The Association Between Alcohol Use and Sexual Risk Behaviors Among African American Women Across Three Developmental Periods: A Review

Abstract

Background:

African American women experience increased rates of sexually transmitted diseases and HIV. The use of alcohol may increase sexual risk behaviors among this population.

Purpose:

This paper provides a review and critique of the literature examining the association between alcohol use and sexual risk behaviors among African American females including a: (a) synthesis of research findings from adolescent, college-aged/early adulthood, and adult samples; (b) methodological critique of the literature; and (c) guidance for future research.

Methods:

We reviewed 32 studies examining the association between alcohol use and sexual risk behaviors among African American females across developmental periods.

Results:

Similar to previous association studies, results suggest that increased use of alcohol is associated with increased sexual risk practices among African American females. Further, even non-abuse levels of drinking among African American females, at all ages, were related to increased sexual risk-taking.

Conclusions:

Future studies should seek to recruit samples that more fully reflect the diversity of African American women’s experiences across the lifespan. Given the association between alcohol use and/or abuse and the prevalence of STI/HIV-associated risk behaviors and adverse biological outcomes (i.e., STIs, including HIV) among African American females across the lifespan, there is a clear need to develop and evaluate prevention research efforts tailored for this subgroup.

INTRODUCTION

HIV is a threat to the health and well-being of many communities in the United States, but for African Americans, the situation is markedly worse. Despite only accounting for 12% of the US population, over half (57%) of HIV/AIDS cases diagnosed in 2009 were among African Americans. Specific to women, HIV was the leading cause of death among African American women 25-29 years of age and relative to Caucasian women, African American women were 23 times more likely to be diagnosed with AIDS [1]. Of African American women diagnosed with HIV during 2001-2004, 78% contracted infection via heterosexual contact [1]. Based on these observations, the CDC has termed the HIV epidemic a “health crisis” for African Americans, especially African American women, and has called for a heightened national response to this glaring racial disparity.

Given the lengthy period between HIV infection and symptoms, many women may have been exposed to HIV and contracted the virus as young adults or adolescents. Thus, understanding the factors that place African American females at increased risk for HIV across the lifespan is imperative. One prominent factor associated with sexual risk-taking and STIs/HIV across a variety of populations, including African American men and women, is alcohol use [2-6]. Alcohol use can interfere with cognitive processing of information (e.g., inhibitory control, cognitive escape) and decrease perceptions of risk, which can in turn influence sexual decision making and increase the likelihood of risky sexual behaviors [7, 8]. Although young African American women may have lower prevalence rates of alcohol use disorders or dependence compared to Caucasian women [8, 9], epidemiological findings suggest that the consequences of alcohol abuse are more severe among African American women [10].

Several reviews have been published on the association between alcohol use and sexual risk-taking across various populations including heterosexual adults, college students, people living with HIV, men who have sex with men (MSM), and adolescent samples [2-6]. However, it is less common for reviews to report on the interactions between gender, race and the alcohol-sexual risk-taking relationship. Given this gap in the literature coupled with the racial disparity in HIV rates in the US, the purpose of this paper is to systematically review and synthesize findings reported in the published literature between 2000 and 2011 on the association between alcohol use and sexual behavior and STIs/HIV among African American females in the United States. Furthermore, many reviews exclusively focus on a particular stage in the lifespan (i.e., adolescents, college age/young adults, adults), therefore limiting the examination of potential developmental patterns or changes in the way alcohol use interacts with sexual behavior across the lifespan. Thus, this review will include studies conducted with African American females either during adolescence, college-age/young adulthood, or adulthood in order to examine developmental patterns in both behaviors and their intersection across the lifespan. To our knowledge there are no longitudinal studies examining the association between alcohol use and sexual risk taking across the life course; thus, we organize study findings by these three developmental periods.

Following a summary of empirical findings, an integrative analysis of the reviewed research will be presented with the following aims: (1) to identify the degree to which theory guided the studies reviewed, (2) to describe the consistency of the existing empirical evidence within and across developmental stages, and (3) to evaluate the extent to which race and gender effects (both independent effects and interaction effects) were assessed in samples comprising mixed genders and races/ethnicities. Finally, conclusions drawn from this review will be used to suggest directions for future research designed to evaluate alcohol related sexual risk behaviors and biological outcomes (i.e., STIs/HIV) in African American females, as well as possible intervention opportunities to reduce sexual risk-taking among alcohol-using African American females.

LITERATURE SEARCH METHOD AND CRITERIA FOR INCLUSION

Database searches of PsycINFO, PsycARTICLES, PubMed, and Medline were conducted to identify published articles in peer-reviewed journals that tested the association between alcohol use and sexual risk behaviors and biological outcomes (STIs/HIV) among African American women. Combinations of the following search terms were used to identify relevant articles: alcohol, alcohol abuse, condom use, sexual risk behavior, sexual behavior, sex, sexually transmitted disease, sexually transmitted infection, HIV, African American, female, adolescent, adolescence, college, adult. A search of references cited in relevant studies was conducted to identify additional articles. English language articles published in peer-reviewed journals were screened for inclusion. Studies were included if they met all of the following criteria: (a) the study examined the association between alcohol and sexual behaviors or biological outcome (STI/HIV); and (b) the study’s sample was composed of at least 25% African American females. As this is a relatively understudied topic, we elected to set the inclusion rate at 25%. Including studies that at least ensured a quarter of their sample (typically this is enough to explore race/ethnic differences) would allow use to include some studies that did explore race/gender differences that directly related to the reviews goal. We also focused the review on the recent empirical literature; thus, only papers published between 2000 and 2011 were included in the review. We chose to focus only on recently published articles in order to provide the most relevant, current patterns of findings in our review. According to these criteria, 32 papers were included. We acknowledge that these search methods may have inadvertently excluded relevant publications. It should be noted that there was one paper that reported subgroup analyses from a larger trial that reported on distinct outcomes from the same sample. Thus, among the 32 papers meeting the inclusion criteria, there were 31 unique studies reviewed.

