Doherty, Irene A. PhD*†; Minnis, Alexandra PhD†; Auerswald, Colette L. MD, MS‡; Adimora, Adaora A. MD, MPH*; Padian, Nancy S. PhD†
NEARLY HALF OF ALL SEXUALLY transmitted infections (STIs) in the United States occur among young people with substantial racial/ethnic disparities.1–3 In 2003, the incidence of chlamydia among Latinos and blacks, aged 15 to 19, was 1,611 and 5,071 cases per 100,000, respectively, compared with 748 per 100,000 among white adolescents.1 The structure of sexual networks is one of several determinants of STI spread.4 Sexual networks that include concurrent sexual relationships (partnerships that overlap in time) permit expedited STI transmission than do sexual networks comprised solely of sequential partnerships.5–7
Although the prevalence of concurrency has been examined in several adult populations,8–13 studies among adolescents have been limited to in-school youth and clinic patients.14–16 In the largest national school-based study of adolescents,17 of the respondents who reported at least 2 sex partners, more than half of them (54%) had a history of concurrent partnerships.14 Among San Francisco youth recruited from a sexually transmitted disease clinic, 31% of those who had at least one serious sex partner during the preceding 6 months and 69% of those with multiple sex partners had concurrent partners.16 Similarly, 26% of black female adolescents seeking care at an adolescent clinic in Baltimore, Maryland, reported concurrent partners during the previous 3 months; 46% suspected that their male partners had other concurrent partners.15 Empiric data describing concurrency remains limited for Latino and other minority youth who are not in school or do not access adolescent clinics.
Latinos represent 41% of California’s adolescent population.18 The largest population of Latinos in San Francisco reside in the Mission District.19 As seen nationally, Latino youth in San Francisco bear a disproportionate burden of STIs; in 2000, for youth aged 14 to 20, the incidence of chlamydia was 1,064.2 cases per 100,000 among Latinos compared to 634.7 cases among non-Latino whites.20
The Mission District is a culturally vibrant yet impoverished neighborhood because 18% of its residents lived below the 2002 federal poverty line (an annual income of <$18,000 for a family of 4).21 Young people are routinely exposed to excessive homelessness, prostitution, the sale of drugs, and the use drugs.22,23 Street gangs in the Mission District also pose an important safety, social, and public health concern. Most youth in San Francisco, who become involved with gangs, join when they are 12 to 14 years old.21 The violence between the 2 largest rival gangs is intense and concentrated. For crimes committed in the Mission District, 80% of homicides, 60% of assaults, and 48% of robberies are gang-related.21
Gang involvement is associated with having high numbers of sex partners, concurrent sex partners, teen pregnancy, and STIs.24–27 Gang members, for example, accounted for 22% of cases in an outbreak of gonorrhea in Colorado Springs, Colorado, although they comprised less than 1% of the population largely because they made up a densely connected sexual network.24 Another analyses of these data concluded that males joined the gang as a means to find sex partners.25
We designed the Mission Teen Health Project, a 2-year longitudinal study, to investigate the effects of social and sexual networks on the occurrence of STIs and unintended pregnancy among adolescents in San Francisco’s Mission District. We recruited a diversity of youth, including undocumented immigrants, gang members, out-of-school youth, and in-school youth. Here, using data from the baseline visit, we report the prevalence of STIs and concurrent sexual partnerships.
Materials and Methods
Recruitment and Eligibility
Many youth who frequent the Mission District for school or socializing do not reside there or are undocumented immigrants and therefore not accessible through clinic-based or household probability sampling methods. Venue-based sampling has been used successfully to enroll individuals from populations that researchers would not have otherwise reached.28–30 It entails first identifying specific areas where the target population spends time and second, conducting systematic outreach in these locations to recruit study participants. We identified 42 venues frequented by Mission youth such as parks, street corners, and transportation stops.31 To gain credibility in the community, younger (ages 19–31) bilingual research assistants were hired and 7 of the 8 were Latino and bicultural. As people entered the designated venues, staff approached those who appeared to be adolescents, described the study, and determined eligibility.
