk
(0) to
high risk
(3).
Trauma
With a checklist of six items (e.g., physically assaulted by family
member), respondents indicated whether they had ever been physically or
sexually assaulted by a spouse/partner, family member, or stranger.
Barriers to Condom Use
From a checklist of 16 items (Meyers et al., 1993), the sexually
unsafe participants were asked to indicate all the reasons they did not use a
condom or barrier during vaginal and anal sex in the last 6 months.
Attitudes Toward the Female Condom
Participants were asked six yes/no items about the female condom.
Internal consistency in the present study was .59.
Community and Program Planning
Finally, in a series of checklists and one open-ended item, respondents
indicated targets for HIV communication they had used and preferred, service
preferences, and barriers to accessing services.
Procedure
The present survey constituted the first step in a comprehensive
assessment of HIV risk behaviors and service among AIs in NYC. A literature
review, key informant interviews, a focus group of HIV educators (described
later), and pilot testing of measures preceded work on the survey.
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
PATTERNS AND PREDICTORS OF HIV RISK 7
Two AI women active and known in the AI community were informed
about the project and trained to distribute the present survey at an AI
gathering (powwow) in an outer borough of NYC in the summer of 1997.
The two workers recruited a convenience sample by encouraging AI men
and women who approached the American Indian Community House (AICH)
HIV outreach table to complete the survey and by circulating at the powwow
and encouraging AI individuals to go to the table to complete the questionnaire.
AI individuals who did not want to approach the table were given the survey
to complete away from the table.
The study’s title (American Indian Wellness Project) and eligibility
criteria (at least age 18 and of AI descent) were clearly printed on the cover
sheet of the questionnaire. The introduction thanked respondents for
volunteering and stated that the information provided would be used to access
funding and to provide services that will better assist AI people. Respondents
were paid $5 for completing the survey, which required less than 30 minutes.
Results
The qualitative themes that emerged from the focus group and key
informant interviews are highlighted below, followed by a discussion of the
quantitative results. To identify any potential subgroups at risk, we ran
bivariate analyses using chi-squares,
t
tests, and Pearson correlation
coefficients with all the general and Indian-specific demographic indicators
and the main variables of interest. The very few that were significant are
controlled in the multivariate analyses.
Qualitative Data
To identify critical community-based needs and themes, we
interviewed key informants, including service recipients and other community
members, and conducted a focus group with eight AI service providers (four
men and four women), including case managers, outreach workers, and
staff. This preliminary qualitative work suggested AI gay men, youth, and
women in violent relationships were at increased risk for exposure to HIV
due to unsafe sexual practices. Sex traders, IDUs, and their partners were
also identified as being at risk. Participants expressed concerns regarding
circular migration; in particular, they cited AI men who come to the city for
job-related concerns and have anonymous sex with other men and then
return to their wives in AI communities upstate. In terms of AOD use, the
outreach coordinators stated that AI youth are likely to engage in sex and
alcohol use simultaneously, thereby increasing their risk for HIV exposure.
They also mentioned culturally specific risk behaviors, including skin piercing,
marking (i.e., tattooing), and the use of indigenous healing practices that
involve blood (i.e., piercing, blood letting). One member noted that during
ceremonies (e.g., at Sun Dance which involves piercing the chest for men
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
8 VOLUME 9, NUMBER 2
and dancing until feet might become bloody for both men and women),
some medicine people are beginning to take precautions to protect against
HIV infection such as cleaning piercing equipment.
Attitudinal Factors: Risk Perception, Self-Efficacy, and HIV
Knowledge
Two survey respondents indicated they were HIV seropositive; others
reported they had no (32%), low (52%), moderate (13%), or high (2%) risk
of getting HIV/AIDS. The mean level of self-efficacy for safer behavior was
4.55 (
SD
= .69) of a possible 5 and was highly negatively skewed. The
mean level of HIV knowledge also was fairly high (19.91 of a possible 23;
SD
= 3.03). Items most frequently missed (and the correct answers) were:
“People who have the AIDS virus quickly get sick” (F), “A negative result on
the HIV test can happen even if somebody has the AIDS virus” (T), “No case
of AIDS was ever caused by social (dry) kissing” (T), “By having just one sex
partner at a time you can protect yourself from AIDS” (F). Four independent-
samples
t
tests indicated HIV knowledge was generally higher among the
81% of the sample who personally knew someone with HIV/AIDS,
t
(98)=
-2.25,
p
<.05; the 41% who had family or close friends with HIV/AIDS,
t
(97)=
-1.87,
p
<.10; the 78% who had considered having a test,
t
(98)=-2.24,
p
<.05;
and the 58% who had actually taken the test,
t
(98)=-2.34,
p
<.05. HIV
knowledge was positively correlated with self-efficacy to engage in safer
behaviors,
r
(98)=.36,
p
<.001, but, contrary to social cognitive models, neither
predicted any drug, trauma, or HIV sexual risk behavior indicator.
