Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Dec 12.
Published in final edited form as: Drug Alcohol Depend. 2020 Jun 12;213:108123. doi: 10.1016/j.drugalcdep.2020.108123

Longitudinal associations between the disruption of incarceration and community re-entry on substance use risk escalation among Black men who have sex with men; a causal analysis

Jonathan Feelemyer 1,*, Typhanye V Dyer 2, Rodman E Turpin 2, Russell A Brewer 3, Christopher Hucks-Oritz 4, Willem F Van Der Mei 5, Charles M Cleland 5, Medha Mazumdar 5, Ellen C Caniglia 5, Amanda Geller 6, Joy D Scheidell 5, Justin M Feldman 5, Kenneth H Mayer 7, Maria R Khan 5
PMCID: PMC8059073  NIHMSID: NIHMS1608190  PMID: 32593152

Abstract

Background and Aims:

While substance use can lead to incarceration, the disruptive effects of incarceration may lead to, or increase psychosocial vulnerability and substance use. Using causal inference methods, we measured longitudinal associations between incarceration and post-release substance use among Black men who have sex with men (BMSM), populations facing disproportionate risk of incarceration and substance use.

Methods:

Using data from the HIV Prevention Trials Network (HPTN 061) study (N=1,553) we estimated associations between past 6-month incarceration and binge drinking, marijuana use, and stimulant use post release (at 12-month follow-up visit). Adjusted models used inverse probability weighting (IPW) to control for baseline (pre-incarceration) substance use and additional risk factors.

Results:

There were 1,133 participants present at the twelve-month follow-up visit. Participants were predominately non-Hispanic Blacks and unemployed. At baseline, 60.1% reported a lifetime history of incarceration, 22.9% were HIV positive and 13.7% had a history of an STI infection. A total of 43% reported a history of depression. In adjusted analyses with IPW, recent incarceration was associated with crack-cocaine (adjusted odds ratio (AOR): 1.53, 95% confidence interval (CI): 1.03, 2.23) and methamphetamine use (AOR: 1.52, 95% CI: 0.94–2.45). Controlling for pre-incarceration binge drinking, incarceration was associated with post-release binge drinking (AOR: 1.47, 95% CI: 1.05, 2.04); in fully adjusted models the AOR was 1.14 (95% CI: 0.81, 1.62). Incarceration was not associated with marijuana use.

Conclusion:

Findings underscore the need to provide substance use treatment in custody and post-release, and to consider alternatives to incarceration for substance using populations.

1. INTRODUCTION

Black men who have sex with men (BMSM) are more likely to use substances compared to non-Hispanic white heterosexual individuals, particularly stimulants and alcohol use; this disparity has existed for decades (Reback et al., 2014; Stall and Wiley, 1988); additionally, it is well documented that BMSM are more likely to use specific substances, including crack cocaine, with one study finding that 94% of MSM crack smokers were Black (McKirnan et al., 2001). Stimulants and alcohol are particularly strong drivers of HIV/STI risk factors, affecting sexual risk behaviors, mental health issues including depression and anxiety, and intimate partner violence (Maulsby et al., 2014; Vosburgh et al., 2012; Wu et al., 2015).

There is a gap in information about the structural-level determinants that increase vulnerability to substance use among BMSM. Improved understanding of these structural-level determinants (including racism, violence, trauma, and co-occurring health disorders) among BMSM is a critical to addressing the HIV epidemic in this population and informing appropriate interventions. Studies have highlighted the race disparity among incarcerated individuals (Mauer, 2011) and sexual minorities are disproportionately represented in the incarcerated populations in the United States; the incarceration rate among sexual minorities in 2012 was 1,882 per 100,000, more than 3 times that of the United States adult population (Meyer et al., 2017). Additionally, there is evidence that sexual minorities are disproportionately represented among those incarcerated; in 2017, 9.3% of men and 35.7% of women in jails identified as sexual minorities, compared to less than 4% prevalence of sexual minorities in the general population (Meyer IH., 2019). Individuals who face dual minority status of being both a racial/ethnic and a sexual minority face disproportionate risk of incarceration; for example, BMSM have twice the incarceration risk compared to their non-racial/ethnic minority MSM counterparts (Millett et al., 2012).

Despite the evidence of relationships between incarceration and substance use among sexual minorities including BMSM, only 10–20% of incarcerated individuals who are drug dependent receive treatment with poor continuity of care (e.g., linkage to drug abuse treatment or referrals upon release), further exacerbating substance abuse after release from correctional settings (Chandler et al., 2009).

Individuals who become incarcerated enter correctional settings with long histories of pre-incarceration substance use and dependence (Calzavara et al., 2003; Karberg and James, 2005), with the majority of currently incarcerated individuals endorsing a history of drug use (Belenko and Peugh, 2005). Drug possession or use is a direct risk factor for incarceration (Winter et al., 2019), and 50% of individuals are currently incarcerated for a drug crime (The Sentencing Project, 2018).