Eligibility criteria included residence in the San Francisco Bay Area, ability to speak either English or Spanish, and being aged 14 to 19. The eligibility of minors (under 18) also required parental consent. To further infiltrate the social networks of study participants, we allowed their friends to enroll if they met the eligibility criteria. We also recruited participants at youth agencies in the neighborhood to legitimize the project and gain the trust of adolescents who received services at these agencies and spent time in the selected venues.
Study visits were conducted at the field office, the participant’s home, or one of the youth agencies with which we collaborated. Because of the serious safety risks for gang-affiliated youth who crossed into the opposing gang’s territory, the field office was located in an area of the Mission District that local gangs considered neutral.22
Informed consent was obtained directly from study participants aged 18 to 19; minors gave their assent after parental consent was obtained by telephone before the visit. Parents were also sent a thank you letter, 2 copies of the consent form, and a stamped envelope in which to return their signed form. The consent and assent form explained the legal requirement for study staff to report cases of sex between minors under 16 and partners at least age 21, cases of physical or sexual abuse, and positive STI test results.
To examine the strengths and weaknesses of audio computer-assisted self-interviewing (ACASI), a portion of the questionnaire was administered through ACASI to 30% of the participants selected randomly. The remainder of participants was interviewed face-to-face for the entire interview. (At follow-up visits, respondents received the opposite administration mode.) Urine samples were tested for chlamydia, gonorrhea, and pregnancy; blood specimens were tested for herpes simplex virus 2 antibody (HSV-2). All participants received free condoms, health education, referrals, and $35 per visit. The study physician arranged for the treatment of participants who tested positive for STIs. The Committee for Human Research at the University of California–San Francisco approved all procedures.
The Social Cognitive Model of Development of Ethnic Identity32,33 and Problem Behavior Theory34 served as the theoretical framework that shaped questionnaire development. Both models (which complement each other) propose that adolescent risk behavior can be explained, in part, by environmental factors and social and ethnic values. Our formative research also informed the development of the questionnaire. The survey instrument assessed factors hypothesized as related to STI, pregnancy risk, and unsafe sexual behavior, including demographic characteristics (education, U.S. generational status), ethnic attitudes, sexual history, gang involvement, substance use, and religiosity.
From the question “How would you describe your race or ethnicity?,” respondents were allowed to choose multiple categories (Latino/a, black, white, Filipino, Pacific Islander, Asian/Asian American, Native American). In this analysis, mixed-race Latinos were considered Latino. Immigrant generation was determined from the respondents’ and their parents’ country of birth. A roster of household members was enumerated to assess the number of guardians living at home.
Familism refers to important Latino cultural values that include identification with, loyalty and attachment to, and solidarity within one’s family. Familism was assessed with a validated instrument specifically designed for use in Latino populations.35 (Cronbach α scores for the 3 components were 0.70, 0.72, and 0.68.) Responses to 13 items were scored with a 5-point Likert scale and lower scores indicated stronger identification with familism.
Religiosity was assessed by church attendance during the previous 6 months (never, special occasions, monthly, weekly).
We dichotomized alcohol and marijuana use as at least monthly over the previous 6 months versus less frequent or no use. Any use of other substances during the previous 6 months was combined into a single variable that included use of ecstasy, crack, cocaine, heroin, speed, mushrooms, or inhalants.
Participants may have been reluctant to report actual gang membership but more willing to report “affiliation,” meaning that they were friends with gang members.22 A positive response to either “Have you ever belonged to a gang or claimed a color?” or “Have you been a gang affiliate but not actually in the gang?” was classified as gang involvement.
Because all participants were at least age 14 at enrollment, early age of coital debut was defined as onset of vaginal intercourse at age 13 years or younger.
Definition of Concurrent Partnerships
The survey instrument asked detailed questions about the respondent’s last 4 sex partners during the previous 6 months, including dates of first and last intercourse, frequency of intercourse, and condom use. To identify concurrent partnerships, the dates that each partnership began and ended were determined. From the question “When was the very last time you two had vaginal sexual intercourse?,” the complete date (i.e., month, day, and year) was estimated, because participants were able to respond in units of days (e.g., “on Saturday”) or weeks (e.g., “at a party 3 weeks ago”). However, because accurate recall of the exact date of first intercourse was unlikely or could be burdensome, we collected the month and year that each sexual partnership began. Then the first of the month was substituted for the day of first sex because using the 15th of the month generated more dates that failed logic checks.
