IN THE UNITED STATES, population surveys have consistently found ethnic differences in the prevalence of sexual behaviors, both in adolescents 1–16 and in adults. 17–24 Some studies have also shown ethnic differences in sexual behavior among high-risk groups such as intravenous drug users and alcoholics. 25–28 In surveys of adolescents and adults in the general population, researchers have often found that African Americans report earlier onset of sexual intercourse and more partners than non-Hispanic whites or Hispanics. 3,6–13,16,18,19,21,22,29 However, African Americans could be more likely than these other ethnic groups to use condoms. 2,6,10,13,16,17,29
Recent studies have highlighted the contradictions and uncertainties often inherent in simple self-reported behaviors (for example, percent condom use). 30,31 In our previous research, we have shown the use of context in clarifying the relationship between behaviors and infection, 32 and in this study, we sought to shed further light on ethnic differences through the evaluation of complex behavioral constructs. In addition, we sought to evaluate attitudes and beliefs in relationship to sexual risk behaviors. Previous studies of the relationships between attitudes and behavior have found a lower prevalence of sexual risk behaviors among people placing greater importance on relationships, expressing greater acceptance of condoms, or perceiving greater vulnerability to disease. 33,34 Most of this research has been performed among adolescents and university students; therefore, with this study population ranging to age 45, we add to the limited information on adults. In addition, our longitudinal study design allows us to evaluate the relationship among attitudes, behaviors, and infections over time.
Finally, a major goal of this analysis was to assess the ethnicity-specific impact of a successful cognitive–behavioral intervention. We have previously described the overall efficacy of this intervention, Project SAFE, to prevent the recurrence of bacterial sexually transmitted diseases (STDs) in low-income minority women with a nonviral STD at baseline. 35 The intervention was tailored separately for Mexican-American and African-American participants. In this analysis, we evaluated these 2 ethnic groups separately to shed light on their behaviors, attitudes and beliefs, infection rates, and intervention efficacy in each group.
Detailed methods have been published previously. 35 Briefly, women eligible for our intervention trial spoke English, were aged 14 to 45, and had a current infection with chlamydia, gonorrhea, trichomonas, or syphilis. Between January 1993 and July 1994, all 947 eligible women attending STD and other public health clinics in San Antonio, Texas, were offered the opportunity to participate. Women enrolling in the study received treatment and counseling and were assigned randomly (after stratification by ethnicity) to the intervention or control group. All subjects were administered questionnaires, lasting approximately 2 hours, to gather clinical and behavioral data at baseline, 6-, and 12-month follow up. In this report, our results relating to behaviors, attitudes, and beliefs are based on 477 women (149 African Americans and 328 Mexican Americans) who attended all 3 study visits. However, when possible in analyses relating only to infection rates, we present results based on 549 women (170 African Americans and 379 Mexican Americans) for whom we obtained data on reinfection over the 12-month study period. Reinfection with chlamydia and/or gonorrhea was determined by amplification tests (GEN-PROBE PACE 2, GEN-PROBE Co., San Diego, CA) performed at the regular 6- and 12-month visits and the nonscheduled “problem visits,” as well as from records of similar off-site amplification tests obtained through the San Antonio Metropolitan Health District. 35
We constructed 5 main dichotomous variables describing sexual behavior leading to increased risk of sexually transmitted infections. The first, compliance, reflects whether each woman refrained from having unprotected sex with an untreated or incompletely treated partner after treatment of her initial infection. The second, douching after sex, reflects whether the woman used a vaginal douche after sexual intercourse. The third, mutual monogamy, reflects whether the woman had one steady, faithful sexual partner or no sexual partner during the study period. The fourth, risky sex, reflects condom use in different contexts: a woman was considered “risky” if she reported a casual partner with whom she never used condoms or any partner with whom she reported both 5 or more unprotected acts in the last 3 months and incorrect or problematic condom use. Finally, the fifth behavior, rapid partner turnover, indicates whether the woman acquired a new sexual partner within 3 months of having sex with a previous partner. More detailed descriptions of these behaviors have been previously published. 32 Additionally, in the current analysis, we have conducted more limited assessments of a sixth variable, concurrency, based on observed overlap between the reported first and last dates of sex with different partners. Fifteen women who did not have a sexual partner during the study period were classified in the low-risk category for each behavioral variable. 32
Attitudes and Beliefs
In addition to evaluating reported prevalences of the 5 main behavioral constructs and their relationship with risk of reinfection, we evaluated possible differences between African-American and Mexican-American participants regarding attitudes about relationships and beliefs regarding partner availability. Women were asked to respond to several relevant measures: how easy or hard it was to find a good man in San Antonio (6-level Likert scale, from very easy to very hard); whether it was a good idea to have simultaneous sexual relationships with more than one man (yes or no); how important it was to be in a relationship with a man (5-level Likert scale); and whether there were more available men or women in San Antonio (5-level Likert scale). In addition, for each of the 5 most recent sexual partners, each woman was asked whether she would still have sex with him if he had sex with other women (yes, no, or conditional). Women who specified that they would only continue a sexual relationship if the man used a condom with her were classified as “no”; therefore, this measure represents a willingness to have unprotected sex with an unfaithful partner. We created dichotomizations for variables assessed on a Likert scale and conducted sensitivity analyses by constructing alternate dichotomizations to test the robustness of the observed associations.