REVIEW OF STUDY FINDINGS AND CRITIQUE OF THE LITERATURE

This portion of the review provides a synthesis of findings regarding the association between alcohol use and sexual risk behaviors among African American women during different developmental periods across the lifespan. The summary is organized by the developmental period of study participants during adolescence, young adulthood/c ollege-aged, and adulthood. Articles in the adolescent age grouping were primarily conducted with school-aged populations, the young adults age group consisted of samples which enrolled primarily individuals who were recently post-high school yet still in their twenties. The adult group consisted of samples of post-high school individuals as well, but the range of individuals included in these samples exceeded 29 years of age. We also describe strengths and limitations of the literature for each of the three developmental periods. Table 1 describes key study characteristics and the reviewed studies’ methodologies. Table 2 reports the data analytic approach used and summarizes study findings.

Table 1.

Summary of Study Methodology: Sample Characteristics, Use of a Theoretical Framework, Study Design, and Core Alcohol and Sexual Behavior Measures

Study Citation	Age Group	Sample Size	% AA	% Female	Theory	Study Design	Alcohol Measures					Sexual Behavior Measures
							FA	BD	AS	AE	AB/AD	UVS	STI	NP	CU	CON	AI	SD
11	A	448	54	50		C	✓	✓	✓		✓		✓	✓	✓
12	A	153	66	37	✓	C	✓		✓			✓			✓
13	A	771	75	100	✓	C	✓
14	A	1,100	55	62		C	✓		✓			✓					✓
15	A	116	82	100		C			✓			✓
16	A	715	100	100	✓	C	✓					✓	✓	✓	✓
17	A	393	100	100	✓	L	✓		✓				✓	✓	✓
18	A	808	26	49		L	✓				✓	✓		✓
19	A	1,034	79	55		L	✓	✓	✓		✓			✓	✓
20	CE	298	35	66	✓	C	✓			✓					✓
21	CE	387	80	70		C		✓	✓					✓	✓
22	CE	848	100	100	✓	L	✓						✓	✓	✓
23	CE	434	33	52		C			✓					✓	✓
24	CE	1,494	43	53	✓	C			✓	✓					✓
25	CE	292	100	60	✓	C		✓				✓
26	AD	1,968	90	28		PC	✓	✓					✓	✓
27	AD	1,525	100	28		PC	✓	✓					✓	✓
28	AD	187	87	100		L	✓	✓						✓	✓
29	AD	685	85	28		C	✓		✓		✓	✓	✓		✓
30	AD	453	84	31		L	✓		✓	✓	✓	✓
31	AD	206	100	62		C		✓								✓
32	AD	973	68	48		C					✓			✓	✓	✓
33	AD	256	97	100		C	✓		✓						✓
34	AD	796	100	33		C	✓						✓
35	AD	1,079	100	38		C					✓		✓
36	AD	4,667	67	100		C	✓											✓
37	AD	495	100	50		C	✓					✓		✓
38	AD	780	100	45		C	✓					✓		✓
39	AD	635	100	100		C	✓	✓	✓			✓		✓				✓
40	AD	258	100	53	✓	C	✓							✓	✓			✓
41	AD	366	84	100		C					✓		✓
42	AD	1,070	100	50		C					✓		✓		✓	✓

Notes: AA: African American; A: Adolescent sample; CE: College-aged/early adult sample; AD: Adult sample; C: Cross-sectional; L: Longitudinal; PC: Prospective cohort; FA: Frequency of alcohol use; BD: Frequency of binge drinking; AS: Alcohol use prior to sex; AE: Alcohol expectancies; AB/AD: Measure of alcohol abuse or alcohol dependence; UVS: Frequency of unprotected vaginal sex; STI: STI diagnosis; NP: Number of sexual partners; CU: Condom use (either frequency or relative frequency of consistent condom use; proportion of condom protected encounters); CON: Concurrent sexual partners; AI: Age of sexual initiation; SD: Exchange sex for drugs.

Table 2.

Data Analytic Approach, Variables Controlled for in Adjusted Analyses, and Study Findings Across Reviewed Studies

