Examining HIV Stigma, Depression, Stress, and Recent Stimulant Use in a Sample of Sexual Minority Men Living with HIV: An Application of the Stigma and Substance Use Process Model

Abstract

Sexual minority men (SMM) with HIV are disproportionately impacted by stigma and mental health disorders. Guided by the Stigma and Substance Use Process Model, we evaluated how HIV stigma impacts mental health outcomes among SMM with HIV. Data were drawn from Thrive With Me, an RCT of an mHealth intervention targeting ART adherence among SMM with HIV. Path analyses tested the relationships between HIV stigma, depression, stress, and recent stimulant use. Overall, 49.1% (194/401) had depression symptoms, 68.8% (276/401) had moderate-to-high stress, and 28.1% (111/401) had detectable stimulant use in urine samples at baseline. In path analyses, baseline internalized HIV stigma was associated with depression and stress 5-months post-baseline and enacted stigma was associated with recent stimulant use 11-months post-baseline. We identified internalized and enacted HIV stigma, but not anticipated stigma, as potentially important intervention targets for stimulant use, depression, and stress among SMM with HIV.

1. Background

Stimulant Use among SMM & PLWH

Substance use, and in particular, stimulant use, is higher among sexual minority men (SMM; also referred to as gay, bisexual, and other men who have sex with men) populations compared to other populations in the United States (1, 2). In addition, a 16-year prospective study (2000-2016) with a general sample of PLWH engaged in care in San Francisco found that stimulant use moderated the association between time engaged in care and viral suppression. Specifically, PLWH reporting less frequent stimulant use (i.e., stimulants were used at half or fewer study visits over 4-6 month follow-up time points) were more likely to be virally suppressed the longer they were engaged in care (3). In contrast, among those PLWH who reported higher levels of stimulant use (i.e., stimulants were used at more than half of study visits), there was no association between time engaged in care and viral suppression (3). Taken together, this indicates that SMM living with HIV who use stimulants may be at greater risk of disease progression and transmission compared those who do not use stimulants (1–6). As such, it is imperative that modifiable determinants of stimulant use in SSM living with HIV be identified in order to promote sustained viral suppression.

HIV Stigma

Stigma was originally defined as an “attribute that is deeply discrediting” and serves to diminish an individual from “a whole and usual person to a tainted, discounted one” (7). Since the emergence of this original definition, the conceptualization of stigma has been more broadly defined by Link and Phelan (2006) as a multilevel, social process by which individuals are devalued based on group membership (8). Subsequent research has demonstrated that there are three main types of stigma that individuals may experience: anticipated, internalized, and enacted (9). Anticipated stigma is the process through which individuals come to expect rejection and negative behaviors from members of larger society in response to their stigmatized identity (9). Internalized stigma, sometimes referred to as self-stigma, is the negative thoughts, feelings, and self-devaluation that result from identifying with a particular stigmatized group (9). And lastly, enacted stigma refers to the direct experiences of rejection and discrimination of individuals from members of larger society as a result of their group membership (9).

HIV stigma - including anticipated, internalized, and enacted forms of this stigma - has been associated with negative mental health and substance use outcomes among persons living with HIV (PLWH), including SMM living with HIV. For example, a study among PLWH in the Southeastern United States found that internalized HIV stigma was associated with both depression and alcohol use, and that the relationship between HIV-related discrimination (i.e., enacted HIV stigma) and depression symptoms was mediated by both discrimination-related stress and internalized HIV stigma (10). Additionally, past research has found that internalized stigma is associated with poorer affective health (e.g. feelings of helplessness) among SMM recently diagnosed with HIV in the United States, and with depression, anxiety, and generalized stress among SMM living with HIV in Australia (11, 12). Finally, qualitative research among 29 PLWH in rural United States, of which 45% identified as SMM, highlighted that anticipated and enacted HIV-, sexual orientation-, and mental health-related stigmas promoted depression, loneliness, and social isolation (13). Greater research is needed, however, to fully elucidate which facets of HIV stigma (i.e., anticipated, internalized, or enacted) are associated with substance use- and mental health- related outcomes, and whether these impacts are sustained over time, among SMM living with HIV.

