Journal of ISSN: 2373-6445JPCPY

Psychology & Clinical Psychiatry
Research Article
Volume 1 Issue 7 - 2014
Modeling High Risk Sexual Behavior in HIV-Negative Gay Men
Mark R Jurek*
Clinical Psychologist, USA
Received: October10, 2014| Published: November 18, 2014
*Corresponding author: Mark R Jurek, Clinical Psychologist, USA, 1122 Clement Street San Francisco, CA 94118 415-987-9347; Email: @
Citation: Jurek MR (2014) Modeling High Risk Sexual Behavior in HIV-Negative Gay Men. J Psychol Clin Psychiatry 1(7): 00042. DOI: 10.15406/jpcpy.2014.01.00042

Abstract

Several predictors of high-risk sexual behavior in gay men (internalized homophobia, locus of control, multiple AIDS-related loss, depression, and substance use) were examined. Eighty-seven HIV-negative gay males (M=36.7 years old, SD=5.9) were recruited through advertisements in the greater San Francisco, CA Bay Area. Structural equation modeling(SEM)analysis revealed that, as predicted, internalized homophobia, multiple AIDS-related losses, and substance use emerged as significant predictors of high-risk sexual behavior. Subsidiary analyses further revealed that multiple AIDS-related losses predicted both internalized homophobia and substance use suggesting that multiple losses experienced by gay men over the course of the AIDS epidemic may lead to other maladaptive behaviors.

Introduction

Since the beginning of the AIDS epidemic a total of 28,793 San Francisco residents have been diagnosed with AIDS, which comprises 18% of California AIDS cases and 3% of cases reported nationally. There have been 19,341 reported AIDS deaths in San Francisco as of December 31, 2010 [1]. Gay and bisexual men continue to bear the greatest burden of HIV infection, accounting for an estimated 61% of new infections [2]. Recent increases in the incidence of high risk sexual behavior [1], underscore an urgent need to identify the factors which, in combination, lead people to place themselves at risk for contracting HIV. Prior studies of high-risk sexual behavior have considered a number of factors; .including: internalized homophobia [3,4], locus of control [5,6], and substance use [7,8]. Others have examined the role of AIDS-related losses and bereavement [9,10], and the link between multiple loss and psychological distress [7,11].
Though informative, much of this research also has been characterized by two limitations. First, most studies have been limited to investigations of single predictors [5,12-13]. Consequently, the interactive effects of multiple factors in contributing to high risk sexual behavior (the most common form of AIDS transmission) have been sparsely examined. Second, much of the research in this area has been empirically derived, rather than theoretically or conceptually guided. Thus, the potentially mediating role of such factors as multiple AIDS-related loss and pessimistic attribution style, for example, has been overlooked. Integrating learned helpless theory [2,14], examination of the interactive role of multiple factors in predicting high-risk sexual behavior among HIV-negative gay males was conducted. Two hypotheses were tested. First, it was predicted that external locus of control, internalized homophobia, and depression would emerge as significant predictors of high-risk sexual behavior. Second, a meditational effect was hypothesized to exist between length of residence in San Francisco (an epicenter in the AIDS epidemic), AIDS-related loss, and pessimistic attributional style. Specifically, it was hypothesized that length of residence in the San Francisco Bay area would predict both AIDS-related loss and pessimistic attributional style, each of which, in turn, would predict high risk sexual behavior. Finally, three covariates (social desirability, substance use, and AIDS risk knowledge) were included to control for their potentially competing influence.

Method

Sample
The sample consisted of 87 HIV-negative gay men (M=36.75 years), residing in the San Francisco Bay Area. Table 1 shows the distribution of demographic characteristics in the sample. All of the participants reported that they had been tested for HIV: 45% had been tested 1-6 months prior to the study, 22% had been tested 7-12 months prior, and 33% more than 12 months prior to the study. Seventy-three percent of the sample had been tested for HIV 1-5 times in the past five years, 23% had been tested 6-10 times, and 4% had been tested more than 10 times.

