Heterosexual contact is the most common mode of HIV transmission among women in the United States.1 Sexual partnerships are the building blocks of networks through which HIV and other sexually transmitted infections (STIs) disseminate. The sequencing, interval between and duration of partnerships, and frequency of sexual interactions influence individual risk of infection and population transmission. Concurrent sexual partnerships (partnerships that overlap in time) can accelerate the spread of HIV infection.2–4 Concurrency confers different levels of risk to the different members of the partnership. Individuals who have concurrent partnerships can more easily transmit infection but have no greater risk of acquiring it than the risk associated with having multiple consecutive partnerships. The partners of individuals who have concurrent partnerships, however, are at increased risk of acquiring infection because of their partner’s concurrency. Partners’ concurrency was an independent risk factor for HIV infection among African American men and women with heterosexually transmitted HIV infection, who lacked traditional high-risk characteristics.5 A number of studies have evaluated the prevalence and correlates of individuals’ participation in concurrency in various populations in the United States and elsewhere,6–8 but partners’ concurrency has received considerably less attention.
In addition to concurrency, other characteristics of partnerships also affect the likelihood of STI transmission to individuals and throughout the population. For example, the interval between partnerships, in addition to the frequency and nature of sexual interactions, is critical. For STIs that have a restricted period of maximum infectiousness (such as gonorrhea and chlamydia), longer gaps between consecutive monogamous partnerships decrease the likelihood that a person who acquires infection from 1 partner will be highly infectious by the time he or she begins a new partnership. Conversely, longer overlaps in partnerships and shorter gaps between consecutive partnerships can dramatically increase the likelihood of transmission. Frequency of intercourse is especially important; more frequent sexual activity provides more opportunities for transmission. The extent of condom use also affects transmission; the protective effect of consistent condom use could conceivably render the sequencing and timing of partnerships inconsequential.
Despite the importance of these parameters of sexual partnerships, data that reflect them can be difficult to collect and are seldom available in 1 database. This information could improve our understanding of the dynamics of population HIV transmission and inform development of interventions, including behavioral interventions that enhance efficacy of preexposure prophylaxis and other biomedical prevention strategies. To address these issues, we conducted a descriptive analysis of sexual partnership characteristics at baseline among participants in HIV Prevention Trials Network (HPTN) 064 [The Women’s HIV Seroincidence HIV Study (ISIS)], a study of women at high risk for HIV infection. Our analysis focused on participants’ concurrency, their partners’ concurrency, and gaps and overlaps between sexual partnerships.
HPTN 064 study methods have been described elsewhere.9 Briefly, HPTN 064 was a multisite, longitudinal cohort study that enrolled eligible women between May 2009 and July 2010 from 10 urban and periurban communities with high prevalences of both poverty and reported HIV infection rates in 6 sites in the Northeastern and Southeastern United States (Atlanta, GA; Baltimore, MD; New York City, NY; Newark, NJ; Raleigh-Durham, NC; and Washington, DC), using venue-based sampling (ie, recruitment of persons who attend specific locations within defined geographic areas). Eligible individuals were aged 18–44 years, self-identified as a woman (transgender individuals were eligible), reported at least 1 episode of unprotected vaginal and/or anal sex with a man in the 6 months before enrollment and an additional personal or partner HIV risk characteristic, resided in an area with high rates of poverty and HIV/AIDS prevalence, and were willing to undergo HIV rapid testing and receive HIV test results. Participants underwent HIV testing and audio computer-assisted self-interviews (ACASI) at entry to the study and at 6-month intervals for up to 12 months. The baseline ACASI, which collected demographic and psychosocial data and information concerning sexual behaviors and drug use, was the data source of this analysis. Institutional review boards at each study site and collaborating institution approved the HPTN 064 study, and a Certificate of Confidentiality was obtained.