A total of 110 respondents reported at least 2 sex partners in the previous 6 months. To identify concurrent partnerships, the partnerships were ordered sequentially starting with the earliest partnership. For a respondent who reported 2 partners, the dates of partner 1 and partner 2 were compared. Respondents who reported 3 partners required 3 comparisons of dates: partners 1 and 2, 1 and 3, 2 and 3. Likewise, for respondents with 4 sex partners, a total of 6 comparisons were necessary to detect concurrency: 1 and 2, 1 and 3, 1 and 4, 2 and 3, and 3 and 4. Partnerships were classified as concurrent if the date of first sex with one partner preceded the date of last sex with a different partner by at least 31 days.
Characteristics of Concurrency.
Each instance of concurrency was then further classified as: 1) transitional, defined as one partnership ending before another partnership began36; 2) contained, meaning that one partnership began and ended within the course of another partnership; or 3) infrequent sex when the respondent reported having sex only one to 2 times with one or both partners. For both transitional and contained concurrent partnerships, intercourse occurred at least once a month with both partners.
Concurrency was the outcome variable for logistic regression analysis. Patterns of concurrent partnerships (e.g., condom use) were also described. Because the overarching goal of this analysis was to characterize the patterns and correlates of concurrent sex partners, we restricted the analysis to respondents who were sexually experienced and therefore members of the sexual networks. For the purpose of comparison to other studies, the prevalence of concurrency is also reported for the subgroup of adolescents who had at least 2 partners.14,16 All analyses were stratified by gender.
Independent variables were retained in multivariable models if they were associated with concurrency in bivariate analysis or of a priori interest. Because the prevalence of concurrency did not differ for participants recruited from venues, agencies, or through referral from other participants (Fisher exact tests: P = 0.40 and = 0.75 for males and females, respectively [data not shown]), it was not necessary to account for recruitment source. Because tabular analysis indicated that higher proportions of participants interviewed through ACASI reported high-risk behaviors (alcohol use, marijuana use, and casual sex partners) than did those interviewed face-to-face (data not shown), we controlled for interview mode in all regression models. In-depth analysis of ACASI use over time is planned for subsequent analyses. We used STATA, version 9.0 for all analyses.37
Description of Sample
Between October 2001 and November 2002, a total of 555 youth were enrolled from street-based venues (n = 351), from community agencies (n = 66), and referrals from other study participants (n = 138). Only 9% of parents refused to allow their child to participate.
The majority of participants identified themselves as Latino (75%) and were born in the United States (70%) (Table 1) Blacks composed 16% of the sample. The sample included more first-generation (immigrant) males (37%) than females (24%) because we intentionally recruited day laborers who are predominantly undocumented immigrant males. Two thirds of the sample (n = 368) were sexually experienced; the majority of those who had ever had intercourse also had at least one sex partner in the past 6 months (n = 321 [87%]). Among sexually experienced youth, the prevalence of chlamydia was 3.3%, <1% for gonorrhea, and 1.4% for HSV-2 (Table 1).
Among youth who had been sexually active in the previous 6 months, significantly more females (62%) reported a single sex partner than did males (41%) (P <0.0001). For each respondent, who had at least one sex partner in the preceding 6 months, we computed the proportion of partners whom they considered casual from the total number reported, which was significantly higher on average for males (45%) than females (24%) (P <0.0001). Slightly more than one third (37%) of the sexually active youth reported consistent condom use with all of their sex partners, but it was more prevalent among males (44%) than females (31%) (P <0.0001).
Prevalence of Concurrency.
The 6-month prevalence of concurrency among sexually experienced youth was 20% overall, 27% among males, and 12% among females (Table 2). Controlling for interview mode, males were 2.6 times as likely as females (odds ratio = [OR] 2.6; 95% confidence interval [CI] = 1.5–4.5; not shown) to have concurrent partnerships. Among youth who had multiple partners in the previous 6 months, concurrency prevalence increased to 64% overall; insufficient power inhibited our ability to detect differences in prevalence between males (65%) and females (48%) (P = 0.4) (Table 2, second row).
