Throughout much of the United States, rates of heterosexually transmitted HIV infection are dramatically higher among African Americans than whites; 74% of new heterosexually transmitted HIV cases in 29 states from 1999-2002 occurred among non-Hispanic blacks,1 although African Americans account for approximately 12% of the population.2 Women are at more risk for heterosexual HIV transmission than men, and black women accounted for almost twice as many of these cases as black men. 1 Racial disparities in heterosexual HIV transmission are especially pronounced in the rural South.3,4 Reasons for African Americans increased rates of heterosexually transmitted HIV infection have remained elusive.
Studies of heterosexual HIV transmission in the United States, typically conducted among higher-risk populations, have identified associations between HIV infection and drug use and risky sexual behaviors.5-10 Few studies, however, have evaluated risk factors for heterosexual transmission among African Americans in depth or examined HIV infection's relationship with socioeconomic factors and sexual network patterns. Moreover, little information is available concerning risk factors for HIV infection among HIV-positive African Americans without high-risk behaviors.
We conducted a population-based case-control study to characterize the epidemic of heterosexually transmitted HIV infection among African Americans in North Carolina, a region with high rates of heterosexual HIV transmission and other sexually transmitted infections. In particular we sought to characterize HIV-positive individuals without high-risk behaviors.
Study participants were African American men and women between the ages of 18-61 years who resided in selected North Carolina counties. The original study area comprised 13 contiguous rural counties with elevated rates of heterosexually acquired HIV infection but, to increase enrollment, subsequently expanded to include almost all of central and eastern North Carolina; much of this region is rural. Enrollment took place during January 1997-March 2000.
Case subjects were individuals whose HIV infection had been reported to the North Carolina HIV/Sexually Transmitted Diseases Prevention and Care Section within the 6 months before their referral to the study and who denied male same-sex activity and injection drug use. After notification of their HIV diagnosis and receipt of post-test counseling from their care provider, they received additional counseling by a North Carolina disease intervention specialist. During this counseling the disease intervention specialist explained the study to eligible clients and obtained signed permission for release of name and contact information so that study staff could contact the client to discuss participation.
Control subjects were randomly selected from 1996 North Carolina driver's license records for all African American men and women listed in the original 13 counties.11 The file contained individuals with a current or recent (within the preceding 6 years) North Carolina driver s license or state identification card. Prospective respondents were selected randomly within strata defined by gender and 5-year age group based on the distribution among cases; respondents who reported injection drug use, male homosexual activity, or HIV infection were excluded from the analysis.
Each prospective study participant was sent a letter with information about the study. A female African American nurse-interviewer then telephoned or, if a working telephone number could not be found, went directly to the participant s home. Diligent attempts to reach prospective participants were continued for up to 6 weeks.
The interviewer requested permission to discuss the study, obtained signed consent, conducted a 1-hour interview in the participant's home or the interviewer's car, obtained a blood specimen for performance of syphilis tests, and provided a $50 cash incentive. The standardized interview questionnaire was developed from questionnaires used in published studies and was pretested and piloted. It covered sexual partnerships, HIV risk behaviors, history of sexually transmitted diseases (STDs), health care access, and demographic information. Information was collected concerning the respondent's sex partners in the past 10 years, with additional detail concerning the last 3 sex partners. A federal certificate of confidentiality was obtained, and interviewers assured participants of absolute confidentiality and that their names would be erased following data cleaning, the only exception being a report to the local health department if their serologic analysis disclosed current syphilis infection.
To determine the prevalence of syphilis among cases and controls, rapid plasma reagin (RPR) testing was performed on blood specimens followed by microhemagglutination testing for Treponema pallidum (MHA-TP) of specimens with reactive RPR results. Presence of syphilis was defined as reactive results on both RPR and MHA-TP. All other test results were considered negative. Respondents with positive serologic tests for syphilis were informed of their results and advised to consult the health department or their health care provider for further evaluation and treatment. Their names were also reported to the state health department, and disease intervention specialists, when necessary, provided counseling and referrals for additional evaluation and treatment. After data entry all personal identifiers were removed from the blood specimens and questionnaire data, and controls blood specimens were tested for HIV infection. The University of North Carolina School of Medicine s Committee on the Protection of the Rights of Human Subjects approved the study.