Data Collection and Measures
The ACASI asked each respondent to list the 3 men with whom she most recently had vaginal or anal intercourse during the 6 months before the interview and to provide the dates (month and year) of first and last (most recent) intercourse with each of these men. The survey asked the respondent about the risk characteristics of each of these male partners (eg, HIV infection status, substance use, etc) and asked whether each man had undergone HIV testing Response choices were yes, no, or do not know. For every partner listed, the survey asked approximately how many times during the past 6 months the respondent and he had: vaginal sex, vaginal sex without a condom, and anal sex. Response choices were more than 10 times, 2–10 times, once, or never. The survey also asked the women whether each of their partners had sex with other people during the last 6 months while he was in a sexual relationship with the respondent. Response choices were definitely did, probably did, probably did not, or definitely did not.
We identified participants’ concurrency by comparing the dates (month and year, not day) of first and last sexual intercourse for the most recent partners described in the ACASI (up to a maximum of 3 partners). Two partnerships were defined as concurrent when the month of first sexual intercourse with 1 partner occurred before the month of last intercourse with another partner.
We calculated the prevalence of participants’ concurrency by dividing the number of women with at least 1 concurrent partnership by the number of women in the study population.
Partners’ concurrency occurred if a participant reported that any of her last 3 partners “definitely did” have sex with another person during the course of her sexual relationship with him. We calculated the prevalence of partners’ concurrency by dividing the number of women with this response by the number of women in the study population with a nonmissing response.
Gaps and Overlaps in Sexual Partnerships
We evaluated the gap or overlap between the last 2 partners reported by women with at least 2 partners during the preceding 6 months. If the date of first intercourse with the second most recent partner preceded the date of first intercourse with the most recent partner, the gap/overlap was defined as the difference in months between the date of first intercourse with the most recent partner and the date of last intercourse with the next most recent partner (positive differences correspond to gaps, negative differences to overlaps). If the date of first intercourse with the second most recent partner occurred after the date of first intercourse with the most recent partner, the overlap was defined as the difference in months between the dates of first and last intercourse with the second most recent partner. Thus, the overlap duration of concurrent partnerships are less than or equal to −1 (because only partnerships that overlapped by more than 1 month were categorized as concurrent). The gaps of all nonconcurrent partnerships are greater than or equal to 0 and can be no greater than 6 months because of the study design.10,11
We examined the associations for participants’ concurrency and partners’ concurrency with age, race/ethnicity, marital status, education, income, use of illicit drugs (other than marijuana) and binge alcohol consumption within the past 6 months, and STI history, as previous research has demonstrated associations between concurrency and these variables.6,8,12,13 We fit multiple log binomial regression models with participants’ concurrency and partners’ concurrency as the dependent variables. Independent variables were removed from the model if they were not associated with the dependent variable (P > 0.1 and their removal did not change the coefficients of other variables by more than 10%).
To evaluate the distribution of risky sexual practices among women with different partnership patterns, we stratified respondents according to their participation in partnerships during the past 6 months: only participants’ concurrency, only partners’ concurrency, both participants’ and partners’ concurrency, and neither. We then compared respondents with respect to frequency and type of intercourse and risk characteristics of partners overall and with each of their last 2 partners. The χ2 statistics were used to compare proportions. The Kruskal–Wallis Test compared median gap lengths among the 4 partnership categories. All statistical analyses were performed with SAS version 9.2 (SAS, Cary, NC).
A total of 2099 women were enrolled in the study. Estimates of participants’ concurrency could not be determined for 471 women; these participants reported having more than 1 sexual partner but did not provide first and last dates of sexual intercourse and were, therefore, excluded from the concurrency analyses. Table 1 outlines characteristics of the overall study population and of the 1628 participants who were eligible for the concurrency analysis. Most respondents were black (85%); smaller proportions were white or Hispanic (7% and 12%, respectively). Only 7% of the study population was currently married; more than half (52%) had never married; 24% were unmarried but living with a sexual partner; and 7% were divorced, widowed, or separated. More than one third (35%) had less than high school education. Nearly half of the participants (46%) reported an annual household income of less than $10,000, 63% were unemployed, and almost half (46%) reported concern about having enough food during the preceding 30 days. Almost one quarter (22%) reported use of illicit drugs other than marijuana during the last 6 months, and 23% reported binge alcohol consumption at least weekly. Participants reported a median of 2 male sex partners in the last 6 months.