Depression and Stress among SMM and PLWH

Research has demonstrated that PLWH experience depression at 2-3 times the rate of those who are not living with HIV (14–16) and, as of 2019, it was estimated that as many as 39% of PLWH also had a diagnosis of depression (17). In addition, anxiety disorders are notably higher among PLWH than the general population, likely due to the chronic stressors experienced by PLWH (e.g., medication side effects, chronic pain, and social stigma) (18). This is of note because depression and stress are associated with negative health outcomes among PLWH including poorer quality of life, reduced social connectedness, suboptimal antiretroviral treatment (ART) adherence, increased risk behaviors (e.g., condomless sex), and substance use (15–20). SMM, a group disproportionately impacted by HIV, may also be particularly vulnerable to depression and stress. As of 2017, rates of depression were estimated to be 17% higher and rates of PTSD were estimated to be double among SMM compared to their heterosexual counterparts (21). Further, experiences of HIV stigma may exacerbate and compound an already elevated risk for depression and stress among SMM living with HIV (13, 22).

The Stigma and Substance Use Process Model

The Stigma and Substance Use Process model was developed to describe how social stigma is associated with behaviors that place individuals at risk for developing substance use disorders (25, 26). In this model, individual manifestations of stigma consist of experiences of social stigma (e.g., anticipated, internalized, and enacted stigma) among individuals within a socially devalued group or with socially devalued characteristics (e.g., SMM living with HIV) (23, 24). The process model hypothesizes that individual-level stigma manifestations impact substance use-related behaviors (e.g., substance misuse) via a series of psychosocial mechanisms (e.g., internalizing symptoms like depression and stress) (23, 24). As such, testing of this process model is needed to provide greater insight on the ways in which HIV stigma may impact mental health and substance use-related outcomes for SMM living with HIV.

The Current Study

SMM living with HIV are disproportionately impacted by HIV stigma, mental health disorders, and substance use, all of which have been found to have negative downstream effects on ART adherence and disease progression. As such, there is a need to better understand the mechanisms through which SMM living with HIV’s experiences with stigma impact mental and physical health-related outcomes so that tailored and effective intervention efforts can be developed for this population. The current study sought to test the pathways presented within the Stigma and Substance Use Process Model (23, 24) to assess whether: (1) greater HIV stigma is significantly associated with future depression symptoms, perceived stress, and recent stimulant use among SMM living with HIV, and (2) whether depression and perceived stress mediate the relationship between HIV stigma and recent stimulant use for this population.

2. Methods

Participants and Procedures

The Thrive With Me (TWM) study is a prospective two-arm randomized controlled trial (RCT) of a virtual behavioral intervention, guided by the information, motivation, and behavioral skills (IMB) model, that seeks to increase ART adherence and viral suppression among SMM living with HIV through peer support, ART information, and ART self-monitoring (25, 26). The TWM study rationale and methodology have previously been described in full (25, 26). Briefly, TWM recruited 401 SMM living with HIV in the New York City area beginning in October 2016. Participants that met inclusion criteria were randomized to either an experimental TWM or control arm. SMM randomized to the experimental arm received the TWM intervention for 5 months and those randomized to the control arm received weekly HIV-related emails over that same time period. In-person computer-assisted self-interviewing (CASI) assessments were conducted with all participants at the Pride Health Research Consortium (PRIDE) offices in New York, by trained research personnel, and occurred at baseline, 5-month, 11-month, and 17-month time points. Participants received US $50 for each in-person assessment completed and all follow-up assessments were completed in August of 2019. All study procedures were approved by the respective IRBs of the University of Minnesota and the Hunter College City University of New York. Additionally, a Certificate of Confidentiality was obtained from NIDA, and a data safety and monitoring board was established to provide oversight of the research practices.

Measures

Sociodemographic Characteristics

At baseline, all participants were asked to self-report common sociodemographic factors including their age (in years), ethnicity (Hispanic/Latinx vs. Not Hispanic/Latinx), race (White, Black, or Other), highest level of education obtained (less than a high school diploma, high school diploma/GED, some college, and college degree), and income level ($0-19,999, $20,000-39,999, $40,000-74,999, and $75,000 or more).