Study Citation	Data Analytic Approach	Variables Controlled for in Adjusted Analyses	Results	Interpretation (Interaction)
11	CS, T, R	Age, race, gender and insurance status	In adjusted analyses, those with alcohol use disorder were more likely to have multiple partners (AOR = 2.29), be inconsistent condom users (AOR = 3.06), and have an STI (AOR = 1.69).	Substance use disorders are prevalent in STI clinic attendees, thus STI clinics may be a good venue for substance use screening and treatment for young people
12	T, R	Age, race, gender, future orientation, condom attitudes, AIDS knowledge; days of marijuana use	More marijuana use than alcohol; in adjusted analysis, alcohol use was not significantly associated with unprotected sex, unprotected sex while high/drunk; or condom use at last sex.	Marijuana was used more often than alcohol, both in general as well as during sexual intercourse. Marijuana and not alcohol use was associated with unprotected sex in this sample.
13	R	Age, household-type (single vs two-parent), race/ethnicity, school year of enrollment	Parenting practices influence daughters’ participation in alcohol and sexual behaviors by increasing daughter’s awareness of parental monitoring, rules, and parental disapproval of such behaviors.*	Findings suggest that both early alcohol use and early sexual initiation can potentially be addressed by enhancing positive parenting practices.
14	R	Gender, race, marijuana use	Early age at first drink was significantly, positively associated with early age at first sex (p < .001) after adjusting for age, gender, race and marijuana use, but the association was not as strong for African American youth	The strength of the association between alcohol use onset and sexual debut varied by race in this sample.
15	R	Relationship quality	Concordant alcohol users (self and partner both using alcohol prior to sex) were 5 times more likely to not use condoms during sex. Discordant partner alcohol users (only self or partner used alcohol; not both) were 3 times more likely to be non-condom users.	Both individual and partner alcohol use should be considered when assessing sexual risk- taking behaviors in adolescents.
16	CS, T, R	Fear of condom negotiation, sexual sensation seeking, peer norms supportive of risk, depressive symptoms	In adjusted analyses, higher frequency alcohol use was associated with number of oral sex partners (past 60 days), and with frequency of oral and vaginal sex (past 60 days), and more drinks per occasion was associated with presence of semen in vaginal fluid (i.e., marker of unprotected sex; AOR = 1.85).	Alcohol use (frequency and amount) was related to increased sexual risk-taking in this sample. Thus, HIV prevention programs should incorporated messages about the intersection xsbetween alcohol and sexual risk.
17	BV, GEE	Age, intervention condition, baseline values for outcome variables	In adjusted analyses, high quantity drinkers at baseline were more likely to test positive for trichomonas, report inconsistent condom use, have sex while high or drunk, have multiple sex partners, and engage in anal sex over the 12-month follow-up.	High quantity alcohol use is predictive of various high risk sexual behaviors, and STI acquistion, in this sample. HIV prevention programs should address the association between alcohol and sexual risk among African American young women.
18	LGC	Gender, income background	Early alcohol use predicted higher level of adolescent drinking, which then predicted likelihood of young adult AUD and risky sexual behavior; this is especially true for youth from low- income backgrounds	Early intervention, especially for disadvantaged youth, may help decrease AUD and risky sexual behaviors (and adverse outcomes of risky sex) in adulthood
19	R	Age, gender, ethnicity, sexual initiation in 7th grade, living arrangements, and perceived maturity	Early alcohol use (reported use in 7th grade) predicted subsequent (10th grade) alcohol problems, unprotected sex, multiple sex partners, being drunk or high during sex, and pregnancy. Also, for females only, early alcohol use predicted sexual debut and recent sex at follow-up.	Prevention programs for urban youth should address both alcohol use, sexual risk, and their intersection.
20	R	Age, gender, social desirability, sensation seeking, assertiveness, alcohol (expectancies, consumption), outcome expectancies, self-efficacy, condom use while sober	Condom use when sober predicted by: social assertiveness, alcohol expectancies (negative association), self-efficacy condom use when intoxicated, positive outcome expectancies of condom use. Condom use when intoxicated predicted by: condom use when sober, alcohol expectancies regarding sexual risk taking, self- efficacy regarding condom use when intoxicated	Condom use with casual partners associated with social assertiveness, alcohol expectancies regarding sexual risk-taking, self-efficacy, and positive outcome expectancies regarding condom use
21	CS, R	Gender, sexual orientation	No differences in # partners during the past 3 months based on alcohol use; At least 1 drink in past 30 days less likely to report not using condom at last sex (OR = 1.74); No association found alcohol/drug use at last sex and condom use at last sex.	No association between alcohol/drug use at last sex and condom use at last sex; Single correlation between any alcohol use in past month and condom use at last sex.
22	BV, GEE	Illicit drug use, intervention group, baseline outcome measures	Women who consumed alcohol were more likely to have multiple partners and risky partners. Alcohol consumption was associated with testing positive for an STI and never using a condom with a casual partner during the past 12 months.	Alcohol use was associated with greater risk behavior engagement and greater likelihood of being diagnosed with an STI.
23	R	Gender, ethnicity, ethnic identity	Among females, no association between ethnicity, ethnic identity and use of alcohol/drugs before or during sex, number of partners, condom use	Among college-aged women, this study suggests that ethnic identity is not a significant predictor of alcohol related sexual behaviors or other sexual risk behaviors
24	R	Gender, amount of alcohol drank before sex, inhibition conflict, alcohol expectancies	Males reported higher levels of condom use than females at last sex; Conflict regarding condom use: negative predictor of condom use for 1st sex with most recent partner; alcohol sexual risk taking expectancy negatively associated with condom use at all 3 occasions; Conflict x expectancy interaction for 1st sex with most recent partner: strong negative relationship between expectancy and condom use among those low in conflict; conflict, expectancy, alcohol interaction predicted condom use at 1st sex and last sex	Alcohol use prior to sexual intercourse led to decreased condom use only among individuals who experienced conflict regarding condom use.
25	R	Gender, current employment, left family home, perceptions of relationship monogamy	No significant gender effect for women in the association between alcohol use and unprotected sex	No association between alcohol use and unprotected sex was identified among rural, Southern African American young adults.
26	R	Age, race, gender, MSM, # injections/day, cocaine use, education, time-on-study and drinks per week reported at the initial visit	Higher alcohol use associated with (past−6-month): non-injection cocaine use, injection drug use, having one or more sex partners 1.08 (1.04, 1.12), MSM, and female-gender 0.87 (0.78, 0.98). Associations were similar for drinks per week and binge drinking, except that higher education was associated with lower odds of binge drinking.	After adjustment for alcohol consumption at study entry and other characteristics, time fixed and time varying participant characteristics were associated with subsequent alcohol consumption. Older age and female gender were both associated with lower alcohol consumption (having one or more sexual partners was independently associated with # drinks per week and binge drinking).
27	MSM	Age at entry, sex, years of education, time-varying drug use, number of sexual partners, and STIs	Hazard ratio (for acquiring HIV) appears to nonlinearly increase with the number of drinks per week. For 1−5, 6−20, 21−50, and 51−140 drinks-per-week respective HRs were: 1.09 (0.6−−1.98), 1.18 (0.66−−2.09), 1.66(.94−−2.93), 2.12 (1.15−−3.90).	Hazard ratio appears to nonlinearly increase with the number of drinks per week. For 1−5, 6−20, 21−50, and 51−140 drinks-per-week respective HRs were: 1.09 (0.6−−1.98), 1.18 (0.66−−2.09), 1.66(.94−−2.93), 2.12 (1.15−−3.90). The independent association between # drinks per-week in the prior 2-years HIV acquisition was muted in the presence of binge drinking.