For the sexual behavior constructs and the questions assessing attitudes and beliefs, we used cumulative measures encompassing the 6- and 12-month follow-up interviews. Exceptions were the question “Are there more available women or men in San Antonio?” (this was asked only at the initial interview), and sex with an untreated partner (this related only to the woman’s behavior immediately after her initial infection and treatment). The cumulative measures reflect the highest-risk response given by the woman during the study follow up.
In several analyses presented here, we make comparisons between Mexican-American and African-American women overall without a separate focus on the intervention effect. Assignment to the intervention or control group was not expected to confound these analyses, because ethnic identity was balanced by design. However, in all analyses, we tested intervention group status as a confounder and as an interaction term with ethnicity to help ensure that any possible ethnic differences in intervention effect were not influencing our estimates of differences between ethnic groups.
We conducted chi-square tests to evaluate differences in demographic characteristics at baseline between African-American and Mexican-American women. In addition, we used chi-square tests to evaluate the relationship between reported attitudes at baseline and reported sexual behaviors over the 12-month follow-up period. We conducted least-squares means regression analyses to estimate the crude and covariate-adjusted proportions of African-American and Mexican-American women who became reinfected or reported behaviors, attitudes, and beliefs of interest over the 12-month follow-up period. Possible confounding factors of interest included intervention study group, age, education, per capita income, and relationship status (living with a spouse vs. separated or unmarried). To assess the power of our measures of sexual behavior in explaining any ethnic differences in STD reinfection, we used logistic regression to model STD reinfection with ethnicity as the main predictor of interest controlling for the 5 complex sexual behaviors described previously. Finally, to assess ethnicity-specific intervention effects on cumulative 12-month measures of sexual risk behaviors and reinfection rates, we used Mantel-Haenszel chi-square tests and logistic regression. All analyses relating to intervention group assignment were performed on the basis of intention to treat. SAS and SPSS statistical software were used for all analyses.
Characteristics of the study population at baseline are provided in Table 1. In our sample, African-American women were more likely to have completed high school than Mexican-American women and reported slightly higher per capita incomes (14.9% of Mexican Americans and 18.1% of African Americans did not provide a household income). More African-American women were over the age of 25; nevertheless, African-American women were much less likely to be living with a spouse than Mexican-American women. By design, all women had at least one nonviral STD at baseline. Gonorrhea, trichomonas, and syphilis were more common in African Americans than in Mexican Americans; however, chlamydia was more common in Mexican Americans.
Ethnic Differences in Infection Rate and Prevalence of Attitudes, Beliefs, and Behaviors
Analyses of attitudes and beliefs about relationships and partner availability, reported over the 12-month study follow up, are summarized in Table 2, adjusting for age, education, and relationship status. Unadjusted results were similar (data not shown). Results suggested that slightly more African Americans thought it was “very hard” to find a suitable partner (P = 0.22) and showed that more African Americans thought there were more available women than men (assessed only at baseline) (P = 0.02). More Mexican Americans than African Americans considered being in a relationship to be one of the most important things in life (P <0.01); however, this difference was attenuated (P = 0.10) when responses were dichotomized as important versus not important. More African-American than Mexican-American women considered it a “good idea” to have concurrent sexual relationships (P <0.01) and would continue having unprotected sex with her partner if he had sex with others (P = 0.03). Consistent with attitudinal data, more African-American than Mexican-American women were involved in concurrent sexual relationships at some time during the study period (27.5% vs. 15.2%, P <0.01).
