Recent surveys suggest that the prevalence of same-sex sexual behavior in women may be increasing in the United States and some other developed countries.1–4 Although women who have sex exclusively with women generally have lower risk for sexually transmitted infections (STIs), most women who have sex with women (WSW) also have sex with men and can be at high risk for STIs, including human immunodeficiency virus (HIV).5–10 Studies about the burden of STIs in representative samples of WSW are rare,4,5 and WSW is often invisible in national surveillance systems for HIV and STIs. WSW are geographically less concentrated and generally less visible than men who have sex men in gay communities, and few HIV and STI services are targeted for WSW.9,10 Understanding the risk behaviors and disease burden in WSW is important for the care and health promotion among WSW in general and for the prevention of STIs.
The National Health and Nutrition Examination Surveys (NHANES) are a series of cross-sectional surveys conducted by the National Center for Health Statistics to compile nationally representative statistics on the health of the US population. Survey participants are selected through complex, multistage probability sampling and are interviewed and undergo a medical examination, during which biologic samples are obtained. Although national estimates of the prevalence of WSW behavior have been reported from the 2002 National Survey of Family Growth (NSFG), the survey did not collect biologic specimens.1 Herpes simplex virus type 2 (HSV-2) infection, 1 of the most common STI, is the main cause of genital herpes. Because HSV-2 is an infection that is almost always sexually transmitted and is not curable, it can serve as a biologic marker for cumulative sexual risk.11
The purpose of this study is to determine the prevalence of same-sex behavior in women 18 to 59 years of age in the general population in the United States and to assess its trend over time. In addition, we describe demographic and behavioral characteristics and the prevalence of HSV-2 in WSW. We also compare selected risk behaviors and HSV-2 prevalence in WSW and non-WSW (women who reported no same-sex partner) and among WSW, by sexual orientation.
Study Population and Survey Methods
Details about the survey methodology of NHANES have been published previously.12 The most recent series began in 1999 and is designed to run continuously with data released every 2 years. Participants were chosen according to a complex, stratified, multistage probability sampling design to select a nationally representative sample of the US civilian noninstitutionalized population and were interviewed and examined by medical professionals.
During NHANES 2001–2006, health examinations were conducted in specially equipped mobile examination centers. The questionnaire about sexual and drug use behaviors was administered in a private room using audio computer assisted self-interview in either English or Spanish. During NHANES 2001–2006, about 6500 female participants of all ages were sampled for interview and examination during each 2-year period. The overall response rate for examination among female participants ranged from 76% to 80%.13
Testing for HSV-2 Infection
Blood samples from survey participants were collected during health examinations at the mobile examination centers. Serum antibodies to HSV-2 were detected using a type-specific immunodot assay. Purified glycoprotein gG-2 of HSV-2 was used as the antigen in the immunodot assay. The immunodot assay is highly sensitive and specific and can discriminate HSV-2 infection from the closely related infection of herpes simplex virus type 1.14,15 In NHANES 2001–2006, the same immunodot assay and the same laboratory were used as for previous NHANES.16
Definitions of WSW
During NHANES 2001–2006, “sex” was defined to include vaginal, oral, or anal sex. Survey participants aged 18 to 59 years were interviewed about the number of male and female partners in lifetime and in the past 12 months. Only women who answered “yes” to the question “have you ever had sex?” were asked about the question “In your lifetime, with how many women have you had sex?” Having ever had sex with women (WSW-ever) was defined as reporting 1 or more female partner in lifetime. Recent same-sex behavior was also assessed (“In the past 12 months, with how many women have you had sex?”). Women who have had sex with a woman in the past year (WSW-pastyear) was defined as reporting 1 or more female partner in the past 12 months. Starting in the year of 2001, a question about sexual orientation was added in NHANES: “Do you think of yourself as … heterosexual or straight (that is, sexually attracted only to men); homosexual or lesbian (that is, sexually attracted only to women); bisexual (that is, sexually attracted to men and women); something else; or you're not sure?”
For this article, the term WSW only applies to behavior (have a female sex partner, ever or in the past year), whereas sexual orientation refers to or be synonymous with self-identification.