Of the 1628 women who were eligible for analysis, 267 (16%) had concurrent partners themselves but believed their partners were monogamous (ie, participants’ concurrency only), 208 (13%) had nonmonogamous partners but did not have concurrent partners themselves (partners’ concurrency only), 389 (24%) had concurrent partners and had nonmonogamous partners (both participants’ and partner’s concurrency), and 764 (47%) had neither concurrent partnerships nor nonmonogamous partners (neither participants’ nor partners’ concurrency). Thus, a total of 656 (40%) women had concurrent partnerships themselves and 597 (36%) believed their partners definitely had other partners during the course of their relationship.
Correlates of Participants’ and Partners’ Concurrency
Participants’ concurrency was strongly associated with relationship status (Table 2). Compared with married women, those who had never married; or were divorced, separated, widowed; or in “other” types of relationships were significantly more likely to engage in concurrent partnerships. Women who were cohabiting were not statistically different from married women in terms of likelihood of concurrency. Substance use, binge alcohol consumption, and history of STI were also associated with concurrency. Age, race/ethnicity, education, and income were not significantly associated with participants’ concurrency. Multivariable analysis revealed an attenuated relationship between participants’ concurrency and STI history, but the relationship between concurrency and other variables persisted (Table 3).
Partners’ concurrency was significantly associated with younger age of the participant (ages 18 through 26 years, compared with age 27 years and older). Partners’ concurrency, similar to participants’ concurrency, was significantly associated with relationship status: compared with married women, those who had never married; or were divorced, separated widowed; or in other types of relationships were more likely to believe their partner had outside sexual partnerships. Cohabiting women were not significantly different from married women in reporting partners’ concurrency. Partners’ concurrency was also associated with history of substance abuse and STI during the past 6 months. There was no relationship between partners’ concurrency and race/ethnicity, education, or income. In multivariable analysis, the relationships between partners’ concurrency and all other covariates remained unchanged (Table 3).
Gaps and Overlaps in Sexual Partnerships
The mean and median overlaps in partnership dates among women who had concurrent partnerships were −17 and −5 months, respectively (Table 4). The mean and median gap in partnerships was about 1 month among women who reported only partners’ concurrency or no involvement with concurrent partnerships.
Sexual Behaviors: Frequency of Unprotected Intercourse and Partner Risk Characteristics
A substantial proportion of women in all concurrency categories reported frequent (more than 10 times) unprotected vaginal intercourse with their most recent partner during the 6 months before the interview (Table 4). Frequent unprotected vaginal intercourse was more likely to be reported by women who were involved with neither type of concurrency (53%) and those who had both concurrent partners and nonmonogamous partners (53%) than women who only had concurrent partners (participants’ concurrency only) (45%) and those who only had nonmonogamous partners (partners’ concurrency only) (42%) (P = 0.0036). Twenty-two percent of participants reported having had anal intercourse at least once with their most recent partner; 14% (235/1628) had anal sex with their most recent partner multiple times, but the distribution of this behavior did not vary by concurrency status.
A respondent was more likely to report either that her last partner had never undergone HIV testing or that she was unaware of whether or not he had been tested if she had both types of concurrent partnerships (45%), only nonmonogamous partners (47%), or only had concurrent partners herself (38%) compared with women who were involved with neither type of concurrency (29%) (P < 0.0001).
Only 2% women believed their most recent partner had sex with men; the prevalence of this belief did not vary by concurrency status (data not shown). There were no significant differences by concurrency status in the prevalence of women who believed that their last partner had injected drugs.