Lifetime Exposure to Trauma
A relatively small number of respondents reported physical assault
by a family member (16%), spouse/partner (16%), or stranger (19%), or
sexual assault by a family member (13%), spouse/partner (7%), or stranger
(15%). However, when responses were collapsed across perpetrator, another,
more disturbing picture emerged: 29% reported physical assault by a
nonpartner, 26% reported sexual assault by a nonpartner, and 20% reported
physical or sexual assault by a partner (i.e., domestic violence). All three
were intercorrelated.
Patterns of Risk Behavior
Lifetime drug and sexual risk behaviors
. From a checklist of 15
lifetime drug and sexual risk behaviors, participants reported engaging in an
average of 3.40 (
SD
=3.31); 76% had engaged in at least one. Percentages
for each behavior are shown in Table 1.
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
PATTERNS AND PREDICTORS OF HIV RISK 9
Sexual risk behaviors in last 6 months
. Analyses indicated 63
participants (39 women and 24 men) reported sexual activity in the last 6
months. Among the sexually active participants, 73% had engaged in
condomless anal or vaginal sex with at least one partner. Rates of unsafe
sex (condomless vaginal, oral, or anal) within each relationship dyad were
as follows: female participants with women (
n
=1) 100%, female participants
with men (
n
=40) 78%, male participants with men (
n
=4) 100%, male
participants with women (
n
=21) 82%. Figure 1 shows rates of unsafe sex
by gender. The two items assessing consistency of condom use indicated
only 17% of participants used condoms all of the time (52% none of the
time) during vaginal/anal sex and 7% used condoms all of the time (85%
none of the time) during oral sex.
Drug risk behaviors in the last 6 months
. In the past 6 months, 43%
of the participants reported AOD use for non-ceremonial purposes. Twenty-
seven percent had used alcohol to the point of being drunk, 22% used
marijuana, 4% used IDUs, and less than 6% used each of the remaining
drugs. Only 3% of respondents said their condom use had been affected
because they were drunk or high (2% said because their partner was drunk
or high).
Table 1
Lifetime Risk Behaviors
% Behavior
67 Condomless vaginal intercourse
51
3
Sex while drunk or high
42
3
Condomless sex while drunk or high
31
2
Sex with a non-monogamous partner
29
2
Sex with a stranger
24 Condomless anal intercourse
22
2
More than 2 sexual partners in a month
20
2
Sex with someone with an STD
19
2
Had an STD
13 Forced condomless sex
10
1
Sex with an injection drug user
8
2
Traded sex for money, drugs, or favors
4
1
Injected nonprescription drugs
4
2
Sex with HIV+ person
3
1
Shared dirty needles
Note:
N
=96–98. STD=sexually transmitted disease. 1=used to calculate the IDU-
sex risk indicator. 2=used to calculate the sex risk indicator. 3=used to calculate the
high-sex indicator.
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
10 VOLUME 9, NUMBER 2
Predictors of Risk Behaviors: Multivariate Analysis
A correlation matrix of bivariate analyses (depicted in Table 2)
indicated trauma and drug use may be correlated with sexual risk behaviors.
To further explore these associations, we created three multivariate sexual
risk behavior models—one for lifetime sex risk behaviors and the other two
for unsafe sex in the past 6 months. Variables were selected for inclusion in
the models based on their significance in the bivariate analyses.