All prevalence estimates were weighted to represent the noninstitutionalized civilian US population and to account for oversampling and nonresponse to the interview and the examination.17,18 We used standard examination weights provided by National Center for Health Statistics without further adjustments.19 SUDAAN software (Release 9.03, Research Triangle Institute, Cary, NC) was used for statistical analyses.20,21 Variance estimates were calculated using a Taylor series linearization that incorporated the complex sample design of the survey.21,22 Confidence intervals (CI) for the prevalence were calculated using a logit transformation with the standard error of the logit prevalence estimated using the δ method. Relative standard errors were computed for each weighted estimate as (standard error (SE)/estimate)*100. An relative standard errors >30% of a prevalence estimate reflects instability and the corresponding estimate should be interpreted cautiously.17
For this article, race-ethnicity was defined by self-report as non-Hispanic white, non-Hispanic black, Mexican American “other”; the “other” category includes all participants who did not belong to the 3 main categories, such as those who reported “multiracial” and persons whose race-ethnicity was missing. Poverty level was defined by US Census Bureau, and is assigned by dividing total family income by the poverty threshold index, then adjusting for number of persons in the household at the year of interview (available at: http://www.census.gov/hhes/www/poverty/definitions.html).
P-values for comparing WSW subgroups by sexual orientation were based on a test for independence using an F-statistic derived from a Wald χ2 for categorical variables.21 Continuous variables were presented as means or medians depending on the symmetry of the observed distribution. Medians are computed through an estimate of the cumulative distribution function based on ungrouped data. Whereas t-tests were used to compute the P-values to compare the central tendency, a logarithmic transformation was applied to positively skewed variables before performing a t-test.
Prevalence of WSW by Age Group
Of all women aged 18 to 59, 95% reported having had sex. Among these, WSW-ever was reported by 7.1% (95% CI, 6.1–8.2) of women (Table 1). From US Census (January 2004), the population counts of noninstitutionalized civilian women aged 18 to 59 years were 84.8 million. Based on the prevalence of WSW-ever at 7.1%, and after taking into account those who had not had sex, the number of WSW-ever was estimated at 5.7 million (95% CI, 4.9–6.6). The prevalence of WSW-pastyear was lower (2.7%; 95% CI, 2.1–3.5) (Table 1).
The prevalence of WSW-ever was significantly different by age, highest in women aged 18 to 29 years and lowest in those aged 50 to 59 years. The prevalence of WSW-ever was significantly lower in women born outside of the United States, but did not differ by education level or poverty level. Similarly, the prevalence of WSW-pastyear also differed significantly by age and birthplace (Table 1).
To further examine the change in WSW prevalence with age, we analyzed the prevalence of WSW behavior by age, after excluding women who reported their birthplace was outside of the 50 US states or Washington, DC. This analysis removes the possible impact by immigrants on the trend of WSW behavior in the United States. In US-born women, the prevalence of WSW-ever decreased significantly with age, from 9.9% in 18- to 29-years-olds to 5.6% in 50- to 59-years-olds (P = 0.01) (Fig. 1); there was a similar trend with age for WSW-pastyear: the prevalence decreased from 4.8% in 18- to 29- years-olds to 1.6% in 50- to 59-years-olds (P = 0.009) (Fig. 1).
Sexual Orientation and Trend Over Time
Among all women aged 18 to 59 years who had had sex, the overall prevalence of heterosexual or straight was 94.6% (95% CI, 93.7–95.4), homosexual or lesbian was 1.4% (95% CI, 1.0–2.0), bisexual was 2.5%, (95% CI, 2.1–3.0), and “not sure” was 1.0% (95% CI, 0.7–1.4). The remaining 0.5% (n = 38) responded as something else, do not know, or refused. The distribution of sexual orientation in women born in the United States differed significantly by age (P = 0.01): Whereas the prevalence of homosexual orientation was relatively low and stable across age group (point estimates ranged from 1.1% to 2.0%), the prevalence of bisexual orientation was negatively correlated with age (point estimates ranged from 1.2% in women aged 50–59 years to 5.2% in women aged 18–29 years, Fig. 1).
Because our survey data were collected over an interval of 6 years (2001–2006), a study participant born in the same year could end up in different age groups depending on the year of interview. To further demonstrate the trend over time, we analyzed sexual orientation in US-born women by birth cohort. In women who had had sex and were born in the United States, the distribution of self-reported sexual orientation was significantly different by year of birth (P = 0.003). The percent of women who reported their sexual orientation as “heterosexual or straight” decreased significantly, from 97.8% in women born in 1940s to 91.7% in those born in 1980s (Fig. 2), whereas the percent of women who reported their sexual orientation as “bisexual” increased from 0.9% in those born in 1940s to 7.3% in those born in the 1980s (Fig. 2). The percent of women who reported their sexual orientation as homosexual or lesbian appeared unchanged, although some estimates may not be reliable because of small sample size (Fig. 2).