We calculated the proportions engaging in risky behaviors and experiencing reinfection, adjusting for age, education, and relationship status at baseline (Table 2). Unadjusted results were similar (data not shown). Only 397 (83%) of the women provided income data; however, results were similar in models with and without this covariate (data not shown). In addition, results were similar between models including and excluding intervention status as a confounder and as an interaction term with ethnicity (data not shown). Accordingly, we present age-, education-, and relationship status-adjusted results for all 477 women in analyses excluding per capita income and intervention status as covariates. African-American women reported higher prevalences of vaginal douching after sex (P <0.01), not being in a mutually monogamous relationship (P <0.01), and engaging in rapid partner turnover (P <0.01). However, the 2 ethnic groups showed no difference in engaging in risky sex (P = 0.73), and Mexican-American women were slightly more likely to have unprotected sex with an untreated partner (P = 0.26). Black women were more likely than Mexican-American women to become reinfected during the 12-month follow up (P <0.01).
Effect of the Cognitive–Behavioral Intervention
Although Mexican-American women had lower rates of reinfection than African American women, the intervention effect was similar in both ethnic groups (Table 3), with 35% to 40% fewer infections in the study groups compared with the control groups. In analyses of 549 women with complete infection data, the crude odds ratio and 95% confidence interval for reinfection, comparing intervention with control women, was 0.54 (0.32–0.91) for Mexican Americans and 0.58 (0.29–1.1) for African Americans. In a model testing the interaction between ethnicity and intervention group status, the odds ratios were shown not to differ between ethnic groups (P value for interaction term = 0.88). Analyses of the 477 women with complete data produced similar results, with crude odds ratios and 95% confidence intervals of 0.62 (0.35–1.1) for Mexican Americans and 0.62 (0.30–1.3) for African Americans, and nonsignificant test for interaction (P = 0.997).
A consistent pattern emerged in which most sexual risk behaviors were less common in study women versus control women in both ethnic groups (Table 3), with the exception of rapid partner turnover in Mexican-American women. Statistically significant intervention effects were observed for mutual monogamy and rapid partner turnover in African Americans only and for risky sex in both ethnic groups. The intervention effect in reducing noncompliance was borderline significant in Mexican Americans (P = 0.06).
Ethnic Differences in the Relationship Between Initial Attitudes/Beliefs and Subsequent Behavior
In analyses of the association between attitudes reported at baseline and behaviors reported during study follow up, we found ethnicity to act as an important confounder and effect modifier of several relationships. Accordingly, results for African-American women are presented in Table 4, and results for Mexican-American women are presented separately in Table 5. For both ethnic groups (borderline significance among Mexican Americans), women reporting it “hard” to find a suitable partner were less likely to engage in rapid partner turnover. Conversely, women considering it a “good idea” to have a man on the side (borderline significance among Mexican Americans) or who reported they would continue having unprotected sex with an unfaithful partner were more likely to engage in rapid partner turnover. In addition, women who reported they would continue having unprotected sex with an unfaithful partner were more likely to engage in risky sex during the follow-up period.
Among African Americans alone (Table 4), women considering it a “good idea” to have a man on the side were somewhat more likely to engage in risky sex (P = 0.08). Among Mexican Americans alone (Table 5), several additional relationships were observed: women reporting it “hard” to find a suitable partner were more likely to be in mutually monogamous relationships. However, Mexican-American women who thought there were more available women than men were less likely to be in mutually monogamous relationships and more likely to engage in risky sex during the follow-up period. Finally, Mexican-American women who thought being in a relationship was “one of the most important things in life” were somewhat more likely to be in a mutually monogamous relationship (P = 0.08) and less likely to have rapid partner turnover, whereas those who reported they would continue having unprotected sex with an unfaithful partner were less likely to be in a mutually monogamous relationship throughout the 12-month study period.