Lifetime Sex Risk Behaviors
Trauma exposure variables (domestic violence and nonpartner sexual
assault) as well as lifetime IDU-sex and high-sex behaviors were explored as
predictors of lifetime sex risk behaviors in a simultaneous least squares
Figure 1
Percentage of Male and Female Participants Engaging in Oral,
Vaginal, and Anal Sex Without a Condom in the Last 6 Months
Females (
n
= 68)
Males (
n
= 32)
Oral Vaginal Anal Any
29%
66%
46%
53%
3%
69%
9%
47%
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
PATTERNS AND PREDICTORS OF HIV RISK 11
Table 2
Intercorrelations Among Main Variables
Variable % 1 2 3456 7 8910
Trauma Variables
(Lifetime)
1. Nonpartner
physical assault
29
2. Nonpartner
sexual assault
26 .51***
3. Domestic
violence 20 .18† .34***
AOD and Sexual
Risk Behaviors
(Lifetime)
4. IDU-sex risk
12 .24* .14 .35***
5. High-sex 51 .19† .21* .10 .25*
6. Sex risk 51 .14 .25** .35*** .30*** .57***
Demographic
Factors
7. Steady partner
status
55 -.02 .04 -.01 -.10 .21* .01
8. Income na -.11 -.11 -.20† -.05 -.06 -.14 .20*
Risk Behaviors
(Last 6 Months)
9. AOD use
43 .09 .05 .07 .05 .28*** .23* .23* -.10
10. Condomless
vaginal/anal sex
73 .06 -.01 .19 .02 .12 .12 .36*** .29* -.06
11. Condom
consistency for
vaginal/anal sex na -.05 -.24† -.21 -.18 -.18 -.17 -.44***-.25† .03-.49***
Note:
N
=97-100, except for #10 and #11 where
N
=58-63. Statistics are Pearson
correlation coefficients.
†
p
<.10. *
p
<.05. **
p
<.01. ***
p
<.005.
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
12 VOLUME 9, NUMBER 2
logistic regression (refer to Walters & Simoni, 1999, for an analysis of women
only). Steady partner status was not included because the model included
risks across the lifetime and only current partner status was assessed. The
simultaneous logistic regression model was significant (
X
2
[4,
N
=95],
p
<.005,
Hosmer-Lemeshow statistic=1.75,
df
=4,
p
=.78), indicating respondents who
had engaged in sex while drunk or high were 14.35 times more likely than
those who had not to engage in risky sexual behaviors (odds ratio [OR]=14.35,
95% Confidence Interval [CI]=4.65, 44.25). Additionally, those who had
experienced domestic violence were 9.26 times more likely than those who
had not to engage in risky sexual behaviors (OR=9.26, 95% CI=1.80, 47.53).
Neither sexual assault nor IDU-sex risk had a significant net effect on lifetime
sexual risk behaviors.
Unsafe Sex in the Last 6 Months
Income and steady partner status were explored as predictors of
condomless vaginal or anal sex in the last 6 months in a simultaneous logistic
regression (trauma and AOD were not significant in bivariate analyses and,
therefore, not included). The logistic regression model was significant (
X
2
[2,
N
=62],
p
<.01, Hosmer-Lemeshow statistic=7.50,
df
=4,
p
=.11), indicating
that respondents in steady partnerships were 4.33 times more likely to engage
in condomless vaginal or anal sex than those without steady partners
(OR=4.33, 95% CI=1.11, 16.77). Monthly income status had no net effect
on condomless vaginal or anal sex.
Consistency of Condom Use in the Last 6 Months
Nonpartner sexual assault, steady partner status, and monthly income
were explored as predictors of consistency of condom use during vaginal
and anal sex in the last 6 months in a simultaneous regression. The model
accounted for 23% of the variance in consistency of condomless sex,
F
(2,52)=7.69,
p
.<005. Having a current steady partner (
B
=-.40,
T
=-.22,
p
<.005) as well as a history of nonpartner sexual assault (
B
=-.27,
T
=-2.25,
p
<.05) were predictors of decreased consistency of condom use. Monthly
income had no net influence.