Demographic Characteristics of WSW
Among WSW-ever, more than half (52.6%; 95% CI, 46.2–59.0) reported their sexual orientation as heterosexual or straight, 28.3% (95% CI, 22.9–34.3) as bisexual, and only 19.1% (95% CI, 14.3–24.9) as homosexual or lesbian. Among WSW-pastyear, a subset of WSW-ever, 23.5 (95% CI, 15.2–34.5) reported their sexual orientation as heterosexual or straight, 42% (95% CI, 21.1–42.0) as bisexual, and 46% (95% CI, 36.8–55.4) as homosexual or lesbian.
Among heterosexual WSW-ever, the mean age was 37.3 years, 74.2% were white, and 67.5% had a high school or more education (Table 2). There were no statistically significant differences in age, race-ethnicity, education level, or poverty level by sexual orientation. However, birthplace and marital status differed significantly by sexual orientation. More than half of heterosexual WSW-ever (53.2%) were married, compared with none of homosexuals and 22.3% of bisexuals (Table 2, P < 0.0001).
Comparison of Risk Behaviors Between Non-WSW and WSW, and by Sexual Orientation Among WSW-Ever
In Table 3, we compared selected risk behaviors reported by non-WSW, WSW-pastyear, and WSW-ever.
Overall, 23.1% of WSW-pastyear and 29.6% of WSW-ever reported having first sex at 14 years of age or younger, significantly higher than the 12.0% reported by non-WSW (P = 0.012 and <0.0001, respectively). Among WSW-ever, only 38.4% (95% CI, 32.1–45.0) reported having 1 or more female partner in the past year (i.e., WSW-pastyear). Nevertheless, when compared to non-WSW, both WSW-pastyear and WSW-ever were more likely to report having a larger number of total partners (lifetime and past year) and having ever used cocaine (Table 3).
Because “WSW-ever” is an inclusive group comprising of diverse risk and behavioral groups, we also compared risk behaviors by sexual orientation. Among WSW-ever, the percent having first sex at 14 years of age or younger differed significantly by sexual orientation: 31.3% in heterosexual WSW-ever, 38.9% in bisexual WSW-ever and 12.9% in homosexual WSW-ever (P = 0.005, Table 3).
Among non-WSW, 87.2% reported 1 or more male partner in the past year (Table 3). The percent of women who had ≥1 male partners in the past year was high among heterosexual and bisexual WSW-ever (88.6% and 79.2%, respectively), but was low (9.6%) among homosexual WSW-ever (P < 0.0001). Among non-WSW, 11.2% reported having had 2 or more sex partners in the past year, compared with 39.6% in WSW-ever (P < 0.0001, Table 3); the percent of women having 2 or more sex partners in the past year was highest among bisexual WSW-ever (52.4%).
Overall, 96.6% of WSW-ever had ever had a male partner (Table 3). Two-thirds of WSW-ever had 10 or more total lifetime partners (male and female), higher compared with non-WSW (67.5% vs. 21.2%, P < 0.0001). Among WSW-ever, this percentage was highest in bisexuals (78.5%) and lowest (56.1%) in homosexuals. The median number of lifetime sex partners (males and females) also differed by sexual orientation (P < 0.0001). Bisexual WSW-ever reported the highest median number of male partners and total partners (Table 3).
Prevalence of HSV-2 Infection
The overall HSV-2 seroprevalence was 23.8% in non-WSW, 30.3% in WSW-pastyear, and 36.2% in WSW-ever (Table 3). The mean age of WSW-pastyear was 34.3 years, younger than non-WSW (38.7 years) and WSW-ever (36.1 year). After adjusting for age, the prevalence of HSV-2 was significantly higher in WSW-past year (P = 0.04) and WSW-ever (P = 0.0002) when compared with non-WSW (data not show). Among WSW-ever, HSV-2 prevalence varied greatly by sexual orientation: 45.6% in heterosexuals, 35.9% in bisexuals, and 8.2% in homosexuals (P = 0.002) (Table 3). Compared with non-WSW, the prevalence of HSV-2 was significantly higher in heterosexual and bisexual WSW-ever (23.8% vs. 42.0%, P = 0.0004). By contract, the prevalence of HSV-2 was significantly lower in homosexual WSW-ever than in non-WSW (8.2% vs. 23.8%, P = 0.002).