28	GEE	(a) Controlling for current IDU or non-IDU status, HIV partner status, male condom as method of birth control, and current mental health diagnosis; (b) Controlling for age, immunosuppression, current IDU or non-IDU status, HIV partner status, male condom as method of birth control, and having more than one sex partner; (c) Controlling for current IDU or non-IDU status and time known HIV positive	Binge alcohol use, in adjusted analysis, associated with (a) no condom use at last vaginal sex encounter 1.51 (1.03, 3.62); (b) inconsistent condom use with partner for vaginal sex 2.60 (1.04, 6.54); and (c) multiple male sex partners 4.76 (1.12, 20.23).	In this study we found a strong relationship between sexual behavior and alcohol consumption patterns. Binge alcohol consumption and ART status were both strongly and positively associated with unsafe sexual practices among this sample of HIV-infected women, lending strength to the equivocal literature on the relationship between both factors and sexual behavior. Furthermore, substantial proportions of women who did not use a condom for last vaginal sex indicated that the partner refused or did not want to use one (39%), and this refusal was significantly associated with her binge drinking pattern—which suggests the importance of binge drinking in the potential negotiation of condom use.
29	R	Where Sexual-Compulsivity was DV: substance use (past 3 months); substance use in sexual situations (past month); DAST scores; alcohol outcome expectancies. Where sexual risk behaviors were DV: age, gender; Sexual-Compulsivity subscales; DAST scores; drug use and alcohol use proximal to sex;	Persons with higher sexual compulsivity scores scored higher on the alcohol (AUDIT). AUDIT was not associated with # of sex partners or unprotected sex with casual partners in adjusted analysis.	Although sexual compulsivity scores were associated with alcohol and other drug use, associations between sexual compulsivity and sexual risks were not accounted for by substance abuse. Gender differences on the composites resulting from the factor analysis suggest that men experience greater personal discomfort related to sexual compulsivity than do women.
30	PA		Alcohol variables were not associated with unprotected sex at the 6-month follow-up among women. For women, sensation seeking was related to alcohol outcome expectancies and AUDIT-QF scores.	The patterns of bivariate correlations among sensation seeking and substance use variables were similar for men and women; however, sensation seeking was not associated with unprotected sex at the 6-month follow-up for women. In fact, for women in the current study, none of the baseline measures were correlated with rates of unprotected sex at the 6-month follow-up.
31	R		Observed bivariate association of binge alcohol use and concurrency POR=2.9 (1.5, 5.2) apparently explained by other variables, and eliminated from multiple logistic model.	The multiple logistic model identified 4 variables as able to account for variation in concurrency prevalence: male sex (odds ratio: 2.5; 95% CI: 1.2–5.2), age (1.4 per 5-year decrease in age; 95% CI: 1.2–1.6), history of smoking crack (2.5; 95% CI: 1.2–5.1), and incarceration of any of the respondent’s last 3 partners (2.7; 95% CI: 1.3–5.6).
32	R	Gender, partner cohabitation, income, DAST score, is partner monogamous (Y/N), lifetime # of partners	In adjusted analyses: Steady/non-steady concurrency was associated with increased AUDIT score 1.04 (1.01 to 1.06); and among patients reporting concurrency, consistent condom use with non-steady partners was more likely among individuals who used less alcohol 0.96 (0.94 to 0.99).	No female-specific results
33	ANOVA, R	Age, partner type at last sex, perceived partner concurrency, knowledge of partner’s HIV status, duration of partnership	Women were significantly more likely to use condoms at last sexual encounter if the encounter was with a casual partner 6.00 (2.80, 12.82); and if alcohol and/or drugs were not used 2.06 (1.06, 4.02).	Alcohol combined w/ other substance use in logistic regression.
34	ANOVA, R		Males initiated alcohol use earlier than females did, and respondents who reported having STIs in the past initiated alcohol use earlier than respondents who reported no STIs.	Both groups initiated most drug use at a similar age except for alcohol: the STI group beginning their alcohol use, on average, one year younger.
35	R	Demographics (gender, education, employment), Psychopathology (depression, antisocial personality, PTSD, sexual dysfunction, other drug use), Sex-related behaviors (STI history, sexual activity, and sex/drug mixing)	Older males were more likely than older females to perceive a greater risk for HIV/AIDS [n.s. in the younger group]. Alcohol abuse was associated 1.96 (1.05–3.64) with older users’ heightened risk perception for HIV/AIDS, [but not alcohol dependence]. In the younger group alcohol dependence was associated 1.91 (1.36–2.68) with heightened risk perception, [but not alcohol abuse.]	Logistic DV is perception of HIV risk. Alcohol dependence was significant among younger group; alcohol abuse was significant among older group.
36	BV, R	Average # sex partners in past 30 days, STIs; frequency of unprotected vaginal/oral sex	A greater proportion of women who exchanged sex for drugs or money used alcohol; mean age for initiating alcohol use (15) was significantly lower among those exchanging sex; women who exchanged sex used alcohol more days in the past month (M_15 days) than women who did not exchange (M_12 days, t (4196) _8) and used alcohol more often during sex.	Sex exchange DV in logistic regression, alcohol not included as predictor variable
37	R		Frequent alcohol and frequent crack use were each associated with almost two additional sex acts with no condom; Frequent alcohol users averaged 2.8 more sex acts than did less than frequent users (x2_16.64, df_1, P_0.0000), and females averaged 1.8 more sex acts than males (x2_6.72, df_1, P_0.0096). Having two or more sex partners was associated with both alcohol and crack use and their interaction (P_0.0255, 0.0063, and 0.0943, respectively).	In this sample, those reporting frequent use of both alcohol and crack were also more likely than those in the other three groups to report alcohol- and crack-impaired sex, two or more sexual partners, greater numbers of sexual acts, and greater numbers of sexual acts without condoms. Those reporting frequent use of both substances had higher risk compared to those reporting frequent use of alcohol or crack only.
38	CS		Female cocaine users showed the same (substantially heightened) sexual risk as male users. Substantially more marijuana users had multiple sex partners than alcohol users.	M/F gender comparisons say little about Female specific risk. The alcohol use only group is likely consuming less alcohol than the crack/cocaine group (#drinks not assessed).
39	R	Homeless, anxiety subscale, Conflict-Tactics subscale, physically abused before age 18, traded sex for drugs in prior 30 days, used drugs/alcohol before age 15, used alcohol to come down from crack, drank alcohol before or during sex, length of average crack run > 24 hr, >10 hits of crack on average day	"Heavy" alcohol users consumed more crack and were more likely to engage in sexual risk behaviors. In bivariate analysis: >1 sexual partner was significant. In multivariate regression: Trading sex for drugs in past 30-days was significant	Findings suggest that simple quantity and frequency questions may be useful for categorizing alcohol use and for identifying women at high-risk of a variety of other problems. For example, those categorized as heavy drinkers were most likely to have histories of childhood trauma and suffer from elevated levels of psychological distress.
40	CA	Sexual context of last encounter, drug use behavior, sociodemographic characteristics, motivation for condom-use, HIV- concealment scale, psychological functioning	The cluster analysis produced three distinct groups based on five HIV drug use and sexual risk behaviors; Alcohol use associated with membership in high-risk groups;	Compared to the other two risk groups [Inconsistent condom-use group, and consistent condom use group], the highest risk group drank alcohol more often, on average almost twice a day.
41	R	Age, and illicit substance use	The prevalence of high scores on the CAGE was 54.5%; 15% of women tested positive for Trichomonas vaginalis. Multivariate logistic regression analyses indicated women with higher CAGE score were more likely to test positive for T vaginalis 2.54 (1.03 to 6.30); have sex with their spouse or steady partner when only they had been drinking 3.44 (1.54 to 7.72) and have sex with their spouse or steady partner when they had both been drinking 3.85 (1.86 to 7.96).	Causality cannot be assessed. However, relatively low CAGE scores were associated with potentially risky sexual behavior and biologically confirmed T vaginalis.
42	R	Other substance use, and Drug Dependency: Texas Christian Drug Screen II (TCUDS)	Couples where the female partner (only) scored positive for alcohol dependency were more likely to report concurrent sexual partners than couples where neither partner scored positive for alcohol dependence (OR = 2.53; 95% CI: 1.27, 5.06).	Contrary to findings from some previous studies [4, 6, 7], alcohol dependency by one or both partners was not associated with testing positive for STIs or self-reported sexual risk indicators except that female alcohol dependency increased the likelihood of concurrent sexual partners.

