The Role of Maladaptive Cognitions in Hypersexuality among Highly Sexually Active Gay and Bisexual Men (2014)

Participants and Procedure

Beginning in February 2011, we began enrolling participants utilizing a combination of recruitment strategies: (1) respondent-driven sampling; (2) internet-based advertisements on social and sexual networking websites; (3) email blasts through New York City gay sex party listservs; and (4), active recruitment in New York City venues, such as gay bars/clubs and sex parties. Participants recruited from the internet or in active venue-based recruitment shifts were pre-screened utilizing a brief survey either via the online survey site Qualtrics (www.qualtrics.com) or a mobile survey via iPod Touch, respectively. This pre-screen assessed number of sex partners in addition to variables relevant to other studies for which we were screening. All participants completed a brief, phone-based screening interview to confirm eligibility, which was defined as: (1) at least 18 years of age; (2) biologically male and self-identified as male; (3) a minimum of nine different male sexual partners in the prior 90 days, with at least two in the prior 30 days; (4) self-identification as gay, bisexual, or some other non-heterosexual identity (e.g., queer); and (5) daily access to the internet in order to complete internet-based assessments (i.e., at-home surveys, daily diary).

Participants were excluded from the project if they demonstrated evidence of serious cognitive or psychiatric impairment that would interfere with their participation or limit their ability to provide informed consent, as indicated by a score of 23 or lower on the Mini-Mental Status Examination (MMSE) (Folstein, Folstein, & McHugh, 1975) or evidence of active and unmanaged symptoms on the psychotic symptoms or suicidality sections of the Structured Clinical Interview for the DSM-IV-IR (SCID) (First, Spitzer, Gibbon, & Williams, 2002).

We operationalized highly sexually active as having at least nine sexual partners in the 90 days prior to enrollment, with at least two of these partners being within the prior 30 days. These cutoffs were based on prior research (Grov et al., 2010; Parsons, Bimbi, & Halkitis, 2001; Parsons et al., 2008), including a probability-based sample of urban MSM (Stall et al., 2002, 2003) that found that 9 partners was more than 2–3 times the average number of sexual partners among sexually active gay and bisexual men. For the purposes of this study, sexual partners were defined as contact with any male partner with whom the participant engaged in sexual activity that had the potential to lead to orgasm, which included, but was not limited to, receptive/insertive anal intercourse, receptive/insertive oral intercourse, receiving or performing anal stimulation manually or orally, and mutual masturbation. All eligibility criteria were confirmed at the baseline appointment, with sex criteria being confirmed using the timeline follow-back interview in which a calendar is used to recall one’s daily sexual behavior (Sobell & Sobell, 1992).

Participation in the study involved both at-home (internet-based) and in-office assessments. After a member of the research staff confirmed participants’ eligibility over the phone, participants were sent a link to complete an internet-based survey at home prior to their first in-office appointment that took approximately one hour to complete. Initial informed consent for completing the at-home survey was obtained as part of the online survey. Participants then completed a series of two baseline appointments at the research site and provided informed consent for their full participation in the year-long project at the beginning of their first face-to-face appointment. All procedures were reviewed and approved by the Institutional Review Board of the City University of New York. This article focuses exclusively on the baseline at-home survey data to examine the psychometric properties of a newly created instrument intended to measure maladaptive cognitions about sex.

Measures

Factor Analyses of the Maladaptive Cognitions about Sex Scale

The results of the CFA are shown in Table 2. We conducted an initial analysis with all items and then made iterative modifications to the scale based on model parameters and modification indices to eliminate psychometric complications such as local dependence (i.e., residual correlations between items) and cross-loading onto multiple factors. Although these problems can be easily dealt with statistically using latent variables, they present difficulties when attempting to use non-latent modeling such as simple linear regression with calculated subscale scores based on average item responses rather than factor analytic results. As such, these decisions were made in order to develop a scale that can be successfully utilized both within and outside of the latent modeling framework.