Of additional interest, we observed discordant associations by ethnic group in women believing that there were more available women than men; among African Americans (Table 4), these women were less likely to have rapid partner turnover (P = 0.06), whereas among Mexican Americans (Table 5), they were more likely (P = 0.054) to have rapid partner turnover. Less notable discordant associations were seen for rapid partner turnover in women who considered being in a relationship “one of the most important things in life” and for mutual monogamy in women who thought there were more women than men available.
Role of Sexual Risk Behaviors in Explaining Ethnic Differences in Infection Rates
Noting the ethnic differences in reinfection, and in reported sexual risk behaviors (Table 2), we fit a logistic regression model predicting reinfection with ethnicity as the main predictor variable. We then controlled for sexual risk behaviors, and intervention study group, to assess the role of these variables as confounders in the association between ethnicity and reinfection. The unadjusted odds ratio (OR) for reinfection in African Americans versus Mexican Americans was 1.8 (95% confidence interval [CI], 1.2–2.8). After adjusting for the 3 behaviors shown in Table 3 to be associated with ethnicity (douching after sex, lack of mutual monogamy, and rapid partner turnover), the OR was reduced to 1.2 (95% CI, 0.72–2.0). Although noncompliance and risky sex were not expected to be strong confounders, they are strong predictors of reinfection; when including all 5 behaviors in the model as covariates, the OR was less attenuated but still insignificant at 1.5 (95% CI, 0.86–2.5). Inclusion of the term for intervention study group had no effect on the estimated association between ethnicity and reinfection in any model examined.
Effect of the Cognitive–Behavioral Intervention
The cognitive–behavioral intervention used in Project SAFE was tailored separately for Mexican-American and African-American women who met in separate groups. A notable finding of this study is that the proportionate reductions in infection rate between control and study women were similar between ethnic groups.
In addition, the intervention reduced most sexual risk behaviors in both ethnic groups, although some reductions were not statistically significant. In particular, among African Americans, the behavioral impact of the intervention was significant for rapid partner turnover and mutual monogamy. These intervention effects were not observed in Mexican Americans. However, controlling for intervention group, African-American women were more likely than Mexican-American women to engage in these behaviors. Therefore, in 2 of the 3 risk behaviors more prevalent in African Americans than Mexican Americans, the intervention effect was significant among African Americans but not among Mexican Americans, bringing the 2 ethnic groups to a more comparable risk profile.
Although the relative impact of the intervention was similar between ethnic groups, African-American women in the study had a substantially higher rate of infection than Mexican-American women. Therefore, the absolute impact of the intervention differed between ethnic groups, and the absolute number of new infections prevented by the intervention was greater in African-American women. However, the crude rates of reinfection, behaviors, and reports of attitudes and beliefs reported here should be interpreted cautiously because of the intervention effects. To varying degrees, these rates should be expected to underestimate the risk profile of this low-income minority population. Nevertheless, valid ethnic comparisons could be made of the prevalences of attitudes, beliefs, and behaviors, because the proportion assigned to the intervention group, and the relative impact of the intervention, were similar in both ethnic groups.
Ethnic Differences in Prevalence of Attitudes, Beliefs, and Sexual Risk Behaviors
Our analyses showed that African-American women were more likely than Mexican-American women to report a dearth of available men and to express attitudes compatible with nonmonogamous relationships (less important to be in a relationship, good idea to have a man on the side, more likely to continue having unprotected sex with an unfaithful partner). During the 12-month follow-up period, African-American women were more likely to report several STD risk behaviors: douching after sex, lack of mutual monogamy, and rapid partner turnover.
However, African-American women were slightly more likely to abstain from sex with an untreated partner after their initial STD infection, and the 2 ethnic groups were equally likely to engage in safer-sex practices. Overall, we therefore conclude that the higher reinfection rates observed in African-American women in our study could be partially attributable to differences in perceived partner availability and relationship decision-making, contributing to higher levels of nonmonogamy and rapid partner turnover. This finding is supported by the results of our logistic regression modeling, which showed that there is no significant association between ethnicity and reinfection after controlling for the sexual risk behaviors described here.