Notes: CS: Chi-square, T: T-tests, R: Regression; LGC: Latent growth curve modeling; BV: Bivariate analyses; GEE: General equation estimating; ANOVA: Analysis of variance; MSM: Marginal structural models; CA: Cluster analysis;

^*: the analysis did not directly associate sex and alcohol use but the authors discuss the connection in the discussion.

Adolescent Studies

Nine studies were identified that explored the association between alcohol use and sexual-risk taking among African American adolescent female adolescents (i.e., mean age of sample was under age 18 years). The adolescent-specific studies were typically cross-sectional in design (n= 6) [11-16] and the rationale for exploring the association between alcohol and sexual behavior was theoretically motivated in less than half of the studies (n =4) [12, 13, 16, 17]. Sample sizes ranged from 116 to over 1,000 participants. The majority of samples (n = 5) [11, 14-17] were recruited from clinic settings (e.g., STI clinics, health clinics, or emergency rooms), and 4 studies recruited all female samples [13, 15-17]; two of which were all African American female samples [16, 17]. Among the adolescent studies, alcohol use disorders or problem drinking were not commonly assessed. Rather, frequency of drinking was the most cited measure of alcohol use employed in the reviewed studies. In regards to sexual behaviors, common sexual behavior measure reported were episodes of unprotected sex (n = 5), condom use (n = 5), number of partners (n = 5), sex after drinking (n = 4), and biological outcomes of risky sex (e.g., STI, pregnancy) were included in 4 studies.

Overall, 7 of the 9 adolescent-specific studies reviewed found significant positive associations between alcohol use and sexual risk behaviors [11, 14-19]. Specific to the 4 studies conducted with exclusively female samples (which were predominately or exclusively African-American samples as well), 3 reported significant positive associations between alcohol use and sexual behavior, as well as association with biological outcomes [15-17]. In the only study that compared racial and gender patterns as they related to the association between alcohol and sexual behavior, they found that the strength of the association between alcohol use onset and sexual debut varied by race, and this was especially true for females (African American verses White girls; with a stronger association observed among White girls) [14]. As a whole, the results reported across the 9 adolescent studies suggest that clinics, especially STI clinics, might be good venues for screening for alcohol use problems among adolescents. Thus, among sexually active African American teens, HIV prevention programs should include education on alcohol use and the intersection between alcohol use and sexual risk-taking. Further, early intervention, especially for youth from disadvantaged backgrounds, may help decrease future alcohol use disorders and risky sexual behaviors. Finally, especially for adolescent females, understanding and assessing partners’ drinking is important when exploring adolescent girls’ sexual behaviors, particularly because men have primary control over condom use.

Several strengths exist among the adolescent studies reviewed. Foremost, there was remarkable consistency in the association between alcohol use and sexual behaviors, despite different measures of both constructs across studies. Additionally, 4 of the 9 studies reviewed enrolled exclusively female adolescent samples, which were either predominately or exclusively African American in composition [13, 15-17]. Three studies found the same pattern of associations as the other 5 mixed race/gender studies both cross-sectionally and longitudinally.

In addition to the strengths, several weaknesses were apparent across the 9 adolescent-specific studies. First, the samples that enrolled sufficiently sized percentages of African-American and female participants tended to be recruited from specialized venues providing services to “high risk” individuals. This limits generalizability of the findings to all adolescent African American females. Next, among studies that recruited multi-racial and mixed-gender samples, very few explicitly examined the association between alcohol use and sexual behavior by gender or race, or explored the interaction effect (gender x race) on the association. Rather, in these studies race and gender were often treated as control variables in the analysis. Finally, alcohol use and sexual behaviors were assessed used in a variety of different measures, and less than half of the studies posited theory or proposed mechanisms to explain their observed associations between alcohol use and sexual behavior among youth.