Reasons for Not Using Condoms
The 48 participants who had unsafe vaginal or anal sex in the past 6
months indicated from a checklist of 16 items all the reasons they had not
used a condom. The majority indicated being with their steady partner
(87%), assuming they were safe (68%), not having the AIDS virus (63%),
and simply not wanting to (57%). They less frequently endorsed their partner
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
PATTERNS AND PREDICTORS OF HIV RISK 13
not wanting to (41%), their partner saying he or she did not have the AIDS
virus (37%), not having a condom at the time (29%), the sex being so exciting
they didn’t want to use one (26%), finding condoms painful or uncomfortable
(14%), using AODs (10%), desiring pregnancy (6%), their partner getting
angry for suggesting using one (6%), and being too embarrassed to get
condoms (4%). Two percent each said not being able to talk about it, not
being able to afford condoms, or being forced to have sex against their will
were reasons for not using condoms.
Attitudes Toward the Female Condom
Respondents reported generally positive attitudes toward the female
condom. Most reported having heard about (88%) or seen (55%) this new
contraceptive and would consider using it (67%) or would like to learn how
to use it (57%). However, far fewer had personally bought or obtained
(11%) or actually used one (6%).
Community and Program Planning Issues
As shown in Figure 2, most respondents indicated they would prefer
to speak about HIV/AIDS to doctors, followed by their spouse/partner and AI
friends. Paired
t
tests of mean responses indicated a higher percentage of
respondents preferred to speak to doctors, their spouse/partner, AI HIV project
staff, nurse, AI health representative, or elders than those who actually did
speak to these targets, whereas a lower percentage of respondents preferred
to talk to non-AI friends than those who actually did.
In answer to an open-ended item on what services would decrease
respondents’ risk of getting HIV/AIDS or would assist them if they had HIV/
AIDS, respondents indicated culturally relevant peer counseling and outreach,
telephone hot line, outreach to youth (especially in the schools), condom
negotiation skills, peer groups, TV programs, free condom distribution, and
public workshops by healthcare providers. Additionally, in response to a
checklist of which HIV services would be most helpful to them, respondents
indicated sex education and condom use to AI youth (77%), HIV education
services at the AICH (75%), cultural approaches to negotiating condom use
(62%), instruction in how to use the female condom (61%), HIV education
services available in your home (52%), and instruction in getting a partner to
use a condom (45%).
Responses to the checklist of means of being informed about HIV/
AIDS services indicated respondents preferred direct, face-to-face encounters
such as educational talks or workshops (68%) and community outreach
workers (36%), followed closely by more anonymous methods such as
television (55%), AICH bulletin (34%), pamphlets (31%), radio (29%), word-
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
14 VOLUME 9, NUMBER 2
Figure 2
Preferred and Actual Sources of HIV Information
% Preferred
% Actual
Aunts or
Uncles
*Elders
Parents
***AI Health
***Nurse
**Non-AI
Friend
***AI HIV
Project
AI Friends
**Spouse/
Partner
***Doctor
Note:
N
=94-100. Statistics were paired
t
tests of mean preferred versus mean
actual percentages for each target.
*
p
<.05. **
p
<.01. ***
p
<.005.
100
500
67
45
36
57
38
53
43
64
40
69
50
68
78
71
43
73
68
85
74
87
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
PATTERNS AND PREDICTORS OF HIV RISK 15
of-mouth (26%), or other (15%). Only eight respondents needed or wanted
information, education, or treatment related to HIV in the last 6 months. As
only four attempted to access these services, we did not conduct analyses of
reported barriers.
Discussion
A survey of 100 AIs at a powwow in NYC provided some of the first
data available on the HIV risk behaviors of urban AIs. Findings indicated
relatively high knowledge and self-efficacy for safer sex behavior, which may
indicate that HIV education and outreach have been successful with this
population. However, elevated knowledge and self-efficacy were not
associated with safer sex behaviors (i.e., consistent condom use), which
may suggest that community education and outreach does not adequately
change community behaviors. Although this study did not adequately test
social cognitive models, these findings suggest the effectiveness and
applicability of such models need to be further considered with respect to
urban AIs.
The majority of respondents knew someone living with HIV/AIDS.
Over half had been tested for the virus, and 2% were HIV-positive. This
seroprevalence is almost three times the estimated rate for NY state of .8%
and corresponds with initial Indian Health Service seroprevalence estimates
of 1-3% for AI populations. Of course, we employed a nonprobability sampling
design, limiting the extent to which we can generalize these findings.