It also is possible that the experience of incarceration may contribute to substance use initiation or escalation. Qualitative studies indicate incarceration introduces individuals to higher-risk networks in which network members are more likely to trade and/or use drugs, and some individuals report experimentation with illicit drugs for the first time as a result of access to these substances while incarcerated (Cope, 2000). While incarceration-related changes in networks may increase exposure and access to drug use, it could also negatively influence mental health which in turn may contribute to self-medication with alcohol and/or drugs (Calzavara et al., 2003).There is literature describing the high rates of overdose risk and mortality after incarceration particularly in the first weeks of community re-entry (Forsyth et al., 2018; Harding-Pink, 1990; Pratt et al., 2006). However, these studies, conducted in samples of individuals with a history of incarceration, cannot examine the independent influence of incarceration on substance use by comparing post-release substance use and risk among those with a recent incarceration to those without this exposure. A previous study in a sample of persons who inject drugs in Vancouver indicated those who had recently been incarcerated face much higher rates of elevated injection drug risk compared to those with no recent incarceration history (DeBeck et al., 2009); however, this study did not take place in the United States, and was not able to examine the important factors unique to urban populations such as BMSM. Longitudinal research examining substance use post incarceration specific to BMSM are extremely limited (Brewer et al., 2014b). Such studies would have the potential for population-level impact given the disproportionate incarceration, HIV and STI risks of BMSM.

In order to address the gaps in the extant literature, the purpose of this analysis was to evaluate the relationship between recent incarceration and substance use among BMSM using a causal analysis approach. In the context of this population’s disproportionate risk of incarceration, elevated risk of substance use post release, and the strong association between substance use and HIV risk, understanding the prospective association between incarceration and substance use post-release can have significant implications not only for substance use intervention, but HIV risk interventions among BMSM as well. Having knowledge of what specific substances of abuse are most prevalent among BMSM post release will allow for tailored drug treatment and harm reduction interventions specifically for BMSM populations in urban cities in the United States.

2. METHODS

2.1. Study Design and Participants

Data from the HIV Prevention Trials Network 061 (HPTN 061) study was used to evaluate the longitudinal relationship between recent incarceration and post-release changes in frequency of alcohol use (i.e., binge drinking) and commonly used drugs in the sample. The study’s enrollment and recruitment methods have been described comprehensively elsewhere (Brewer et al., 2014a). HPTN 061 sought to examine the feasibility and efficacy of interventions to prevent the acquisition and transmission of HIV predominantly among BMSM. Enrollment took place from 2009–2010 in six US metropolitan cities: Atlanta, Boston, Los Angeles, New York City, San Francisco, and Washington D.C. Participants were recruited directly from the community or as sexual network partners referred by index participants.

Individuals were eligible to participate in the study if they were Black, African American, Caribbean Black, or multiethnic Black; identified as a man or assigned male at birth; were at least 18 years old; reported ≥1 instance of unprotected (i.e. without a condom) insertive or receptive anal intercourse (UAI) with a man in the prior six months; resided in the metropolitan area of the research clinic; did not plan to move away during the study period; and provided informed consent for the study. Individuals were ineligible if they were enrolled in any other HIV interventional research study, had been a participant in an HIV vaccine trial, or were a community-recruited participant in a category that had already reached its enrollment cap. Prescreening to determine eligibility was performed either in person or over the telephone.

Institutional review boards at all participating institutions approved the study. Participants completed an audio computer-assisted self-interview (ACASI) at baseline, six-month, and twelve-month follow-up that assessed demographic information, HIV risk behaviors, experiences of violence, internalized homophobia, and other characteristics. At these visits, participants were tested for STIs using urine tests and rectal swabs (throat swabs were not used), and for HIV using HIV rapid testing with confirmation of results via Western blot testing at the sites. Additional HIV testing was performed retrospectively for quality assurance at the HPTN Laboratory Center. During the six-month visit, participants were also asked if they had been incarcerated in the previous six months (i.e., during the time since baseline).

2.2. Substance Use:

Participants reported on their substance use frequency in the six months prior to the study visit at baseline, six- and twelve-month visits. Substances inquired about included marijuana, crack cocaine, powder cocaine, methamphetamine, and alcohol (defined as binge drinking, defined by NIAAA (National institute on Alcohol Abuse and Alcoholism, 2019)). Heroin was also asked about, but less than 4% of participants reported use of heroin, so we did not conduct further analysis on this substance due to the small number of users. Participants were also asked about injection drug use, but this was reported by less than 1% of the sample and was also not examined further.

Participants indicated frequency of drug use with an eight-level response scale ranging from (8)daily, (7)several times a week, (6)weekly, (5)several times a month, (4)monthly, (3)a few times, (2)once, or (1)never. For alcohol, participants indicated the number of binge drinking events in the last month (defined by NIAAA), use on a five-level response scale ranging from never to (5)daily/almost daily, (4)weekly, (3)monthly, (2)less than monthly, or (1)never. Due to small cell counts for some frequency response levels, we combined the different levels into three response categories for the purpose of this analysis.

For substance use, the updated frequency categories were:

  1. Frequently (includes original scale items (8) (daily), (7) (several times a week), (6) (weekly); and (5) (several times a month));

  2. Occasionally (includes original scale items (4) (monthly) and (3) (a few times));

  3. Rarely/never (includes original scale items (2) (once) and (1) (never))

For binge drinking (alcohol use, greater than six drinks at one time), the updated frequency categories were:

  1. Frequently (includes original scale items (5) (daily or almost daily));

  2. Occasionally (includes original scale items (4) (weekly) (3) (monthly), and (2) less than monthly);

  3. Never (includes original scale item (1) (never)).

2.3. Incarceration Status:

At six-month follow-up, participants reported the number of times they spent one or more nights incarcerated in the previous six months. We defined recent incarceration as having experienced at least one incarceration in the six months prior to the six-month follow-up visit.

2.4. Substance Use Outcomes:

At twelve months, the utilization of the following substances was assessed: binge drinking, marijuana, cocaine, crack cocaine and methamphetamine.