The initial factor loadings column in Table 2 displays both the unstandardized and standardized results of the CFA with all 17 items entered onto their respective factors. As can be seen in Table 2, the initial model did not fit the data well—the CFI and TLI were both less than 0.95 and the RMSEA was above 0.06. There were several sources of misfit for the original model. Items 8, 13, and 14 loaded poorly onto the Disqualified Benefits subscale relative to the other items and were thus removed from future iterations. Item 1 was removed due to a high residual correlation with Item 2 and Item 4 was removed due to residual correlation with several other items on the Magnified Necessity subscale. The presence of residual correlations suggests that, in addition to the factor of interest, the items shared another unmeasured construct in common which resulted in remaining covariation that was unexplained by the model which can bias non-latent uses of the scale that do not take their covariation into account. Item 12 was removed as a result of cross-loading onto the Minimized Self-Efficacy subscale as well as potential residual correlations with several items on that subscale.

The final CFA model had significantly improved fit, with all indices other than the chi-square test statistic indicating strong fit to the data based on established thresholds. The Magnified Necessity subscale contained Items 2, 3, 5, 6, and 7; the Disqualified Benefits subscale contained Items 9–11; the Minimized Self-Efficacy subscale contained Items 15–17. The resultant factors were also improved by the removal of items—for example, the variance of the Disqualified Benefits factor more than doubled. Interestingly, the correlations of the Magnified Necessity and Disqualified Benefits subscales with the Minimized Self-Efficacy subscale did not change appreciably between the original and final models. The hypothesized lack of correlation between the Necessity and Benefits subscales was supported by the model. When allowed to vary freely and be estimated by the model, the correlation was estimated to be 0.07, was non-significant, and worsened fit of the overall model.

Modeling the Association Between MCAS Subscales and Hypersexuality

Having confirmed the best-fitting structure for the three MCAS subscales, we next sought to test the structural relationships among them and hypersexuality screening results. Results of the SEM analysis are shown in Fig. 1. The SEM analysis confirmed a cognitive model of hypersexuality consistent with self-regulatory efficacy models of behavior, as described in the Discussion. Model fit was excellent, with all indicators exceeding minimum criteria for good fit. Both the Magnified Necessity and Disqualified Benefits subscales had significant direct effects on the Minimized Self-Efficacy subscale, suggesting that higher levels on these two factors were associated with more minimizing of one’s sexual self-efficacy; the Magnified Necessity subscale was a considerably stronger predictor of Minimized Self-Efficacy than was the Disqualified Benefits subscale. All three subscales significantly predicted screening positive for hypersexuality and explained 45% of the variation in screening results. The influence of Magnified Necessity and Disqualified Benefits on screening positive for hypersexuality was partially mediated by Minimized Self-Efficacy–both had significant direct effects through Minimized Self-Efficacy. In all, Magnified Necessity was the strongest predictor of screening positive for hypersexuality with a total effect of 0.55 compared to 0.32 for Disqualified Benefits and 0.26 for Minimized Self-Efficacy.

Demographic Differences in the MCAS Subscales

Using a one-way ANOVA with Fisher’s least significant difference (i.e., LSD) post-hoc tests, we found significant differences in scores on the Disqualified Benefits subscale by racial/ethnic background. Black men had higher scores on the Disqualified Benefits subscale than Latino (p = .004), White (p = .02), and men of unknown background (p = .01); Latino men had lower scores than multiracial men (p = .04) in addition to Black men; men who were multiracial had higher scores than men of unknown background (p = .03) in addition to Latino men. No significant racial/ethnic differences were found with regard to the Magnified Necessity or Minimized Self-Efficacy subscales and we did not identify any differences in the three MCAS subscales by HIV status, employment, educational attainment, or relationship status.

Bivariate Association of the MCAS Subscales with Relevant Psychosocial Variables

We next explored the bivariate correlations between the three MCAS subscales and other psychosocial variables that have been theoretically or empirically proposed to impact hypersexuality. As can be seen in Table 3, we found similar patterns of associations across the three subscales, with each having a significant and positive correlation with impulsivity, emotional dysregulation, depression/anxiety, and sexual compulsivity. The Magnified Necessity and Minimized Self-Efficacy subscales were significantly and positively associated with sexual excitation while the Disqualified Benefits subscale had a coefficient of nearly zero. All three MCAS subscales were significantly and positively associated with the Sexual Inhibition subscale corresponding to inhibition due to the threat of performance failure (i.e., Sexual Inhibition I), while only the Disqualified Benefits subscale was association with the Sexual Inhibition subscale related to inhibition resulting from the threat of performance consequences (i.e., Sexual Inhibition II). Many of the psychosocial variables also had strong associations with each other.