College-Aged and Young Adult Studies

Six cross-sectional studies examined the association between alcohol use and sexual risk behaviors among college-aged or young adult African American women [20-25]. The majority (n = 4) utilized a theoretical framework (e.g., Health Belief Model, Social Cognitive Theory) to guide the study [20, 22, 24, 25]. Sample sizes ranged from 30 to 1,494 participants with a median sample size of 300 participants. Two studies were conducted with only African-American participants [22, 25]; African-Americans represented between 33% and 80% of the samples in the other four studies [20, 21, 23, 24]. Only one study was conducted with an all female sample [22]; the remaining four studies included both males and females [20, 21, 23-25]. Recruitment took place in college campuses (n = 3) [20, 21, 23], by a respondent driven sampling approach (n = 1) [25], by random digit dialing (n = 1) [24], and by telephone from a database of medical center patients (n = 1) [22]. Frequency and quantity of alcohol use during a specific time frame (e.g., during the past month) were the most common alcohol measures; alcohol use disorders or problem drinking were not commonly assessed. To measure sexual behaviors, all studies assessed condom use (n = 6) [20-25], while three studies assessed number of sexual partners [21-23]; other sexual behavior measures included age of sexual debut, frequency of sex, and partner type.

Findings were mixed regarding the association between alcohol use and sexual risk behaviors among college-aged young women. In the four studies that examined the alcohol risky-sex relationship directly, two found no association between alcohol use and condom use and one study found a global association between increased alcohol use and decreased condom use [21-23, 25]. A third study found that individuals who drank alcohol during the past 30 days were more likely to report not using a condom at last sex [21]. Lastly, a fourth study found that alcohol use during the past month was associated with having more sexual partners and riskier partners and past alcohol use was correlated with increased likelihood of testing positive for an STI and not using condoms with a casual sexual partner during the previous 12 months [22]. Two studies examined mediators of the link between alcohol use and sexual risk behaviors and identified social assertiveness, outcome expectancies, and conflict surrounding condom use as potential mediators [20, 24]. No studies directly examined the role of race as it relates to alcohol use and sexual risk behaviors. Thus, the association between alcohol use and sexual risk behaviors among African American college-aged individuals still remains unclear. In addition, there is a need to examine drinking and sexual behavior dynamically (e.g., when the partner may also be drinking) and the context in which the drinking occurs. However, this research does highlight potential mediating factors in the association between drinking and sexual risk behaviors. Thus, interventions that increase young women’s ability to navigate conflict surrounding condom use, stressing assertive communication and fostering positive outcome expectancies regarding condom use may foster sexual risk reduction behavior change.

In the college-aged and young adult literature, there were several strengths of the reviewed studies. First, studies typically recruited relatively large samples. A few of the studies also examined important mediating variables (e.g., alcohol expectancies) in the association between alcohol use and sexual risk behaviors. In addition, multivariate analyses to examine potential mediating factors in the alcohol-sexual risk association were sound and included relevant control variables.

There were a number of limitations within this literature. Across the reviewed studies there were inconsistent findings in the association between alcohol and sexual risk behaviors. The inconsistent findings may be the result of measurement issues. For example, the measurement of alcohol use was often crude (e.g., use of any alcohol during the past month). Additionally, few studies examined the proximal effects of alcohol use on sexual behaviors. Most studies did not recruit diverse participant samples and relatively few studies explored the association between alcohol and sexual risk behaviors by race, gender, or examined potential interaction effects between race and gender. Instead, race and gender were treated as control variables. Within this developmental period, there was also a lack of studies with non-college attending African American young adults. Lastly, while the use of theory was often discussed in the introduction, studies typically included a limited subset of constructs from the theory and the theoretical frameworks did not guide the statistical analyses conducted.

Adult Studies

Seventeen studies were identified that explored the association between alcohol use and sexual-risk among African American adult females. Two studies reported on the AIDS Link to Intravenous Experience (ALIVE) cohort (a 20-year prospective study [26, 27]; two included short-term follow-up [28, 29]; and the remainder (n=13) were cross-sectional in design [3, 30-41]. The rationale for exploring the association between alcohol and sexual-risk was linked to the disinhibiting effects of alcohol or its association with childhood trauma and other stressful life events; however few (n = 3) referenced particular theories (e.g., Alcohol Expectancy Theory, Alcohol Myopia Theory) [29, 30, 32] or applied theoretical frameworks when selecting measures [29, 30].

Studies used purposive methods to compose samples of crack and injection-drug users [26, 27, 34-39]; persons with concurrent sexual relationships [33]; persons with HIV/AIDS [28, 31, 40-42]; and persons seeking STI prevention or treatment services [29, 30, 32]. Sample sizes ranged from 187 to 4,667 with most (n=9) including 500-2000 participants [26, 27, 30, 32, 34, 35, 38, 39, 42]. Samples were recruited from a state health department; STI clinics; through participation in risk-reduction RCTs; and using respondent driven sampling or purposive snowball methodologies. All samples were either 100% (n=9) or predominately African American. Five studies recruited only women; the remainder included 28% to 62% women.

Among the adult studies, problem drinking or alcohol use disorders were assessed using the Alcohol Use Disorders Identification Test (AUDIT; n=2); Cutting down, Annoyed by criticism, Guilty, Eye opener scores (CAGE; n=2); and the NIMH’s Diagnostic Interview Schedule for alcohol abuse and alcohol dependence (n=1). Five studies used standardized binge drinking measures. Prevalence and frequency measures were applied to varying timeframes and used to create continuous variables (e.g., the product of typical drinking days per week and typical drinks per drinking day), ordinal variables (e.g., low to high frequency groups), nominal classifications (e.g., “heavy drinker”) and drug/alcohol combined groups. Studies also assessed alcohol use or intoxication proximal to sex (n=6); alcohol use expectancies (n=2) and age of initiation (n=2).