Baseline HIV Stigma

The primary independent variable of interest for the current study, HIV stigma at baseline, was measured with the HIV Stigma Mechanism Scale (12). The HIV Stigma Mechanism Scale consists of 24 items across 3 subscales that measure lifetime anticipated HIV stigma, internalized HIV stigma, and enacted HIV stigma (12). The anticipated HIV stigma subscale consists of 9 items exemplified by the following, “[Because of my HIV status]… Healthcare workers will treat me with less respect.” The internalized HIV stigma subscale consists of 6 items including, “I feel ashamed of having HIV.” The enacted HIV stigma scale consists of 9 items and includes items like the following, “[Because of my HIV status]… Family members have avoided me.” All items were scored on a 5-point Likert-type scale with higher scores representing higher levels of stigma. Positive items were reverse coded and then all items were averaged to create composite scores for each subscale. Each subscale demonstrated good reliability, both in previous research (12) and in the current sample (α = .89-.92).

Psychosocial Mechanisms at 5-Month Follow-Up

A potential mediating variable of interest for the current study, depression symptoms at 5-month follow-up, was measured using the 10-item Center for Epidemiologic Studies Depression Scale (CESD-10) (27). The CESD-10 asks participants to report on the depressive symptomology they may have experienced in the past 7 days (27). This scale consists of items like the following, “I was bothered by things that usually don’t bother me” and all items were scored on a 4-point Likert-type scale ranging from 0 (none of the time) to 3 (most of the time) (27). All items were summed to create a composite depression score, with total scores ranging from 0 to 30, and dichotomized with a score of ≥ 10 indicating depressive symptomology (27, 28). Further the CESD-10 has demonstrated good reliability in previous research with PLWH populations (29) as well as among our sample (α = .80).

A second potential mediating variable of interest for the current study, perceived stress at 5-month follow-up, was measured with the 10-item Perceived Stress Scale (PSS) (30). The PSS asks participants to report on perceived stress experienced in the past 30 days. This measure consists of items like the following, “In the last month, how often have you felt that you were unable to control the important things in your life?” and items were scored on a 5-point Likert-type scale ranging from 0 (never) to 4 (very often). Positive items were reverse coded and all items were summed to create a composite perceived stress score, with total scores ranging from 0 to 40 (30). This scale has demonstrated good reliability both in existing literature (31) and within the current sample (α = .85). Further, PSS scores were dichotomized into two categories for the current study: (0) low perceived stress (PSS score = 0-13) vs. (1) moderate-to-high perceived stress (PSS score ≥ 14).

Recent Stimulant Use at 11-month Follow-Up

The primary outcome of interest, stimulant use at 11-month follow-up, was evaluated through a urine screen panel. Participants completed urine screens at each assessment time point to measure the use of the following substances: cocaine, amphetamines, and methamphetamine. Urine screens were completed with the Integrated E-Z Split Key Cup II-5 panel (Innovation Laboratories) (32). This test is able to detect use of these drugs from 1 to 4 days after use (32). For the current study, urine screen results were combined into a dichotomous composite measure of stimulant use at 11-month follow-up: (0) no stimulant use detected vs. (1) any cocaine, amphetamine, and/or methamphetamine use detected.

Statistical Analyses

Descriptive statistics were calculated for participant sociodemographic characteristics, anticipated, internalized, and enacted HIV stigma scores, depression scores, perceived stress scores, and recent stimulant use at baseline. Next, preliminary bivariate logistic regression analyses were conducted to assess the associations between baseline sociodemographic characteristics, intervention condition, baseline anticipated, internalized, and enacted HIV stigma scores, and the following outcomes: depression and perceived stress scores at 5 months post-baseline, and recent stimulant use at 11 months post-baseline. Preliminary multivariable logistic regression analyses were then conducted to further test the associations between anticipated, internalized, and enacted HIV stigma with the outcomes of depression, perceived stress, and recent stimulant use while controlling for sociodemographic factors found to be associated with the variables of interest in the bivariate analyses. Both bivariate and multivariable analyses were employed as part of a model building approach, and preliminary findings informed, and identified covariates for inclusion in, the path analysis. Additionally, given that intervention condition was not associated with any of the independent or outcome variables of interest in the bivariate analyses, this variable was not included in the multivariable or path analyses.

