Abstract
Little is known about the sexual health of male veterans. This study used nationally representative data from the 2011 to 2013 National Survey of Family Growth to compare sexual behaviors and history of sexually transmitted infections (STIs) between male veterans and nonveterans. The sample included 3,860 men aged 18 to 44 years who reported ever having sex with a man or woman. The key independent variable was veteran status. Sexual behavior outcomes included ≥6 lifetime female partners, ≥10 lifetime partners of either sex, ≥2 past-year partners of either sex, having past-year partners of both sexes, and condom nonuse at last vaginal sex. STI outcomes included past-year history of chlamydia, gonorrhea, or receiving any STI treatment; lifetime history of herpes, genital warts, or syphilis; and an aggregate measure capturing any reported STI history. Logistic regression models were used to evaluate associations between veteran status and each outcome. In models adjusting for age, race/ethnicity, education, income, and marital status, veterans had significantly greater odds than nonveterans of having ≥6 lifetime female partners (OR = 1.5, 95% CI [1.02, 2.31]). In models adjusting for age and marital status, veterans had significantly greater odds of having partners of both sexes in the past year (OR = 4.8, 95% CI [1.2, 19.8]), and gonorrhea in the past year (OR = 3.2, 95% CI [1.2, 8.5]). Male veterans were thus significantly more likely than nonveterans to have STI risk factors. Health care providers should be aware that male veterans may be at higher risk for STIs and assess veterans’ sexual risk behaviors.
Keywords
Introduction
Sexually transmitted infections (STIs) are a serious health issue in the United States, with an estimated 20 million new STI cases each year that result in roughly $16 billion in direct annual medical costs (Owusu-Edusei et al., 2013). Preventing and containing STIs is a considerable public health challenge (Workowski, 2015). National STI screening guidelines and policies are primarily focused on women due to the serious reproductive consequences of these infections in women (Workowski, 2015; Workowski & Bolan, 2015; Zakher, Cantor, Pappas, Daeges, & Nelson, 2014). Consequently, research efforts are also largely focused on the sexual health of women, not men (Sonfield, 2002). However, understanding more about STIs and the sexual health behaviors and risks of men could be a means to not only prevent and contain infections in men but also prevent infection and subsequent sequelae of STIs in women.
While there is limited research about the sexual health of men in general, there is a particular dearth of information about the sexual health of male veterans, a population that may be at higher risk for STIs. Higher rates of STIs have been documented in male and female active-duty military populations as well as female veterans compared with general U.S. populations (Armed Forces Health Surveillance Center, 2013; Gaydos, Howell, Quinn, McKee, & Gaydos, 2003; Goyal, Mattocks, & Sadler, 2012; Lehavot et al., 2014; Seña et al., 2000). Multiple studies have reported that military service members engage in high-risk sex behaviors, including having multiple partners, having concurrent partners, not using condoms or inconsistent use of condoms, and engaging in sexual activity under the influence of drugs or alcohol (Aldous et al., 2011; Jordan, Lee, Nowak, Johns, & Gaydos, 2011; Lehavot et al., 2014; Meyers et al., 2008). In one study of health behaviors in a representative sample of all active duty service members, both male and female, across all branches, over a third reported seldom or never using condoms with new sexual partners (Barlas, Higgins, Pflieger, & Diecker, 2013). Veterans also have higher prevalence of substance use disorders and mental health issues, including posttraumatic stress disorder (PTSD), which have been linked to sexual risk behaviors and STIs (Goulet et al., 2014; Hoerster et al., 2012; Hutton, McCaul, Santora, & Erbelding, 2008). There is also evidence that indicates that male veterans have a higher lifetime prevalence than nonveterans of having extramarital sex and paying for sex (London, Allen, & Wilmoth, 2013; London & Wilmoth, 2015). These findings indicate that male veterans may engage in more risky sex behaviors than their civilian counterparts.
Currently, there are no published studies directly comparing sexual risk behaviors and incidence of STIs in male veterans and nonveterans. This study sought to address this gap in the literature using nationally representative data from the National Survey of Family Growth (NSFG).