Patterns of Concurrent Partnerships
Statistical power was also not sufficient to identify other differences in patterns of concurrent partnerships by gender (Table 2). Respondents with 2 or more overlapping partnerships could have more than one type of concurrency (i.e., transitional, contained, infrequent sex), thereby causing sum of column percentages to exceed 100% (Table 2). Males had a greater tendency to be in transitional concurrent partnerships (62%) than did females (35%). Most concurrent partnerships for females (57%) involved one or 2 episodes of sex with one or both partners in contrast to 38% of males. More males reported consistent condom use with both partners (58%) than did females (35%).
Correlates of Concurrency
In unadjusted analysis, males who received the ACASI survey were more than 3 times as likely as those interviewed face-to-face (OR = 30.64; 95% CI = 1.79–7.42) to have concurrent sex partners (Table 3). Concurrency among females, however, was not affected interview mode.
Third-generation males (i.e., U.S.-born with U.S.-born parent[s]) were 4 times as likely as first-generation males (OR = 4.2; 95% CI = 1.75–10.13 to have concurrent partnerships. The likelihood of concurrency was also greater for males who had early onset of sexual intercourse (OR = 2.85; 95% CI = 1.42–5.72), had increasing numbers of sexual partners (OR = 1.46; 95% CI = 1.24–1.71), or smoked marijuana once a month or more frequently (OR = 2.63; 95% CI = 1.33–3.55) (Table 3). However, living with at 2 less adults (OR = 0.39; 95% CI = 0.18–0.85) and weekly church attendance (OR = 0.34; 95% CI = 0.12–0.93) lowered the likelihood of males having concurrent partnerships. In the multivariable model, U.S.-born males with early onset of sexual intercourse and more sex partners persisted as independent risk factors associated with concurrency (Table 4). In a separate analysis (not shown) of the subsets of males who had at least one sex partner in the past 6 months, those who were below grade level for their age were 2 times more likely to have concurrent partners (OR = 2.2; 95% CI = 0.98–4.7).
Among females, the likelihood of concurrency increased with age (OR = 1.4; 95% CI = 1.0–1.9), number of sex partners (OR = 5.9; 95% CI = 3.0–11.4), alcohol use (OR = 2.9; 95% CI = 1.2–7.0), and use of illegal substances (other than marijuana) (OR = 3.9; 95% CI = 1.6–9.7) (Table 3). Conversely, adoption of familism reduced the risk of concurrency (OR = 3.0; 95% CI = 1.0–9.0).
Among sexually experienced females, familism scores and the number of sex partners were correlated (correlation coefficient 0.28, P = 0.0001); females who did not identify with the values related to familism also had more sex partners. When both familism scores and number of sex partners were included in the model, the effect of familism disappeared. Therefore, 2 multivariable models are presented in Table 4 to assess the association of each factor with concurrency. In each model, females’ age and substance use remained associated with having concurrent sex partners (Table 4). The odds of concurrency increased with increasing numbers of sex partners (model 1; OR = 5.6; 95% CI = 2.84–10.9). Likewise, the lack of identity with familism increased the likelihood of concurrency (model 2; OR = 3.85; 95% CI = 1.19–12.4).
Although the relationship between concurrency and gang exposure was not associated with concurrency in the adjusted model for females, gang exposure itself was strongly associated with alcohol and illegal substance use (P <0.0001, data not shown). Among females who drank alcohol at least monthly and/or used illegal substances, 76% were also gang members or affiliates (OR for gangs = 5.5; 95% CI = 2.6–11.6;, not shown). Thus, the adjusted measure of effect between substance use and concurrency in both multivariable models may have obscured the effect of gang exposure. Furthermore, the power to detect an effect for gang exposure was compromised, because only 13 females reported it.
Through the Mission Teen Health Project, we studied sexual behaviors of adolescents who resided in or frequented a predominantly Latino neighborhood of San Francisco. Two thirds of respondents (66%) were sexually experienced, which substantially surpasses the national estimate of 46%.38,39 The majority of sexually experienced youth (87%) had at least one sex partner during the previous 6 months. Approximately one third (37%) of the sexually active youth reported consistent condom use with all of their sex partners.