 

% Respondents

Ethnic group

Black

3

Hispanic

7

White

83

Other

7

Educational level

High school diploma

2

Junior/business college

2

Some college

19

College degree

45

Graduate/professional degree

27

Other

5

Employment

Not employed

6

Employed part-time

21

Employed full-time

70

Other

3

Annual income

$ 5,000 - $25,000

32

$26,000 - $45,000

32

$46,000 - $65,000

20

$66,000 - $85,000

7

>$85,000

9

Table 1: Sample Demographic Characteristics.
Modelling high risk sexual behavior in HIV-negative gay men 2
Procedure: Participants were recruited through ads in SF Bay area gay newspapers which briefly described the study and included an 800 number where participants could leave their name, address, and telephone number, in order to be sent a copy of the survey packet. Participants completed and returned surveys containing: (a) a demographic questionnaire; (b) the Attributional Style Questionnaire [15]; (c) the Multi-Axial Gay Inventory- Men’s Short Version MAGI- MSV [16]; (d) the Beck Depression Inventory [17]; (e) the Brief Michigan Alcoholism Screening Test [18]; (f) substance use items from the Millon Clinical Multiaxial Inventory-III [19]; (g) the UCSF Center for AIDS Prevention Studies Sexual Behavior Questionnaire [20]; (h) the Adult Nowicki-Strickland Internal-External Control Scale [21]; (i) a modified version of the Death Experience Inventory [22]; G) the AIDS Risk Knowledge Questionnaire [23]; (k) the Profile of Mood States [24]; (1) the Index of Homophobia [25]; and (m) the Marlowe-Crowne Social Desirability Scale [26].

Results

Preliminary Analyses
The AIDS Risk Knowledge (ARK) Questionnaire showed poor reliability (α=.47) which could not be improved through the selective deletion of problematic items. Examination of total ARK scores
Revealed a likely ceiling effect: The highest possible score was 40, yet our sample mean was 35.9 (SD=2.16), with a range of 10 points. Therefore, ARK scores were excluded from further analysis. The remaining subscales demonstrated good reliability (α ranged between .71 and .93).
Modelling High Risk Behavior: Data were analyzed using Analysis of Moment Structures [27,28]. Examinations of the relationships between (a) depression, a latent variable with two indicators (BDI and POMS Depression-Dejection scores); (b) internalized homophobia, a latent variable with two indicators (IHP and MAGI-MSV scores); (c) internality/externality, a latent variable with two indicators (ANS-IE and ASQ scores); (d) AIDS exposure, a latent variable with two indicators (AIDS-related loss [DEI score] and length of residence in San Francisco [in years]); and (e) an observed, dependent variable (high-risk sexual behavior) were conducted. Social desirability (an indicator) and substance use, a latent variable with two indicators (BMAST scale and MCMI-III item scores) were included as covariates.
Model Estimation: Model estimation proceeded in three stages. First, the model was assessed and determined to be over identified with 51 degrees of freedom. Second, a test of the independence model was performed and easily rejected, (χ2 [45, N= 87] =124.7, p< .001). Third, a Chi-square test of differences between the sample and estimated population model was significant (χ2 [51, N= 87] =140.5, p<.001), indicating that population and sample covariance’s differed.
Because the initial CFI was low (.05), regression weights for individual indicators were examined. As shown in Table 2, two indicators failed to reach significance (length of residence in San Francisco and locus of control). Therefore, both variables were removed and the model re-estimated with the remaining indicators.
As relative Chi-square values below 3 indicate good fitting models [29,30], the Relative Chi -square obtained for the modified model (1.7) indicated that it fit the data reasonably well. Other indices confirmed that the modified model’s goodness-of-fit had improved (i.e., GFI = .91; AGFI =.80; RMR =.07; CFI =.80). Finally, one hundred bootstrap replications [31] were performed. No significant differences emerged, indicating that the distribution of estimated parameter estimates was no wider than the expected estimates under assumptions of multivariate normality.
Modelling high risk sexual behavior in HIV-negative gay men 3
Hypothesis 1: Determining Direct Predictors of High
Risk Sexual Behavior: Next, multiple regression analysis was conducted, in which all indicators and latent variables were simultaneously entered as predictors. As shown in Figure 1, 20% of the variance in high risk sexual behavior was accounted for by the predictors (p<.01). Examination of the latent variables revealed that internalized homophobia (R=.26, p<.04), AIDS-related loss (R=.16, p<.04), and substance use (R=.30, p<.04) were significant predictors of high-risk sexual behavior. The remaining variables (depression, internality/externality, social desirability) were non-significant.
Hypothesis 2: Testing a Mediational Model of High Risk Sexual Behavior: Finally, to determine whether a meditational path emerged from the significant latent variables (AIDS-related loss, internalized homophobia, and substance use) to the dependent variable, high-risk sexual behaviour, two steps were completed. First, AIDS-related loss was regressed on the two latent variables to determine meditational effects; then, the direct effects of internalized homophobia and substance use were examined for significance.