Method
Data Source and Study Population
All data analyzed in this study were from the cross-sectional survey, the 2011 to 2013 NSFG. The survey is designed by the Centers for Disease Control and Prevention’s National Center for Health Statistics to provide nationally representative data on reproductive and sexual health measures (Lepkowski, Mosher, Davis, Groves, & Van Hoewyk, 2010). The 2011 to 2013 NSFG uses a stratified, multistage, national probability sample of women and men 15 to 44 years of age living in households in the United States and oversamples Blacks, Hispanics, and teens (Lepkowski et al., 2010). Data are collected via face-to-face computer-assisted interviews or by Audio Computer-Assisted Self-Interviewing for more sensitive questions (e.g., sexual history information). The 2011 to 2013 data set included a total of 4,815 men, with a response rate of 72.1% (Copen, Chandra, & Febo-Vazquez, 2016). For this analysis, those who had never had sex with either a man or woman were excluded as were men younger than 18 years, since they could not have served in the armed forces. The current sample thus included all men aged 18 to 44 years who reported ever having had vaginal sex with a woman or oral or anal sex with a man.
Study Variables
The key independent variable of interest was veteran status, which was based on the response to the question “Have you ever been on active duty in the Armed Forces for a period of 6 months or more?” Covariates included age, race/ethnicity, education, income, marital status, and sexual orientation.
Behavioral outcomes included binge drinking (defined as having five or more drinks within a couple of hours, consistent with U.S. Department of Health & Human Services [2015] definition for men) at least once in the past year, ≥6 lifetime female sexual partners, ≥10 lifetime sexual partners (female or male), ≥2 past-year sexual partners (female or male), having partners of both sexes in the past year, and condom nonuse at last vaginal sex. STI outcomes included a positive self-reported history of chlamydia (past 12 months), gonorrhea (past 12 months), herpes (lifetime), genital warts (lifetime), or syphilis (lifetime), and an aggregate measure capturing reported history of at least one of the five aforementioned STIs. Self-reported receipt of STI treatment in the past 12 months was also assessed.
Because number of sexual partners was capped in the NSFG data set (e.g., for lifetime number of sexual partners, all participants with 50 or more partners were assigned a value of 50) for privacy protection reasons, number of partners could not be treated as a continuous variable in these analyses. In the absence of standardized definitions of the number of partners that constitutes “high-risk” behavior (Jonsson, Karlsson, Rylander, Gustavsson, & Wadell, 1997; Karlsson et al., 1995; Kjaer et al., 1997; Ley et al., 1991; Mosher, Chandra, & Jones, 2005; Santelli, Brener, Lowry, Bhatt, & Zabin, 1998; Vaccarella et al., 2006), cut points were modeled on other studies examining STI risk (Burk et al., 1996; Centers for Disease Control and Prevention, 2010; Datta et al., 2007; Fenton et al., 2001; Finer, Darroch, & Singh, 1999; Hariri et al., 2011; Karlsson et al., 1995; Kjaer et al., 1997; Ley et al., 1991; Mosher et al., 2005; Santelli et al., 1998; Stahlman et al., 2014; Vaccarella et al., 2006).
Statistical Analysis
Sociodemographic characteristics of the study sample, as well as behavioral and STI outcomes, were evaluated, by veteran status, using t tests for age and Rao–Scott corrected chi-squared tests for the remaining variables, which were all categorical. This correction was applied to account for the complex sampling design of these survey data. Sociodemographic characteristics were also compared for veterans versus nonveterans, using the same tests, but stratified by whether participants reported a history of STI. Weighted logistic regression analyses were then used to evaluate the bivariate associations between veteran status and each of the outcomes to calculate an unadjusted odds ratio (OR). Two sets of multivariable logistic regression analyses were performed. Outcomes with sufficient numbers of events (i.e., ≥150 events across both veterans and nonveterans), with an event defined as a single respondent who met the criteria for a variable (e.g., having had at least one STI), were adjusted for age, race/ethnicity, education, income, and marital status. Although sexual orientation was examined as a potential covariate, adjustment for this variable was not performed due to small numbers of participants identifying as not heterosexual or straight. Moreover, this variable did not differ by veteran status. Outcomes with fewer than 150 events were adjusted only for age and marital status, which were chosen a priori based on previously observed associations between these variables and sexual behaviors and STIs (Stamm et al., 1984; Taylor, Adimora, & Schoenbach, 2010).
Statistical analyses were performed in Stata SE software (version 14.1, StataCorp, College Station, TX). All estimates used survey weighting provided by the NSFG to reflect the national household population of men aged 15 to 44 years, with the use of the “subpop” command to restrict analysis to men who met the eligibility criteria. The University of Pittsburgh Institutional Review Board approved this analysis of the NSFG data set.
