Abstract
Objective:
Mounting evidence suggests a syndemic relation between methamphetamine use and depression to increase sexual risk-taking (i.e., HIV transmission risk behavior) among men who have sex with men. This prospective analysis of longitudinal data collected from an outpatient methamphetamine abuse treatment program for gay and bisexual men assessed whether symptoms of depression mediated and/or moderated the associations between methamphetamine use and unprotected insertive/receptive anal intercourse.
Methods:
From November 2005 through October 2007, 167 treatment-seeking gay and bisexual men (63% HIV-positive) enrolled in and attended a 16-week methamphetamine abuse outpatient treatment program. Participants’ depressive symptoms, biomarker-confirmed methamphetamine use, and self-reported sexual risk-taking were assessed at baseline and follow-up evaluations. Path analysis tested the mediating and moderating effects of depression on the associations between methamphetamine use and unprotected insertive/receptive anal intercourse.
Results:
Methamphetamine use during the treatment period had a significant indirect (Coef. = −.15; 95% CI = −.23 - −.06), but no direct (Coef. = .11; ns) or total effect (Coef. = −.04; ns) on participants’ sexual risk-taking after accounting for the significant mediating (Coef. = .56; 95% CI = .33 - .78) and moderating (Coef. = −.03; 95% CI = −.04 - −.02) effects of depression.
Conclusions:
Depression fully mediated and weakly moderated associations between methamphetamine use and sexual risk-taking in this sample. Interventions and treatment programs to reduce sexual risk-taking among gay and bisexual men should simultaneously address methamphetamine use and depression to optimize health outcomes.
Keywords: depression, methamphetamine, HIV/AIDS, sexual risk-taking, gay and bisexual men
Evidence demonstrates that methamphetamine use (Bowers, Branson, Fletcher, & Reback 2012; Shoptaw & Reback 2006) and psychiatric distress (Blashill, O’Cleirigh, Mayer, Goshe, & Safren, 2012; Halkitis et al., 2013) are associated with increased sexual risk-taking among men who have sex with men (MSM). Further, evidence suggests a syndemic (i.e., synergistically reinforcing) relation between methamphetamine use and depression to increase sexual risk behaviors among MSM (Halkitis et al., 2013; Safren et al., 2011). Mediation and moderation analyses are thus warranted to estimate the direct, indirect, and total effects of methamphetamine use on sexual risk-taking among MSM when accounting for depressed mood. This study employed path analytic modeling to test the mediating and moderating effects of depression on methamphetamine use and sexual risk-taking among gay- or bisexually identified MSM enrolled in treatment for methamphetamine abuse. It was hypothesized that depression would mediate and moderate associations between methamphetamine use and sexual risk-taking.
Method
Participants
Participants (N = 171) seeking outpatient treatment for methamphetamine use were recruited from November 2005 – October 2007. Eligibility criteria were self-identified gay/bisexual man, 18-65 years, who met abuse or dependence criteria for methamphetamine (DSM-IV). Participants were excluded if they had a psychiatric condition that required a more intensive intervention or failed to comply with research requirements.
Procedures
All participants provided informed consent at intake. Treatment included thrice weekly gay-specific cognitive behavioral therapy groups coupled with individual-based contingency management for eight weeks, followed by a weekly continuing care group coupled with contingency management from week nine through 16 (for more detail see: Reback & Shoptaw, 2014). Participants provided up to 24 urine drug screens testing for methamphetamine use during the eight week intensive treatment period. Follow-up evaluations were conducted at 8 weeks, 16 weeks, and 6-months post baseline. The Friends Research Institute Institutional Review Board approved the study and provided oversight for all study activities.
Assessments
Admission Form.
Administered at baseline, this measure (Rawson et al., 1995) collects information on demographics, drug-use, treatment history, family status, legal status, HIV status, HIV-medical care (if appropriate) and general health and psychiatric status.
Behavioral Questionnaire-Amphetamine (BQA):
The BQA assessed sexual risk behaviors (i.e., unprotected anal intercourse) of participants in the past 30 days (Chesney et al., 1997); it was modified in consultation with the developers for use with gay/bisexual men.
Beck Depression Inventory (BDI).
The BDI (Beck et al., 1968) was administered once a week for 16 weeks and at the 6-month follow-up evaluation. The BDI was used as an index of depression and to monitor participants’ safety, particularly suicidal ideation. Scores during the treatment period were averaged to create an index of depression symptoms during treatment.