We examined the association of HIV infection with demographic variables, including socioeconomic indicators (history of incarceration of respondent or a sex partner, educational attainment, recent concern about hunger), history of STD (syphilis, gonorrhea, Chlamydia infection, and genital herpes), traditional risk variables for acquisition of HIV and other STDs (partner number, partners' behaviors, substance use, partners' STD and substance abuse history), and sexual network-related variables (concurrent partnerships of respondent and respondent's partner or partners). A respondent was identified as having a concurrent partnership if the start and end dates of ≥2 of the respondent s 3 most recent partnerships overlapped.12-14 A respondent was categorized as having a partner who was not monogamous (ie, a partner who had concurrent partnerships) if the respondent believed that the partner "definitely had sex with other people" during his or her sexual relationship with the respondent (as opposed to "probably did," "probably did not," or "definitely did not"). Similarly, respondents were categorized as having partners with other risk behaviors (eg, crack cocaine use) if the respondent believed it "very likely" (as opposed to "somewhat likely," "somewhat unlikely," "or very unlikely") that the partner had ever engaged in the behavior. We also identified a "lower-risk" subset of cases and controls in an attempt to identify the transmission risk of people who did not report any high-risk behaviors. These respondents were defined as those who denied STD diagnosis within the past year; trading sex for money, drugs, or shelter; use of crack or cocaine; binge alcohol consumption; a partner who had injected drugs or smoked crack; and sexual contact with a man who had sex with men.
Differences between case and control subjects were tested with χ2 and Wilcoxon tests. Exposure odds ratios (ORs) with 95% CIs and Breslow-Day tests for effect modification of the OR were determined using PROC FREQ in SAS version 8. SAS PROC LOGISTIC (SAS, Inc., Cary, NC) was used for multiple logistic regression analysis to examine associations controlling for differences in other explanatory variables. Variables that were associated with HIV infection with P<0.05 in bivariate analyses were tested in multiple logistic models. Age and sex were included in all models because controls were broadly frequency matched on these variables. Diagnosis of STD within the preceding year was also included in the overall model because of STD s potential role as a confounder: a recent diagnosis of STD is a common reason for HIV testing. Number of partners was log transformed.
Of 493 apparently eligible African American HIV cases (based on age, HIV reporting date, and risk factors), disease intervention specialists interviewed 482 (98%). Of these, 49% (235/482) declined to release contact information to the study, and 4 were not referred for unknown reasons. Of the 243/482 cases (50%) referred to the study, 17 ultimately proved ineligible (11 subsequently reported a history of injection drug use or male same-sex activity, 5 were referred >6 months after their HIV infection was reported, and 1 had uncertain HIV status). Ten could not be located, 7 were unavailable for reasons such as illness or incarceration, and 3 declined to participate after learning more about the study. The remaining 206 were interviewed, and their data were included in the study. Distributions of age, gender, screening risk characteristics, and method of case detection were generally similar among consenting case subjects and those who did not participate. Consent rates were similar for women (49%) and men (44%) and declined slightly with age (53% for case subjects <25 years of age, 47% for case subjects aged ≥40 years). Consenting case subjects were slightly more likely than those who refused participation to report sex with an injection drug user (9% vs. 4%). However, those who consented and those who refused were equally likely to report exchange of sex for drugs or money (18% vs. 15%) or sex with a person with AIDS or person with HIV of unknown risk (33% vs. 35%). The prevalence of sexual contact with a bisexual man was the same (4%) among women who participated and those who did not.
Women who participated were slightly more likely to have been diagnosed through prenatal screening (9%) than those who did not (6%), but consenting and nonconsenting case subjects otherwise had similar modes of case detection. For example, 53% of those who consented and 58% of those who refused were tested in non-health department facilities; 12% of both consenting and refusing case subjects were tested because they were named as contacts to a reported HIV case; 12% of participants and 14% of those who refused were tested at a public STD clinic. Reactive syphilis serology precipitated HIV testing among similar proportions of consenting (4%) and nonconsenting (3%) cases. Seven percent of participants and 5% of those who refused were tested during institutional screening (eg, in prisons, jails, or psychiatric hospitals).