2.5. Statistical Analysis:

Baseline demographic information and substance use were summarized for the participants. We examined the prevalence of substance use among participants stratified for baseline and 12-month follow-up time periods. In order to evaluate the association between incarceration at six months and substance use at 12-month follow-up, we estimated ordered logistic regression models controlling for baseline drug use among the participants.

We utilized multiple imputation by chained equations for missing data (Van Buuren, 2018) in the questionnaire responses yielding 40 imputed datasets using predictive mean matching. Out of the 1,133 subjects present at the 12-month visit, 416 (36.7%) were missing information on either a covariate or the exposure. Quality of the imputations was assessed via visual inspection of density plots of the imputed variables. Values of the imputed variables were checked to make sure that only plausible values were imputed. Proportional odds test were conducted for each of the different substance use ordered variables (including binge drinking, marijuana, crack, cocaine and methamphetamine) and satisfied the proportional odds/parallel assumption (data not shown).

We used inverse probability weighting (IPW) (Cole and Hernán, 2008) with stabilized weights in fully adjusted analysis to adjust for baseline confounding (variables used to estimate the weights were measured at baseline and included transgender identity, any STI, unstable housing, high school education (high school or less; some college or more), hard drug use, weekly marijuana use, incarceration history, HIV testing history, insufficient income, whether the subject also has sex with women, alcohol use disorders identification test (AUDIT) score, experience of violence, either buying or selling sex, multiple partnership, concurrent partnership, city, cohabitation status, health coverage, HIV status at baseline, age, center for epidemiologic studies depression scale (CES-D) score, social support scale, perceived homophobia scale, perceived racism scale, and internalized homophobia scale).

Predicted probabilities, also known as propensity scores, of recent incarceration at six months (the main exposure) were extracted and converted into inverse probability weights, which were stabilized in regard to the probability of the observed exposure (Hernán et al., 2006). Logistic regression with the Ridge penalty was conducted for each of the forty imputed datasets, to create forty sets of propensity scores, which were used to create forty sets of inverse probability weights. These weights were then used in the fully adjusted analysis to assess the relationship between six-month incarceration status and each of the individual substances (including binge drinking, marijuana use, crack use, cocaine use, and methamphetamine use). For all fully adjusted models, we controlled for baseline use of the particular substance outcome (for instance, in assessing the relationship between incarceration at six months and binge drinking at twelve-month follow-up, we controlled for baseline frequency of binge drinking in addition to variables included in the IPW model). Ordered logistic regression with robust standard errors was conducted for each substance use outcome in each of the forty imputed datasets. Parameter estimates and variances were extracted from each model, and were pooled to obtain unweighted and weighted odds ratios and standard errors for the association between incarceration and different substances of use following Rubin’s rules (Rubin, 2004).

All statistical analysis was conducted in R (Team, 2017).

3. RESULTS

3.1. Demographic Characteristics:

A total of 1,553 participants were included at baseline, 1,169 were present and responded to the incarceration question at the six-month visit, and 1,133 participants were included in the twelve-month follow-up analysis. Participant demographics are described in Table 1. At baseline, the majority of participants were non-Hispanic Blacks and unemployed; approximately 47.3% had at least a high school education, over 55.4% made less than $20,000 per year, and 9.6% reported unstable housing. At baseline, 60.1% reported a lifetime history of incarceration, 22.9% were HIV positive and 13.7% had a history of an STI infection. A total of 43% reported a history of depression. Characteristics were similar among follow-up participants, although there was a notably higher prevalence of past six months incarceration among baseline participants compared to follow-up participants (39.9% vs 15.4%), and slightly higher prevalence of depression (43.2% vs 38.9%).

Table 1:

Sample Demographic, Socioeconomic and Health Background among Baseline and Follow-up 061 Participants

Total Sample N Total Sample % Follow-up N (12 months) Follow-up %
1553 1133
Age
18–30 517 33.4 396 34.9
31–50 812 52.4 596 52.7
50 and over 220 14.2 140 12.4
Ethnicity
Non-Hispanic 1430 92.3 1067 92.4
Hispanic 119 7.7 88 7.6
Education
Greater than High School 732 47.3 585 50.7
High School 816 52.7 569 49.3
Insufficient Income
No 690 44.6 529 45.8
Yes 858 55.4 626 54.2
Unstable Housing
No 1401 90.5 1054 91.3
Yes 148 9.6 101 8.7
City of Residence
Washington DC 227 14.6 220 18.7
Atlanta 292 18.8 256 21.8
Boston 237 15.3 179 15.3
Los Angeles 283 18.2 159 13.5
New York City 310 19.9 207 17.6
San Francisco 204 13.1 153 13.0
Health Coverage
No 613 39.6 456 39.5
Yes 936 60.4 699 60.5
Incarcerated
No 607 60.1 970 84.6
Yes 914 39.9 176 15.4
HIV Serostatus
Negative 1167 77.1 920 80.1
Positive 347 22.9 228 19.9
STI (Any)
Negative 1245 86.3 855 89.4
Positive 198 13.7 101 10.6
Depression
No 759 56.8 608 61.1
Yes 578 43.2 387 38.9
Open in a new tab
*