Like found in other studies of adults and adolescents, the 6-month prevalence of concurrency was 20% among sexually experienced respondents8,14–16,40 and 3 times as likely among males as females (27% males vs. 12% females). Consistent with another study of San Francisco youth, concurrency prevalence rose to 58% for the subset of youth who reported at least 2 partners during the previous 6 months.16
Correlates of concurrency differed by gender. U.S.-born males, who had their first vaginal sexual experience before age 13 and had more sex partners were most likely to have concurrent partnerships. Most of their concurrent partnerships involved ongoing sex with both partners usually as they transitioned from one partnership to another partnership. Additionally, male youth with recent sexual activity (in the past 6 months) and below grade level were also more likely to have concurrent partnerships. A small group of older females who used drugs were probably involved with gangs, and did not identify with familism, were most at risk for having concurrent partnerships characterized by one to 2 episodes of sexual intercourse with one partner during a partnership with someone else.
This study has limitations. Venue-based sampling is not representative of the larger population from which the sample was drawn or Latino youth elsewhere.41 Our objective, however, was to describe features of sexual networks in an urban Latino neighborhood among youth who are often inaccessible through recruitment in schools or clinics. Venue-based sampling allowed us to tap into these hidden networks and describe them.
Several factors may have led to an underestimate of concurrency. First, laws requiring us to notify Child Protective Services of sexual activity between individuals less than 16 years old and adults 21 years or older may have made participants reluctant to report some of their partnerships. Second, we collected data on a maximum of 4 partnerships during the preceding 6 months, but 24 respondents reported more than 4 partners. Concurrent partnerships were identified from the data available for 3 of the 4 females and 11 of the 20 males who reported at least 5 sex partners. Third, the precision and accuracy of dates of first and last sex were probably compromised based on the frequency of sex and the time that elapsed between the dates of sex and the interview date. A simulation study that used data from 5 studies to estimate the effects of reporting errors on measurement of concurrent partnerships suggested that 80% of concurrent partnerships are detected.42
Last, the findings presented here were derived from egocentric data and assessment of the prevalence of concurrency at the individual level is merely a crude proxy of the underlying structure of sexual networks. Ideally, a sociometric study that depicts all sexual links in the target population would provide a more complete understanding of STI transmission. Despite this inherent flaw in egocentric studies, individual-level analyses of concurrency prevalence has the potential to enhance the understanding of its role in STI transmission. These types of analyses, for example, have been used to explain the growing epidemic of heterosexual HIV transmission among blacks in the southeastern United States.9,10,43–45
The combination of rapid partnership change, substantial concurrency, inconsistent condom use, and gang-related activity suggest ample opportunities for rapid STI dissemination. Nonetheless, the STI prevalence in this sample was low. Why might this be the case? First, STI prevalence among young people tends to be highest for youth in late adolescence or early adulthood.1 Participants were generally younger; the median age was 16 (not shown). Second, in addition to individual characteristics, the likelihood of STI acquisition depends on the underlying disease prevalence, the period of infectivity, the risk profile of others in the population, and the properties of the sexual networks. The intrinsic prevalence of STI in the Mission District was sufficiently low relative to other neighborhoods in San Francisco20 and levels of condom use, although not extensive, were apparently adequate to curtail transmission among these young people.46 The high rates of concurrent partnerships, however, suggest that an influx of STI into the community would markedly increase STI prevalence among Mission District youth.
1. Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance, 2003. Atlanta: US Department of Health and Human Services, 2004.
2. Miller WC, Ford CA, Morris M, et al. Prevalence of chlamydial and gonococcal infections among young adults in the United States. JAMA 2004; 291:2229–2236.
3. Weinstock H, Berman S, Cates W Jr. Sexually transmitted diseases among American youth: Incidence and prevalence estimates, 2000. Perspect Sex Reprod Health 2004; 36:6–10.
4. Aral S. Sexual network patterns as determinants of STD rates: Paradigm shift in the behavioral epidemiology of STDs made visible. Sex Transm Dis 1999; 26:262–264.
5. Koumans EH, Farley TA, Gibson JJ, et al. Characteristics of persons with syphilis in areas of persisting syphilis in the United States: Sustained transmission associated with concurrent partnerships. Sex Transm Dis 2001; 28:497–503.