We evaluated frequency of vaginal intercourse with both partners among the 1193 women who reported 2 or more partners during the 6 months preceding the interview (Table 5). Substantial proportions of women reported unprotected intercourse with at least 2 of their most recent partners on more than 1 occasion. A woman with multiple partners was more likely to report multiple episodes of unprotected vaginal intercourse with 2 or more of these partners if she reported both types of concurrency (60%), participants’ concurrency only (50%), or partners’ concurrency only (33%) compared with a woman who was involved with neither type of concurrency (14%) (P < 0.0001). Of women with both types of concurrency, 16% reported unprotected vaginal intercourse with at least 2 partners on more than 10 occasions in the last 6 months. Similarly, 10% of women with only participants’ concurrency, 7% of those with only partners’ concurrency, and 3% of those of those who had neither type of concurrency reported more than 10 episodes of unprotected vaginal intercourse with at least 2 partners with both types of concurrency were most likely to report that both of their last 2 partners had not undergone HIV testing or were unaware that they had been tested (48%), followed by women with only participants’ concurrency (41%) or only partners’ concurrency (40%), whereas women who did not participate in any type of concurrency were least likely (17%) (P < 0.0001).
In this cohort of women at risk for HIV infection who were recruited from 10 communities in the United States, both participants’ and partners’ concurrency were frequently reported. During the 6 months before the baseline interview, 40% of participants had concurrent partnerships, and 36% believed their partners had concurrent partnerships; 24% of all respondents had concurrent partnerships themselves and strongly believed their partners did as well. Marital status, substance use, and history of STI were associated with participants’ and partners’ concurrency. Among the 47% of women who had neither concurrent nor nonmonogamous partners, the median gap between partnerships was 1 month, which is short enough to allow transmission of several STI pathogens, including acute HIV infection. Compared with women who had neither concurrent nor nonmonogamous partners, those with any type of concurrency were more likely to report multiple episodes of unprotected vaginal intercourse with at least 2 of their last 3 partners, and this was especially evident for women who had both concurrent and nonmonogamous partners. These findings suggest substantial opportunities for sexual transmission of HIV and other STIs among the women in this study.
Other studies have also revealed short gaps, but differences in study design make it difficult to directly compare results. Our study examined partnerships that occurred during the last 6 months. Therefore, the maximum possible gap is 6 months, which may not be comparable with other studies that measured partnerships over shorter or longer intervals. In a representative sample of the U.S. population, among women aged 15–44 years with multiple consecutive partners, the mean gap varied with age from as short as 8 months among the youngest women to 18 months among women aged 30–44 years and was longer among non-Hispanic black women (15 months) than non-Hispanic white women (12 months). Of note, roughly 11%–14% women had gaps less than 1.8 months.10 The 18- to 39-year-old male and female Seattle respondents in a random digit dial survey reported a mean gap length of 60.8 days.11 Gap lengths among male and female patients with consecutive partners in a Malawi STI clinic averaged 21 days.14 These short gaps would especially facilitate HIV transmission during acute infection when HIV viral load is extremely high.15
The duration of overlap in concurrent partnerships varied considerably among respondents in our study, as evidenced by the marked differences between the median of 5 months and the mean of 17 months, indicating that a small proportion of women had partnerships that overlapped by at least 17 months. Variation in duration of overlap has been observed in other study populations.11,14,16 Long-term concurrency can provide increased opportunities for HIV transmission.17
Estimates of per contact risk of HIV acquisition through vaginal intercourse vary from 0.001 to as high as 0.1; risk is especially high if the HIV-uninfected individual has an STI (as did 11% of our participants) or if the index contact has early or late stage HIV infection with high viral load.18 Anal intercourse further heightens HIV transmission risk.18 A substantial minority (14%) of all participants in this study had anal intercourse multiple times with their most recent partner. Moreover, women who had both concurrent and nonmonogamous partners were especially likely to have had unprotected vaginal intercourse with multiple partners and to report unknown HIV status of those partners. Unprotected intercourse during the course of concurrent partnerships has been previously reported. More than one third (35%) of men with concurrent partners in a representative sample of the U.S. population did not use condoms during last intercourse with either of their partners.16 These findings suggest a need for increased availability, promotion, and acceptability of condom use; increased diagnosis and treatment of STIs, other HIV prevention strategies such as preexposure HIV prophylaxis; and expansion of HIV testing, as recommended by the U.S. National HIV/AIDS Strategy.19
A key strength of this study is its collection of extensive details concerning sexual partnerships among women at high risk for HIV infection. We believe these findings may be generalizable to women with similar characteristics in the United States. However, it is important to note that our study population is not representative of any racial/ethnic or demographic group. Rather, participants were recruited because of their high-risk characteristics. Although they may be representative of women at high risk for HIV infection, they are not representative of poor women in general.