Covariates with missing values may not add up to total N

Several socio-demographic variables, depression, and STI diagnosis were associated with reporting previous or current use of substances at follow-up in bivariate analysis (Table 2). When examining stimulants (which include crack cocaine, cocaine, and methamphetamine), we noted several important associations. Older age groups (OR range: 3.92 – 12.94), insufficient income (OR: 1.56, 95% CI: 1.13, 2.15), unstable housing (OR: 2.81, 95% CI: 1.77, 4.45), having health coverage (OR: 1.73, 95% CI: 1.23, 2.43), incarceration history (OR: 3.71, 95% CI: 2.49, 5.53) and depression (OR: 1.51, 95% CI: 1.08, 2.11) were positively associated with crack cocaine use, while having an STI infection was negatively associated with crack cocaine use (OR: 0.29, 95% CI: 0.15, 0.59). Older age groups (OR range: 1.24 – 1.33), insufficient income (OR 1.54, 95% CI: 1.05, 2.27) and those with incarceration history (OR: 3.07, 95% CI: 1.93, 4.88) were more likely to report cocaine use. Finally, Hispanic ethnicity (OR: 2.57, 95% CI: 1.29, 5.11), and incarceration history (OR: 2.86, 95% CI 1.53, 5.32) were positively associated with methamphetamine use.

Table 2:

Odds Ratios from Ordered Logit models for associations with frequency of use of each substance at follow-up and select demographic characteristics (N=1553)

Covariate Binge Drinking1 Marijuana2 Crack2 Cocaine2 Methamphetamine2
OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Age
18–30 (Reference)
31–50 0.90 (0.69, 1.17) 0.61 (0.47, 0.79) 3.92 (6.06, 22.55) 1.24 (0.81, 1.89) 1.04 (0.60, 1.81)
50 and over 0.64 (0.43, 0.98) 0.58 (0.39, 0.86) 12.94 (6.19, 27.04) 1.33 (0.73, 2.46) 1.18 (0.53, 2.63)
Ethnicity
Non-Hispanic (Reference)
Hispanic 1.51 (0.97, 2.34) 1.43 (0.92, 2.22) 0.88 (0.48, 1.62) 0.81 (0.38, 1.71) 2.57 (1.29, 5.11)
Education
Greater than High School (Reference)
High School 1.01 (0.79, 1.27) 0.86 (0.68, 1.08) 0.63 (0.46, 0.87) 0.89 (0.62, 1.30) 0.92 (0.56, 1.51)
Insufficient Income
No (Reference)
Yes 1.24 (0.97, 1.58) 1.34 (1.06, 1.69) 1.56 (1.13, 2.15) 1.54 (1.05, 2.27) 1.14 (0.69, 1.89)
Unstable Housing
No (Reference)
Yes 1.10 (0.73, 1.68) 1.59 (1.05, 2.42) 2.81 (1.77, 4.45) 0.94 (0.48, 1.87) 1.31 (0.58, 2.96)
City of Residence
Washington DC (Reference)
Atlanta 1.07 (0.70, 1.61) 1.19 (0.78, 1.82) 1.83 (0.94, 3.58) 3.24 (1.50, 6.97) 0.19 (0.02, 1.71)
Boston 1.48 (0.95, 2.31) 2.24 (1.44, 3.52) 4.47 (2.34, 8.52) 2.39 (1.05, 5.48) 1.59 (0.44, 5.76)
Los Angeles 0.42 (0.26, 0.65) 1.56 (1.02, 2.39) 1.33 (0.66, 2.69) 1.01 (0.41, 2.51) 6.15 (2.10, 17.98)
New York City 0.88 (0.59, 1.31) 1.75 (1.17, 2.62) 2.10 (1.10, 3.99) 2.66 (1.23, 5.72) 0.49 (0.11, 2.24)
San Francisco 1.01 (0.64, 1.57) 2.56 (1.62, 4.04) 3.11 (1.58, 6.09) 2.50 (1.08, 5.76) 8.16 (2.76, 24.08)
Health Coverage
No (Reference)
Yes 0.96 (0.77, 1.23) 1.12 (0.88, 1.43) 1.73 (1.23, 2.43) 0.73 (0.50, 1.06) 1.03 (0.62, 1.72)
Incarcerated
No (Reference)
Yes 1.10 (0.86, 1.41) 1.78 (1.39, 2.27) 3.71 (2.49, 5.53) 3.07 (1.93, 4.88) 2.86 (1.53, 5.32)
HIV Serostatus
Negative (Reference)
Positive 0.71 (0.52, 0.98) 0.81 (0.60, 1.09) 1.19 (0.82, 1.76) 0.73 (0.44, 1.23) 0.73 (0.37, 1.47)
STI (Any)
Negative (Reference)
Positive 0.83 (0.57, 1.19) 1.04 (0.74, 1.48) 0.29 (0.15, 0.59) 0.95 (0.53, 1.68) 0.71 (0.30, 1.69)
Depression
No (Reference)
Yes 1.36 (1.05, 1.77) 1.27 (0.98, 1.63) 1.51 (1.08, 2.11) 1.27 (0.84, 1.91) 1.37 (0.79, 2.34)
Open in a new tab
1

Original coding scheme for alcohol included frequency scale from 1 (never) to 5 (daily/almost daily), and was collapsed to three levels: frequently, occasionally, and rarely/never.

2

Original coding scheme for drugs included frequency scale from 1 (never) to 8 (daily), and was collapsed to three levels: frequently, occasionally, and rarely/never.