Construct

Indicators

Beta Weight

p

AIDS exposure

DEI

.94

.01

Length of SF residence

- .23

ns

Depression

BDI

- .26

.04

POMS

1.08

.005

Internality/Externality

ANS-IE

- .10

ns

ASQ

1.02

.001

Internalized homophobia

IHP

.42

.05

MAGI-MSV

.77

.01

Substance use

BMAST

- 1.07

.02

MCMI-111

1.12

.03

Table 2: Standardized Regression Weights for Indicators Predicting Latent Constructs.
Modeling High Risk Sexual Behavior in HIV-Negative Gay Men 5
As shown in Figure 2, AIDS-related loss was significant in mediating a path both to internalized homophobia (β =.34, p= .03) and to substance use (β =.36, p=.03). Subsequently, internalized homophobia predicted high-risk sexual behavior (β =.19, p=.04) and substance use predicted a trend in the same direction (β =.18, p=.06). Eight percent of the variance in high-risk sexual behavior (p=.01) was accounted for by this analysis.

Conclusion

The present findings suggest that many psychosocial variables (e.g. internalized homophobia, multiple AIDS-related loss, and substance use) are of importance in understanding the practice of high-risk sexual behavior among HIV-negative gay men. As predicted, internalized homophobia, AIDS-related loss, and substance use emerged as significant predictors. By contrast, pessimistic attributional style, external locus of control, and depression were not significant. The application of the present findings for future research includes: utilizing a domain-specific Attributional Style Questionnaire (ASQ) in conjunction with the generalized ASQ for an investigation of state versus trait-like phenomena in the prediction of high-risk sexual activity, utilizing the three subscales of the ASQ (internality, stability, and globality) to test for their individual salience in predicting risky behavior, and conducting cross-validation studies with gay men from other urban epicenters.
Figure 1: Modified Model. Modelling High Risk Sexual Behavior in HIV- Negative Gay Men 6
Figure 2: Mediational Model.
The construct of internalized homophobia, while emerging as a significant predictor of engagement in high-risk sexual behavior, warrants further study. In particular, whether sexual risk-taking behavior could be reduced by enhancing gay self-acceptance requires closer examination of the definition of internalized homophobia. The MAGI-MSV proved reliable, yet future studies of internalized homophobia should employ a factor analysis to investigate the measures’ underlying structure. Such an examination could lead to more informed clinical understanding and interventions. Internalized homophobia may be a product of self-loathing, societal response to AIDS, moral and religious acceptability, or some combination of factors [32].
That AIDS-related loss was predictive of internalized homophobia and substance use suggests that the repeated, uncontrollable losses many gay men have experienced over the course of the AIDS epidemic may lead to the development of maladaptive attitudes and behaviors. Consistent with learned helpless theory [2,14], these findings suggest that repeated exposure to AIDS­ related loss may engender a certain degree of passivity or fatalism, which, in turn, sets the stage for engagement in high risk sexual activity, or, at the very least, a decrease in proactive, risk avoidant behavior. It is important to note that, as these findings with the ARK scale suggest, such linkages occur.
Modelling high risk sexual behavior in HIV-negative gay men 4
Although the present study utilized structural equation modeling to test hypotheses, a conventional regression analysis was employed to predict a single observed variable as a linear combination of four latent variables and two covariates. Future research could be meaningfully enhanced by the use of more exploratory path analyses. Such analyses could theoretically propose a sequential model of variables whose collective influence mediates a path of increasing variation in high-risk sexual behavior. For example, as previously stated, multiple AIDS-related losses may best be viewed as a setting condition which influences attributional style, which in turn influences other variables in the prediction of high-risk behavior. Indeed, the present study did find that certain variables were predictive of others, and yet a significant path connecting all variables did not emerge in subsidiary analyses. Future studies should incorporate other variables as predictors which might provide a missing link or links in the sequential change of psychosocial factors influencing engagement in risky behavior.
Further, the reduction in variance accounted for in testing the mediational hypothesis (from 20% to 8%) suggests that the obtained model may be most useful in predicting behavior at the extreme end of the sexual risk continuum (i.e., repeated, unprotected anal intercourse with multiple partners). Additional studies are needed to determine the (multiple) predictors of behavior which, though risky, does not fall at the far end of that continuum. Such understanding is necessary for clinicians and health providers to provide more focused interventions targeted towards harm-reduction and AIDS prevention.