We have previously noted in this population 32 that the impact of condom use varies in ways dependent on context. Clearly, for example, if both partners are uninfected, then the use of a condom will not alter the likelihood of STD transmission. We constructed our “risky sex” variable to reflect contextual aspects of condom use, and our variable construction therefore differs in significant ways from previous studies. Indeed, the analysis reported here showed no differences in “risky sex” between African Americans and Mexican Americans, whereas most previous surveys have found that African Americans were more likely than Hispanics to use condoms. 17,20,21,23,24 Consistent with our findings, previous surveys have shown that African-American women were more likely than Hispanic women to have multiple sexual partners 18,19,21,22,36–38; in addition, more African-American women might believe that their sexual partner has other partners. 23 Taken together, the results of prior studies, suggesting overall that African-American women have more sexual partners but may use condoms more consistently, provide support and additional validation of our conceptualization of condom use as a behavior whose impact is highly dependent on the context of the sexual relationship.
What Produces Ethnic Differences in Behavior?
Observed ethnic differences in behavior could in some cases be attributable to differences in demographic characteristics such as marital status, educational attainment, and urban residence. For example, one nationally representative survey 21 found that African-American women reported more sexual partners during the last 3 months than Hispanic or non-Hispanic white women. However, after controlling for marital status and other factors, African-American women had slightly fewer partners than non-Hispanic whites or Hispanics (not statistically significant). Similarly, in a survey of women at high risk for STDs because of risky sexual practices or drug use, 18 investigators found that most differences among Hispanics, African Americans, and non-Hispanic whites vanished after controlling for factors such as age, education, marital status, and comfort in discussing sex. In our analyses, the observed ethnic differences did not substantially change after controlling for age, education, and relationship status. However, ethnic differences could certainly be a misleading perspective for comparing patterns of sexual behavior. Researchers should incorporate important explanatory covariates into analyses of sexual behavior and avoid simple bivariate analyses that could preclude a more full understanding of the behavior differences between groups.
Despite potential difficulties in interpretation, however, true ethnic differences are likely to exist. Previous researchers have noted that cultural differences in gender relationships, gender role expectations, and male dominance could result in ethnic differences in sexual behavior and relationship decision-making. 18,21 Another factor relating to gender relationships is perceived partner availability. In Bexar County, Texas (including San Antonio), the 2000 U.S. Census counted over 500,000 Hispanic or Latino adults but less than 70,000 black or African-American adults. 39 Compounding this local demographic reality for African-American women, national data show significant ethnic differences in the rate of male incarceration. In 1996, the incarceration rates among African-American, Hispanic, and non-Hispanic white men in the United States were 3.1%, 1.3%, and 0.4%, respectively. Among men aged 25 to 29, corresponding rates were 8.3%, 2.6%, and 0.8%. 40 Furthermore, the 1996 correctional supervision rates (jail, prison, probation, or parole) for African-American men aged 18 to 19, 20 to 24, and 25 to 29 were 16.2%, 29.4%, and 28.9%, respectively; corresponding rates for non-Hispanic white men were 4.4%, 8.0%, and 7.1%, respectively. 41 Given the high rates of incarceration among young African American men in the United States, and the small proportion of San Antonio residents who are African American, it is not surprising that African-American women in our study reported a perception of fewer available partners than their Mexican-American counterparts. In addition, this perception could partly explain the higher tolerance of nonmonogamy reported by African-American women. For example, tolerance of nonmonogamy could reflect strategies adopted by some African-American women to maintain heterosexual relationships in the face of limited choice. It is important to underscore that African-American women in the intervention group enacted behavior change in this area, where they were in most need; that is, they made significant reductions in rapid partner turnover and significant increases in mutual monogamy.
In summary, we found that African-American women in our study were more likely to report risky sexual behaviors (rapid partner turnover, lack of mutual monogamy, and douching after sex), and were more likely to experience reinfection over a 12-month follow-up, in comparison to Mexican-American women. These differences in reinfection and behavior were not attributable to the behavioral intervention, because both ethnic groups were evenly distributed between the study and control groups. Rather, regression models suggested that risky sexual behaviors largely explained the different rates of reinfection, whereas ethnic differences in partner availability, and attitudes about partner selection and the acceptability of nonmonogamy, appeared to partially explain the ethnic differences in behavior. Notably, despite the substantial observed ethnic differences in attitudes, behaviors, and reinfection rates, the cognitive–behavioral intervention used in Project SAFE resulted in similar proportionate reductions in the rate of reinfection among both ethnic groups, comparing study women with control women. This accomplishment is encouraging, in light of the disproportionate burden of sexually transmitted disease borne by low-income minority populations in the United States.