Among the 63 participants who were sexually active in the last 6
months, rates of condomless sex were high and related to having a steady
partner. Inconsistent condom use was related to having a steady partner
and nonpartner sexual assault. These findings support targeting couples for
safer sex condom interventions. Future research will need to identify the
specificity of the HIV risk associated with condomless sex among different
types of steady partner relationships (i.e., monogamous or nonmonogamous).
Because approximately one-third of all AI children are estimated to be at risk
for becoming victims of child abuse, with AI girls incurring disproportionate
sexual and physical abuse (National Indian Justice Center, 1990), it is
imperative that future studies identify the cumulative effect of sexual trauma
on condom use among AIs. The finding also highlights the importance of
integrating sexual trauma survivor issues into HIV prevention strategies.
Over two-thirds of the sample engaged in at least one unsafe sexual
or AOD risk behavior in their lifetime. Moreover, exposure to trauma, AOD
use, and sexual risk behaviors were intercorrelated, suggesting these factors
needed to be considered in concert to decrease HIV risk among urban AIs.
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
16 VOLUME 9, NUMBER 2
These initial findings on the relationship between AOD use and
exposure to trauma are supported by previous research. Preliminary studies
of both non-AI male and female samples suggest that there is a causal
connection between childhood victimization and development of drinking
problems (Flanigan, 1990). Kovach (1986) postulates that the link between
childhood abuse and adult alcohol abuse might be mediated by a delayed
onset of posttraumatic stress disorder (PTSD) symptoms in adulthood with
which the individuals attempt to cope through alcohol abuse (i.e., self-
medication hypothesis). Other studies also supported this linkage with respect
to exposure to other forms of violent crime and trauma (Stewart, 1996).
Kilpatrick, Acierno, Resnick, Saunders, and Best (1997) found that the severity
of trauma symptoms is positively associated with comorbid alcoholism across
a variety of traumatic events. Future research with AI populations will need
to identify the role of PTSD or other traumatic symptomatology as a factor
related to AOD use in sexual encounters. Previous trauma might have
implications for current sexual behavior, negotiation skills in sexual encounters,
and alcohol use as a form of self-medication (Stewart, 1996) among AI
populations.
Future research will need to discern how the patterns of sexual risk
behaviors are temporally associated with AOD use and exposure to trauma.
For example, the self-medication hypothesis postulates that AOD use dampens
cognitive symptoms (Stewart, 1996). Individuals exposed to trauma may be
at increased risk for learning to drink to reduce tension. Consistent with a
postcolonial framework (Duran & Duran, 1995), alcohol may be used as a
coping mechanism for those who have experienced prolonged, cumulative
trauma, although future research will need to empirically substantiate this
possibility. Although there is evidence of the comorbidity of alcohol-related
problems and trauma among AI populations, studies of comorbidity fail to
resolve the temporal ordering of AOD alcohol use, trauma exposure and
subsequent HIV sexual risk behaviors. Future research will need to identify
the temporal patterns of exposure to trauma (including delayed traumatic
reactions) and alcohol-drug use (including drinking styles) to ultimately discern
the mechanisms by which alcohol and trauma may act as covariates in sexual
risk-taking. Moreover, the mixing and phasing of AOD while simultaneously
engaging in risky sexual behaviors reinforces the importance in future research
of identifying the contextual use of alcohol in relation to sexual expectations
(i.e., alcohol-sex expectancy).
The data on services have implications for community-based HIV
preventive program planning among urban AIs. Findings that 57% of the
respondents were willing to learn how to use the female condom indicate
the eagerness of AIs to incorporate new HIV prevention strategies. AIs in
the study also identified preferred targets for HIV communication (e.g.,
doctors), but we need further information as to the preferred gender, age,
ethnicity, and tribe (if these preferences exist). Some AIs may not want to
access a local tribal-affiliated organization for fear that via the “moccasin
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
PATTERNS AND PREDICTORS OF HIV RISK 17
telegraph” their questions or concerns will not remain confidential. Participants
also indicated a strong preference for direct, face-to-face encounters for
public dissemination of HIV education, suggesting the possibility of home
visits. In fact, in a study of domestic violence among urban AIs, Norton and
Manson (1995) successfully used home visits. Perhaps a culturally appropriate
home-visit intervention could address domestic violence and HIV risk
simultaneously.
There are several important methodological limitations to these data.
First, because we employed a non-probability convenience sample (of mainly
heterosexual, educated women), generalizability of the findings is not possible.