Maladaptive Cognitions about Sex and Sexual Minority Male Development

Gay and bisexual men are significantly more likely to report maladaptive cognitions, such as low self-worth and hopelessness, across the life course than heterosexual men (e.g., Hatzenbuehler, 2009; Hatzenbuehler, McLaughlin, & Nolen-Hoeksema, 2008; Safren & G., 1999). Gay and bisexual men might experience more cognitive biases specifically about sex given their disproportionate exposure to childhood sexual abuse, minority stressors around their sexual orientation, and the secrecy and shame that often surrounds an emerging gay or bisexual identity across much of early development (D’Augelli, 2002; Lelutiu-Weinberger et al., 2011; Pachankis & Bernstein, 2012; Parsons et al., 2012; Stall et al., 2003). For example, childhood sexual abuse is associated with cognitive distress and rumination (Briere & Elliott, 2003), which in turn partially mediate the relationship between childhood sexual abuse and consumption behaviors, like eating and substance use, to cope with distress (Sarin & Nolen-Hoeksema, 2010). Further, hiding a core aspect of one’s identity, such as one’s sexual orientation, across an important period of development has been shown to powerfully shape one’s self-concept and health behavior (Pachankis & Hatzenbuehler, 2013). While not directly tested here, a model that locates the source of maladaptive thoughts about sex in adolescent development is consistent with developmental models of minority stress and other health behaviors. Inclusion of a measure of maladaptive cognitions about sex in studies of gay and bisexual men’s development can further elucidate the role of cognition in models of gay and bisexual men’s sexuality and the consequences of minority stress experiences.

Clinical Implications

Our findings regarding the contribution of magnified benefits, disqualified drawbacks, and minimized self-efficacy in a predictive model of hypersexuality were consistent with existing case studies and clinical guidance for treating this phenomenon (e.g., Shepherd, 2010; Weiss, 2004) as well as approaches to treating other sexual problems, such as exhibitionism and fetishism (Murphy & Page, 2008; Wincze, 2000). Cognitive approaches in these treatments facilitate accurate appraisals of the potential consequences of a given sexual activity and foster self-efficacy for controlling one’s problematic sexual behavior. Further, treatment approaches for other behavioral excess problems (e.g., substance abuse, pathological gambling) employ cognitive restructuring techniques ranging from abstractly construing tempting stimuli (e.g., Hofmann, Deutsch, Lancaster, & Banaji, 2010) to interfering with the automatic processing of temptations (e.g., Wiers, Rinck, Kordts, Houben, & Strack, 2010). These techniques ultimately build self-efficacy for behavior change, more adaptive beliefs about the problem behavior, and self-control (Marlatt & Gordon, 1985). An intervention that aimed to facilitate insight into self-justifications for recent unprotected anal sex among men who have sex with men yielded a 60% reduction in unprotected anal sex among recipients compared to no change among a group who received standard HIV risk-reduction counseling (Dilley et al., 2007). The results of numerous relapse prevention studies examining other health-risk behaviors demonstrate that interventions that change cognitions about one’s problematic behavior can, in fact, lead to reductions in that behavior.

Because our study could not establish causality, clinical implications must be drawn with caution. While reductions in maladaptive cognitions might precede reductions in hypersexual behavior, we cannot rule out the possibility that maladaptive cognitions might follow problematic behavior or that an unmeasured third variable might explain the relationship between cognition and behavior. Still, the results of the present study suggest that high levels of maladaptive thoughts about sex, especially disqualified benefits of sex, co-occur with more problematic hypersexuality. In fact, it is possible that the primary factor differentiating highly sexual active gay and bisexual men who screen positive and negative for hypersexuality may be the level of cognitive distress experienced by gay men with problematic hypersexuality although this possibility awaits empirical examination. Our results were also consistent with the possibility that a healthy cognitive perspective on sexuality might be inconsistent with recurrent, hard to control sexual fantasies, urges, and behaviors associated with significant personal distress and adverse consequences. Thus, our results suggest that treatment approaches that induce negative attitudes toward sexuality fail to highlight the benefits of sex and encourage the belief that one is not in control of his sexual behavior may unintentionally serve to perpetuate, rather than reduce, hypersexuality.