3.2. Substance Use, Baseline vs. 12 Month Follow-up:

Table 3 presents data on substance use by time period (baseline vs follow-up). At baseline versus 12-month follow-up, the substances with the highest prevalence of use included marijuana (55.9% vs 51.7%, p=0.034), crack (26.2% vs 18.7%, p<0.001), cocaine (18.5% vs 13.9%, p=0.002), and methamphetamine (9.5% vs 7.5%, p=0.074). We observed reductions all illicit drug use from baseline to 12-month follow-up, with the largest reductions seen for stimulants. However, we did see an increase in binge drinking use from baseline to follow-up (52.6% to 62.1%, p<0.001).

Table 3:

Frequency of Substance Use at Baseline and Follow-up

Binge drinking1 Marijuana2 Cocaine2 Crack2 Meth2
% % % % %
Baseline
Percent Using Frequently 14.8 36.3 6.1 15.1 3.6
Percent Using Occasionally 32.6 15.2 9.1 9.3 4.3
Percent Using Rarely or Never 52.6 48.5 84.8 75.6 92.2
12-month follow-up
Percent Using Frequently 9.9 33.4 3.9 9.4 3.1
Percent Using Occasionally 27.9 14.3 7.5 7.7 3
Percent Using Rarely or Never 62.2 52.4 88.6 82.9 93.9
Open in a new tab
1

Original coding scheme for alcohol included frequency scale from 1 (never) to 5 (daily/almost daily), and was collapsed to three levels: frequently, occasionally, and rarely/never.

2

Original coding scheme for drugs included frequency scale from 1 (never) to 8 (daily), and was collapsed to three levels: frequently, occasionally, and rarely/never.

3.3. Six-month incarceration and associations with substance use at 12 months:

Table 4 presents substance use at 12-month follow-up stratified by report of recent incarceration at six months. Among those reporting recent incarceration, frequent substance use was highest for marijuana (37.1%) followed by binge drinking (18.9%) and crack cocaine use (18%), while among those not incarcerated, frequent substance use was high for marijuana (32.6%) but was less than 10% for all other substances, including binge drinking. In adjusted analysis controlling for baseline drug use, report of recent incarceration at six months was associated with crack use at 12 months (AOR: 1.44, 95% CI: 1.01, 2.05), which remained significant in adjusted weighted analysis with multiple imputation (AOR: 1.53, 95% CI: 1.03, 2.23), and binge drinking (AOR: 1.47, 95% CI: 1.05, 2.04). There was also an association between reported recent incarceration and methamphetamine use (AOR: 1.52, 95% CI: 0.94–2.45), but the precision for this estimate was low due to the number of participants who reported this substance use at twelve-month follow-up (N=10).

Table 4:

Substance Use Stratified by Incarceration Status at 12-month follow-up & Adjusted Analysis of Incarceration Association with Selected Substances of Use

Reported recent Incarceration at six-month visit Binge Drinking1
N (%)		 Marijuana2
N (%)		 Crack2
N (%)		 Cocaine2
N (%)		 Methamphetamine2
N (%)
Yes (N=165)
Rarely or Never 66 (56.9) 67 (46.9) 99 (71.2) 117 (86.1) 125 (93.3)
Occasionally 28 (24.1) 23 (16.1) 15 (10.8) 12 (8.9) 2 (1.5)
Frequently 22 (18.9) 53 (37.1) 25 (18.0) 7 (5.2) 7 (5.2)
No (N=1004)
Rarely or Never 498 (63.4) 473 (53.5) 729 (84.4) 767 (89.5) 812 (94.6)
Occasionally 227 (28.8) 123 (13.9) 62 (7.2) 59 (6.9) 25 (2.9)
Frequently 61 (7.8) 288 (32.6) 73 (8.5) 31 (3.6) 21 (2.5)
Odds Ratio (95% CI) 1.68 (1.18, 2.38) 1.267 (0.91, 1.77) 2.21 (1.47, 3.33) 1.39 (0.82, 2.36) 1.29 (0.62, 2.72)
Adjusted Odds Ratio (95% CI)* 1.47 (1.05, 2.04) 1.01 (0.71, 1.43) 1.44 (1.01, 2.05) 1.18 (0.81, 1.71) 1.43 (0.95, 2.12)
Adjusted Weighted Odds Ratio (95% CI)** 1.14 (0.81, 1.62) 1.04 (0.75, 1.43) 1.53 (1.05, 2.23) 1.19 (0.77, 1.84) 1.52 (0.94, 2.45)
Open in a new tab
#

p<0.05 by chi square test

*

ordered logistic regression controlling for baseline drug use

**

ordered logistic regression with weighted propensity score and imputation for missing data

1

Original coding scheme for alcohol included frequency scale from 1 (never) to 5 (daily/almost daily), and was collapsed to three levels: frequently, occasionally, and rarely/never.

2

Original coding scheme for drugs included frequency scale from 1 (never) to 8 (daily), and was collapsed to three levels: frequently, occasionally, and rarely/never.

4. DISCUSSION

In this sample of BMSM, we observed that recent incarceration reported at six months was associated with use or increased use of a range of substances post release, after controlling for baseline substance use and other risk factors prior to incarceration. In adjusted analysis using IPW, we noted higher odds of increased frequency of crack cocaine use and a positive trend toward increased frequency of methamphetamine use. Additionally, by including an inverse probability weighted model with multiple imputation, we were able to achieve balance on several important key confounding factors among this BMSM population and more clearly identify the causal association between incarceration and substance use. These findings highlight potential shifts in substance use among this population from marijuana use to crack cocaine and methamphetamine use (Kuo I., 2017). It also highlights the need for comprehensive and integrated care within correctional settings and the importance of continuity of care and treatment post-release among participants in this study. It is well aligned with state efforts to implement alternatives to incarceration for those with substance use histories and problematic substance use (Kearley and Gottfredson, 2017).