Future studies might incorporate modified probability sampling methods,
such as multiplicity sampling (Rothbart, Fine, & Sudman, 1989). Second,
because the data were self-reported, the findings may be subject to social
desirability. Future research comparing self-administered versus interview
survey methodologies among AIs are warranted to decrease cultural
insensitivity and social desirability. The results from the focus group and key
informant interviews support matching respondents by gender for survey
interviewing techniques given the charged nature of the HIV sexual risk
behavior questions. Third, our assessment of trauma, AOD use, and sex risk
variables was primarily dichotomous and did not provide depth and breadth
or the social context related to such behaviors. For example, drinking styles
(abusive vs. moderate), chronicity, frequency, and quantity of AOD use were
not addressed. Additionally, we did not identify the role of the steady partner
and meaning of that relationship to the respondent (again it was a dichotomous
variable) which would give us a better understanding of what steady partner
status means in terms of inconsistent condom use. Further quantitative and
qualitative methods for psychometric refinement regarding trauma, AOD use/
abuse, and HIV sexual risk behaviors clearly need to be cultivated, validated,
and refined in collaboration with AI populations. Finally, our cross-sectional
design precludes any causal interpretations.
Despite these methodological limitations, the study has provided
some preliminary data on an understudied and potentially at-risk population.
Future preventive efforts should consider the strengths and resources of
those who are using condoms consistently. Incorporating the role of resilient
AI community members into the development of research and community
intervention would not only facilitate construction of culturally meaningful
interventions, but would also reinforce the current resilience and strength of
urban AI communities.
Karina L. Walters
Columbia University School of Social Work
622 W. 113
th
Street
New York, NY 10025
E-Mail: kw81@columbia.edu
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
18 VOLUME 9, NUMBER 2
References
Bandura, A. (1986). Social diffusion and innovation. In A. Bandura (Ed.),
Social
foundations of thought and action: A social cognitive theory
(pp. 142-
181). Englewood Cliffs, NJ: Prentice Hall.
Bartholow, B. N., Doll, L. S., Joy, D., & Douglas, J. M. (1994). Emotional, behavioral,
and HIV risks associated with sexual abuse among adult homosexual and
bisexual men.
Child Abuse and Neglect, 18
(9), 747-761.
Beauvais, F. (1992). An annotated bibliography of papers on drug abuse among
Indian youth by staff of the Tri-ethnic center for prevention research.
American Indian and Alaska Native Mental Health Research: The Journal of
the National Center, 5
, 51-67.
Beltrane, T., & McQueen, D. V. (1979). Urban and rural drinking patterns: The
special case of the Lumbee.
The International Journal of the
Addictions,14
(4), 533-548.
Brassard, P., Smeja, C., & Valverde, C. (1996). Needs assessment for an urban
Native HIV and AIDS prevention program.
AIDS Education and Prevention,
8
(4), 343-351.
Bullock, S. L., Myers, T., Calzavara, L. M., Cockerill, R., Marshall, V. W., & Burchell,
A. (1996).
Unprotected intercourse and the meanings ascribed to sex by
Aboriginal people living on-reserve in Ontario, Canada
. Presentation at the
XI International Conference on AIDS, Vancouver, BC.
Bureau of the Census. (1991).
1990 Census counts of American Indians,
Eskimos, Aleuts, and American Indian and Alaska Native areas
. Washington,
DC: Racial Statistics Branch, Population Division.
Calzavara, L. M., Burchell, A. N., Myers, T., Bullock, S. L., Escobar, M., & Cockerill,
R. (1998). Condom use among Aboriginal people in Ontario, Canada.
International Journal of STD and AIDS, 9
(5), 272-279.
Conner, J. L., & Conner, C. N. (1992). Expected benefits of alcohol use on
sexual behavior: Native American adolescents.
Psychological Reports, 70
,
91-98.
Cunningham, R. M., Stiffman, A. R., Dore, P., & Earls, F. (1994). The association
of physical and sexual abuse with HIV risk behaviors in adolescence and
young adulthood: Implications for public health.
Child Abuse and Neglect,
18
(3), 233-245.
Duran, E., & Duran, B. (1995).
Native American postcolonial psychology
. Albany,
NY: State University of New York.