1. Ellen JM, Aral SO, Madger LS. Do differences in sexual behaviors account for the racial/ethnic differences in adolescents’ self-reported history of a sexually transmitted disease? Sex Transm Dis 1998; 25: 125–129.
2. Graves KL, Hines AM. Ethnic differences in the association between behavior with a new partner: An event-based analysis. AIDS Educ Prev 1997; 9: 219–237.
3. Coker AL, Richter DL, Valois RF, McKeown RE, Garrison CZ, Vincent ML. Correlates and consequences of early initiation of sexual intercourse. J Sch Health 1994; 64: 372–377.
4. Weller NF, Tortolero SR, Kelder SH, Grunbaum JA, Carvajal SC, Gingiss PM. Health risk behaviors of Texas students attending dropout prevention/recovery schools in 1997. J Sch Health 1999; 69: 22–28.
5. Neumark-Sztainer D, Story M, French S, Cassuto N, Jacobs DR Jr, Resnick MD. Patterns of health-compromising behaviors among Minnesota adolescents: Sociodemographic variations. Am J Public Health 1996; 86: 1599–1606.
6. Kann L, Kinchen SA, Williams BI, et al. Youth risk behavior surveillance—United States, 1999. State and Local YRBSS Coordinators. J Sch Health 2000; 70: 271–285.
7. Benson MD, Torpy EJ. Sexual behavior in junior high school students. Obstet Gynecol 1995; 85: 279–284.
8. Singh S, Darroch JE. Trends in sexual activity among adolescent American women: 1982–1995. Fam Plann Perspect 1999; 31: 212–219.
9. Upchurch DM, Levy-Storms L, Sucoff CA, Aneshensel CS. Gender and ethnic differences in the timing of first sexual intercourse. Fam Plann Perspect 1998; 30: 121–127.
10. Ford K, Norris AE. Knowledge of AIDS transmission, risk behavior, and perceptions of risk among urban, low-income, African-American and Hispanic youth. Am J Prev Med 1993; 9: 297–306.
11. Fortenberry JD. Number of sexual partners and health lifestyle of adolescents. Use of the AMA Guidelines for Adolescent Preventive Services to address a basic research question. Arch Pediatr Adolesc Med 1997; 151: 1139–1143.
12. Strunin L. Adolescents’ perceptions of risk for HIV infection: Implications for future research. Soc Sci Med 1991; 32: 221–228.
13. Ford K, Norris AE. Sexually transmitted diseases: Experience and risk factors among urban, low income, African American and Hispanic youth. Ethn Health 1996; 1: 175–184.
14. Norris AE, Ford K, Shyr Y, Schork MA. Heterosexual experiences and partnerships of urban, low-income African-American and Hispanic youth. J Acquir Immun Defic Syndr 1996; 11: 288–300.
15. Brown EJ. AIDS-related risk behavior of young college students. Abnf J 2000; 11: 37–43.
16. Kann L, Warren CW, Harris WA, et al. Youth risk behavior surveillance—United States, 1993. MMWR Morb Mortal Wkly Rep 1995; 44: 1–56.
17. Hines AM, Graves KL. AIDS protection and contraception among African American, Hispanic, and white women. Health Soc Work 1998; 23: 186–194.
18. Quadagno D, Sly DF, Harrison DF, Eberstein IW, Soler HR. Ethnic differences in sexual decisions and sexual behavior. Arch Sex Behav 1998; 27: 57–75.
19. Caetano R, Hines AM. Alcohol, sexual practices, and risk of AIDS among African Americans, Hispanics, and whites. J Acquir Immun Defic Syndr 1995; 10: 554–561.
20. Catania JA, Coates TJ, Golden E, et al. Correlates of condom use among African American, Hispanic, and white heterosexuals in San Francisco: The AMEN longitudinal survey. AIDS in Multi-Ethnic Neighborhoods survey. AIDS Educ Prev 1994; 6: 12–26.