Finally, a path analysis using the maximum likelihood estimation approach was conducted with the PROC CALIS procedure in SAS (33) to test the following hypothesized pathways presented in the Stigma and Substance Use Process model: (1) the direct effects of baseline anticipated, internalized, and/or enacted HIV stigma on depression symptoms 5 months post-baseline, perceived stress 5 months post-baseline, and recent stimulant use 11 months post-baseline, as well as the direct effects of depression symptoms and perceived stress 5 months post-baseline on stimulant use 11 months post-baseline (i.e., the a, b, and c paths), (2) the indirect effects of baseline anticipated, internalized, and/or enacted HIV stigma on recent stimulant use 11 months post-baseline through depression symptoms 5 months post-baseline, and (3) the indirect effects of baseline anticipated, internalized, and/or enacted HIV stigma on stimulant use 11 months post-baseline through perceived stress 5 months post-baseline (See Figure 1 for the conceptual model tested in the current study). The PROC CALIS procedure partitions the total effects of anticipated, internalized, and enacted HIV stigma at baseline on stimulant use 11-months post-baseline via depression symptoms and perceived stress 5-months post-baseline into direct and indirect (i.e., mediated) effects and generates standardized regression coefficients, t-values, and estimated standard errors for each (33). Additionally, all variables were screened prior to analysis to ensure the statistical assumptions required for path analyses (e.g., low multicollinearity) were met. Missing cases were excluded from all analyses and all analyses were conducted in SAS On Demand for Academics (SAS Institute Inc., Cary, North Carolina, USA).

Figure 1.

Hypothesized pathways, informed by the Stigma and Substance Use Process Model, between HIV stigma, depression, stress, and recent stimulant use.

3. Results

Sociodemographic Characteristics

Overall, 401 SMM living with HIV participated in the TWM RCT baseline interview. The average age of participants was 39.1 years old (Standard Deviation [SD] = 10.8). Additionally, the majority of participants identified as Black/African American (59.9%), with smaller proportions identifying as White (29.4%) or another race (10.7%). A little over 40% of participants had a four-year college degree (41.8%), 34.0% had postgraduate education, 17.3% had a high school diploma, and 7.0% had less than a high school diploma. Mean anticipated, internalized, and enacted HIV stigma scores were 2.0 (SD = 0.9), 2.2 (SD = 1.1), and 1.5 (SD = 0.7), respectively. Additionally, approximately half (49.1%) of participants reported experiencing depression symptoms in the past week, 68.8% reported experiencing moderate-to-high stress in the past 30 days, and 28.1% had recently used stimulants (52.3% of which had used methamphetamine, 46.8% of which had used cocaine, and 40.5% of which had used amphetamines) at baseline. See Table 1 for a summary of baseline participant characteristics.

Table 1.

Sociodemographics, HIV stigma, depression, perceived stress, and recent stimulant use at baseline among SMM living with HIV (n = 401).

Variable	n(%)
Age (Mean, SD)	39.08(10.76)
Race
White	113(29.43)
Black/African American	230(59.90)
Other	41(10.68)
Education
Less than a high school diploma	28(7.00)
High school diploma	69(17.25)
Four-year college degree	167(41.75)
Any post-graduate studies	136(34.00)
Income
$0-19,999	226(58.55)
$20,000-39,999	85(22.02)
$40,000-74,999	50(12.95)
$75,000 or more	25(6.48)
Condition
Intervention	202(50.37)
Control	199(49.63)
HIV Stigma
Anticipated Stigma (Mean, SD)	2.01(0.85)
Internalized Stigma (Mean, SD)	2.19 (1.07)
Enacted Stigma (Mean, SD)	1.51(0.70)
Depression Symptoms
Yes	194(49.11)
No	201(50.89)
Perceived Stress
Low Stress	125(31.17)
Moderate to High Stress	276(68.83)
Recent Stimulant Use
Yes	111(28.10)
No	284(71.90)