Urine Drug Screening:
Urine drug screening for methamphetamine metabolites (Phamatech, 2014) was conducted thrice weekly during the first eight weeks, then weekly during weeks 9 through 16, and then again at the 6-month follow-up evaluation. A count of each participant’s clean urine samples submitted during treatment is used for analysis.
Statistical Analysis
Participants who completed a baseline assessment but never attended an intervention session (n = 4) were removed from all statistical analyses (analytical sample n = 167). A total of 86 (51.5%) participants failed to complete at least one BDI assessment during the 8-week intensive treatment intervention. Depression scores from the 8-week intensive treatment period were thus averaged within participants to generate a single index of depression for each participant; the 51.5% of participants who did not complete a depression assessment during the 8-week intensive treatment period were marked as missing. There were also missing data (n = 36; 21.6%) in the self-reported number of unprotected anal sex partners at treatment completion. Little’s (1988) test indicated that depression and sexual risk data were missing completely at random (χ2df=6 = 7.88; p = 0.25) and was not covariate dependent on participants’ baseline characteristics (i.e., age, race, sexual identity, HIV status, or years of heavy methamphetamine use; χ2df=36 = 34.71; p = 0.53). As such, path analysis employing the maximum likelihood with missing values estimation method was used for hypothesis testing (Duncan & Duncan, 1994); Figure 1 provides a diagram of this path analytic model. All estimates were derived using robust variance/covariance matrices (Huber, 1967; White, 1980) and were run using Stata 13SE.
Figure 1.
Proposed Path Analytic Model
Results
Most participants were White, identified as gay, and self-reported a HIV-positive sero-status. The mean participant age was 40.0 years, while the mean length of heavy methamphetamine use at baseline was 3.8 years. Participants’ self-reported unprotected receptive or insertive anal intercourse in the past 30 days reduced significantly from baseline to 8-week follow-up, as did mean depression scores. The percentage of participants reporting unprotected anal intercourse reduced from baseline (76%) to follow-up (52%), and participants provided a mean of 7.5 (SD = 9.6) methamphetamine metabolite-free urine samples during the 8-week intensive treatment period.
Table 2 provides the direct, indirect, and total effects of the path analytic model proposed in Figure 1. Baseline depression scores and the count of methamphetamine-metabolite free urine samples submitted during the treatment period were both significantly associated with participant depression during the treatment period. Engagement in unprotected insertive or receptive anal intercourse at baseline was significantly associated with reports of unprotected anal intercourse at treatment completion, as were mean BDI scores during the treatment period. The interaction term between participants’ mean depression scores and the number of methamphetamine-metabolite free urine samples was also associated with engagement in unprotected anal intercourse at completion of the 8-week intensive treatment period.
Table 2.
Robust Path Analytic Estimates of Direct, Indirect, and Total Effectsa (N = 167)
| Outcome Variable | Predictor Variable(s) | Direct Effects | Indirect Effects | Total Effects | |||
|---|---|---|---|---|---|---|---|
| Coef. (95% CI) | p | Coef. (95% CI) | p | Coef. (95% CI) | p | ||
| Baseline BDI Score | |||||||
| Lifetime # Years Heavy Meth Use | −.22 (−.48 - .04) | .095 | (no path) | - | −.22 (−.48 - .04) | .095 | |
| var(e) | 96.05 (76.30 - 120.91) | - | - | - | - | - | |
| Baseline # Times Unprotected Anal Intercourse | |||||||
| Lifetime # Years Heavy Meth Use | −.30 (−.86 - .27) | .300 | (no path) | - | −.30 (−.86 - .27) | .300 | |
| var(e) | 432.28 (245.53 – 761.06) | - | - | - | - | - | |
| Treatment Period # Methamphetamine-Metabolite Free Urine Samples | |||||||
| Lifetime # Years Heavy Meth Use | −.23 (−.57 - .10) | .174 | (no path) | - | −.23 (−.57 - .10) | .174 | |
| var(e) | 89.80 (77.85 – 103.58) | - | - | - | - | - | |