We estimated that there are 5.7 million women who had ever had sex with women in the United States. Of WSW-ever, only 19.1% (95% CI, 14.3–24.9) reported their sexual orientation as homosexual or lesbian. Among all women aged 18 to 59 years who had had sex, the prevalence of homosexual and bisexual orientation was 1.4% and 2.6%, respectively. The prevalence of WSW behavior is increasing among younger women. Our study suggests that WSW, together as a group, reported higher risk behaviors and had a higher burden of HSV-2 infection than non-WSW, but there were significant differences among subsets of WSW; self-reported sexual orientation is an important predictor of risk behaviors and burden of STIs.
Based on our population-based data, the prevalence of WSW is about 7% in sexually experienced women aged 18 to 59 years and was higher in younger women. The prevalence of WSW was assessed in the NSFG, a nationally representative sample of people aged 15 to 44 years. In NSFG conducted in 2002, 11.2% of women reported that they had had “a sexual experience with another female at some time in their lives.”1 The standard error for this estimate is 0.5, and therefore, the estimated 95% CI is 10.2% to 12.2%, significantly higher than our estimate of 7.1% (95% CI, 6.1–8.2) from NHANES 2001–2006. However, this difference should not be surprising because prior research shows that in women, the criteria or definitions of homosexual contact (reflected by the different wording in the surveys) can profoundly affect the reported prevalence in population-based surveys.2 In NSFG 2002, although the relationship between age and WSW prevalence was not linear (women 35–39 years of age reported the second highest prevalence, at 12.3%), the prevalence of WSW-ever was highest among women 20 to 29 years of age (14.1%) and lowest in women 40 to 44 years of age (7.8%). Furthermore, in NSFG 2002 there was a clear trend for increasing prevalence of bisexual orientation with younger age, (from 2.0% in 35–44 year olds to 7.4% in 18–19-years-olds),1 a similar trend we found using data from NHANES 2001–2006.
The observed increases in the proportions of women who report sex with other women (ever or in the past year) and in the prevalence of bisexual orientation in more recent birth cohorts of US women need further confirmation and exploration. The higher prevalence in younger women may be in part because of higher willingness to report as a result of lessened social stigma associated with disclosure of homosexual behavior. Furthermore, studies in recent years have shown that sexual identity categories are more fluid in women, especially in young people,23,24 making it difficult to interpret trend data. However, it is conceivable that the practice of same-sex behavior is increasing in women in the US in light of the changes in social norms and public attitude in the last 2 decades.25 Higher prevalence of bisexuality in younger women was observed in population-based surveys in the United Kingdom and in Australia.4,23 The relative stability over time in the prevalence of homosexual/lesbian women has also been reported,4,23 suggesting that changing social norms and public attitude has little influence on this group of women.
The limitations of population-based surveys include the small number of women who reported same-sex sexual contact because this condition is relatively rare and may be geographically concentrated in large cities.26 We had to combine data from 3 NHANES survey cycles for more reliable analyses. In addition, reporting bias resulting from differences in respondents' willingness to report socially censured behaviors such as WSW is a concern. The use of audio computer assisted self-interview instead of face-to-face interview in the NHANES is likely to increase the willingness to report.27
Findings from our population-based surveys add to the literature concerning the heterogeneity within the group of women defined by generic terms such as “WSW.”5,6,28–30 Because the study participants were drawn from the general population, our findings indicate that high risk sexual practices such as younger age at sexual debut and multiple sex partners are prevalent among some, but not all, subgroups of women who have sex with other women. Among WSW-ever, although demographic characteristics, including age, race-ethnicity, and education level, were comparable between the groups defined by self-reported sexual orientation, behavioral characteristics differed. Heterosexual and bisexual women had higher sexual risk and higher HSV-2 prevalence than homosexual women. Medical care providers and public health practitioners should consider sexual behavior and sexual orientation when assessing risks for some STIs and determining appropriate care and prevention services for these women.
1.Mosher WD, Chandra A, Jones J. Sexual behavior and selected health measures: Men and women 15–44 years of age, United States, 2002. Adv Data 2005; 362:1–55.
2.Pedersen W, Kristiansen HW. Homosexual experience, desire and identity among young adults. J Homosex 2008; 54:68–102.
3.Grulich AE, de Visser RO, Smith AM, et al. Sex in Australia: Homosexual experience and recent homosexual encounters. Aust N Z J Public Health 2003; 27:155–163.
4.Mercer CH, Bailey JV, Johnson AM, et al. Women who report having sex with women: British national probability data on prevalence, sexual behaviors, and health outcomes. Am J Public Health 2007; 97:1126–1133.