The results of this study approach, but largely circumvent, an important nomenclatural issue with clinical implications. Specifically, the reification of problematic hypersexuality in a standard diagnostic nomenclature and research agenda could be argued to pathologize a healthy aspect of human life. This argument may be especially important for gay and bisexual men, a group of individuals whose sexuality has been variably pathologized across modern history, a social problem which continues today (Gallup, 2012). However, the presence of extremely rigid or inaccurate thoughts about sex among gay and bisexual men represents a clinical problem in and of itself, potentially even a pathognomonic symptom of problematic hypersexuality, regardless of any argument for and against the moral or social value of intense sexual fantasies, urges, or behaviors. As a result, the identification and treatment of maladaptive thought content and associated cognitive processes about sex using valid measures and conceptual models represents a key mental health priority regardless of its association with a specific mental health problem. This study suggests that reducing the cognitive distress faced by men who experience problematic hypersexuality, rather than reducing levels of sexual behavior, may itself reduce problematic hypersexuality.

Limitations

Two notable limitations of this study were the sampling approach and cross-sectional design. Although we were able to recruit a diverse sample of highly sexually active gay and bisexual men, all of these men lived in the New York City metropolitan area, were required to have access to the internet, and were highly educated. Future studies are needed to determine whether samples of non-urban or less-educated men who are highly sexually active maintain varying profiles of maladaptive cognitions that manifest potentially different associations with hypersexuality. A larger sample, additionally, would have yielded more power to detect significant predictors in our multivariable logistic model. Further, the cross-sectional approach used in the present study limited our ability to determine whether maladaptive cognitions about sex were a cause, an outcome, both, or neither of problematic hypersexuality. A longitudinal design that follows highly sexually active gay and bisexual men over a critical period before the development of problematic hypersexuality would provide the means necessary for identifying the temporal role of maladaptive cognitions about sex. As mentioned previously, these associations are likely to operate in feedback with each other and future work should utilize designs that are able to investigate co-occurring changes in sexual behavior, maladaptive cognitions, and hypersexuality. Further, ecological momentary sampling of cognitions before and after sexual encounters would allow for the identification of fluctuations in maladaptive cognitions about sex and their temporal influence on sexual behavior.

Finally, the Board of Trustees of the American Psychiatric Association decided against including Hypersexual Disorder either as a formal diagnosis or in the section of the manual for further study. However, ongoing research is needed to investigate the possible criteria of problematic hypersexuality as well as the instrument proposed to assess it, the Hypersexual Disorder Screening Inventory, our primary outcome measure. For the current analyses we focused on a self-report version of the scale rather than a clinician-administered scale. It is currently unknown whether differing modes of assessment meaningfully impact the scale’s ability to classify hypersexuality. Investigations seeking to establish the most accurate measurement approach to problematic hypersexuality are need to establish hypersexuality as a valid diagnostic taxon.

This project was supported by a research grant from the National Institute of Mental Health (R01-MH087714; Jeffrey T. Parsons, Principal Investigator). H. Jonathon Rendina was supported, in part, by a National Institute of Mental Health Ruth L. Kirchstein Individual Predoctoral Fellowship (F31-MH095622). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors would like to acknowledge the contributions of the Pillow Talk Research Team: Ruben Jimenez, Joshua Guthals, and Brian Mustanski. We would also like to thank CHEST staff who played important roles in the implementation of the project: Chris Cruz, Fran Ferayorni, Sitaji Gurung, and Chris Hietikko, as well as our team of research assistants, recruiters, and interns. Finally, we thank Chris Ryan, Daniel Nardicio, and Stephan Adelson and the participants who volunteered their time for this study.