| Treatment Period Average BDI Score | |||||||
| Lifetime # Years Heavy Meth Use | (no path) | - | .00 (−.13 - .14) | .969 | .00 (−.13 - .14) | .969 | |
| Baseline BDI Score | .27* (.13 - .41) | < .001 | (no path) | - | .27* (.13 - .41) | < .001 | |
| Treatment Period # Clean Urine Samples | −.26* (−.42 - −.11) | .001 | (no path) | - | −.26* (−.42 - −.11) | .001 | |
| var(e) | 33.56 (20.81 - 54.12) | - | - | - | - | - | |
| 8-Week Follow-up # Times Unprotected Anal Intercourse | |||||||
| Lifetime # Years Heavy Meth Use | (no path) | - | −.09 (−.27 - .08) | .294 | −.09 (−.27 - .08) | .294 | |
| Baseline BDI Score | (no path) | - | .15* (.07 - .23) | < .001 | .15* (.07 - .23) | < .001 | |
| Baseline # Times Unprotected Anal Intercourse | .24* (.08 - .39) | .003 | (no path) | - | .24* (.08 - .39) | .003 | |
| Treatment Period # Clean Urine Samples | .11 (−.01 - .23) | .078 | −.15* (−.23 - −.06) | .001 | −.04 (−.15 - .07) | .485 | |
| Treatment Period Average BDI Score | .56* (.33 - .78) | < .001 | (no path) | - | .56* (.33 - .78) | < .001 | |
| Interaction Effect: # Clean Urines X BDI Score | −.03* (−.04 - −.02) | < .001 | (no path) | - | −.03* (−.04 - −.02) | < .001 | |
| var(e) | 17.64 (9.38 - 33.16) | - | - | - | - | - | |
None of the statistical controls produced significant coefficient estimates; results are untabulated.
p < .01, two-tailed
Though methamphetamine use had no direct effect on unprotected anal intercourse at 8-week follow-up, decomposition of indirect effects revealed that increased methamphetamine-metabolite free urine screens were significantly indirectly associated with reduced engagement in unprotected anal intercourse at 8-week follow-up. When effects were aggregated, however, there was no total effect of methamphetamine use on sexual risk-taking at 8-week follow-up.
Modification indices indicated a significant missing pathway from baseline engagement in unprotected anal intercourse to methamphetamine-metabolite free urine samples during the treatment period. Though temporally valid, this pathway runs counter to the expected causal sequence and was thus not included. There were no other temporally valid significant improvements. The path model’s overall coefficient of determination was .405; due to the use of robust standard error estimation techniques, additional diagnostic tests were not available.
Discussion
Previous research demonstrating strong, direct associations between methamphetamine use and sexual risk-taking have set the foundation for the study of health risks facing MSM in the United States. Results presented here serve only to extend prior research by demonstrating that depression can interact with and/or be influenced by methamphetamine use, and that such effects have significant impacts on sexual risk behaviors in this population. Future research should attempt to further specify the unique and combined contributions of methamphetamine use and depression on the sexual risk behaviors of MSM.
This study is limited by its small sample size, use of self-reported HIV status and sexual risk-taking, lack of a control group or assessment of partner serostatus, and use of participants self-enrolled in a methamphetamine treatment program. Results may not be generalizable to non-treatment seeking MSM, to methamphetamine-using MSM who do not identify as gay or bisexual, or to samples from rural areas. This study was further limited by the amount of missing data (though data was missing completely at random, reducing concerns of bias), and the related need to average depression scores during the 8-week intensive treatment period.
These results suggest that HIV risk-reduction interventions designed for gay and bisexual men should simultaneously and synergistically address the issues of substance use, sexual risk behaviors, and depression (e.g., Mimiaga et al., 2012; Carrico et al., 2014). Researchers and service providers working with methamphetamine-using gay and bisexual men should, when fiscally and programmatically possible, seek to integrate methamphetamine use treatment, mental health treatment, and behavioral HIV sexual risk reduction interventions in concert.
Table 1.