These results share the limitations of studies that rely on self-reported data, including those involving social desirability, recall, understanding, and communication,20,21 and we used a relatively insensitive measure of coital frequency. In addition, concurrency—particularly partners’ concurrency—is difficult to measure. Because our survey only queried women about their last 3 partners, we may have missed concurrency among women who had 4 or more partners. In addition, our concurrency definition (which we adopted to avoid misclassifying respondents with short gaps between consecutive partnerships) may have missed concurrent partnerships whose overlap only occurred within 1 calendar month.
Our comparison of first and last dates of sexual intercourse is considered one of the more robust strategies for measuring concurrency.22 The UNAIDS Reference Group on Estimates, Modeling, and Projections recommends determination of the 6-month point prevalence of concurrency (ie, the prevalence of concurrency exactly 6 months before the interview).23,24 This definition is useful as a standard for comparison of concurrency rates across populations. However, our goal was not to evaluate population prevalence of concurrency; we evaluated the cumulative proportion of concurrency to fully capture the characteristics of concurrent partnerships.
Determination of partners’ concurrency is problematic, as people are reporting the usually unobserved behavior of others. Although some studies have noted poor agreement between individuals’ reports of their partners’ concurrency and the partners’ reports of their own concurrent partnerships,25–27 the poor agreement stems largely from respondents’ failure to identify their partners’ lack of monogamy and not from overreports of the partners’ concurrency.28
In contrast to previous nationally representative studies, in which the crude and adjusted prevalence of concurrency among black women exceeded that among whites and Hispanics,6,8 there were no racial differences in concurrency among our respondents. Our study is notable for the exceptional poverty of the participants. Almost half survived on an annual income of less than $10,000, and all resided in an area with high rates of poverty. Given the importance of contextual factors in sexual network patterns,29,30 it is likely that the adverse economic context shared by participants of all race/ethnicities may have contributed to the distribution of the observed network patterns.
This study identified several sexual partnership characteristics that contribute to HIV and STI acquisition and transmission, including short time gaps between consecutive partnerships, partners’ concurrency, long-term participants’ concurrency, unprotected intercourse with concurrent partners, and lack of awareness of partner HIV serostatus among individuals whose network position placed them at high risk of acquiring or transmitting STIs, including HIV. These findings can improve our understanding of partnership dynamics and help target behavioral and biomedical interventions to prevent HIV infection.
The authors thank the study participants, community stakeholders, and staff from each study site. In particular, they acknowledge Lynda Emel, Jonathan Lucas, Nirupama Sista, Kathy Hinson, Elizabeth DiNenno, Ann O’Leary, Lisa Diane White, Waheedah Shabaaz-El, Quarraisha Abdool-Karim, Sten Vermund, Edward E. Telzak, Rita Sondengam, Cheryl Guity, Tracy Hunt, Khadijah Abass, Eileen Rios, Irene Kuo, Christopher Chauncey Watson, Christopher Walker, Oluwakemi Amola, and LeTanya Johnson-Lewis.
2. Morris M, Kretzschmar M. Concurrent partnerships and transmission dynamics in networks. Soc Networks. 1995;17:299–318.
3. Watts CH, May RM. The influence of concurrent partnerships on the dynamics of HIV/AIDS. Math Biosci. 1992;108:89–104.
4. Morris M, Goodreau S, Moody J. Sexual networks, concurrency, and STD/HIV. In: Holmes KK, et al., eds. Sexually Transmitted Diseases. New York, NY: McGraw Hill Medical; 2008:109–125.
5. Adimora AA, Schoenbach VJ, Martinson FEA, et al.. Heterosexually transmitted HIV infection among African Americans in North Carolina. J Acquir Immune Defic Syndr. 2006;41:616–623.
6. Adimora AA, Schoenbach VJ, Bonas DM, et al.. Concurrent sexual partnerships among women in the United States. Epidemiology. 2002;13:320–327.