We sampled 1063 potential subjects from the driver's license file. Of these, 697 (67%) could not be located, largely because of outdated (371) or incorrect (35) addresses. An additional 22 potential control subjects were unavailable for reasons such as incarceration or illness, and 17 were ineligible (5 because of age, 4 not African American, 6 homosexual men, 1 reported injection drug use, and 1 reported HIV infection). Of the 327 potential subjects who could be found and were eligible, 101 refused to participate, and 226 (69%) were interviewed. All controls had negative HIV tests.
Case and control subjects were similar in age (Table 1). Case subjects were more likely to be unmarried, to have less than a high school education, and to report annual income of <$16,000 (OR 4.8; 95% CI: 3.0, 7.7). Case subjects also more frequently reported indices associated with poverty, such as lack of health insurance, homelessness, and concern about having enough food during the past month. Both male and female case subjects were also much more likely than control subjects to have been incarcerated for at least 24 hours (OR 5.7; 95% CI: 3.3, 9.7).
As expected, substantial proportions of case subjects had high-risk behaviors related to sexual activity and substance use. Almost all (86% of cases and 86% controls) reported unprotected sex with at least 1 of their last 3 partners on ≥10 occasions. Case subjects were more likely than control subjects to report younger age at 1st sexual intercourse, greater numbers of sexual partners during their lifetime and in the past year, and concurrent partnerships during the preceding 1 and 5 years (Table 2). Case subjects were more likely to report (heterosexual) anal intercourse, exchanging sex for drugs or money, an STD diagnosis, and binge alcohol consumption. Crack use was a prominent risk factor (OR 8.6; 95% CI: 4.4, 16.9). Among respondents who denied having a partner who had injected drugs or had sex with men, 41 case subjects (20%) and 9 control subjects (4%) reported having smoked crack; more than half of these respondents (24 of 41 cases and 5 of 9 controls) traded sex.
Number of Partners
Case subjects reported more partners than did control subjects during their lifetime and in the past year, and men had more partners than women. Male case and control subjects reported, respectively, a median of 20 and 15 partners during their lifetime, whereas female case and control subjects reported a median of 8.5 and 5 partners, respectively. Male case subjects had a median of 2 partners during the preceding year (25% reported ≥4 partners). Male control subjects had a median of 1 partner (25% reported ≥2). Similarly, the median number of partners during the past year was 2 for female case subjects (25% reported ≥3) and 1 for female control subjects (25% had ≥2).
Case subjects reported riskier partnerships than control subjects and were more likely to have had a recent partner they believed had injected drugs, used crack cocaine, been incarcerated, or had sex with others while in a sexual relationship with the respondent. Female case subjects (10%, n = 13) were more likely than female control subjects (1%, n= 2) to have had a male partner who had had sex with other men. Almost half of the case subjects (48%) believed that at least 1 of their last 3 partners smoked crack; moreover, 32 case subjects (15%) and 21 control subjects (9%) denied using crack themselves or having a partner who had either injected drugs or had sex with men, but reported that at least 1of their last 3 partners had smoked crack.
Results of syphilis testing were available for 201 case and 197 control subjects. Nine male (11%) and 12 female (9%) case subjects had positive RPR and MHA-TP results. An additional male case subject had a reactive RPR (1:32), but MHA-TP was inadvertently not tested. Two male (3%) and 2 female (1%) control subjects had positive RPR and MHA-TP results.
Multiple Logistic Analysis
Of the crude associations of HIV infection with the variables in Tables 1 and 2, the associations with low educational attainment (adjusted OR: 3.0; 95% CI: 1.8, 5.2), number of lifetime sex partners (1.5 per log10; 95% CI: 1.1, 1.9), personal history of crack use (3.7; 95% CI: 1.7, 8.1), and a sex partner who had either injected drugs (3.0; 95% CI: 1.2, 7.7) or smoked crack (2.1; 95% CI: 1.2, 3.8) persisted in multiple logistic regression models (Table 3).