Participant Sociodemographics, Methamphetamine Use History, Unprotected Anal Intercourse, and Depression
| Baseline (N = 167) | Treatment Completion (n = 131) | |||
|---|---|---|---|---|
| Characteristic | Mean [SD] or N (%) | Mean [SD] or N (%) | Sig.a | |
| Age | Years | 40.0 [8.2] | - | - |
| Race | ||||
| Caucasian/white | 107 (64.1%) | - | - | |
| Non-Caucasian/non-white | 60 (35.9%) | - | - | |
| Sexual Identity | ||||
| Gay-identified | 156 (93.4%) | - | - | |
| Not gay-identified | 11 (6.6%) | - | - | |
| Educational Attainment | Years | 15.0 [2.7] | - | - |
| Housing Statusb | ||||
| Own/Rent House or Apt. | 127 (77.0%) | - | - | |
| Staying with Family/Friends | 28 (17.0%) | - | - | |
| Group Housing/Sober Living | 5 (3.0%) | - | - | |
| No Current Address | 5 (3.0%) | - | - | |
| Income (yearly) | ||||
| Less than $15,000 | 54 (32.9%) | |||
| $15,001 - $30,000 | 32 (19.5%) | |||
| $30,001 - $60,000 | 42 (25.6%) | |||
| More than $60,000 | 36 (22.0%) | |||
| HIV-Positive Status | 105 (62.9%) | |||
| Heavy Methamphetamine Use | Years | 3.8 [4.9] | - | - |
| Unprotected Receptive/Insertive Anal Intercourse in the Past 30 Days | # Times | 12.1 [20.9] | 3.5 [6.8] | *** |
| Depressiond | BDI Score/Mean BDI Score | 17.3 [9.9] | 8.5 [6.6] | *** |
p ≤ 0.001
n = 165
n = 164
Baseline (n = 164) vs. treatment period (n = 81)
References
- Beck AT, Ward CH, Mendelson J, Mock J, & Erbaugh J (1968). The beck depression inventory. Archives of General Psychiatry 4:561–571 [DOI] [PubMed] [Google Scholar]
- Blashill A, O’Cleirigh C, Mayer K, Goshe B, & Safren S (2012). Body mass index, depression and sexual transmission risk behaviors among HIV-positive MSM. AIDS and Behavior, 16(8), 2251–2256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bowers JR, Branson CM, Fletcher JB, & Reback CJ (2012). Predictors of HIV sexual risk behavior among men who have sex with men, men who have sex with men and women, and transgender women. International Journal of Sexual Health, 24(4), 290–302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Carrico A, Flentje A, Gruber V, Woods W, Discepola M, Dilworth S, . . . Siever M (2014). Community-based harm reduction substance abuse treatment with methamphetamine-using men who have sex with men. Journal of Urban Health, 91(3), 555–567. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chesney MA, Chambers DB, & Kahn JO (1997). Risk behavior for HIV infection among participants in preventive HIV vaccine trials. Journal of Acquired Immune Deficiency Syndromes, 16, 266–271. [DOI] [PubMed] [Google Scholar]
- Duncan SC, & Duncan TE (1994). Modeling incomplete longitudinal substance use data using latent variable growth curve methodology. Multivar Beh Res, 29(4), 313–338. [DOI] [PubMed] [Google Scholar]
- Halkitis P, Moeller R, Siconolfi D, Storholm E, Solomon T, & Bub K (2013). Measurement model exploring a syndemic in emerging adult gay and bisexual men. AIDS and Behavior, 17(2), 662–673. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Huber PJ (1967). The behavior of maximum likelihood estimates under nonstandard conditions. In Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability Berkeley, CA: University of California Press, vol. 1, 221–233. [Google Scholar]
- Little RJA (1988). A test of missing completely at random for multivariate data with missing values. Journal of the American Statistical Association, 83(404), 1198–1202. [Google Scholar]
- Mimiaga MJ, Reisner SL, Pantalone DW, O’Cleirigh C, Mayer KH, & Safren SA (2012). A pilot trial of integrated behavioral activation and sexual risk reduction counseling for HIV-uninfected men who have sex with men abusing crystal methamphetamine. AIDS Patient Care and STDs, 26(11), 681–693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Phamatech. (2014). Phamatech lab and diag. July/16/2014, http://www.phamatech.com/index.html [Google Scholar]
- Rawson RA, Shoptaw SJ, Obert JL, McCann MJ, Hasson AL, Marinelli-Casey PJ, … Ling W (1995). An intensive outpatient approach for cocaine abuse treatment: The matrix model. J Subst Abuse Treat. 12, 117–27. [DOI] [PubMed] [Google Scholar]
- Reback CJ, & Shoptaw S (2014). Development of an evidence-based, gay-specific cognitive behavioral therapy intervention for methamphetamine-abusing gay and bisexual men. Add Beh. 39, 1286–1291. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Safren SA, Reisner SL, Herrick A, Mimiaga MJ, & Stall R (2011). Mental health and HIV risk in men who have sex with men. JAIDS, 55(Supl. 2), S74–S77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shoptaw S, & Reback CJ (2006). Associations between methamphetamine use and HIV among men who have sex with men. J Urban Health, 83(6), 1151–1157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- White H (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 48: 817–830. [Google Scholar]