21. Kost K, Forrest JD. Am women’s sexual behavior and exposure to risk of sexually transmitted diseases. Fam Plann Perspect 1992; 24: 244–254.
22. Anderson JE, Dahlberg LL. High-risk sexual behavior in the general population. Results from a national survey, 1988–1990. Sex Transm Dis 1992; 19: 320–325.
23. Eversley RB, Policar M, White V, et al. Self-reported sexually transmitted diseases among family planning clients: Ethnic differences in sexual risk behavior and HIV risk reduction. Ethn Dis 1993; 3: 181–188.
24. Johnson EH, Jackson LA, Hinkle Y, et al. What is the significance of African American–white differences in risky sexual behavior? J Natl Med Assoc 1994; 86: 745–759.
25. Nyamathi AM, Lewis C, Leake B, Flaskerud J, Bennett C. Barriers to condom use and needle cleaning among impoverished minority female injection drug users and partners of injection drug users. Public Health Rep 1995; 110: 166–172.
26. Grella CE, Annon JJ, Anglin MD. Ethnic differences in HIV risk behaviors, self-perceptions, and treatment outcomes among women in methadone maintenance treatment. J Psychoactive Drugs 1995; 27: 421–433.
27. Schilling RF, el-Bassel N, Schinke SP, Nichols S, Botvin GJ, Orlandi MA. Sexual behavior, attitudes toward safer sex, and gender among a cohort of 244 recovering IV drug users. Int J Addict 1991; 26: 859–877.
28. Scheidt DM, Windle M. The alcoholics in treatment HIV risk (ATRISK) study: Gender, ethnic and geographic group comparisons. J Stud Alcohol 1995; 56: 300–308.
29. Trends in sexual risk behaviors among high school students—United States, 1991–2001. MMWR Morb Mortal Wkly Rep 2002; 51: 856–859.
30. Zenilman JM, Weisman CS, Rompalo AM, et al. Condom use to prevent incident STDs: The validity of self-reported condom use. Sex Transm Dis 1995; 22: 15–21.
31. Fishbein M, Jarvis B. Failure to find a behavioral surrogate for STD incidence—What does it really mean? Sex Transm Dis 2000; 27: 452–455.
32. Shain RN, Perdue ST, Piper JM, et al. Behaviors changed by intervention are associated with reduced STD recurrence: The importance of context in measurement. Sex Transm Dis 2002; 29: 520–529.
33. Morrill AC, Ickovics JR, Golubchikov VV, Beren SE, Rodin J. Safer sex: Social and psychological predictors of behavioral maintenance and change among heterosexual women. J Consult Clin Psychol 1996; 64: 819–828.
34. Zimet GD, Bunch DL, Anglin TM, Lazebnik R, Williams P, Krowchuk DP. Relationship of AIDS-related attitudes to sexual behavior changes in adolescents. J Adolesc Health 1992; 13: 493–498.
35. Shain RN, Piper JM, Newton ER, et al. A randomized, controlled trial of a behavioral intervention to prevent sexually transmitted disease among minority women. N Engl J Med 1999; 340: 93–100.
36. Neff JA, Crawford SL. The Health Belief Model and HIV risk behaviours: A causal model analysis among Anglos, African-Americans and Mexican-Americans. Ethn Health 1998; 3: 283–299.
37. Catania JA, Coates TJ, Stall R, et al. Prevalence of AIDS-related risk factors and condom use in the United States. Science 1992; 258: 1101–1106.
38. Dolcini MM, Catania JA, Coates TJ, et al. Demographic characteristics of heterosexuals with multiple partners: The National AIDS Behavioral Surveys. Fam Plann Perspect 1993; 25: 208–214.
40. Gilliard DK, Beck AJ. Prisoners in 1997. Washington, DC: US Department of Justice, Bureau of Justice Statistics (NCJ 170014), 1998.
© Copyright 2004 American Sexually Transmitted Diseases Association
41. Beck AJ. Trends in US correctional populations. In: Haas KC, Alpert GP, eds. The Dilemmas of Corrections: Contemporary Readings, 4th ed. Prospect Heights, IL: Waveland Press, 1999.