Auerswald, Colette L. MD, MS*; Muth, Stephen Q. BA†; Brown, Beth MS, MPA‡; Padian, Nancy PhD‡§∥; Ellen, Jonathan MD¶
From the *Division of Adolescent Medicine, Department of Pediatrics, the ‡Department of Obstetrics and Gynecology, and the ∥Center for Reproductive Health Research, University of California at San Francisco, San Francisco, California; †independent consultant, Colorado Springs, Colorado; the §School of Public Health, University of California, Berkeley, Berkeley, California; and the ¶Division of General Pediatrics and Adolescent Medicine, Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, Maryland
The authors thank Ali Minnis, John Potterat, Mary Ott, and Matthew Kramer for their editorial suggestions. This endeavor would not have been possible without the cooperation of our adolescent subjects, for which the authors are grateful.
This research was supported by an award from NIAID (U01 AI47639) and from NICHD (1K23 HD01490). Research assistants Edward Cruz, Shira Rutman, and Drina Boban, MHA, interviewed and tracked the hard-to-reach subjects.
Correspondence: Colette L. Auerswald, MD, MS, Assistant Professor of Pediatrics, Division of Adolescent Medicine, University of California at San Francisco, Box 0503, San Francisco, CA 94143-0503. E-mail: firstname.lastname@example.org
Received for publication August 2, 2005, and accepted December 13, 2005.
AFRICAN AMERICAN ADOLESCENTS EXPERIENCE DISPROPORTIONATE rates of sexually transmitted infections (STIs) relative to their white and Latino counterparts.1 Furthermore, African American females 15 to 19 years of age suffer disproportionately from STIs relative to their male counterparts. Nationally, African American females 15 to 19 years old have almost twice the rate of gonorrhea than their male peers (3495 vs. 1794 cases per 100,000).1 In San Francisco, the sex differential is similar (3363 vs. 1476 cases per 100,000).2 Although young females' increased biologic susceptibility to STIs has been documented,3,4 differences in sexual behavior are generally postulated to explain differences in rates of STIs. However, in the case of African American adolescents, differences in risky sexual behavior do not appear to explain differential morbidity by gender. The 2003 Youth Risk Behavior Survey found that 9th to 12th grade African American males were more likely than their female peers to have had sexual intercourse (73.8% vs. 60.9%), to have initiated sexual activity before the age of 13 years (31.8% vs. 6.9%), to have had greater than 3 lifetime sexual partners (41.7% vs. 16.3%), and to be currently sexually active (54.0% vs. 44.2%).5 These results would suggest a priori that African American adolescent males should have higher rates of STIs than their female counterparts rather than the opposite.
One explanation for this apparent paradox is that the difference in rates is the result of a difference in partner characteristics.6 Demographic and behavioral characteristics of partners have been shown to increase the risk for STIs. Having an older sex partner is correlated with earlier coiarche,7,8 decreased condom use,9,10 and increased risk of STIs11,12 and pregnancy.10,13,14 Given that incarceration15–17 and gang membership18 are associated with increased STI risk, individuals with partners who have been incarcerated or are members of a gang are at increased risk themselves for STIs as a result of greater likelihood of exposure to an infected partner.
Similarly, partner race has been postulated to affect likelihood of exposure and thus risk for STIs. Laumann and colleagues have suggested that African American adults are at higher risk for STI acquisition relative to adults of other ethnic groups because of greater proximity to members of core sexual networks characterized by multiple sexual partners.19 Ellen et al have suggested that the differential risk for STIs in African American adolescents in San Francisco may be the result of their proximity to core network members.20 Finally, Potterat et al have suggested that youth who select partners nonassortatively (i.e., teens who select partners of a different race) were at highest risk for an STI.21
Studies of partner selection and STIs in adolescents to date have only been conducted in school-based or high-risk clinic samples22,23 and have rarely been linked to biologically confirmed STIs. Little is known about whether sex differences in sexual partner characteristics explain sex differences in rates of STIs among a nonclinic sample of school-aged youth. Knowledge of the contribution of sexual partner characteristics to STI risk could inform interventions to prevent STIs among school-aged youth residing in high-prevalence communities. To the extent that partner characteristics matter, then efforts to reduce STIs in youth residing in high-prevalence communities would need to focus on sexual partner selection and on sex partners and their networks rather than continuing to only focus on individual behavior change such as condom use.