SD: Standard Deviation

HIV Stigma & Depression

In separate preliminary bivariate analyses, anticipated (Odds Ratio [OR] = 2.08, 95% Confidence Interval [CI] = 1.58-2.74, p < .001), internalized (OR = 2.09, 95% CI = 1.65-2.73, p < .001) and enacted (OR = 1.84, 95% CI = 1.33-2.55, p < .001) HIV stigma were prospectively associated with depression symptoms, such that participants with higher HIV stigma scores at baseline were more likely to have depressive symptoms at 5 months post-baseline. In a preliminary multivariable model controlling for age and race, baseline internalized HIV stigma remained prospectively associated with depression symptoms at 5-month follow-up (Adjusted Odds Ratio [aOR] = 1.64, 95% CI = 1.25-2.15, p < .001), but baseline anticipated (aOR = 1.50, 95% CI = 0.99-2.26, p = 0.057) and enacted (aOR = 1.19, 95% CI = 0.76-1.88, p = .447) HIV stigma did not. See Table 2 for a summary of preliminary bivariate and multivariable results.

Table 2.

Bivariate and multivariable associations between sociodemographics and HIV stigma at baseline, depression symptoms and perceived stress at 5-month follow-up, and recent stimulant use at 11-month follow-up among SMM living with HIV (n = 401).

	Depression Symptoms (5-Month Follow-Up)		Perceived Stress (5-Month Follow-Up)		Recent Stimulant Use (11-Month Follow-Up)
	OR (95% CI)	aOR (95% CI)	OR (95% CI)	aOR (95% CI)	OR (95% CI)	aOR (95% CI)
Demographics at Baseline
Age (Mean (SD))	0.97 ** (0.95-0.99)	0.96 ** (0.94-0.99)	0.98 * (0.96-0.99)	0.99 (0.97-1.01)	1.01 (0.99-1.03)	--
Race
White	2.04 * (1.25-3.34)	2.45 ** (1.41-4.26)	1.53 (0.93-2.54)	--	0.96 (0.56-1.64)	--
Black (Ref)	--	--	--	--	--	--
Other	1.33 (0.65-2.74)	1.09 (0.48-2.45)	1.10 (0.53-2.29)	--	0.66 (0.28-1.43)	--
Education
Less than HS Diploma	2.36 (0.90-6.21)	--	5.47 * (1.55-19.31)	4.35 (1.17-16.21)	2.93 * (1.17-7.31)	3.12 * (1.15-8.47)
HS Diploma/GED	0.96 (0.51-1.80)	--	1.49 (0.78-2.84)	1.57 (0.77-3.17)	1.53 (0.72-3.23)	1.53 (0.70-3.34)
Some College	0.88 (0.54-1.43)	--	1.16 (0.71-1.89)	1.16 (0.68-1.99)	1.81 * (1.03-3.18)	1.79 (0.99-3.23)
College (Ref)	--	--	--	--	--	--
Income
$0-19,999 (Ref)	--	--	--	--	--	--
$20,000-39,999	0.61 (0.36-1.03)	--	0.78 (0.46-1.34)	--	0.67 (0.37-1.21)	--
$40,000-74,999	0.77 (0.39-1.52)	--	0.56 (0.28-1.10)	--	0.47 (0.21-1.09)	--
$75,000 or more	0.28 (0.10-0.79)	--	0.48 (0.20-1.19)	--	0.36 (0.28-2.03)	--
Condition
Intervention	1.18 (0.78-1.80)	--	1.33 (0.87-2.06)	--	1.24 (0.77-1.97)	--
Control (Ref)	--	--	--	--	--	--
HIV Stigma at Baseline
Anticipated	2.08 *** (1.58-2.74)	1.50 (0.99-2.26)	1.40 *** (1.41-2.48)	1.41 (0.92-2.15)	1.12 (0.84-1.47)	--
Internalized	2.09 *** (1.65-2.63)	1.64 *** (1.25-2.15)	1.85 *** (1.46-2.34)	1.52 ** (1.15-1.99)	1.00 (0.80-1.26)	--
Enacted	1.84 *** (1.33-2.55)	1.19 (0.76-1.88)	1.56 ** (1.13-2.23)	1.06 (0.67-1.68)	1.41 * (1.01-1.98)	1.36 (0.96-1.94)
Psychosocial Mechanisms at 5-Month Follow-Up
Depression
Yes	--	--	--	--	1.02 (0.63-1.65)	--
No (Ref)	--	--	--	--	--	--
Perceived Stress
Moderate-High	--	--	--	--	1.48 (0.89-2.46)	--
Low (Ref)	--	--	--	--	--	--

Abbreviations: OR = Odds Ratio; aOR = Adjusted Odds Ratio; CI = Confidence Interval; SD = Standard Deviation

^*p < .05;

^**p < .01;

^***p < .001.

