Introduction

Black sexual and gender minority individuals are disproportionately impacted by HIV in the United States (US) [1,2,3,4,5,6,7,8,9]. In 2017, Black men who have sex with men (BMSM) nationally accounted for more than a quarter of all new HIV diagnoses in the US, with a majority of new HIV diagnoses occurring among younger BMSM aged 13–34 [1]. Among transgender individuals, an estimated 14% of transgender women were living with HIV in 2017, with the highest prevalence estimate (i.e., 44%) being among Black transgender women (BTW) [10].

High incarceration rates have also been identified among BMSM and BTW in the US [11, 12]. A landmark longitudinal study, the HIV Prevention Trials Network (HPTN) 061 study conducted in six urban US cities (i.e., District of Columbia, San Francisco, Los Angeles, Boston, Atlanta, and New York City), highlighted the disproportionate burden of incarceration among sexual and gender minority populations. Consisting of primarily cisgender BMSM, the HPTN 061 study showed an elevated lifetime history of incarceration (60%) and high incarceration incidence (35%, 95% CI = [31%, 38%]) among sexual and gender minority participants. Furthermore, BTW had a greater odds (3.26) of incarceration history compared with cisgender BMSM [12]. However, to date, there is not a comprehensive understanding of the impact of incarceration among these priority populations and there remains a dearth of information about the connection between criminal justice involvement and HIV among BMSM and BTW who sit at the intersection of multiple marginalized identities (i.e., Black race, gender identity, sexual orientation) [3, 5]. Thus, in the current systematic review, we sought to expand understanding of the burden of incarceration as well as the relationship between HIV and criminal justice involvement among Black sexual and gender minority populations in the US [11, 12].

Previous studies have shown that network characteristics (e.g., high sexually transmitted infection [STI] rates among sexual partners), psychosocial characteristics (e.g., mental health and substance use/misuse), and socio-structural factors (e.g., unstable housing, racism/stigma, unemployment) contribute to HIV vulnerability and poor access to HIV services for Black sexual and gender minority populations [1, 13,14,15,16,17,18,19,20,21,22,23,24,25]. What is less understood are the reasons surrounding why these characteristics may or may not be more pronounced and/or exacerbated among criminal justice–involved (CJI) BMSM and BTW. To address this gap, we focus our systematic review on quantitative studies that have examined the various intersections between criminal justice involvement, HIV, and key characteristics (e.g., network, psychosocial, socio-structural) known to be associated with HIV vulnerability and poor access to HIV services (i.e., HIV care and treatment). We use the term CJI to include those who are currently detained or under criminal justice supervision (probation or parole) or who have a history of arrest, incarceration, or supervision. We also highlight epidemiological studies focused on YBMSM given that they account for the majority of new HIV diagnoses among all men who have sex with men (MSM) and the need to fully engage this population in order to reach HIV elimination [1, 14].

Pre-detention, detention, and post-release incarceration periods have been identified as important intervention opportunities for evidence-based and informed HIV interventions [15, 16]. A recent systematic review found that the most common interventions within criminal justice settings consisted of HIV/STI screenings as well as educational interventions [17]. In addition, relatively few interventions were focused on MSM and TW, perhaps, because this approach may expose these populations to unwarranted attention and violence within criminal justice settings [17]. Of the 58 studies identified in this review, only one was focused specifically on CJI among BMSM [17]. The information gathered from the current systematic review can be used to inform gaps in existing research focused on CJI BMSM and BTW in the US and guide the development of future research and interventions.

Methods

The final systematic review search was completed on September 6, 2018, and included all articles up to that date. Relevant keywords and Medical Subject Headings (MeSH) terms based on a priori knowledge and manuscript goals were used to search the PubMed, PsychINFO, and SocINDEX databases. The finalized terms inputted into these databases are available in the appendix. The search yielded a total of 393 articles that were then exported into an Endnote library [18]. Seventy-four (74) duplicate articles were removed and the remaining articles were assessed for eligibility.