About one-quarter of case subjects denied all of the above risks (history of trading sex, a sexual partner who injected drugs, use of crack or inhaled heroin or cocaine, frequent binge alcohol consumption within the past year, a recent partner who had smoked crack, a male partner who had sex with other men, and diagnosis of an STD within the past year) and had negative syphilis serologic results. Among these lower-risk respondents (58 case and 157 control subjects), case subjects were more likely than control subjects to be young and unmarried and to report less than high school education, annual income <$16,000, recent concern about hunger, and history of incarceration (Table 4).
Lower-risk respondents reported fewer partners during their lifetime and in the past year than did respondents with high-risk behavior. In this lower-risk subset, case subjects reported more partners than did control subjects, and men had more partners than women. However, the association between partner number and HIV infection differed by gender: Among lower-risk men there were no significant differences between case and control subjects in lifetime number of partners or number of partners in the past year. In contrast, lower-risk female case subjects had more partners during their lifetime and in the past year than did lower-risk female control subjects.
Although there was no difference between lower-risk case and control subjects in age at 1st sexual intercourse, frequency of unprotected sex with 1 of the last 3 partners, or history of anal intercourse, lower-risk case subjects were more likely to report a history of STD; concurrent partnerships during the past 5 years; and a recent partner who was not monogamous during the relationship with the respondent or had been incarcerated.
The associations of HIV infection with low educational attainment (OR 5.0; 2.2, 11.1), hunger (3.7; 1.5, 8.9), and nonmonogamous partner(s) (2.9; 1.3, 6.4) persisted in multiple logistic regression models among the lower risk respondents.
This study, one of the few population-based studies of heterosexual HIV infection among African Americans in the United States, identified poverty, crack cocaine, STDs, and sexual network patterns as key contributors to the markedly elevated heterosexually transmitted HIV rates in this ethnic group. Multivariate analysis confirmed strong associations of HIV infection with less than high school education, history of crack use, increased partner number, and a recent partner who injected drugs or smoked crack.
However, more than one-quarter of case subjects denied high-risk behavior, including partnerships with persons who injected drugs or smoked crack. Among these apparently lower-risk respondents, risk for HIV infection was associated with less than high school education, recent concern about personal hunger, and a partner who had concurrent partnerships during the course of the relationship with the respondent. Poverty is the common denominator for both of these groups in the growing entrenchment of heterosexually transmitted HIV infection in this population.
This study has several limitations. Selection bias may have resulted from the high refusal rate among case subjects. However, although we cannot exclude the possibility that case subjects who participated differed in important ways from those who refused, distribution of gender, age, some important risk characteristics, and mode of case detection were similar among those who consented to be in the study and those who did not. In addition, we were unable to locate a large proportion of the potential control subjects identified by the sampling frame. However, the median income of our control subjects was similar to that of blacks in eastern North Carolina, and prevalence of crack cocaine use among them (5%) was similar to the rate among blacks in the general US population (5.7%).15,16 We necessarily relied on respondents' own report of their sexual behavior. Although the validity of self-reported sexual behavior has been a source of concern,17 some studies have revealed good agreement among male and female heterosexual partners concerning measures of sexual behavior (eg, number of episodes of intercourse, condom use).18 A related concern is that people's report of their partners' behaviors may be inaccurate. This could be a reason why partnerships with injection drug users and partnerships between women and men who had sex with other men were not commonly reported in this study, although they were associated with increased HIV infection risk. In addition, these analyses excluded male subjects who reported having had male sex partners, but it is possible that some male respondents had same-sex partnerships and declined to disclose them. This study cannot evaluate risk of HIV infection or behaviors associated with transmission among these men. Although use of the driver's license file to sample controls may have resulted in selection bias due to out-of-date addresses, it remains a better frame for sampling rural African Americans from the general population than voter registration lists, random digit dialing, telephone directories, and Medicare and utility listings.11 Finally, it is unclear to what extent our findings from a largely rural southern state are generalizable to heterosexual HIV transmission among blacks in the rest of the United States. However, poverty is a feature of life for many African Americans.