HIV Stigma & Perceived Stress

Similarly, in preliminary bivariate analyses, anticipated (OR = 1.40, 95% CI = 1.41-2.48, p < .001), internalized (OR = 1.85, 95% CI = 1.46-2.34, p < .001), and enacted (OR = 1.56, 95% CI = 1.13-2.23, p = .008) HIV stigma was prospectively associated with perceived stress, such that participants with higher baseline HIV stigma scores were more likely to have moderate-to-high perceived stress 5 months post-baseline. In a preliminary multivariable model controlling for age and education level, baseline internalized HIV stigma remained prospectively associated with moderate-to-high perceived stress at 5-month follow-up (aOR= 1.52, 95% CI = 1.15-1.99, p = .003), but baseline anticipated (aOR = 1.41, 95% CI = 0.92-2.15, p = .112) and enacted (aOR = 1.06, 95% CI = 0.67-1.68, p = .807) HIV stigma did not.

HIV Stigma and Stimulant Use

In preliminary bivariate analyses, baseline enacted HIV stigma was prospectively associated with stimulant use at 11-month follow up (OR = 1.41, 95% CI = 1.01-1.98, p = .044), but baseline anticipated (OR = 1.12, 95% CI = 0.84-1.47, p = .442) and internalized (OR = 1.00, 95% CI = 0.80-1.26, p = .978) HIV stigma was not. Depression symptoms (OR = 1.02, 95% CI = 0.63-1.65, p = .947) and perceived stress (OR = 1.48, 95% CI = 0.89-2.46, p = .134) at 5-month follow-up were also not associated with stimulant use 11 months post-baseline. In a preliminary multivariable model controlling for education level, baseline enacted HIV stigma was only marginally associated with stimulant use at 11-month follow-up (aOR = 1.36, 95% CI = 0.96-1.94, p = .088).

Path Analyses

The fit indices for the current analysis indicated “perfect fit” and the likely saturation of the path model (χ² = 0.00; CFI = 1.00). Given that the main aim of the current analysis was to test the paths within the Stigma and Substance Use Process model, however, we did not evaluate alternative, more parsimonious models. Depression symptoms: In path analyses controlling for covariates determined by the preliminary analyses (i.e., age, race, education level, baseline depression symptoms, perceived stress, and stimulant use), neither baseline anticipated HIV stigma (β = 0.09, SE = 0.08, p = .257) nor enacted HIV stigma (β = 0.05, SE = 0.07, p = .460) had a direct effect on depression symptoms at 5-month follow-up. Baseline internalized HIV stigma, however, had a positive direct effect on depression symptoms 5 months post-baseline (β = 0.28, SE = 0.06, p < .001). Perceived stress: Similarly, neither baseline anticipated HIV stigma (β = 0.12, SE = 0.08, p = .126) nor enacted HIV stigma (β = 0.004, SE = 0.07, p = .952) had a direct effect on perceived stress at 5-month follow-up. Baseline internalized HIV stigma, however, had a positive direct effect on perceived stress 5 months post-baseline (β = 0.22, SE = 0.06, p < .001). Stimulant use: In addition, neither baseline anticipated HIV stigma (β = −0.04, SE = 0.08, p = .627) nor internalized HIV stigma (β = −0.08, SE = 0.07, p = .228) had a direct effect on recent stimulant use at 11-month follow-up (see Figure 2). Baseline enacted HIV stigma, however, had a positive direct effect on recent stimulant use 11 months post-baseline (β = 0.14, SE = 0.07, p = .045). Lastly, neither depression symptoms (β = −0.06, SE = 0.07, p = .373) nor perceived stress (β = 0.13, SE = 0.07, p = .053) at 5-month follow-up had a direct effect on recent stimulant use 11 months post-baseline.

Figure 2.