6. Morris M, Kretzschmar M. Concurrent partnerships and the spread of HIV. AIDS 1997; 11:641–648.
7. Potterat JJ, Zimmerman-Rogers H, Muth SQ, et al. Chlamydia transmission: Concurrency, reproduction number, and the epidemic trajectory. Am J Epidemiol 1999; 150:1331–1339.
8. Adimora AA, Schoenbach VJ, Bonas DM, Martinson FE, Donaldson KH, Stancil TR. Concurrent sexual partnerships among women in the United States. Epidemiology 2002; 13:320–327.
9. Adimora AA, Schoenbach VJ, Martinson F, Donaldson KH, Stancil TR, Fullilove RE. Concurrent sexual partnerships among African Americans in the rural south. Ann Epidemiol 2004; 14:155–160.
10. Adimora AA, Schoenbach VJ, Martinson FE, Donaldson KH, Stancil TR, Fullilove RE. Concurrent partnerships among rural African Americans with recently reported heterosexually transmitted HIV infection. J Acquir Immun Defic Syndr 2003; 34:423–429.
11. Fenton KA, Mercer CH, McManus S, et al. Ethnic variations in sexual behaviour in Great Britain and risk of sexually transmitted infections: A probability survey. Lancet 2005; 365:1246–1255.
12. Johnson AM, Mercer C, Erens B, et al. Sexual behavior in Britain: Partnerships, practices, and HIV risk behaviors. Lancet 2001; 358:1835–1842.
13. Manhart LE, Aral SO, Holmes KK, Foxman B. Sex partner concurrency: Measurement, prevalence, and correlates among urban 18–39-year-olds. Sex Transm Dis 2002; 29:133–143.
14. Ford K, Sohn W, Lepkowski J. American adolescents: Sexual mixing patterns, bridge partners, and concurrency. Sex Transm Dis 2002; 29:13–19.
15. Jennings J, Glass B, Parham P, Adler N, Ellen JM. Sex partner concurrency, geographic context, and adolescent sexually transmitted infections. Sex Transm Dis 2004; 31:734–739.
16. Rosenberg M, Gurvey J, Adler N, Dunlop M, Ellen J. Concurrent sex partners and risk for sexually transmitted diseases among adolescents. Sex Transm Dis 1999; 26:208–212.
17. Resnick MD, Bearman PS, Blum RW, et al. Protecting adolescents from harm. Findings from the National Longitudinal Study on Adolescent Health. JAMA 1997; 278:823–832.
18. Brindis C, Ozer E, Adams S, et al. Health Profile of California’s Adolescents: Findings From the 2001 Health Interview Survey. Los Angeles: UCLA Center for Health Policy Research, 2004.
19. Guzman B. The Hispanic Population Census 2000 Brief. US Census Bureau, May 2001, Report No. C2KBR001-3.
20. Sexually Transmitted Disease Annual Summary. San Francisco Department of Public Health, 2001.
21. Bennett R, Abrams B, Reyes P, Tharp J, Bennett P. San Francisco’s Gang-Free Communities Initiative Assessment Report. San Francisco: San Francisco Department of Children, Youth and Their Families, San Francisco Juvenile Probation Department, Mayor’s Office of Criminal Justice, 2002.
22. Auerswald C, Greene K, Minnis A, Doherty IA, Ellen J, Padian N. Qualitative assessment of venues for purposive sampling of hard-to-reach youth: An illustration in a Latino community. Sex Transm Dis 2004; 21:133–138.
23. Epele ME. Gender, violence and HIV: Women’s survival in the streets. Cult Med Psychiatry 2002; 26:33–54.
24. Centers for Disease Control and Prevention. Gang-related outbreak of penicillinase-producing Neisseria gonorrhoeae
and other sexually transmitted diseases—Colorado Spring, Colorado, 1989–1991. Morbid Mortal Wkly Rep 1993; 42.
25. Palmer C, Tilley C. Sexual access to females as a motivation for joining gangs: An evolutionary approach. J Sex Res 1995; 32:213–217.
26. Voisin DR, Salazar LF, Crosby R, DiClemente RJ, Yarber WL, Staples-Horne M. The association between gang involvement and sexual behaviours among detained adolescent males. Sex Transm Infect 2004; 80:440–442.