STDs facilitate spread of HIV19-21 and likely enhanced HIV transmission in this population. STDs were common among both case and control subjects, including among lower-risk respondents who denied high-risk behaviors. Because STD evaluation is an important stimulus for detection of HIV, concern for selection bias complicates interpretation of the association between STD diagnosis and HIV infection in our study. However, other studies in similar populations have also found strong associations between other STDs and heterosexual HIV infection.9 If STD is associated with a 3-fold risk of acquiring HIV21 and the 5-year prevalence of STD in the general African American population is 10% (likely a low estimate, considering that herpes simplex virus-2 seroprevalence among African Americans is 46%22), then the population-attributable risk23 of STD for heterosexual transmission of HIV in our study population would be 17%.
Case subjects reported more sex partners than did control subjects; men reported substantially more partners than women; and the partner distribution was strongly skewed, especially among male case subjects. Male case subjects a median of 20 partners during their lifetime, whereas 10% reported >200 partners. In contrast, women case subjects overall reported a median of 8.5 partners during their lifetime, with 10% reporting ≥30 or more. Lower-risk female case subjects reported a median of only 6 partners during their lifetime. The importance of core groups in heterosexual HIV transmission has been extensively documented.24 Our findings support this concept and suggest that a core group, comprised predominantly of men, contributes disproportionately to heterosexual HIV transmission within this population.
The 5-year prevalence of concurrent (overlapping) partnerships among men (53%) and women (31%) from this study s control population was higher than that among black women from the general US population sampled by the National Survey of Family Growth (21%)12,14; among case subjects, the 5-year concurrency prevalence was even greater (63% in male cases, 58% in female cases).13 In addition, having a partner who had concurrent partnerships was also reported widely by case and control subjects, including the "lower-risk" subset. Such extensive concurrency (on the part of respondents and their partners) speeds population HIV transmission25,26 and suggests high network density, a feature of networks with high HIV prevalence.27
More than one-third of cases used crack, and an additional 15% denied using crack themselves but reported that 1 of their last 3 partners used the drug. More than half of the case subjects who used crack had also traded sex. The associations of crack cocaine with exchange of sex for drugs28,29 and with HIV infection6,8,9,30 have been well documented. Our results illustrate crack cocaine's propensity to influence sexual networks and bridge high-risk subpopulations to each other and to the general population.
Among the most striking findings was the association of HIV infection in both the study population as a whole and in the low-risk subset with indicators of low socioeconomic status, such as low income, lower educational attainment, and recent concern about having enough food. Other investigators have observed associations between poverty and HIV infection,6,9,31-39 which may be linked through a variety of pathways.40,41 For example, poverty results in decreased health care access with resultant increased duration of treatable STDs,42 which in turn facilitates HIV transmission. Targeted marketing of crack cocaine43 to neighborhoods of low socioeconomic status increases residents risk of crack use and exchange of sex for drugs. Residential segregation concentrates poverty, drug use, and other deleterious influences within the neighborhood,44 shrinking the pool of available partners to those with a high prevalence of risk behaviors and HIV infection.45 Indeed, residence in specific neighborhoods was a risk factor for HIV infection in a rural Florida community with extensive heterosexual HIV transmission.9 Qualitative research among black men and women from the study region indicates that African Americans extremely adverse socioeconomic circumstances and the relative scarcity of black men influence partner selection, women's bargaining power and availability, and both genders participation in sexual risk behaviors.46 Mounting evidence indicates that on both the population31 and individual levels, economic inequality is a determinant of HIV infection.
In summary, blacks with heterosexually acquired HIV infection are more likely than blacks in the general population to report risk factors for infection, particularly crack use and sexual partnerships with individuals who smoke crack. The substantial proportion of African Americans with heterosexually acquired HIV infection who do not report high-risk behavior suggests that heterosexual HIV infection has become endemic among blacks in the southern United States. Potential interventions include successfully combating the crack cocaine epidemic that has ravaged black and other impoverished communities, implementation of more effective STD diagnosis and treatment programs, and behavioral interventions to decrease numbers of partners and increase condom use. However, effective long-term public health strategies for decreasing HIV rates among African Americans and others in the United States will require addressing economic inequalities that promote risk for infection.