Full path model with anticipated, internalized, and enacted HIV stigma at baseline, depression symptoms and perceived stress at 5-month follow-up, and recent stimulant use at 11-month follow-up controlling for age, race, education level, baseline depression, and baseline stress (n = 309).

Note: All presented path estimates are standardized regression coefficients. * p < .05, ** p < .01, *** p < .001.

Overall, the relationship between HIV stigma (anticipated and internalized) and recent stimulant use was not mediated by either psychosocial mechanism (i.e., depression or perceived stress). More specifically, the mediated path results indicated that anticipated HIV stigma at baseline did not have an indirect effect on recent stimulant use at 11-month follow-up through depression symptoms (β_IND = 0.003, SE = 0.006, p = .581) or perceived stress (β_IND = 0.004, SE = 0.006, p = .484) at 5-month follow-up. Similarly, the path analysis results indicated that internalized HIV stigma at baseline did not have an indirect effect on recent stimulant use at 11-month follow-up through depression symptoms (β_IND = 0.003, SE = 0.006, p = .562) or perceived stress (β_IND = 0.005, SE = 0.005, p = .334) at 5-month follow-up. Lastly, path analysis results indicated that enacted HIV stigma at baseline did not have an indirect effect on recent stimulant use at 11-month follow-up through depression symptoms (β_IND = 0.001, SE = 0.004, p = .799) or perceived stress (β_IND = −0.0003, SE = 0.002, p = .985) at 5-month follow-up.

4. Discussion

The primary aim of the current study was to longitudinally test the hypothesized pathways presented within the Stigma and Substance Use Process Model (23, 24) among a diverse sample of SMM living with HIV from New York City, NY. The results presented herein provide evidence supporting a number of the pathways within this theoretical process model. For example, increases in lifetime internalized HIV stigma (i.e., self-devaluation) at baseline were found to be significantly associated with increases in the psychosocial outcomes of depression symptoms and perceived stress at 5-month follow-up. In addition to corroborating the Stigma and Substance Use Process Model (23, 24), these findings are in line with previous research that found internalized stigma was associated with poorer affective health (e.g. feelings of helplessness) among SMM recently diagnosed with HIV in the United States, and with elevated mental health concerns (e.g., depression, anxiety, and generalized stress) among SMM living with HIV in Australia (11, 12). Given that SMM living with HIV may be particularly vulnerable to stigma, depression, anxiety, and stress (11, 13, 21), and that these factors negatively impact disease progression and treatment adherence (15, 16, 18, 19), combination HIV and mental health services could be important sites for stigma screening and intervention. Furthermore, the integration of interventions that promote engagement in care and that can be leveraged to support reductions in internalized HIV stigma – like the integrating ENGagement Adherence Goals upon Entry (iENGAGE) intervention, an intervention targeting retention in HIV care for PLWH newly entering treatment – with successful mental health intervention techniques (e.g., individual, in-person cognitive behavioral therapy sessions) for SMM living with HIV is critically needed (34, 35).

Additionally, increases in lifetime enacted HIV stigma (e.g., experiences of discrimination) at baseline were significantly associated with an increase in the odds of recent stimulant use at 11-month follow-up for our sample of SMM living with HIV. These findings align with the Stigma and Substance Use Process Model and expand existing research that has found that HIV stigma and discrimination is associated with hazardous drinking and alcohol use severity among PLWH in the Southeastern US (10, 36). Taken together, these findings indicate that experiences of HIV discrimination are enduring and may have long-lasting effects on substance use for SMM living with HIV. As such, future research should explore whether specific forms of social support, coping strategies, and/or strategies for building resilience can be employed to buffer the longitudinal associations between enacted HIV stigma and subsequent substance use behaviors for SMM living with HIV (11, 37–40). Additionally, given that enacted HIV stigma is a function of sociostructural forces, combined structural- and individual-level interventions may be needed to address HIV discrimination (41–44). Further, such multi-level approaches will likely be vital to address intersecting forms of stigma (e.g., HIV-, sexual orientation-, mental health-related stigmas) for SMM living with HIV (45–47).