Among the adult studies five used biologically confirmed HIV/STI as the primary outcome, or a risk-indicator. Commonly assessed sexual-risk behaviors included condom use (n=8) and number of sexual partners or partner concurrency (n=13) [33]. Other sexual-risk variables included: exchange of sex for money and or drugs; sex with injection drug user (IDU); sex when partner was drinking; cohabitation; and sexual dysfunction. Additionally, studies assessed perceptions of HIV-risk and partner monogamy; knowledge of HIV and partner’s HIV-status; and age differences among couples. Personality dispositions oriented toward stimulation, sexual preoccupation and poor impulse control were also examined as predictors of alcohol use and sexual risk behavior [29, 30].

Results from adult studies revealed multiple alcohol-behavior associations across high-risk sub-groups. Among crack and IDUs research indicated that, controlling for gender, initiating alcohol use at a younger age was associated with history of STIs [34] and commercial sex work (CSW) [36]. However, the CSW-alcohol relationship was inconsistent. CSWs were found to use alcohol more days in the past month and more often during sex in a mixed-race sample [36]; but among African American women, CSW was associated with decreased odds of being a moderate or heavy drinker [39]. Other studies indicated that frequent alcohol use was associated with increased number of sex-acts and sex-partners [26, 37] and inconsistent condom use [37, 38, 40]. Although research among drug users seems to show that the alcohol-drug interaction is an important HIV-risk indicator [36-39]; the independent effects of alcohol are difficult to parse. Evidence from the ALIVE cohort is perhaps the strongest support for an independent relationship between alcohol and HIV-risk among IDUs [26, 27] and in standard adjusted analysis having one or more sex partner was the only significant measured sexual-risk variable [26].

In other studies exploring alcohol use and sexual-risk among people living with HIV/AIDS and other high-risk populations, there were also alcohol-behavior associations [28, 31, 42]. Among women in these studies multiple measures of alcohol abuse (binge-drinking, AUDIT, and lifetime dependency) were associated with concurrency [31, 32, 42]. Binge-drinking and AUDIT were also each associated with inconsistent condom use [28, 32]; and alcohol consumption at non-abuse levels (CAGE score >1) was associated with biologically confirmed STI and other sexual-risk indicators [28, 40, 41].

Finally, among studies examining mechanisms including personality dispositions, sexual preoccupation and poor impulse control, “sensation seeking” and “sexual compulsivity” were each correlated with alcohol abuse, alcohol-expectancies, and use of alcohol in sexual contexts. However, significant gender differences were evident —indicating that these factors are less salient in women [29, 30]. Among women in the prospective study no baseline measures were correlated with rates of unprotected sex at follow-up [29].

Several strengths exist among the adult studies reviewed. Foremost, Howe and colleagues provide the strongest evidence to-date for a dose-response relationship between alcohol and HIV acquisition –reporting an apparent nonlinear increase in the HIV-acquisition hazard ratio with drinks-per-week among African American IDUs. The research design of the ALIVE study and weighted analysis enabled examination of the alcohol-HIV relationship with the presumption that sexual activity and drug use are time varying covariates, affected by prior alcohol consumption. Howe et al. concluded that standard adjustment for time-updated sexual activity and drug use may underestimate alcohol’s HIV-risk contribution. Other studies tested underlying pathways and mechanisms in the alcohol-HIV relationship, with mixed results. Alcohol consumption at abuse and non-abuse levels were each associated with sexual behavior among African American women. Studies consistently showed associations between alcohol use and condom use, and concurrency or number of sexual partners.

Another strength of the reviewed studies was the enrollment of large numbers of high-risk African Americans –several samples were either predominately or exclusively African American women. The reviewed studies indicate that although alcohol use may be less prevalent among African American women, but they may be more sensitive to alcohol’s effects within the contexts of sero-discordant relationships, CSW situations, and other high-risk environments.

In addition to the strengths, several weaknesses were apparent across the 17 adult-specific studies. First, the findings have very limited generalizability to all African American women, given that about half of the studies purposively sampled illicit drug-users. Moreover, the measurement of alcohol and sexual-behavior variables across studies was inconsistent and within-study comparisons by race and gender were limited. Finally, several studies discussed the potentially time varying and bidirectional effects of alcohol on HIV-risk behaviors but did not reference particular theories. Advances in theory could guide a necessary disaggregation of distal, proximal, and time-varying effects of alcohol on both interpersonal and intrapersonal outcomes.

FINAL REVIEW CONCLUSIONS AND AGENDA FOR FUTURE STUDIES

African Americans experience elevated rates of sexually transmitted infections and HIV [43, 44]. One factor that may result in increased sexual risk taking is the use or abuse of alcohol. To examine the association between alcohol use and sexual risk behaviors among African American females, we synthesized findings from 32 studies stratified across three distinct developmental periods. The bulk of the reviewed studies were cross-sectional (n = 25) in design and examined the global association between overall patterns of alcohol use and high-risk sexual practices. Similar to previous global association studies [45], results suggest that increased use of alcohol is associated with increased sexual risk practices among African American females. However, few studies examined the proximal effects of alcohol use on sexual risk practices. Future studies should employ study methodologies that can better examine the alcohol-risky sex link; such approaches could include the use of daily diary reporting or event-level analyses. Related to this, research utilizing better measurement approaches (for both sexual behavior and alcohol use) are needed. In particular, distinguishing between occasional alcohol use (e.g., once per month) relative to problem drinking or alcohol abuse is needed for this population. Additionally, few studies employed a theoretical framework to guide study design and hypotheses. As such, important constructs that may either mediate or moderate the association between alcohol and sexual behaviors (e.g., alcohol expectancies, impulsivity, abuse) were not included in the analyses. Similarly, few studies examined contextual and relationship factors that may influence this association. For example, the type of partner has been shown to influence the degree to which alcohol is associated with risky sex [46, 47]. To advance this literature, future research should utilize a theoretical framework to select measures and guide study analyses. Such an approach would also facilitate an examination of the extent to which the association between alcohol use and sexual risk behavior and underlying mechanisms differs based on race and related constructs (e.g., experiences of racism, cultural norms related to alcohol use and sexual behaviors).