References

  1. San Francisco Department of Public Health (2010) HIV Epidemiology Section, HIV/AIDS Epidemiology Annual Report.
  2. Abramson LY, Seligman ME, Teasdale JD (1978) Learned helplessness in humans: Critique and reformulation. J Abnorm Psychol 87(1): 49-74.
  3. Folkman S, Chesney MA, Pollack L, Phillips C (1992) Stress, coping and high-risk sexual behavior. Health Psychol 11(4): 218-222.
  4. Perkins DO, Leserman J, Murphy C, Evans DL (1993) Psychosocial predictors of high-risk sexual behavior among HIV-negative homosexual men. AIDS Educ Prev 5(2): 141-152.
  5. Aspinwall LG, Kemeny ME, Taylor SE, Schneider SG, Dudley JP (1991) Psychosocial predictors of gay men's AIDS risk­ reduction behavior. Health Psychol 10(6): 432-444.
  6. Kelly JA, St Lawrence JS, Brasfield TL (1991) Predictors of vulnerability to AIDS risk behavior relapse. J Consult Clin Psychol 59(1): 163-166.
  7. Martin JL (1988) Psychological consequences of AIDS-related bereavement among gay men. J Consult Clin Psychol 56(6): 856-862.
  8. Meyer IH, Dean L (1995) Patterns of sexual behavior and risk taking among New York City gay men. AIDS Educ Prev 7(Suppl): 13-23.
  9. Folkman S, Chesney M, Collette L, Boccellari A, Cooke M (1996) Postbereavement depressive mood and its prebereavement predictors in HIV+ and HIV-gay men. J Pers Soc Psychol 70(2): 336-348.
  10. Johnston WI (1995) HIV-Negative: How the uninfected are affected by AIDS. New York: Insight Books, Plenum Press.
  11. Martin JL, Dean L (1993) Effects of AIDS related bereavement and HIV-related illness on psychological distress among gay men: A 7-year longitudinal study, 1985-1991. J Consult Clin Psychol 61(1): 94-103.
  12. Mulry G, Kalichman SC, Kelly JA (1994) Substance use and unsafe sex among gay men: Global versus situational use of substances. Journal of Sex Education and Therapy 20(3): 175-184.
  13. Vinke J, Bolton R, Mak R, Blank S (1993) Coming Out and AIDS-Related High-Risk Sexual Behavior. Arch Sex Behav 22(6): 559-587.
  14. Peterson C, Maier SF, Seligman MEP (1993) Helplessness: A Theory for the Age of Personal Control. Oxford University Press, Inc., New York, USA.
  15. Peterson C, Semmel A, von Baeyer C, Abramson LY, Metalsky GL, et al. (1982). The attributional Style Questionnaire. Cognitive Therapy and Research 6(3): 287-299.
  16. Shidlo A, Hollander G (1996) Multi-Axial Gay Inventory-Men’s Short Version
  17. Beck AT (1978 Beck Depression Inventory Manual. Psychological Corporation, Texas, USA.
  18. Pokorny AD, Miller BA, Kaplan HB (1972) The brief Michigan alcoholism screening test. Am J Psychiatry 129(3): 342-345.
  19. Millon T (1993) Millon Clinical Multiaxial Inventory (4th edn), National Computer Systems, Minneapolis, USA.