Other hypothesized pathways within the Stigma and Substance Use Process Model (23, 24), however, were not confirmed within the current analyses. Specifically, lifetime anticipated HIV stigma was not significantly associated with depression symptoms, perceived stress, or recent stimulant use at follow up when controlling for covariates. This indicates that anticipated HIV stigma may not be as critical of an intervention target for SMM living with HIV, and that internalized and enacted HIV stigma are potentially more consequential determinants of mental and behavioral health for this population. Additionally, neither depression symptoms nor perceived stress were significantly associated with recent stimulant use at 11-month follow-up, and these two psychosocial variables did not significantly mediate the relationship between lifetime HIV stigma at baseline (anticipated, internalized, and enacted) and recent stimulant use at 11-month follow-up. These findings correspond to previous research indicating that emotional dysregulation does not mediate the relationship between internalized HIV stigma and stimulant use at the within-person level among SMM living with HIV in New York City (48). Further, this information can be employed to inform and refine the pathways presented within Stigma and Substance Use Process Model (23, 24) and to guide the development of tailored intervention efforts for SMM living with HIV. Future research is needed, however, to more fully test the Stigma and Substance Use Process Model and to clearly elucidate the relationships between HIV stigma, other forms of stigma (e.g., race-, sexual orientation-, and mental health-related stigma), intersectional stigma, depression, perceived stress, and substance use across varying timepoints (e.g., with follow-up visits at alternative time intervals) and with alternative measures of substance use (e.g., use of other substance types, frequency of use, and substance misuse).

Limitations

This study has a number of limitations to consider. First, though the current study employed both active (e.g., venue-based direct outreach) and passive (e.g., internet-based advertising) recruitment approaches, the sample only included SMM living with HIV that could attend in-person assessments in New York, NY. Additionally, the sample was predominately Black/African American (60%) and highly educated (76% had at least some college education). As such, the representativeness of the sample and the generalizability of the findings may be limited. Despite this, however, the current sample closely resembles the racial/ethnic composition of national HIV diagnoses in the US (49). Further, the use of urinalysis to determine recent stimulant use 11-months post-baseline, though rigorous, may have failed to capture other patterns of stimulant use (e.g., stimulant use from more than 4 days prior to measurement). This coupled with the length of time between baseline and follow-up time points could have potentially reduced our ability to detect significant associations for the specified paths in the Stigma and Substance Use Process Model. Due to the use of longitudinal designs, attrition, and the potential bias introduced by resulting missingness, is a limitation for these analyses. For the current study, a total of 351 (88%) and 324 (81%) of the original 401 participants completed the 5-month and11-month follow-up surveys, respectively. Though retention rates were within acceptable limits, it is still possible that bias was introduced due to participant attrition (50). Despite the potential for attrition bias, however, the prevalence of depression (baseline: 49.1%; 5-months: 47.6%; 11-months: 45.8%), stress (baseline: 68.8%; 5-months: 61.4%; 11-months: 61.4%), and stimulant use (baseline: 28.1%; 5-months: 28.3%; and 11-months: 30.6%) remained relatively stable over time. In addition, given that the TWM study employed an informational control group (vs. a pure control group), it is possible that participant depression symptoms, perceived stress, and stimulant use at 5- and 11-months post-baseline were influenced by study participation. Lastly, intervention randomization was not included as a covariate within the multivariable and path analysis models. Given that HIV stigma, depression symptoms, perceived stress, and stimulant use were not the primary targets for the TWM intervention, however, randomization should be evenly distributed across groups. Further, the TWM intervention did not have a significant impact on HIV stigma, depression symptoms, or perceived stress in bivariate results.

5. Conclusion

The current study identified that internalized HIV stigma is significantly associated with depression symptoms and perceived stress 5-months later, and that experiences of enacted HIV stigma are significantly associated with recent stimulant use nearly a year later for SMM living with HIV. These findings reinforce and expand existing literature by delineating how specific forms of HIV stigma negatively impact mental and physical health, and how these effects are sustained over time. This work provides partial evidence for the validity of Stigma and Substance Use Process Model and also contributes information for the potential refinement of this theory. Further, the results of this study have important implications for the development of tailored interventions and treatment efforts targeting stigma, mental health, and disease progression among SMM living with HIV.

Availability of Data, Material, and Code:

The datasets generated and analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.