Research examining changes in substance use from pre- to post-incarceration among BMSM using a causal analysis approach are lacking in the literature, with most of the research conducted in the context of treatment studies or cross-sectional studies without examination of temporal relationships (Kouyoumdjian et al., 2015). This study reports on longitudinal changes in substance use among a sample of BMSM in six cities in the United States with moderate prevalence of not only incarceration history but recent incarceration, with rates similar to other reports on incarceration among BMSM populations (Brewer et al., 2014a, b). Previous studies have noted important ramifications associated with incarceration including possible loss of employment (Holzer et al., 2005), changes in overall physical and mental health (Massoglia and Pridemore, 2015), and increased use of alcohol and illicit substances (Boys et al., 2002). In our study we specifically address the issues surrounding changes in alcohol and illicit substance use in the context of recent incarceration and how incarceration may change patterns and frequency of substance use.

A previous study among criminal justice-involved individuals documented reductions in substance use frequency and dependence from pre-incarceration to one-year post release (Tangney et al., 2016). We did not however observe these reductions among this sample of BMSM. This begs the question of what specific factors could be driving substance use or increases in substance use post incarceration among our BMSM participants. There are possible factors not measured in our questionnaire that could further tease apart the relationship between incarceration and changes in substance use, including incarceration-specific trauma, changes in resilience post-incarceration, changes in continuity of care and treatment, and differences in social structure support.

In unadjusted and adjusted analysis using IPW, recent incarceration reported at six months was associated with increases in binge drinking and crack use at twelve-month follow-up. Additionally, in adjusted analysis using IPW, incarceration at six months was associated with increased use of crack cocaine and there was a positive trend for other stimulants (cocaine and methamphetamine) at twelve-month follow-up. These results are consistent with previous studies documenting high levels of stimulant and alcohol use compared to other substances among incarcerated individuals. Currently, cocaine/crack cocaine use ranks second to only marijuana use among those incarcerated in the United States (Bronson et al., 2017).

The results of this study have several implications for BMSM who experience recent incarceration. First, harm reduction programs need to more adequately address the dual epidemic of stimulant/alcohol use among minority BMSM, with special focus on BMSM with a history of recent incarceration. Given the high HIV prevalence (22.9%) and STI prevalence (13.7%) in this population, interventions should emphasize HIV and STI risk reduction, especially as polysubstance use among BMSM is associated with increased risk for condomless anal intercourse and being diagnosed with a recent STI infection (Baral et al., 2013; Santos et al., 2014). There should also be increased access to and provisions for pre-exposure prophylaxis (PrEP) to help reduce HIV transmission (Parsons and Cox, 2019). Second, study findings underscore the importance of continuity of care and treatment services for participants enrolled in this study. Correctional settings offer an opportunity to provide these services, however, post-release continuity of care and treatment (e.g., harm reduction services, access to substance use treatment) is critical.

Among BMSM in our sample, 46% of those with incarceration at six-month follow-up were more likely to report depression at 12-month follow up compared to 37% reporting depression among those without recent incarceration; these rates are similar to rates seen among other studies among minority BMSM (Yu et al., 2015). As depression is often associated with illicit substance use (Davis et al., 2008) and recent incarceration (Porter and Novisky, 2017), interventions for BMSM in this sample should not only focus on reducing drug use and drug related harms, but also on treatment for mental health illness, which can be associated with reduction in illicit substance use (Dodge et al., 2005) and improved quality of life. These interventions should be culturally appropriate and should specifically address psychological distress.

Our study advances the literature in several ways. First, we utilized a large cohort of BMSM collected at several sites across the United States. Our use of a longitudinal study design with prospective cohort data allows for the assessment of temporality; this is notably important given the bidirectional association between incarceration and substance use. Our analyses accounted for a wide range of confounders, utilizing propensity score modeling while maintaining sufficient statistical power and uses a causal analysis approach. Finally, our study fills a gap in extant literature on how incarceration can impact risk of future substance use among BMSM.

There are important limitations to consider in the interpretation of our results. We only had one pre- and one post-measurement for substance use, and there is a possibility that variability in substance use during follow-up occurred; indeed, new (after baseline assessment) or unreported substance use may have led to incarceration rather than incarceration leading to substance use. Many of the questions that were asked of participants were based upon self-report; as a result, baseline and follow-up measurements could suffer from recall or social desirability bias, particularly for questions related to stigmatized risk behaviors such as drug injection and risky sexual risk behaviors. Finally, we do note that there was loss to follow-up among the BMSM in our sample, however, we assumed this data was missing at random (MAR), and through multiple imputation, we were able to impute the missing data for the purposes of the analysis (analysis showed no significant differences in characteristics among those included and not included in the follow-up analysis). It should be noted that among those that were loss to follow-up, there is the possibility that some of the participants were incarcerated during the follow-up period, which is why they were not able to attend follow-up interviews for the study.

5. CONCLUSION

To the best of our knowledge, this study is one of the first to specifically assess changes in substance use over time with the identification of potential shifts in drug of choice over time among incarcerated BMSM in the United States using a causal analysis approach. The results from this study clearly indicate a need for enhanced culturally appropriate interventions for minority BMSM, with special focus on stimulants and alcohol use. Correctional settings may offer a contact point in linking BMSM substance users to services, and integration of HIV prevention and harm reduction services will help to reduce untoward harm and further HIV transmission after incarceration release

Highlights.