Across studies the measurement of alcohol use and sexual risk behaviors typically focused on general frequency measures. When measuring alcohol use, a paucity of studies assessed problem drinking, alcohol abuse, and/or alcohol dependence. Instead, measurement typically focused on the frequency and/or quantity of alcohol consumed during a particular time frame (e.g., number of drinking occasions in the past month). Additionally, none of the reviewed studies included non self-reported measures of alcohol use (e.g., Blood Alcohol Concentration). In future studies, alcohol measurement should be more precise, with an emphasis on alcohol use consumed in relationship to specific sexual encounters. Measurement of sexual risk practices typically focused on overall frequency of condom use or number of sexual partners across a discrete time period. Surprisingly, only a handful of studies examined sexual behaviors subsequent to when alcohol was consumed by the respondent and/or her partner. In addition, a minority of studies included biological measures (e.g., STI testing) to assess sexual risk behaviors. Moving forward, studies should move beyond measuring the general association between alcohol use and sexual risk behaviors.

Since the relationship between alcohol use and sexual behaviors may vary across distinct periods of life across the lifespan, we examined this association during adolescence, college-aged/early adulthood, and adulthood. The majority of the studies examining the association between alcohol and sexual risk-taking among African American females were conducted with especially high-risk samples (e.g., STI clinic attendees, IV drug users) or convenience samples (e.g., college students). The adolescent and adult literatures tended to conduct their studies more often with high-risk samples, whereas the college/young adult studies relied more on convenience samples. Thus, findings from all age groups reviewed may not be representative of and therefore may not generalize to the larger population of African American females at each developmental period. Future studies should seek to recruit samples that more fully reflect the diversity of African American women’s experiences across the lifespan.

There is also a clear need for longitudinal studies to better examine the association between alcohol use and sexual behaviors across the lifespan. In addition, among studies that recruited multi-racial and mixed-gender samples, few examined the association between alcohol and sexual behavior by gender or race, or explored the interaction effect on the association. For studies with diverse participants it will be important to examine the role of gender and race explicitly rather than treating them as control variables in analyses.

IMPLICATIONS FOR STI/HIV PREVENTION INTERVENTIONS

Given the association between alcohol use and/or abuse and the prevalence of STI/HIV-associated risk behaviors and adverse biological outcomes (i.e., STIs, including HIV) among African American females across the lifespan, there is a clear need to develop and evaluate the efficacy of prevention research efforts tailored for this subgroup. The findings from this review suggest that even non-abuse levels of drinking among African American females, at all ages, are related to increased sexual risk-taking, thus, incorporating messages about the intersection of alcohol use and sexual decision making into STI/HIV prevention programs for African American women could prove an effective strategy for reducing STIs in this vulnerable population.

Many efficacious STI/HIV risk-reduction programs have been developed and evaluated across a variety of venues, including schools, community centers, clinics and specialized locations such as detention facilities, yet few have been designed specifically for alcohol-using or abusing African American females. Many existing efficacious STI/HIV prevention programs could be programmatically and methodologically enhanced and translated to alcohol-using subgroups of women. For example, the SiHLE program designed for female African American teens [48], could be modified to include information specifically relevant for alcohol using girls and their risk behaviors and then employed with alcohol-using African American adolescent females.

Optimizing STI/HIV prevention efforts in the future will require prioritizing the development and evaluation of innovative, theory-based, empirically-derived and rigorously designed research specially tailored to alcohol-using African American women. This includes a focus, not only alcohol use, but on women’s ethnicity, culture, developmental stage, and other relevant socio-demographic characteristics. Future STI/HIV prevention research for alcohol-using African American females, at various stages in the lifespan, will need to emphasize commonalities between this population and other populations, as well as identify differences between populations, differences that could be useful in designing and tailoring of STI/HIV prevention programs specifically to address the needs of alcohol-using African American females.

Future Research Questions.

• Similar to previous global association studies, results suggest that increased use of alcohol is associated with increased sexual risk practices among African American females. However, few studies examined the proximal effects of alcohol use on sexual risk practices. Future studies should employ study methodologies that can better examine the alcohol-risky sex link; such approaches could include the use of daily diary reporting or event-level analyses.

• The majority of the studies examining the association between alcohol and sexual risk-taking among African American females were conducted with especially high-risk samples or convenience samples. Thus, findings from all age groups reviewed may not be representative of and therefore may not generalize to the larger population of African American females across the lifespan. Future studies should seek to recruit samples that more fully reflect the diversity of African American women's experiences across the lifespan.

• Given the association between alcohol use and/or abuse and the prevalence of STD/HIV-associated risk behaviors and adverse biological outcomes (i.e., STDs, including HIV) among African American females across the lifespan, there is a clear need to scale up prevention research efforts tailored for this subgroup.

Key Learning Objectives.

• African Americans experience elevated rates of sexually transmitted diseases and HIV. One factor that may result in increased sexual risk taking is the use or abuse of alcohol. Similar to previous association studies, results suggest that increased use of alcohol is associated with increased sexual risk practices among African American females across the lifespan.

• Among studies that recruited multi-racial and mixed-gender samples, few examined the association between alcohol and sexual behavior by gender or race, or explored the interaction effect on the association. For studies with diverse participants it will be important to examine the role of gender and race explicitly rather than treating them as control variables in analyses.

• The findings from this review suggest that even non-abuse levels of drinking among African American females, at all ages, are related to increased sexual risk-taking, thus, incorporating messages about the intersection of alcohol use and sexual decision making into STD/HIV prevention programs for African American women could prove an effective strategy for reducing STDs in this vulnerable population.

• The majority of the studies examining the association between alcohol and sexual risk-taking among African American females were conducted with especially high-risk samples (e.g., STI clinic attendees, IV drug users) or convenience samples (e.g., college students). The adolescent and adult literatures tended to conduct their studies more often with high-risk samples, whereas the college/young adult studies relied more on convenience samples. Thus, findings from all age groups reviewed may not be representative of and therefore may not generalize to the larger population of African American females at each developmental period.