Duran, E., Duran, B., Yellow Horse Braveheart, M., & Yellow Horse-Davis, S.
(1998). Healing the American Indian soul wound. In Y. Danieli (Ed.),
Intergenerational handbook of multigenerational legacies of trauma
(pp.
341-354). NY: Plenum.
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
PATTERNS AND PREDICTORS OF HIV RISK 19
Fischbach, R. L., & Herbert, B. (1997). Domestic violence and mental health:
Correlates and conundrums within and across cultures.
Social Science and
Medicine, 45
(8), 1161-1176.
Fishbein, M., & Ajzen, I. (1975).
Belief attitude, intention, and behavior: An
introduction to theory and research
. Reading, MA: Addison-Wesley.
Flanigan, B. J. (1990). The social context of alcohol consumption prior to
female sexual intercourse.
Journal of Alcohol and Drug Education, 36
(1),
97-113.
Gregory, D. (1992). Much remains to be done.
American Indian and Alaska
Native Mental Health Research: The Journal of the National Center, 4
(3),
89-94.
Grossman, D. C., Krieger, J. W., Sugarman, J. R., & Forquera, R. A. (1994).
Health status of urban American Indians and Alaska Natives.
Journal of the
American Medical Association, 271
(11), 845-850.
Hall, R. L., Wilder, D., Bodenroeder, P., & Hess, M. (1990). Assessment of AIDS
knowledge, attitudes, behaviors, and risk level of northwestern American
Indians.
American Journal of Public Health, 80
(7), 857-877.
Havens, J. F., Mellins, C. A., & Pilowski, D. (1996). Mental health issues in HIV-
affected women and children.
International Review of Psychiatry, 8
(2-3),
217-225.
Jemmott, L. S., & Jemmott, J. B. (1991). Applying the theory of reasoned
action to AIDS risk behavior: Condom use among Black women.
Nursing
and Research, 40
, 228-233.
Kelly, J. A. (1995).
Changing HIV risk behavior: Practical strategies
. NY: Guilford
Press.
Kelly, J. A., St. Lawrence, J. S., Hood, H. V., & Brasfield, T. L. (1989). An
objective test of AIDS risk behavior knowledge: Scale development,
validation, and norms.
Journal of Behavioral Theory and Experimental
Psychiatry, 20
, 227-234.
Kilpatrick, D. G., Acierno, R., Resnick, H. S., Saunders, B. E., & Best, C. L. (1997).
A 2-year longitudinal analysis of the relationship between violent assault
and substance use in women.
Journal of Consulting and Clinical Psychology,
65
(5), 834-847.
Kovach, J. A. (1986). Incest as a treatment issue for alcoholic women.
Alcoholism
Treatment Quarterly, 3
, 1-15.
May, P. A. (1996). Overview of alcohol abuse epidemiology for American Indian
populations. In G. D. Sandefeur, R. R. Rundfuss, & B. Cohen (Eds.),
Changing
numbers, changing needs, American Indian demography and public health
(pp. 235-261). Washington, DC: National Academy Press.
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
20 VOLUME 9, NUMBER 2
Metler, R., Conway, G. A., & Stehr-Green, J. (1991). AIDS surveillance among
American Indians and Alaska Natives.
American Journal of Public Health,
81
(11), 1469-1471.
Meyers, T., Calzavara, L. M., Cockerill, R., Marshall, V. W., & Bullock, S. L. (1993).
Ontario First Nations AIDS and Healthy Lifestyle Survey, National AIDS
Clearinghouse
. Canadian Public Health Association: Ottowa, Ontario.
Miller, M., & Paone, D. (1998). Social network characteristics as mediators in the
relationship between sexual abuse and HIV risk.
Social Science & Medicine,
47
(6), 765-777.
National Indian Justice Center. (1990).
Child abuse and neglect in American
Indian and Alaska Native communities and the role of the Indian Health
Service
. Unpublished final report, U.S. Department of Health and Human
Services.
Norton, I. M., & Manson, S. M. (1995). A silent minority: Battered American
Indian women.
Journal of Family Violence, 10
(3), 307-318.
NYC Department of Health. (1997).
Epidemiological profile of HIV/AIDS in
NYC
. Bureau of Disease Intervention Research. NYC Department of Health.