  • 1553 Black men who have sex with men were recruited from six cities in the US

  • Frequent drug use lowest for methamphetamine and highest for marijuana

  • Recent incarceration associated with binge drinking in adjusted analysis

  • Recent incarceration associated with crack use in adjusted weighted analysis

  • Need to consider alternatives to incarceration for substance using populations

Acknowledgments

We would like to acknowledge the following: HPTN 061 grant support is provided by the National Institute of Allergy and Infectious Disease (NIAID), National Institute on Drug Abuse (NIDA) and National Institute of Mental Health (NIMH): Cooperative Agreements UM1 AI068619, UM1 AI068617, and UM1 AI068613.

Role of Funding Source Nothing declared

Footnotes

Conflict of Interest

No conflict of interest declared

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Baral SD, Poteat T, Stromdahl S, Wirtz AL, Guadamuz TE, Beyrer C, 2013. Worldwide burden of HIV in transgender women: a systematic review and meta-analysis. Lancet Infect Dis 13(3), 214–222. [DOI] [PubMed] [Google Scholar]
  2. Belenko S, Peugh J, 2005. Estimating drug treatment needs among state prison inmates. Drug Alcohol Depend 77(3), 269–281. [DOI] [PubMed] [Google Scholar]
  3. Boys A, Farrell M, Bebbington P, Brugha T, Coid J, Jenkins R, Lewis G, Marsden J, Meltzer H, Singleton N, Taylor C, 2002. Drug use and initiation in prison: results from a national prison survey in England and Wales. Addiction 97(12), 1551–1560. [DOI] [PubMed] [Google Scholar]
  4. Brewer RA, Magnus M, Kuo I, Wang L, Liu TY, Mayer KH, 2014a. Exploring the relationship between incarceration and HIV among black men who have sex with men in the United States. J Acquir Immune Defic Syndr 65(2), 218–225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Brewer RA, Magnus M, Kuo I, Wang L, Liu TY, Mayer KH, 2014b. The high prevalence of incarceration history among Black men who have sex with men in the United States: associations and implications. Am J Public Health 104(3), 448–454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bronson J, Stroop J, Zimmer S, Berzofsky MJW, DC: United States Department of Justice, 2017. Drug use, dependence, and abuse among state prisoners and jail inmates, 2007–2009. [Google Scholar]
  7. Calzavara LM, Burchell AN, Schlossberg J, Myers T, Escobar M, Wallace E, Major C, Strike C, Millson M, 2003. Prior opiate injection and incarceration history predict injection drug use among inmates. Addiction 98(9), 1257–1265. [DOI] [PubMed] [Google Scholar]
  8. Chandler RK, Fletcher BW, Volkow ND, 2009. Treating drug abuse and addiction in the criminal justice system: improving public health and safety. JAMA 301(2), 183–190. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cole SR, Hernán M, 2008. Constructing inverse probability weights for marginal structural models. Am J Epidemiol 168(6), 656–664. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cope N, 2000. Drug Use in Prison: the experience of young offenders. Drugs: Education, Prevention and Policy, 7(4), 355–366. [Google Scholar]
  11. Davis L, Uezato A, Newell JM, Frazier E, 2008. Major depression and comorbid substance use disorders. Curr Opin Psychiatry 21(1), 14–18. [DOI] [PubMed] [Google Scholar]
  12. DeBeck K, Kerr T, Li K, Milloy MJ, Montaner J, Wood E, 2009. Incarceration and drug use patterns among a cohort of injection drug users. Addiction 104(1), 69–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dodge R, Sindelar J, Sinha R, 2005. The role of depression symptoms in predicting drug abstinence in outpatient substance abuse treatment. J Subst Abuse Treat 28(2), 189–196. [DOI] [PubMed] [Google Scholar]
  14. Forsyth SJ, Carroll M, Lennox N, Kinner SA, 2018. Incidence and risk factors for mortality after release from prison in Australia: a prospective cohort study. Addiction 113(5), 937–945. [DOI] [PubMed] [Google Scholar]
  15. Harding-Pink D, 1990. Mortality following release from prison. Med Sci Law 30(1), 12–16. [DOI] [PubMed] [Google Scholar]
  16. Hernán MA, Lanoy E, Costagliola D, Robins J, 2006. Comparison of dynamic treatment regimes via inverse probability weighting. Basic Clin Pharmacol Toxicol 98(3), 237–242. [DOI] [PubMed] [Google Scholar]
  17. Holzer HJ, Offner P, Sorensen E, 2005. Declining employment among young black less-educated men: The role of incarceration and child support. J Pol Anal Manage 24(2), 329–350. [Google Scholar]
  18. Karberg J, James D, 2005. Substance dependence, abuse, and treatment of jail inmates, 2002 (Publication No. NCJ-209588), Bureau of Justice Statsitics. [Google Scholar]
  19. Kearley B, Gottfredson D, 2017. Long term effects of drug court participation: evidence from a 15-year follow-up of a randomized controlled trial. Journal of Experimental Criminology, 1–21. [Google Scholar]