The objective of this study was to determine whether the demographic and behavioral characteristics of adolescents' sexual partners mediate the sex differences in rates of STIs.
Materials and Methods
From July 2000 to August 2001, we recruited a random digit dial telephone household sample of 14- to 19-year-old African American youth who reside in the Bayview Hunter's Point District of San Francisco, California, a predominantly African American, economically marginalized and sociogeographically segregated district of San Francisco. The sample of telephone numbers was generated using a procedure called list-assisted random digit sampling. Approximately 40,000 phone numbers were dialed, of which 22,000 were identified as households. Ninety percent of households were excluded because they were out of the target area or did not host an eligible adolescent. A final sample of 673 teens living in 470 eligible households was identified. After excluding 93 duplicated respondents, 580 household youth were offered the interview. Three hundred fifty of the 580 household youth were interviewed (60.3% response rate). The distribution of the nonresponse rate among RDD index teens was: 20% guardian refusal, 13% respondent refusal, 4% informant refusal, and 3% respondent never available. Male adolescents and younger adolescents had higher refusal rates.
The focus of this article is on the adolescents who had had sex in the 3 months before recruitment. Of the 350 youth (205 females) in the cohort, 166 (99 female) had had sex in the prior 3 months, 133 (83 female) were virgins, 50 (23 female) had had sex outside of the 3 months, and one male youth declined to answer the question.
Consent for participation was obtained from parents for all minor participants and directly from participants 18 years or older for participation in an adolescent health study that included interviews and STD testing for all participants. Minor study participants gave assent. Participating youth were interviewed and tested for Neisseria gonorrhea (GC) and Chlamydia trachomatis (CT). Interviews were conducted by telephone using computer-assisted telephone interview (CATI) technology. Eligible adolescents not reached by phone were located and interviewed by a field team using home visits, word-of-mouth referral, or information from public records, key informants, and community leaders. Field team interviews were also administered using CATI technology. Youth undergoing STI testing were met in the field by study staff who instructed youth on the use of tests and provided written instructions. Testing took place at a convenient location of the participant's choice such as their home, school, or workplace. Youth were reimbursed $25 for completion of the CATI interview and $10 for sexually transmitted disease testing.
The University of California at San Francisco Institutional Review Board approved all study procedures.
Participants reported their age and the number and names of up to 6 of their sexual partners in the 3 months before being interviewed. Youth were asked to provide the names of partners because partners were subsequently offered participation in the study. Participants had the option of not giving a partner's name or not allowing a partner to be contacted. Only data from the original participants, not partner data, are reported here. Number of partners reported by the participant was treated as a continuous variable.
Partner age was reported by the participant. Age differences between the participants and each partner were calculated and dichotomized as greater than or equal to 5 years older (partner older than participant) or absolutely less than 5 years older (including cases when the partner is younger than the participant). The literature defines an older partner variously as greater than 2 years older and 5 years older or greater. Clinically, we are more concerned about the 15 year old partnering with the 20 year old than with a 17 year old so we chose the latter definition. Analyses performed with an age difference of 2 years yielded the same results.
Partner race was reported by the participant and then dichotomized as African American or not. This dichotomization was performed to measure nonassortative mixing by our all-African American index sample.
Partner incarceration was reported by the participant by asking: “Was [partner] ever in jail or juvenile hall?”
Partner gang membership was reported by the participant by asking: “Was [partner] ever in a gang or neighborhood crew?” “Neighborhood crew” is another term for gangs in the community.
Level of sexual activity of the partners was determined by asking: “Did [partner] have any other sex partners during the last 3 months?”
CT and GC infection were determined by obtaining a first-void urine from males or a self-administered vaginal swab from females, which were tested by ligase chain reaction (LCx; Abbott). STI tests were not obtained from 12.1% of the respondents because they could not be located or refused to provide a specimen after being interviewed by phone or in person. These missing results were recoded as negatives to bias effect estimates toward the null (results were the same when analyses were conducted with the missing tests deleted). Participants who tested positive for CT or GC were referred for care to their primary care provider or to a low cost or free adolescent-friendly clinic.
Approach to Multiple Dyads per Participant.
Because we were interested in the summation of risk at the individual level, not the risk of individual partnerships at the dyadic level, we created a single record per index participant. The single record summarized the number of partners, the number of partners that were reported to have the specified characteristic (race of partner, older partner, perceived gang membership, perceived incarceration, and having other partners in specified period or not). We then recoded this to record whether the index had any partner with the specified characteristic.
To test our hypothesis that the relationship between gender (our independent variable) and STI rates (our dependent variable) was mediated by partner characteristics (such as partner race, age, perceived incarceration or gang membership, or level of sexual activity), we performed the following 4-step regression approach to testing mediation effects as per Holmbeck24 using logistic regression. Our approach is also summarized in Figure 1.
1. We determined that there were differences in the rate of STI diagnosis by gender even once the analyses were controlled for the potential confounding effect of number of sex partners;
2. We determined that there were gender differences in sexual partner characteristics (i.e., a difference by gender in having any partner with the characteristic of interest) and in number of partners;
3. We determined which sex partner characteristics were significantly associated with the diagnosis of STIs, controlling for number of sex partners; and
4. We demonstrated that the difference in the rate of STI diagnosis by gender is decreased when controlling for having any partner with a significant partner characteristic, even when controlling for the effect of partner number.
The cohort included 99 sexually active female participants and 67 sexually active male participants of similar age (mean age, 17.6 years for females and 17.4 for males). All cohort members were African American. The mean number of partners reported by the males over the 3 months before being interviewed was 2.0 (range, 1–7). This was significantly greater than the mean of 1.3 partners reported by the females (range, 1–5; Wilcoxon 2-sample exact t test, P <0.0001). Only one youth reported more than 6 partners. No youth reported same sex sexual activity.
Gender Differences in Sexually Transmitted Infections
The overall rate of positive STIs was 9.2%. As stated above, we were not able to collect tests from 12.2% of the sample. The rate of STIs was higher among females (12.1%; 12 of 99) than among males (4.5%; 3 of 67). The difference was not statistically significant (odds ratio [OR] = 1.31, 95% confidence interval [CI] = 0.95–1.81). However, when we controlled for number of sex partners, the difference bordered on significance (OR = 1.39, 95% CI = 0.98–1.96; P = 0.07).
Gender Differences in Sexual Partner Characteristics
Partner demographics and high-risk characteristics by gender are summarized in Table 1. There were significant sex differences in demographic characteristics of partners. Female respondents were significantly more likely than male respondents to have an older partner (OR = 1.48; 95% CI = 1.08–2.03) and significantly less likely to have a non-African American partner (OR = 0.63; 95% CI = 0.51–0.77). Furthermore, female respondents were nearly twice as likely as male respondents to report that at least one sexual partner had been incarcerated (OR = 1.86; 95% CI = 1.50–2.29). There was no sex difference in rates of partners perceived to have a history of gang membership or in partners perceived to have had other partners.
Sexual Partner Characteristics and Sexually Transmitted Infections
As shown in Table 2, the only partner characteristic associated with STI diagnosis was history of incarceration (OR = 6.56; 95% CI = 1.77–24.27). Number of sexual partners was not associated with STI diagnosis (OR = 1.09; 95% CI = 0.69–1.73).
Impact of Sexual Partner Characteristics on Gender Differences in Sexually Transmitted Infections
Because the only partner characteristic to be associated with both gender and STI diagnosis was history of incarceration, we did not test the mediation effects of the other partner characteristics. Controlling for partner incarceration and number of partners, the sex difference in STI rates was eliminated (OR = 1.07; 95% CI = 0.70–1.63).
This study showed that, although there were several gender differences in sexual partner characteristics, only having had a partner with a history of incarceration mediated gender differences in rates of STIs. This finding supports, although does not prove, our hypothesis that the higher rate of STIs among adolescent females relative to adolescent males, despite males' greater sexual risk behaviors, is related to partner selection.
In network terms, this suggests that the risk for contracting an STI is determined more by the risk of the partner than by the risk of the participant. Incarceration is associated with an increased risk of STIs and HIV.15–17,25,26 Incarceration may be a marker for STI risk and, in particular, for being a member of a core group for infection. Incarceration was correlated with having concurrent partners in a study of the partners of rural African Americans.27 As stated earlier, Laumann et al and Ellen et al have suggested that African American adults and adolescents are at higher risk for STI acquisition because of greater proximity to members of core sexual networks.19,20 This study supports these conclusions based on a random digit dial household sample of African American adolescents.
There are several limitations to our findings. First, the gender difference in the rate of STIs controlling for the number of sexual partners was of borderline significance. We report this finding nevertheless because of the small sample and the exploratory nature of this analysis. Second, the data regarding the partners are reported by the participants and not by the partners themselves. Third, the reporting may be subject to recall bias because of the time period (3 months) over which our adolescent participants were asked to report their activities. There may also be reporting bias as a result of the stigmatizing nature of the behaviors under study such as gang membership and incarceration of partners. Although the population-based sample is a strength of the study, youth without phones in the home could not be included in this random digit dial sample. However, the extensive telephone coverage of the Bayview District leads us to believe that the effect of these missing youth was not substantial. Finally, there were missing STI testing data on 12.1% of youth and, as mentioned previously, 40% of eligible adolescents failed to participate. The greater participation by females reflects the greater likelihood that females would be found at home versus males and the greater likelihood of male versus female refusals. Without data on the risk behavior of the male refusals, it is difficult to guess how this may have biased our findings. These data may have changed the overall findings, although as we have stated, we believe that by coding missing tests as negative, we biased our effect estimates toward the null hypothesis.
In conclusion, our findings suggest that African American females in an economically deprived community report significantly different patterns of partner selection relative to their male age mates, differences that appear to partially mediate their risk of STIs. Differences in individual behavior such as partner number did not alter the sex difference in STI rates. Sex differences in STI rates among African American adolescents may in part be the result of social contextual factors such as differences in partner characteristics. These differences may provide clues to interventions to improve the health of all African American adolescents.
1. Centers for Disease Control, Division of Sexually Transmitted Disease. Sexually Transmitted Disease Surveillance Report, 2001: STDs in Racial and Ethnic Minorities. Atlanta: Centers for Disease Control, 2002.
2. Division of STD Control. San Francisco Sexually Transmitted Disease Annual Summary, 2001. San Francisco: San Francisco Department of Public Health, July 2002.
3. Critchlow C, Wolner-Hanssen P, Eschenbach D, et al. Determinants of cervical ectopia and of cervicitis: Age, oral contraception, specific cervical infection, smoking and douching. Am J Obstet Gynecol 1995; 173:534–543.
4. Ellen J, Lammel C, Shafer M, et al. Cervical secretory immunoglobulin A in adolescent girls. J Adolesc Health 1999; 25:150–154.
5. Centers for Disease Control and Prevention. Youth risk behavior surveillance—United States, 2003. MMWR Morb Mortal Wkly Rep 2004; 53(S5-2):17–19.
6. Zenilman JM. Gonorrhea, chlamydia and the sexual network [Editorial]. Sex Transm Dis 2000; 27:224–225.
7. Rickert V, Wiemann C, Berenson A. Health risk behaviors among pregnant adolescents with older partners. Arch Pediatr Adolesc Med 1997; 151:276–280.
8. Marin B, Coyle K, Gomez C, et al. Older boyfriends and girlfriends increase risk of sexual initiation in young adolescents. J Adolesc Health 2000; 27:409–418.
9. DiClemente R, Wingwood G, Crosby R, et al. Sexual risk behaviors associated with having older sex partners: A study of African American adolescent females. Sex Transm Dis 2002; 29:20–24.
10. Miller KS, Clark L, Moore J. Sexual initiation with older male partners and subsequent HIV risk behavior among female adolescents. Fam Plann Perspect 1997; 29:212–214.
11. Catania JA, Binson D, Stone V. Relationship of sexual mixing across age and ethnic groups to herpes simplex virus-2 among unmarried heterosexual adults with multiple sexual partners. Health Psychol 1996; 15:362–370.
12. Aral SO. Sexual mixing patterns in the spread of gonococcal and chlamydial infections. Am J Public Health 1999; 89:825–833.
13. Darroch J, Landry D, Oslak S. Age differences between sexual partners in the United States. Fam Plann Perspect 1999; 31:160–167.
14. Taylor D, Chavez G, Adams E, et al. Demographic characteristics in adult paternity for first births to adolescents under 15 years of age. J Adolesc Health 1999; 24:251–258.
15. Pack RP, Diclemente RJ, Hook EW III, et al. High prevalence of asymptomatic STDs in incarcerated minority youth: A case for screening. Sex Transm Dis 2000; 27:175–177.
16. Oh K, Cloud G, Wallace L, et al. Sexual behavior and sexually transmitted diseases among male adolescents in detention. Sex Transm Dis 1994; 21:127–132.
17. Mertz K, Finelle L, Levine W, et al. Gonorrhea in male adolescents and young adults in Newark, New Jersey. Sex Transm Dis 2000; 27:201–207.
18. Joseph-Di Caprio J. Comparisons and determinants of high-risk behaviors among incarcerated and public school adolescents. Poster presented at the Society for Adolescent Medicine Annual Meeting, 1995.
19. Laumann E, Youm Y. Racial/ethnic group differences in the prevalence of sexually transmitted diseases in the United States: A network explanation. Sex Transm Dis 1999; 26:250–261.
20. Ellen J, Hessol N, Kohn R, Bolan G. An investigation of geographic clustering of repeat cases of gonorrhea and chlamydial infection in San Francisco, 1989–1993: Evidence for core groups. J Infect Dis 1997; 175:1519–1522.
21. Potterat J, Muth S, Rothenberg R, et al. Sexual network structure as an indicator of epidemic phase. Sex Transm Infect 2002; 78(suppl I):i152–i158.
22. Ford K, Sohn W, Lepkowski J. American adolescents: Sexual mixing patterns, bridge partners and concurrency. Sex Transm Dis 2002; 29:13–19.
23. Rosenberg MD, Gurvey JE, Adler N, et al. Concurrent sex partners and risk for sexually transmitted diseases among adolescents. Sex Transm Dis 1999; 26:208–212.
24. Holmbeck G. Toward terminological, conceptual and statistical clarity in the study of mediators and moderators: Examples from the child—clinical and pediatric psychology literatures. J Consult Clin Psychol 1997; 65:599–610.
25. Kahn R, Voigt R, Swint E, et al. Early syphilis in the United States identified in corrections facilities: 1999–2002. Sex Transm Dis 2004; 31:360–364.
26. Hightow L, Miller W, Leone P, et al. Predictors of repeat testing and HIV seroconversion in a sexually transmitted disease clinic population. Sex Transm Dis 2004; 31:455–459.
27. Adimora A, Schoenbach V, Martinson F, et al. Concurrent sexual partnerships among African Americans in the rural south. Ann Epidemiol 2004; 14:155–160.