27. Wingood GM, DiClemente RJ, Crosby R, et al. Gang involvement and the health of African American female adolescents. Pediatrics 2002; 110:e57.
28. MacKellar D, Valleroy L, Karon J, Lemp G, Janssen R. The Young Men’s Survey: Methods for estimating HIV seroprevalence and risk factors among young men who have sex with men. Public Health Rep 1996; Suppl 1:138–144.
29. van Griensven F, Thanprasertsuk S, Jommaroeng R, et al. Evidence of a previously undocumented epidemic of HIV infection among men who have sex with men in Bangkok, Thailand. AIDS 2005; 19:521–526.
30. Weir SS, Pailman C, Mahlalela X, Coetzee N, Meidany F, Boerma J. From people to places: Focusing AIDS prevention efforts where it matters most. AIDS 2003; 17:895–903.
31. Minnis A, Auerswald C, Doherty IA, Ellen J, Shiboski S, Padian N. Qualitative and quantitative methods for developing a venue-based sampling approach for a study of Latino adolescents’ social and sexual networks. In: The Network Paradigm in Research on Drug Abuse, HIV, and Other Blood-Borne and Sexually Transmitted Infections: New Perspectives, Approaches, and Applications: NIDA/CAMCODA Working Meeting, May 3–4, 2001; 2002, National Institute of Drug Abuse, National Institutes of Health.
32. Knight GP, Bernal ME, Cota MK, Garza CA, Ocampo KA. Family socialization and Mexican American identity and behavior. In: Martha E, Bernal E, George P, et al, eds. Ethnic Identity: Formation and Transmission Among Hispanics and Other Minorities. State University of New York Press, 1993:viii, 311.
33. Knight GP, Bernal ME, Garza CA, et al. Family socialization and the ethnic identity of Mexican-American children. Journal of Cross-Cultural Psychology 1993; 24:99–114.
34. Jessor R, Jessor SL. Problem Behavior and Psychosocial Development: A Longitudinal Study of Youth. New York: Academic Press, 1977.
35. Sabogal F, Marin G, Otero-Sabogal R. Hispanic familism and acculturation: What changes and what doesn’t? Hispanic Journal of Behavioral Sciences 1987; 9:397–418.
36. Gorbach PM, Stoner BP, Aral SO, WL HW, Holmes KK. ‘It takes a village’: Understanding concurrent sexual partnerships in Seattle, Washington. Sex Transm Dis 2002; 29:453–462.
37. StataCorp. Stata Statistical Software, Release 9.0. College Station, TX: StataCorp., 2005.
38. Amba J, Martinez G, Mosher W, Dawson B. Teenagers in the United States: Sexual activity, contraceptive use, and childbearing, 2002. Vital Health Stat 2004; 23.
39. Biddlecom AE. Trends in sexual behaviours and infections among young people in the United States. Sex Transm Infect 2004; 80(suppl 2): ii74–ii79.
40. Manhardt L, Aral S, Holmes K, Foxman B. Sex partner concurrency measurement, prevalence, and correlates among urban 18–39-year-olds. Sex Transm Dis 2002; 29:133–143.
41. Sarmiento OL, Miller WC, Ford CA, et al. Disparities in routine physical examinations among in-school adolescents of differing Latino origins. J Adolesc Health 2004; 35:310–320.
42. Brewer DD, Rothenberg RB, Muth SQ, Roberts JM Jr, Potterat JJ. Agreement in reported sexual partnership dates and implications for measuring concurrency. Sex Transm Dis 2006; 33:277–283.
43. Adimora A, Schoenbach V, Doherty I. Sexual network components, social context, and HIV among African Americans in the southern United States. Sex Transm Dis 2006; 33:S39–S45.
44. Adimora A, Schoenbach V, Martinson F, et al. Heterosexual HIV transmission among African Americans in North Carolina. J Acquir Immun Defic Syndr 2006; 41:617–623.
45. Adimora AA, Schoenbach VJ, Doherty IA. HIV and African Americans in the southern United States: Sexual networks and social context. Sex Transm Dis 2006; 33(suppl):S39–S45.
46. Rothenberg RB, Potterat JJ, Woodhouse DE. Personal risk taking and the spread of disease: Beyond core groups. J Infect Dis 1996; 174(suppl 2):S144–S149.