  20. Kouyoumdjian FG, McIsaac KE, Liauw J, Green S, Karachiwalla F, Siu W, Burkholder K, Binswanger I, Kiefer L, Kinner SA, Korchinski M, Matheson FI, Young P, Hwang SW, 2015. A systematic review of randomized controlled trials of interventions to improve the health of persons during imprisonment and in the year after release. Am J Public Health 105(4), e13–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kuo I, P.R., Opoku J, Rawls A, Magnu M,. Peterson J, Kharfen M, Greenberg A, 2017. Changing pattern of crystal meth use in black & white MSM, Washington, DC, 2008–2014, CROI. Seattle Washington. [Google Scholar]
  22. Massoglia M, Pridemore WA, 2015. Incarceration and Health. Annual Review of Sociology 41(1), 291–310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Mauer MJ, 2011. Addressing racial disparities in incarceration. The Prison Journal 91(3_suppl), 87S–101S. [Google Scholar]
  24. Maulsby C, Millett G, Lindsey K, Kelley R, Johnson K, Montoya D, Holtgrave D, 2014. HIV among Black men who have sex with men (MSM) in the United States: a review of the literature. AIDS Behav 18(1), 10–25. [DOI] [PubMed] [Google Scholar]
  25. McKirnan D, Vanable PA, Ostrow DG, Hope B, 2001. Expectancies of sexual “escape” and sexual risk among drug and alcohol-involved gay and bisexual men. Journal of Substance Abuse 13, 137–154. [DOI] [PubMed] [Google Scholar]
  26. Meyer IH, Flores AR, Stemple L, Romero AP, Wilson BD, Herman JL, 2017. Incarceration Rates and Traits of Sexual Minorities in the United States: National Inmate Survey, 2011–2012. Am J Public hHealth 107(2), 267–273. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Meyer IH., 2019. Lesbians, Gay Men, and Bisexuals in U.S. Jails and Prisons, Policy Research Associates; (Ed.). [Google Scholar]
  28. Millett GA, Peterson JL, Flores SA, Hart TA, Jeffries W.L.t., Wilson PA, Rourke SB, Heilig CM, Elford J, Fenton KA, Remis RS, 2012. Comparisons of disparities and risks of HIV infection in black and other men who have sex with men in Canada, UK, and USA: a meta-analysis. Lancet 380(9839), 341–348. [DOI] [PubMed] [Google Scholar]
  29. National institute on Alcohol Abuse and Alcoholism, 2019. Drinking Levels Defined: Binge Drinking. https://www.niaaa.nih.gov/alcohol-health/overview-alcohol-consumption/moderate-binge-drinking. (Accessed September 20, 2019).
  30. Parsons J, Cox C, 2019. PrEP in Prisons: HIV prevention in incarcerated populations. Int J Prison Health. 2019 Nov 26;16(2):199–206 [DOI] [PubMed] [Google Scholar]
  31. Porter LC, Novisky MA, 2017. Pathways to depressive symptoms among former inmates. Justice Quarterly 34(5), 847–872. [Google Scholar]
  32. Pratt D, Piper M, Appleby L, Webb R, Shaw J, 2006. Suicide in recently released prisoners: a population-based cohort study. Lancet 368(9530), 119–123. [DOI] [PubMed] [Google Scholar]
  33. R Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/. [Google Scholar]
  34. Reback CJ, Fletcher JBJA, Behavior, 2014. HIV prevalence, substance use, and sexual risk behaviors among transgender women recruited through outreach. AIDS Behav 18(7), 1359–1367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Rubin DB, 2004. Multiple Imputation for Nonresponse in Surveys. Wiley. [Google Scholar]
  36. Santos GM, Rapues J, Wilson EC, Macias O, Packer T, Colfax G, Raymond HF, 2014. Alcohol and substance use among transgender women in San Francisco: prevalence and association with human immunodeficiency virus infection. Drug Alcohol Rev 33(3), 287–295. [DOI] [PubMed] [Google Scholar]
  37. Stall R, Wiley J, 1988. A comparison of alcohol and drug use patterns of homosexual and heterosexual men: the San Francisco Men’s Health Study. Drug Alcohol Depend 22(1–2), 63–73. [DOI] [PubMed] [Google Scholar]
  38. Tangney JP, Folk JB, Graham DM, Stuewig JB, Blalock DV, Salatino A, Blasko BB, Moore KE, 2016. Changes in Inmates’ Substance Use and Dependence From Pre-Incarceration to One Year Post-Release. J Crim Justice 46, 228–238. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. The Sentencing Project, 2018. Trends in U.S. Corrections, U.S. State and Federal Prison Population, 1925–2017, in: The Sentencing Project; (Ed.). Washington, D.C. [Google Scholar]
  40. Van Buuren S, 2018. Flexible imputation of missing data. Chapman and Hall/CRC. [Google Scholar]
  41. Vosburgh HW, Mansergh G, Sullivan PS, Purcell DW, 2012. A review of the literature on event-level substance use and sexual risk behavior among men who have sex with men. AIDS Behav 16(6), 1394–1410. [DOI] [PubMed] [Google Scholar]
  42. Winter RJ, Stoové M, Agius PA, Hellard ME, Kinner SA, 2019. Injecting drug use is an independent risk factor for reincarceration after release from prison: A prospective cohort study. Drug Alcohol Rev 38(3), 254–263. [DOI] [PubMed] [Google Scholar]
  43. Wu E, El-Bassel N, McVinney LD, Hess L, Fopeano MV, Hwang HG, Charania M, Mansergh G, 2015. The association between substance use and intimate partner violence within Black male same-sex relationships. J Interpers Violence 30(5), 762–781. [DOI] [PubMed] [Google Scholar]
  44. Yu G, Wall MM, Chiasson MA, Hirshfield S, 2015. Complex drug use patterns and associated HIV transmission risk behaviors in an Internet sample of U.S. men who have sex with men. Arch Sex Behav 44(2), 421–428. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES