Self-Esteem in HIV-Positive and HIV-Negative Gay and Bisexual Men: Implications for Risk-Taking Behaviors with Casual Sex Partners

Abstract

Research suggests that self-esteem in gay and bisexual men may be linked with sexual risk-taking behaviors. As part of a larger investigation into the sexual practices of gay and bisexual men, we assessed serostatus, self-esteem, condom use, and HIV disclosure to sexual partners. Among HIV-negative men, no relationships were found between their self-esteem and tendency to discuss their and their partners’ HIV status. However, among HIV-positive men, there was a positive relationship between self-esteem and disclosure to receptive and insertive anal sex partners. These results suggest greater attention to the self-esteem of HIV-positive men by attending healthcare workers and social support groups.

INTRODUCTION

Stigma and social support have been concepts widely studied as influential over the sexual risk practices of gay and bisexual men¹. Though related to stigma and social support², self-esteem has been featured in far fewer research^3–7 studies as an independent contributor to such behaviors. Self-esteem is “the valuative experience of oneself as a social object…one’s chronic, overall sense of worth as an individual with social significance”⁸; it is how much one likes himself. This construct has been negatively correlated to HIV-stigma and positively correlated with social support in a previous study². As such, high or low self-esteem may prove to be influential agents over the tendencies to protect oneself from acquiring HIV or from further spreading the disease (i.e., variations in condom use and HIV disclosure). Hypothetically, discomfort with, or dislike of, oneself may translate into a decreased likelihood to discuss the disease with others or to use condoms, even if nondisclosure or condom disuse may contribute to the spread of HIV.

As mentioned, there have been few studies that explored self-esteem as an independent predictor of sexual risk-taking behaviors. Furthermore, these mostly used subgroups of gay and bisexual men as samples. For example, Adam, Husbands, Murray, and Maxwell³ found qualitative evidence showing an association between decreased self-esteem and lower condom use among barebackers (i.e., men who actively eschew condoms). Rosario, Schrimshaw, and Hunter⁴ found similar, but less direct effects for gay and bisexual teenagers not identifying as barebackers; the number of sexual partners was a mediating factor. Preston, D’Augelli, Kassab, and Starks⁵ found that in rural men who have sex with men, increased stigma, related to low self-esteem, increased sexual risk-taking behavior. Similar results were found for African American men who have sex with men⁶. As for HIV disclosure, Dowshen, Binns, and Garofalo² found no association between self-esteem and “disclosure concerns” among HIV-positive young (16–24 year old) gay and bisexual men. Zea, Reisen, Poppen, Bianchi, and Echeverry⁷ found in Latino gay and bisexual men increased disclosure rates to friends, family, and main partners to be associated with increased self-esteem; however, they did not explore these effects with casual male partners.

Three goals for this current study emerge from the available literature. First, previous research^3–6 indicates there may be a trend towards condom disuse in men reporting lower self-esteem. Yet further research, using a larger sample of gay and bisexual men (and not just specific subgroups; e.g., barebackers, Latinos, teens) is warranted. A larger, more representative sample will increase the generalizability of findings to more gay and bisexual men. Second, neither Dowshen and colleagues² nor Zea and colleagues⁸ specifically explored the correlation between self-esteem and actual disclosure to casual sexual partners. Conducting such additional analyses would add subsequent value to their findings. Third, no studies to our knowledge have explored the degree to which HIV serostatus acts as a moderator between self-liking (i.e., high or low self-esteem) and risk-taking behaviors. These three goals justify further exploration of the relationships between self-esteem and risk behavior in HIV-positive and HIV-negative gay and bisexual men. We hypothesized that, despite HIV serostatus, self-esteem would be negatively related to condom use and HIV-disclosure during receptive and insertive anal intercourse with casual partners.

METHODS

Procedures

Attendees at two events (the International Mr. Leathermen Competition in Chicago, Illinois, and PrideFest in Milwaukee, Wisconsin) were asked to participate in a brief paper survey. Specifically, the research staff solicited all male individuals passing by the booth at both events, regardless of conspicuous sexual orientation (i.e., not merely those wearing rainbow flags). If they consented, the men were given a survey on a clipboard and a foldout chair on which to sit. Because we solicited all men, there was virtually no experimenter bias. The research staff used the same communication, “Take our 5–10 minute survey on sexual health and receive $5.00.” If any of the men wanted further explanation, they were directed to the study coordinator who further explained the survey and gave background information on the affiliated institution. Participants were given $5.00 cash upon completing the survey.

Participants

This study focused on the sexual health, risk practices, and behaviors of gay and bisexual men. Women were not surveyed. The final sample (N = 1,468) was comprised of 1,382 gay men (94.1%) and 86 bisexual men (5.9%).

Measures

HIV Status

Participants indicated their HIV-status by self-report (i.e., HIV-positive, HIV-negative, or unknown). Too few men reported an unknown status; thus, only men reporting an HIV-negative or HIV-positive status were included in the analyses.

Self-Esteem

We used the self-liking dimension of the Self-Liking/Self-Competence Scale-Revisited Version (SLCS-R)³ to assess participants’ degrees of self-esteem (α = .91). This 8-item measure asked participants to rate their agreement with statements like, “I never doubt my personal worth,” or, “I feel great about who I am” (1 = “strongly disagree”, 7 = “strongly agree”). Higher scores indicated higher self-esteem.

Condom Use

Participants were asked, “In the past 90 days, how often did you use condoms during receptive anal intercourse and (as a separate question) insertive anal intercourse with casual male partners?” They signified their specific use of condoms (e.g., 1 = “I never wear condoms,” 2 = “I rarely wear condoms,” 3 = “I sometimes wear condoms,” 4 = “I usually wear condoms,” or 5 = “I always wear condoms”). The scale was treated continuously in the analyses. Alternatively, men could select, “I have never had this sort of partner,” if they had not participated in receptive or insertive anal intercourse in the previous 90 days. Such men were omitted from the respective analyses.

Disclosure to Casual Partners

As when measuring condom use, we separated HIV disclosure by partner type: partners where the participants were receptive and (as a distinct measure) partners where the participants were insertive during anal intercourse. We asked, “In the past 90 days, how many receptive anal sex partners have you had? (Please exclude your main partner).” This was followed by, “with how many did you discuss your HIV-status? (Please exclude your main partner).” A ratio then was created in which number of disclosed partners was divided by number of total partners (e.g., a score of 1 signified 100% disclosure, a score of 0 signified 0% disclosure). Men who reported never having any partners with whom they were receptive or insertive during anal intercourse were omitted from the analyses.

Statistical Analysis

All analyses were conducted using JMP statistical software, version 5.1.2. The data were analyzed using stepwise multiple moderated regression models to assess the associations between the variables. The sample size varied across tests; individuals could report not having had anal sex in the past 90 days. We used stepwise statistical methods to control for certain key variables. Specifically, we controlled for age and survey location in all the regressions (due to their statistically significant, negative relationships with condom use), which were entered on the first step. On step two, we controlled for receptive disclosure (for receptive condom use) and insertive disclosure (for insertive condom use) because the tendency towards condom disuse could have been explained by a high degree of disclosure⁹. On step three, we controlled for number of receptive partners (for receptive condom use) and number of insertive partners (for insertive condom use) to control for the mediation found in previous research between self-esteem, number of partners, and increased sexual risk behaviors⁵. The independent (self-esteem) and moderator variables (HIV-status) were entered on step four. The interaction term, created by multiplying the given moderator (HIV-status) with the independent variable (self-esteem)¹⁰, was entered last. When the interaction term proved to be significant, we deconstructed the relationship between HIV status and self-esteem to measure the strength of the association by serostatus (i.e., HIV-negative versus HIV-positive men).

After we tested condom use as the dependent variable, we repeated the same statistical processes using disclosure when the receptive partner and disclosure when the insertive partner as the dependent variables. We controlled for receptive and insertive condom use on step two, following the same procedures previously described.

RESULTS

Sample

The sample of gay and bisexual men ranged in age from 18 to 73 (M = 38.86, SD = 11.97). Seventeen men reported an unknown serostatus and were omitted from the following analyses. Of the remaining men, 1,169 were HIV-negative (80.6%) and 282 were HIV-positive (19.4%). Most respondents were White (76.4%) and most completed college, some graduate school, or had a graduate degree (62.5%). On average, the men disclosed their serostatus when receptive during anal intercourse with 73.74% (SD = 41.36%) of partners and with 74.31% (SD = 41.25%) of partners when insertive. The men “usually” used condoms when receptive during anal intercourse (M = 3.73, SD = 1.46, 1–5) and when insertive (M = 3.71, SD = 1.43, 1–5). The men also reported moderately high levels of self-esteem (M = 4.99, SD = 1.25, 1–7).

Self-Esteem

No differences were found between HIV-positive (M = 5.03, SD = 1.27) and HIV-negative men (M = 4.98, SD = 1.25) regarding self-esteem. We tested if self-esteem and the HIV/self-esteem interaction could predict receptive and insertive condom use. Self-esteem was not related to either receptive or insertive condom use as a main effect or as an interaction effect (i.e., HIV-negative and HIV-positive men did not vary differently in their condom use depending on self-esteem.)

Following, we tested a model in which self-esteem and its interaction with HIV-status predicted disclosure when the receptive partner. This was significant, F(7, 550) = 2.07, p = .04, R² = .03. Receptive condom use was not related to receptive disclosure, t(550) = 1.73, p = .08, β = .08. Self-esteem produced a main effect, where gay and bisexual men’s tendencies towards disclosure increased with self-esteem, t(550) = 2.52, p = .01, β = .12. This effect was not mediated by number of receptive partners. Furthermore, the HIV serostatus and self-esteem interaction was significant, t(550) = 2.02, p = .04, ΔR² = .01, β = .09. Subsequent analyses indicated that no relationship existed for self-esteem and receptive HIV-disclosure (p = .66, β = .02) for HIV-negative men; however, as shown in the top portion of Figure 1, a positive relationship existed for self-esteem and receptive HIV-disclosure (p < .01, β = .22) for HIV-positive men.

Figure 1.

HIV-positive men’s self-esteem predicting HIV disclosure when the receptive partner and when the insertive partner. The topmost graph represents disclosure when the receptive partner (p < .01, β = .22). The bottommost graph represents disclosure when the insertive partner (p < .001, β = .30). Both graphs represent the relationships among HIV-positive men after controlling for the variables: age, survey location, number of receptive partners and receptive anal condom use (for disclosure when the receptive partner), and number of insertive partners and insertive anal condom use (for disclosure when the insertive partner). Horizontal lines in the graphs represent the mean disclosure rates for the sample.

Last, we tested self-esteem and its interaction with HIV-status on disclosure when the insertive partner. This also was significant, F(7, 546) = 4.72, p < .001, R² = .06. Insertive condom use was positively related to insertive disclosure, t(546) = 3.73, p < .001, β = .17. Self-esteem produced a main effect, where gay and bisexual men’s tendencies towards disclosure increased with self-esteem, t(546) = 3.26, p < .01, β = .16. This effect was not mediated by number of insertive partners. Furthermore, the HIV serostatus and self-esteem interaction was significant, t(546) = 2.10, p = .04, ΔR² = .01, β = .10. Subsequent analyses indicated that no relationship existed for self-esteem and insertive HIV-disclosure (p = .31, β = .05) for HIV-negative men; however, as shown in the bottom portion of Figure 1, a positive relationship existed for self-esteem and insertive HIV-disclosure (p < .001, β = .30) for HIV-positive men.

DISCUSSION

Our results indicated that self-esteem was an important independent predictor of HIV disclosure, but was not influential over condom use. Specifically, positive main effects were found between self-esteem and HIV disclosure during both penetrative roles. It should be noted that these main effects have dubious generalizability to all gay and bisexual men given the subsequent interaction effects shown to be significant. The relatively strong, positive relationship between self-esteem and insertive/receptive HIV disclosure found in HIV-positive men (and not HIV-negative men) was ultimately the impetus for the main effect. Thus, the results showed that self-esteem signifies little about the sexual risk-taking practices of HIV-negative men; however, for HIV-positive men, nondisclosure could be a symptom of self-dislike. Numbers of receptive or insertive partners were not found to mediate this relationship, as found in previous research⁵.

HIV serostatus did not predict variations in self-esteem, which suggested that acquiring the disease did not systematically ensure the adoption of self-disliking attitudes. Among those HIV-positive men holding negative attitudes about themselves, their tendency towards nondisclosure could protect from the potential sexual rejection that might increase their low self-esteem. It might be an ego-defensive tactic. Alternatively, the tendency towards disclosure in HIV-positive men with high self-esteem might represent amplified self-security. Even if disclosure might lead to a rejection by the sexual partner, such men’s self-security might be sufficiently elevated as to make the rejection event insignificant. These merely are speculations. Future qualitative interviews of HIV-positive men might certify the validity of these claims.

We conducted this research with three goals: to increase the sample size and its representativeness of gay and bisexual men beyond particular ethnic, racial, or social groups (e.g., Latino men, barebackers), to assess previously overlooked disclosure rates to actual casual partners, and to examine HIV as a moderator of the potential association between self-esteem and negative sexual risk practices. In contrast to previous studies on condom use and self-esteem in barebacking⁴, teenage⁵, rural⁶, and African American⁷ gay and bisexual men, the large population of mostly white, middle-aged, and educated gay and bisexual men we surveyed showed no significant results. This suggests that self-esteem may be related less with safer sex and more with the social ramifications that communicating an HIV-positive status produce (e.g., alienation, stigma, negative impressions, or intrapersonal discomfort with being HIV-positive). However, when exploring actual disclosure rates to partners, we found similar and significant results to those documented by Zea and colleagues⁸ (i.e., a positive association between self-esteem and disclosure rates to friends, family, and main partners by HIV-positive Latino gay and bisexual men). Finally, in satisfying our third goal, we showed self-esteem to be a relevant factor only for HIV-positive men and only when predicting actual disclosure.

This research was not without its limitations. We cannot be certain of the reversibility of the relationships. It may be that self-esteem influences disclosure; or disclosure (i.e., being comfortable and having enough self-efficacy to discuss the disease even when rejection might be a corollary) contributes to increased self-esteem. Conversely, it ultimately may be that nondisclosure in HIV-positive men contributes to regret and remorse that, over time, may underlie low self-liking. Other measures of associated psychological states such as guilt or remorse would have increased the scope of the results and added insight towards such alternative hypotheses. Additionally, we only measured one dimension of self-esteem (i.e., self-liking). Other dimensions of self-esteem, such as self-competency³, may either have strengthened or weakened the results.

The rate of actual disclosure by the participants may be a limitation. This rate was unusually high when compared against previous studies. For example, the sample reported disclosing to roughly 75% of casual partners; studies^9,11–12 consistently show this rate to be more like 25% to 48% of partners. This could suggest a tendency towards social desirability by our sample, which unfortunately was not measured by the original paper survey. Our measurement of disclosure also did not take into account indirect communication methods (e.g., inclusion of one’s serostatus on an internet profile, disclosure by friends). It did not take into account familiarity with the casual partners or the greater informational and psychological contexts that accompany disclosure (e.g., testing rates, knowledge of serostatus, disclosure expectancies). We solely focused on the active discussion of serostatus by the sex partners. Granted disclosure rates were still quite high, this may be a limitation.

Despite these limitations, the results suggested some important implications for individuals who counsel, research, or work with HIV-positive men. HIV health managers might initiate more comprehensive discussions about HIV disclosure to casual partners, particularly when interacting with an individual exhibiting or expressing low self-esteem. Behavior modification techniques such as interpersonal skill training and scenario-based learning might help make men feel more comfortable disclosing their serostatus to casual partners. In addition, researchers more accurately might define the association between self-esteem and disclosure in HIV positive men. Bolstering self-esteem in HIV-positive men is important for both men’s psychological and physiological health. If it were found that increased HIV disclosure ultimately caused increases in self-esteem (i.e., the reverse hypothesis), new interventions could be applied to increase self-esteem. These might include increasing men’s self-efficacy to disclose through role-play exercises initiated by health managers, or group counseling for HIV-positive men by HIV-positive men (i.e., community groups) to discuss successful ways to disclose. Even a passive, information-only intervention may be helpful—for example, a media campaign educating men living in high HIV prevalence areas that disclosure not only prevents the spread of HIV, but also is psychologically therapeutic for those already infected. Overall, our research suggests further evidence for the link between psychological affect and behavior enactment in HIV-positive men. The relationships described in this study should be noted when constructing and implementing future interventions to increase HIV disclosure or to increase self-esteem.