  20. Chesney MA (1997) Clinical trial of coping effectiveness training for HIV- positive men. University of California, San Francisco, Center for AIDS Prevention Studies.
  21. Nowicki S, Duke MP (1974) A locus of control scale for college as well as noncollege adults. Journal of Personality Assessment 38(2): 136-137.
  22. O’Riordan S (1993) The effects of death education on death anxiety and death depression. Unpublished doctoral dissertation, Western Graduate School of Psychology, USA.
  23. Kelly JA, St Lawrence JS, Hood HV, Brasfield TL (1989) An objective scale of AIDS risk behavior knowledge: Scale development, validation, and norms. J Behav Ther Exp Psychiatry 20(3): 227-234.
  24. McNair DM, Lorr M, Droppleman LF (1971) Manual for the Profile of Mood States. Educational and Industrial Testing Service, San Diego, USA.
  25. Hudson WW, Ricketts WA (1980) A strategy for the measurement of homophobia. J Homosex 5(4): 357-372.
  26. Crowne DP, Marlowe D (1960) A new scale of social desirability independent of psychopathology. J Consult Psychol 24: 349-354.
  27. Arbuckle JL (1989) AMOS: Analysis of moment structures. The American Statistician 43: 66-67.
  28. Arbuckle JL (1994) Computer announcement AMOS: Analysis of moment structures. Psychometrika 59(1): 135-137.
  29. Carmine, EG, McIver JP (1983) An introduction to the analysis of models with unobserved variables. Political Methodology 9(1): 51-102.
  30. Norris AE (1997) Structural equation modelling. In: BH Munro (Ed.), Statistical methods for health care research (3rd edn), Lippincott-Raven Publishers Philadelphia, USA, pp. 368-396.
  31. Efron B (1987) Better bootstrap confidence intervals. Journal of the American Statistical Association 82: 171-185.
  32. Ross MW, Rosser BR (1996) Measurement and correlates of internalized homophobia: A factor analytic study. J Clin Psychol 52(1): 15-21.
  33. Kelly JA, St Lawrence JS, Brasfield TL, Lemke A, Amidei T, et al. (1990) Psychological factors that predict AIDS high- risk versus AIDS precautionary behavior. J Consult Clin Psychol 58(1): 117-120.
  34. Peterson, Christopher (1982) Learned helplessness and health psychology. Health Psychology 1(2): 153-168.
© 2014-2016 MedCrave Group, All rights reserved. No part of this content may be reproduced or transmitted in any form or by any means as per the standard guidelines of fair use.
Creative Commons License Open Access by MedCrave Group is licensed under a Creative Commons Attribution 4.0 International License.
Based on a work at http://medcraveonline.com
Best viewed in Mozilla Firefox | Google Chrome | Above IE 7.0 version | Opera |Privacy Policy