The authors dedicate this manuscript to the late Stephanie Betran, RN, for her invaluable contribution to the project. We thank Drs. Amy Lansky and Joanne Earp for their assistance with questionnaire development; Ms. Evelyn Faust, Ms. Judy Owen-O' Dowd, and the NC HIV-STD Control Section and Disease Intervention Specialists for their assistance with recruitment; UNC Hospitals Clinical Microbiology/Immunology Laboratory for performance of syphilis serologic assays; Dr. Susan Fiscus for performance of HIV tests; Drs. Sevgi Aral, Ward Cates, Paul Godley, Mike Cohen, and Greg Samsa for their insightful comments; Ms. Kathryn Donaldson Simmons for assistance with data analysis, Ms. Merritha Williams and the late Stephanie Betran for recruiting and interviewing the respondents; and the respondents for their participation in this research.
1. Centers for Disease Control and Prevention. Heterosexual transmission
: 29 states, 1999-2002. Morb Mortal Wkly Rep
3. Centers for Disease Control and Prevention. Characteristics of persons living with AIDS and HIV
, 2001. HIV/AIDS Surveillance Suppl Rep
. 2003;9:1-27. Available at: http://www.cdc.gov/hiv/stats/hasrsupp.htm
. Accessed June 13, 2004.
4. Hader SL, Smith DK, Moore JS, et al. HIV
infection in women in the United States: status at the millennium. JAMA
5. Chirgwin KD, Feldman J, Dehovitz JA, et al. Incidence and risk factors for heterosexually acquired HIV
in an inner-city cohort of women: temporal association with pregnancy. J Acquir Immune Defic Syndr Hum Retrovirol
6. Wohl AR, Lu S, Odem S, et al. Sociodemographic and behavioral characteristics of African-American women with HIV
and AIDS in Los Angeles County, 1990-1997. J Acquir Immune Defic Syndr Hum Retrovirol
7. Wohl AR, Johnson DF, Lu S, et al. HIV
risk behaviors among African American men in Los Angeles County who self-identify as heterosexual. J Acquir Immune Defic Syndr
8. Ellerbrock TV, Lieb S, Harrington PE, et al. Heterosexually transmitted human immunodeficiency virus infection among pregnant women in a rural Florida community. N Engl J Med
9. Ellerbrock TV, Chamblee S, Bush TJ, et al. Human immunodeficiency virus infection in a rural community in the United States. Am J Epidemiol
10. Misegades L, Page-Shafer K, Halperin D, et al. Anal intercourse among young low-income women in California: an overlooked risk factor for HIV
11. Adimora A, Schoenbach V, Martinson F, et al. Driver's license and voter registration lists as population-based sampling frames for rural African Americans
. Ann Epidemiol
12. Adimora A, Schoenbach V, Bonas D, et al. Concurrent sexual partnerships among women in the United States. Epidemiology
13. Adimora AA, Schoenbach VJ, Martinson FE, et al. Concurrent partnerships among rural African Americans
with recently reported heterosexually transmitted HIV
infection. J Acquir Immune Defic Syndr
14. Adimora AA, Schoenbach VJ, Martinson F, et al. Concurrent sexual partnerships among African Americans
in the rural south. Ann Epidemiol
17. Catania JA, Gibson DR, Chitwood DD, et al. Methodological problems in AIDS behavioral research: influences on measurement error and participation bias in studies of sexual behavior. Psychol Bull
18. Harvey SM, Bird ST, Henderson JT, et al. He said, she said: concordance between sexual partners. Sex Transm Dis
19. Wasserheit JN. Epidemiological synergy: interrelationships between human immunodeficiency virus infection and other sexually transmitted diseases. Sex Transm Dis
20. Cohen MS, Hoffman IF, Royce RA, et al. Reduction of concentration of HIV
-1 in semen after treatment of urethritis: implications for prevention of sexual transmission of HIV
21. Fleming DT, Wasserheit JN. From epidemiologic synergy to public health policy and practice: the contribution of other sexually transmitted diseases to sexual transmission of HIV
infection. Sex Transm Infect
22. Fleming DT, McQuillan GM, Johnson RE, et al. Herpes simplex virus type 2 in the United States, 1976 to 1994. N Engl J Med
23. Kahn HA. An Introduction to Epidemiologic Methods
. New York: Oxford University Press; 1983.