7. Adimora AA, Schoenbach VJ, Doherty IA. Concurrent sexual partnerships among men in the United States. Am J Public Health. 2007;97:2230–2237.
8. Adimora AA, Schoenbach VJ, Taylor EM, et al.. Concurrent partnerships, nonmonogamous partners, and substance use among women in the United States. Am J Public Health. 2011;101:128–136.
9. Hodder SL, Justman J, Hughes JP, et al.. HIV acquisition among women from selected Areas of the United States: a cohort study. Ann Intern Med. 2013;158:10–18.
10. Kraut-Becher JR, Aral SO. Gap length: an important factor in sexually transmitted disease transmission. Sex Transm Dis. 2003;30:221–225.
11. Foxman B, Newman M, Percha B, et al.. Measures of sexual partnerships: lengths, gaps, overlaps, and sexually transmitted infection. Sex Transm Dis. 2006;33:209–214.
12. 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. 2003;34:423–429.
13. Adimora AA, Schoenbach VJ, Martinson F, et al.. Concurrent sexual partnerships among African Americans in the rural south. Ann Epidemiol. 2004;14:155–160.
14. Powers KA, Hoffman IF, Ghani AC, et al.. Sexual partnership patterns in Malawi: implications for HIV/STI transmission. Sex Transm Dis. 2011;38:657–666.
15. Pilcher CD, Tien HC, Eron JJ, et al.. Brief but efficient: acute HIV infection and the sexual transmission of HIV. J Infect Dis. 2004;189:1785–1792.
16. Doherty IA, Schoenbach VJ, Adimora AA. Condom use and duration of concurrent partnerships among men in the United States. Sex Transm Dis. 2009;36:265–272.
17. Morris M, Epstein H, Wawer M. Timing is everything: international variations in historical sexual partnership concurrency and HIV prevalence. PLoS One. 2010;5:e14092.
18. Powers KA, Poole C, Pettifor AE, et al.. Rethinking the heterosexual infectivity of HIV-1: a systematic review and meta-analysis. Lancet Infect Dis. 2008;8:553–563.
20. Warnecke RB, Sudman S, Johnson TP, et al.. Cognitive aspects of recalling and reporting health-related events: Papanicolaou smears, clinical breast examinations, and mammograms. Am J Epidemiol. 1997;146:982–992.
21. Fenton KA, Johnson AM, McManus S, et al.. Measuring sexual behaviour: methodological challenges in survey research. Sex Transm Infect. 2001;77:84–92.
22. Lurie MN, Rosenthal S. Concurrent partnerships as a driver of the HIV Epidemic in sub-Saharan Africa? The evidence is limited. AIDS Behav. 2010;14:17–24; discussion 25–28.
23. UNAIDS Reference Group on Estimates, Modelling, and Projections: Working Group on Measuring Concurrent Partnerships. HIV: consensus indicators are needed for concurrency. Lancet. 2010;375:621–622.
25. Seal DW. Interpartner concordance of self-reported sexual behavior among college dating couples J Sex Res. 1997;34:39–55.
26. Stoner BP, et al.. Avoiding risky sex partners: perception of partners’ risks v partners’ self reported risks. Sex Transm Infect. 2003;79:197–201.
27. Helleringer S, Kohler H-P, Kalilani-Phiri L, et al.. The reliability of sexual partnership histories: implications for the measurement of partnership concurrency during surveys. AIDS. 2011;25:503–511.
28. Drumright LN, Gorbach PM, Holmes KK. Do people really know their sex partners? Concurrency, knowledge of partner behavior, and sexually transmitted infections within partnerships. Sex Transm Dis. 2004;31:437–442.
29. Adimora AA, Schoenbach VJ. Contextual factors and the black-white disparity in heterosexual HIV transmission. Epidemiology. 2002;13:707–712.
30. Adimora AA, Schoenbach VJ. Social context, sexual networks, and racial disparities in rates of sexually transmitted infections. J Infect Dis. 2005;191(suppl 1):S115–S122.
Keywords:© 2014 by Lippincott Williams & Wilkins
HIV; epidemiology; concurrent partnerships; perceived partner concurrency; sexual networks; indirect concurrency