Results
Participant Characteristics
The current sample consisted of a total of 3,860 men: 409 veterans and 3,451 nonveterans. The sociodemographic characteristics of the study sample, as well as the unweighted sample sizes, are reported in Tables 1 and 2. Compared with nonveterans, veterans were slightly older, had higher income, and were more likely to be married or cohabiting. There were no significant differences between veterans and nonveterans with regard to race, education level, or sexual orientation. Among those reporting no history of STI, who constituted the majority of the sample, comparisons between veterans and nonveterans were similar. Among respondents reporting a history of at least one STI, veterans were more likely than nonveterans to be married or cohabiting, but there were no other significant differences.
Sociodemographic Characteristics of Adult Males Who Had Ever Had Sex, by Veteran Status (N = 3,860).
Note. p Values based on Rao–Scott chi-squared tests for independence.
Household income reported as percentage of Federal Poverty Level (FPL) as defined by the U.S. Census Bureau, adjusted for family size. Each interview year uses the FPL from the previous calendar year (e.g., 2011 interviews use 2010 FPL thresholds; National Center for Health Statistics, 2014).
Sociodemographic Characteristics of Adult Males Who Had Ever Had Sex, by STI Status (N = 3,860).
Note. STI = sexually transmitted infection. STI (−) refers to respondents who did not report a history of any STI; STI (+) refers to respondents who reported a history of at least one STI. p Values based on Rao–Scott chi-squared tests for independence.
Household income reported as percentage of Federal Poverty Level (FPL) as defined by the U.S. Census Bureau, adjusted for family size. Each interview year uses the FPL from the previous calendar year (e.g., 2011 interviews use 2010 FPL thresholds; National Center for Health Statistics, 2014).
Sexual Behaviors and STIs
Results of bivariate analyses are reported in Table 3. The mean age at first sex (not reported in Table 3) was 16.8 in both groups (SD = 3.3 for veterans, 3.6 for nonveterans). For behavioral outcomes, veterans were significantly more likely than nonveterans to have had ≥6 lifetime female partners (66.0% vs. 52.1%, p < .01), to have not used a condom at last vaginal sex (73.3% vs. 61.6%, p < .01), and to have had sex with both female and male partners in the past year (4.4% vs. 1.1%, p = .02). For STI outcomes, veterans were more likely than nonveterans to report ever having been diagnosed with genital warts (5.3% vs. 2.1%, p = .04).
Sexual Behaviors and History of Sexually Transmitted Infections of Adult Males Who Had Ever Had Sex, by Veteran Status (N = 3,860).
Note. p Values based on Rao–Scott chi-squared tests for independence.
Defined as having had ≥5 drinks within a couple of hours ≥1 time in the past 12 months.
The results of the unadjusted and adjusted logistic regression analyses are reported in Table 4. In adjusted regression models, veterans had greater odds than nonveterans of having ≥6 lifetime female partners (adjusted OR = 1.54, 95% confidence interval [CI 1.02, 2.31]), of having had sex with both female and male partners in the past year (adjusted OR = 4.80, 95% CI [1.16, 19.79]), and of having had gonorrhea in the past 12 months (adjusted OR = 3.15, 95% CI [1.17, 8.47]). It is worth noting that the point estimates of the ORs for all other sexual behaviors and STIs examined were also greater than 1, indicating that veterans trended toward greater risk, although these associations were statistically nonsignificant.
Unadjusted and Adjusted Logistic Regression Models (N = 3,860) With OR Expressed as Odds for Veterans Versus Nonveterans.
Note. OR = odds ratio; CI = confidence interval; STI = sexually transmitted infection. Full models adjusted for age, race/ethnicity, education, income, and marital status. Partial models adjusted for age and marital status.
Discussion
In this study using nationally representative data from the 2011 to 2013 NSFG, male veterans were identified as being more likely than nonveterans to have had six or more lifetime female partners and to have had gonorrhea in the past year. Veterans were also reported to be more likely than nonveterans to have had both female and male partners in the past year. These findings indicate that, like their female veteran counterparts, male veterans seem to engage in more risky sex behaviors than nonveterans (Lehavot et al., 2014).
The finding that veterans are more likely to have six or more lifetime female partners is significant because a higher number of sexual partners has consistently been associated with greater STI acquisition (Datta et al., 2007; Fenton et al., 2001; Jonsson et al., 1997; Mosher et al., 2005). In addition, compared with nonveterans, male veterans also had higher odds of having gonorrhea in the past 12 months. Though the reasons for this are unclear, potential explanations are that from 2012 to 2013 the rate of reported cases of gonorrhea in men increased over 4% and that treatment of gonorrhea infections has become challenging due to antimicrobial resistance, an issue not occurring to the same degree in other STIs (Centers for Disease Control and Prevention, 2014). Given that prior STI diagnosis is associated with risk for acquiring new STIs, these findings suggest that veterans may have a higher risk of contracting new STIs (Peterman et al., 2006; Santelli et al., 1998).
Similar to their female veteran counterparts, male veterans were identified as being more likely than nonveterans to have had both male and female sex partners (Lehavot et al., 2014). Having sex partners of both sexes has implications for sexual health screening efforts. Men who have sex with men are at high risk for infection with a number of STIs, including HIV (Workowski, 2015; Workowski & Bolan, 2015). Currently, the Centers for Disease Control and Prevention recommend that sexually active men who have sex with men be screened annually for HIV, syphilis, chlamydia, and gonorrhea (Workowski & Bolan, 2015). However, there are no consensus guidelines for routine STI screening of men who have sex with women, apart from HIV screening for all sexually active men aged 13 to 64 years (Kalmuss & Tatum, 2007; Workowski & Bolan, 2015). Given that the different recommendations are based on differences in sexual behavior rather than sexual orientation, it is important for health care providers to assess their male veteran patients’ sexual history and sexual behaviors in order to determine if they should be screened for STIs. Although health care providers may be less familiar with discussing sexual health issues with male patients than with female patients, it remains the only way to assess risk. Yet less than half of sexually active men aged 20 to 44 years report receiving any sexual or reproductive health care, including counseling or advice about STIs, in a given year (Kalmuss & Tatum, 2007).
Proactively and routinely eliciting sexual histories in an open and nonjudgmental manner improves risk assessment, helps build trust and patient satisfaction, and can promote self-efficacy and safer sex and thus better sexual health outcomes (Nusbaum & Hamilton, 2002). This may be a particular challenge because of the lack of consensus guidelines for men’s sexual health (The Alan Guttmacher Institute, 2002; Marcell & Male Training Center for Family Planning and Reproductive Health, 2014; Sonfield, 2002), and that men seek fewer preventive health services (Courtenay, 2000; Pinkhasov et al., 2010; Springer & Mouzon, 2011; Williams, 2003). Nonetheless, periodically performing risk assessment, risk reduction counseling, and screening of men who may be at higher risk for STIs, such as male veterans, has the potential to benefit the sexual health of men and their partners.
There are a number of possible reasons that male veterans may engage in riskier sexual behaviors than their civilian counterparts. An important factor historically implicated in high-risk sexual behavior and STIs is alcohol use (Boden, Fergusson, & Horwood, 2011; Connor, Kydd, & Dickson, 2015; Cooper, 2002, 2006; Halpern-Felsher, Millstein, & Ellen, 1996; Kiene, Barta, Tennen, & Armeli, 2009; Leigh, 2002). Interestingly, no significant association between veteran status and binge drinking at least once in the past year was identified. This is consistent with findings in female veterans who were also reported to have higher rates of risky sex behaviors and STIs but not higher rates of binge drinking, suggesting that observed behavioral differences among veterans may be rooted in other psychosocial antecedents (Lehavot et al., 2014). One possibility is that the men who choose to serve in the military have higher risk tolerance or greater sensation seeking at baseline or are more likely to engage in sexual risk behaviors than those who do not join the military, and that these behavioral patterns continue during and after their service (Goyal et al., 2012; Thompson, Kao, & Thomas, 2005). No difference in age at first sex was identified, though this finding alone does not rule out a “higher baseline risk” explanation. It is also possible that the sexual environment that service members are exposed to in the military, in which there is a high prevalence of partner concurrency, inconsistent condom use, and having sex under the influence of alcohol, leads to the adoption of riskier sexual practices that continue into their lives as veterans (Barlas et al., 2013; Goyal et al., 2012; Harbertson et al., 2015; Seña et al., 2000; Stahlman et al., 2014).
Another potential explanation for these findings is that psychological effects associated with military service may carry over into veterans’ lives and affect their sexual risk taking behaviors. Mental health disorders, which are disproportionately observed in veteran populations, have been linked to riskier sexual behaviors (Berenson, Breitkopf, & Wu, 2003; Lehavot, Hoerster, Nelson, Jakupcak, & Simpson, 2012; Marshall et al., 2013; Seal, Bertenthal, Miner, Sen, & Marmar, 2007; Stahlman et al., 2014). In particular, rates of PTSD are substantially higher among veterans than civilians, and research indicates that individuals with PTSD not only engage in riskier sexual behavior but also use sex as a coping mechanism (Cohen et al., 2012; Hoerster et al., 2012; Howard, 2007; Killgore et al., 2008; Stecker, Fortney, Owen, McGovern, & Williams, 2010). Unfortunately, the NSFG does not assess mental health conditions.
The higher risk for STIs among veterans observed in this study could also be due to factors not examined in this study such as differences in patterns of sexual activity (e.g., having concurrent sexual partners or exchanging sex for drugs or money) or of sexual networks (Harbertson et al., 2015). In addition, preexisting cultural norms about masculinity may be more concentrated in the military environment, which may promote greater risk-taking behaviors, potentially leading to long-term behavior change (Morris, Smith, Farooqui, & Surís, 2014). Although this study cannot distinguish between these possibilities, when interpreted in light of other research, these findings suggest that military service has or is linked to lasting effects on veterans’ health and may contribute, in varying and interconnected ways, to higher risk sexual behavior and increased risk for STIs. Future research is needed to determine the underlying reasons why veterans engage in higher risk sexual behaviors than their nonveteran counterparts.
This analysis has some important limitations to consider. First, the NSFG is a cross-sectional study, so it does not include data about sexual behaviors across the life course or even in relation to the timing of military service. The lack of this temporal information prevents determination of the possibility of a causal relationship between military service and the risks identified here. Second, although the NSFG is nationally representative, the survey weights were not created with veteran status in mind and thus, this sample may be less representative of male veterans than it is of all men aged 18 to 44 years. It is also important to note that this sample includes only men younger than age 45, who comprise about 19% of all veterans (U.S. Department of Veterans Affairs Office of the Actuary, 2014). Third, the study’s outcomes are based on self-report, which may lead to underreporting of the number of sexual partners or STI histories. However, the majority of the outcomes were obtained through Audio Computer-Assisted Self-Interviewing rather than face-to-face interviews, which should, in theory, limit the effects of social desirability bias. Fourth, the study is limited by the behavioral variables included in the NSFG questionnaire, which did not allow for further exploration of other specific sexual practices that could also contribute to risk, such as anonymous sex or group sex. Fifth, the aggregate STI measure combines two past-year diagnoses and three ever diagnoses, rendering it neither a wholly past-year nor a lifetime measure. Nevertheless, the measure may more accurately reflect morbidity as gonorrhea and chlamydia may be acquired (and treated) multiple times, while herpes and genital warts are incurable and often cause recurrent symptoms. Sixth, this study did not consider the potential effects of sexual assault on STI transmission or care-seeking behavior (Morris et al., 2014). Although the NSFG does include two questions about lifetime history of sexual assault, because the timing is unknown and because of the concern for potential differences in underreporting between veterans and nonveterans (Morris et al., 2014), they were not included. Finally, this study did not consider the possible effects of deployment on STI transmission (Harbertson et al., 2015), as the NSFG does not ask about deployment history.
In summary, this is one of the first studies to compare male veterans’ sexual behaviors and STI histories to those of nonveterans in a nationally representative sample. It identifies higher rates of sexual risk behaviors and prior STI in veterans. Health care providers need to be aware that male veterans may be at higher risk for STIs and routinely assess their sexual behaviors, screen for STIs accordingly, and perform risk reduction counseling during clinical encounters. In addition, future research is necessary to further understand the ways in which the experiences of military service and the sexual environment therein potentially influence veterans’ sexual behaviors and risk for STIs.
Footnotes
Acknowledgements
The authors would like to thank Dan Winger of the Clinical and Translational Science Institute (CTSI) at the University of Pittsburgh, who provided statistical consultation.
Authors’ Note
Some of the data included in this article were presented as an abstract poster presentation at the Society of General Internal Medicine’s 2016 Annual Meeting, May 11 to 14, 2016.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Mr. Evans was funded by the Clinical Scientist Training Program at the University of Pittsburgh School of Medicine. Dr. Borrero’s effort on this study was supported by grant # R21 HD076327. CTSI support funded by National Institutes of Health Clinical and Translational Science Award program, grants UL1 RR024153 and UL1TR000005.