O’Hara, P., Parris, D., Fichtner, R. R., & Oster, R. (1998). Influence of alcohol and
drug use on AIDS risk behavior among youth in dropout prevention.
Journal
of Drug Education, 28
(2), 159-168.
Paul, J. P., Stall, R., & Davis, F. (1993). Sexual risk for HIV transmission among
gay/bisexual men in substance abuse treatment.
AIDS Education and
Prevention, 5
, 11-24.
Rothbart, G. S., Fine, M., & Sudman, S. (1989). On finding and interviewing the
needles in the haystack: The use of multiplicity sampling. In E. Singer & S.
Presser (Eds.),
Survey research methods: A reader
(pp. 18-31). Chicago,
IL: The University of Chicago Press.
Singer, N. (1995). Understanding sexual risk behavior from drug users’ accounts
of their life experiences.
Qualitative Health Research, 5
(2), 237-249.
Smith, K. W., McGraw, S. A., Costa, L. A., & McKinlay, J. B. (1996). A self-
efficacy scale for HIV risk behaviors: Development and evaluation.
AIDS
Education and Prevention, 8
(2), 97-105.
Somlai, A. M., Kelly, J. A., Wagstaff, D. A., & Whitson, D. P. (1998). Patterns,
predictors, and situational contexts of HIV risk behaviors among homeless
men and women.
Social Work, 43
, 7-20.
Stevens, S., & Estrada, A. L. (2000). Editorial: HIV and AIDS Among American
Indians and Alaska Natives.
American Indian and Alaska Native Mental Health
Research: The Journal of the National Center, 9
(1), v-ix. Available at:
www.uchsc.edu/sm/ncaianmhr/jover1.htm.
Stewart, S. H. (1996). Alcohol abuse individuals exposed to trauma: A critical
review.
Psychological Bulletin, 120
, 83-112.
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
PATTERNS AND PREDICTORS OF HIV RISK 21
Tafoya, N., & Delvecchio, A. (1996). Back to the future: An examination of the
Native American holocaust experience. In M. McGoldrick & J. Giordano
(Eds.),
Ethnicity and Family Therapy (2
nd
edition)
(pp. 45-54). NY: Guilford
Press.
Valdiserri, R. O., Aultman, T. V., & Curran, J. W. (1995). Community planning: A
national strategy to improve HIV prevention programs.
Journal of Community
Health, 20
(2), 87-100.
Walker, R. D., Lambert, D. M., Walker, P. S., Kivlahan, D. R., Donovan, D. M., &
Howard, M. O. (1996). Alcohol abuse in urban Indian women: A longitudinal
study for assessment and risk evaluation.
American Indian and Alaska Native
Mental Health Research: The Journal of the National Center, 7
(1), 1-47.
Walters, K. L., & Simoni, J. M. (1999). Trauma, substance use, and HIV risk
among urban American Indian women.
Cultural Diversity and Ethnic Minority
Psychology: Effects of HIV/AIDS among Ethnic Minority Women, Couples,
Families, and Communities [Special Issue], 5,
236-248.
Weaver, H. N. (1998). Indigenous people in a multicultural society: Unique
issues for human services.
Social Work, 43
(3), 203-212.
Wingood, G. M., & DiClemente, R. J. (1998). Partner influences and gender-
related factors associated with noncondom use among young adult African
American women.
American Journal of Community Psychology, 26
(1), 29-
51.
Yellow Horse Braveheart, M. (1998). The return to the sacred path: Healing
the historical trauma and historical unresolved grief response among the
Lakota through a psychoeducational group intervention.
Smith College
Studies in Social Work, 68
(3), 288-305.
Authors’ Note
This study was supported in part by the New York City American Indian
Community House. Parts of this paper were presented at the meeting of the
American Psychological Association, San Francisco, August, 1998.
We acknowledge research assistants Sandra B. Adames, Joe F.
Kerwin, Dan S. Lichtman, Mary G. Martone, Nafisseh Saroudi, and Carrie B.
Sendrow.
Karina Walters is an enrolled citizen of the Choctaw Nation of
Oklahoma, and Curtis Harris is San Carlos Apache.
American Indian and Alaska Native Mental Health Resea ch: The Journal of the National Center r
American Indian and Alaska Native Programs, University of Colorado at Denver and Health Sciences Center (www.uchsc.edu/ai)