5.Tao G. Sexual orientation and related viral sexually transmitted disease rates among US women aged 15 to 44 years. Am J Public Health 2008; 98:1007–1009.
6.Shotsky WJ. Women who have sex with other women: HIV seroprevalence in New York State counseling and testing programs. Women Health 1996; 24:1–15.
7.Bevier PJ, Chiasson MA, Heffernan RT, et al. Women at a sexually transmitted disease clinic who reported same-sex contact: Their HIV seroprevalence and risk behaviors. Am J Public Health 1995; 85:1366–1371.
8.Marrazzo JM, Koutsky LA, Handsfield HH. Characteristics of female sexually transmitted disease clinic clients who report same-sex behavior. Int J STD AIDS 2001; 12:41–46.
9.Bauer GR, Welles SL. Beyond assumptions of negligible risk: Sexually transmitted diseases and women who have sex with women. Am J Public Health 2001; 91:1282–1286.
10.Hader SL, Smith DK, Moore JS, et al. HIV infection in women in the United States: Status at the millennium. JAMA 2001; 285:1186–1192.
11.Corey L, Wald A. Genital Herpes. In: Holmes KK, Sparling FP, Stamm WE, et al, eds. Sexually Transmitted Diseases, 4th ed. New York, NY: McGraw-Hill, 2008:399–437.
12.National Center for Health Statistics. Plan and operation of the third National Health and Nutrition Examination Survey, 1988–1994. Vital Health Stat 1 1994:1–407.
14.Lee FK, Pereira L, Griffin C, et al. A novel glycoprotein (gG-1) for detection of herpes simplex virus specific antibodies. J Virol Methods 1986; 14:111–118.
15.Ashley RL, Militoni J, Lee F, et al. Comparison of western blot (immunoblot) and glycoprotein G-specific immunodot enzyme assay for detecting antibodies to herpes simplex virus type 1 and type 2 in human sera. J Clin Microbiol 1988; 26:662–667.
16.Xu F, Sternberg MR, Kottiri BJ, et al. National trends in herpes simplex virus type 1 and type 2 in the United States: Data from the National Health and Nutrition Examination Survey (NHANES). JAMA 2006; 296:964–973.
17.Mohadjer L, Montaquila J, Waksberg J. National Health and Nutrition Examination Survey III: Weighting and examination methodology. Prepared by Westat for National Center for Health Statistics, Hyattsville, MD. February 1996.
19.Sternberg MR, Hadgu A. A comparison of methods to further adjust for non-response due to missing lab data in National Health and Nutrition Examination Survey (NHANES). In: Proceedings of the American Statistical Association, section on Statistical Computing [CD-ROM]. Alexandria VA: American Statistical Association, 2007.
20.Shah BV, Barnwell BG, Bieler GS, et al. SUDAAN Users Manual, release 7.0. Research Triangle Park, NC: Research Triangle Institute; 1996.
21.Korn EL, Barry GI. Analysis of Health Surveys. New York, NY: Wiley, 1999.
22.Wolter K. Introduction to Variance Estimation. New York, NY: Springer-Verlag, 1985.
23.Jorm AF, Dear KB, Rodgers B, et al. Cohort difference in sexual orientation: Results from a large age-stratified population sample. Gerontology 2003; 49:392–395.
24.Savin-Williams RC, Ream GL. Prevalence and stability of sexual orientation components during adolescence and young adulthood. Arch Sex Behav 2007; 36:385–394.
25.Hicks GR, Lee TT. Public attitudes toward gays and lesbians: Trends and predictors. J Homosex 2006; 51:57–77.
26.Michaels S. The prevalence of homosexuality in the United States. In: Cabaj RP, Stein TS, eds. Textbook of Homosexuality and Mental Health. Washington, DC: American Psychiatric Press, 1996:43–63.
27.Turner CF, Ku L, Rogers SM, et al. Adolescent sexual behavior, drug use, and violence: Increased reporting with computer survey technology. Science 1998; 28:867–873.
28.Lindley LL, Barnett CL, Brandt HM, et al. STDs among sexually active female college students: Does sexual orientation make a difference? Perspect Sex Reprod Health 2008; 40:212–217.
29.Young RM, Meyer IH. The trouble with “MSM” and “WSW”: Erasure of the sexual-minority person in public health discourse. Am J Public Health 2005; 95:1144–1149.
30.Goodenow C, Szalacha LA, Robin LE, et al. Dimensions of sexual orientation and HIV-related risk among adolescent females: Evidence from a statewide survey. Am J Public Health 2008; 98:1051–1058.