24. Plummer FA, Nagelkerke NJ, Moses S, et al. The importance of core groups in the epidemiology and control of HIV
-1 infection. AIDS
. 1991;5(Suppl 1):S169-S176.
25. Morris M, Kretzschmar M. Concurrent partnerships and transmission dynamics in networks. Soc Networks
26. Kretzschmar M, Morris M. Measures of concurrency in networks and the spread of infectious disease. Math Biosci
27. Rothenberg RB, Long DM, Sterk CE, et al. The Atlanta Urban Networks Study: a blueprint for endemic transmission. AIDS
. 2000; 14:2191-2200.
28. Fullilove RE, Fullilove MT, Bowser BP, et al. Risk of sexually transmitted disease among black adolescent crack users in Oakland and San Francisco, Calif. JAMA
29. Fullilove MT, Golden E, Fullilove RED, et al. Crack cocaine use and high-risk behaviors among sexually active black adolescents. J Adolesc Health
30. Edlin BR, Irwin KL, Faruque S, et al. Intersecting epidemics: -crack cocaine use and HIV
infection among inner-city young adults. Multicenter Crack Cocaine and HIV
Infection Study Team. N Engl J Med
31. Zierler S, Krieger N, Tang Y, et al. Economic deprivation and AIDS incidence in Massachusetts. Am J Public Health
32. St. Louis ME, Conway GA, Hayman CR, et al. Human immunodeficiency virus infection in disadvantaged adolescents: findings from the US Job Corps. JAMA
33. Krueger LE, Wood RW, Diehr PH, et al. Poverty
seropositivity: the poor are more likely to be infected. AIDS
34. Diaz T, Chu SY, Buehler JW, et al. Socioeconomic differences among people with AIDS: results from a Multistate Surveillance Project. Am J Prev Med
35. Fife D, Mode C. AIDS incidence and income. J Acquir Immune Defic Syndr
36. Simon PA, Hu DJ, Diaz T, et al. Income and AIDS rates in Los Angeles County. AIDS
37. Hu DJ, Frey R, Costa S, et al. AIDS rates and sociodemographic variables in the Newark, New Jersey, metropolitan area. 9th International Conference on AIDS
, Abstract no. WS-C04-82.Berlin, Germany; June 6-11, 1993.
38. Farmer P. Women, poverty
, and AIDS. In: Farmer P, ed. Women, Poverty and AIDS. Monroe
, ME: Common Courage Press; 1996.
39. Farmer P. Infections and Inequalities: The Modern Plagues
. Berkeley and Los Angeles, CA: University of California Press; 1999.
40. Adimora AA, Schoenbach VJ. Social context, sexual networks, and racial disparities in sexually transmitted infection rates. J Infect Dis
. 2005;191(Suppl 1):S115-S122. Review.
41. Adimora AA, Schoenbach VJ. Contextual factors and the black-white disparity in heterosexual HIV
42. Aral S, Holmes K. Social and behavioral determinants of the epidemiology of STDs: industrialized and developing countries. In: Holmes K, Per-Anders M, Sparling PF, et al, eds. Sexually Transmitted Diseases
. New York: Mc-Graw-Hill; 1999.
43. Krieger N. Embodying inequality: a review of concepts, measures, and methods for studying health consequences of discrimination. Int J Health Serv
44. Massey DS, Denton NA. American Apartheid: Segregation and the Making of the Underclass
. Cambridge, MA: Harvard University Press; 1993.
45. Ickovics J, Beren S, Grigorenko E, et al. Pathways of risk: race, social class, stress, and coping as factors predicting heterosexual risk behaviors for HIV
among women. AIDS Behav
46. Adimora AA, Schoenbach VJ, Martinson FE, et al. Social context of sexual relationships among rural African Americans
. Sex Transm Dis