Article Eligibility

The systematic review was focused on quantitative epidemiological studies conducted in the US that provided data for individuals 18 and older. Thus, qualitative and non-US studies were excluded. Studies had to (1) report criminal justice involvement estimates (e.g., previous incarceration history or incidence) that were specific to BMSM and/or BTW participants and/or (2) contain CJI estimates specific to BMSM and/or BTW within the context of at least one of the following conditions: (a) HIV, (b) STIs, (c) sexual risk behaviors, (d) mental health outcomes, (e) substance use/misuse, and (f) socio-structural experiences (including stigma/discrimination, housing status, education levels, and employment history). Articles were reviewed for inclusion or exclusion in an Endnote library. Detailed information on article identification and selection, including exclusion criteria, is available in Fig. 1 (PRISMA diagram). Each ineligible article was categorized by a single exclusion reason. An initial review of titles, abstracts, and full texts of articles was conducted by a member of the research team, and uncertainties related to eligibility were resolved with input from the primary author. In total, 47 articles were identified for inclusion.

Fig. 1
figure 1

PRISMA flow diagram: systematic review of US-based studies among CJI BMSM and BTW

Full size image

Data Abstraction

General article characteristics (e.g., location, type of study, primary outcomes) are described in Table 1. All relevant point estimates from the eligible articles were abstracted and described in Tables 2–4. Within Tables 2–4, each article with qualifying data is presented along with article-specific estimates and organized according to variables of interest. Table 2 includes articles with HIV/STI estimates. Table 3 includes articles involving sexual behavior, substance use, and mental health. Table 4 includes all articles that assessed and provided information about socio-structural characteristics to include stigma/discrimination, education, employment, and housing.

Table 1 Characteristics of included studies focused on criminal justice–involved (CJI) Black men who have sex with men (BMSM) and Black transgender women (BTW), 2000–2018
Full size table

Quality Assessment

Eligible studies were assessed for quality using a ten-item risk of bias tool developed by Hoy et al. [19] for population-based prevalence studies (range: 0, worst–10, best). The risk of bias tool included domains such as the method of sample selection, non-response bias, and reliability and validity of the study instrument. Some limitations with applicability of the scale existed due to its intended use to rate observational studies of clinical disease prevalence. The scale was modified to more appropriately assess the quality of included articles based on the senior investigator’s knowledge of epidemiological studies of sexual and gender minorities and based on previous work [26]. Initially, 5 articles were scored independently to assess interrater reliability. A member of the research team generated a quality score for each article using the scale items with input from senior investigators.

We organize our findings in the following order: (1) a summary of the studies included in the systematic review; (2) studies describing criminal justice involvement estimates among BMSM and BTW; (3) studies focused on the intersection of HIV and criminal justice involvement to include the HIV care continuum; (4) studies focused on criminal justice involvement and STIs; (5) studies focused on criminal justice involvement and sexual risk behaviors; (6) studies focused on criminal justice involvement and substance use/misuse; (7) studies focused on criminal justice involvement and mental health outcomes; and finally (8) studies focused on criminal justice involvement and socio-structural factors.

Results

Summary of Included Articles

A total of 47 eligible articles published over a span of 18 years were identified for inclusion. Aggregate characteristics of the included articles are summarized in Table 1 to include studies by region, design, and primary outcomes (e.g., HIV, STIs, mental health, and substance use). Tables 2–4 containing all abstracted article data are available as part of the supplementary materials. The mean rating for article quality based on the risk of bias tool [19] was 7.04 (SD = 2.00), with scores ranging from 4 to 9.

Notably, 25 published articles (53.2%) were analyzed utilizing datasets similar to at least one other article (i.e., based on the same local cohort or from the same multi-site study such as the HPTN 061 study and the National HIV Behavioral Surveillance survey). Information on usage of similar datasets is detailed in a separate column for each applicable article in Table 4. Nearly all studies were implemented in urban settings, with just one article providing specific estimates for a rural location [27]. All studies in the Midwest (n = 5) were conducted in Chicago, IL. More than a quarter (n=13) of all studies were conducted in the West Coast, primarily in Los Angeles County, CA.

Criminal Justice Involvement Estimates Among BMSM and BTW

All included articles described criminal justice involvement estimates for BMSM and BTW. However, the type (e.g., prior history, arrest, currently detained) and measures used varied substantially across studies. The most common measure, reported by 28 studies (60%), was prevalence of incarceration history. Of these, 19 articles reported lifetime incarceration history estimates [11, 12, 28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44]. Eleven of these articles had estimates focused on BMSM of all ages (i.e., samples of all or primarily BMSM) [11, 28,29,30,31, 34,35,36, 40, 42, 43], five particularly among YBMSM [33, 37,38,39, 44], one among BTW only [41], and two among both BMSM and BTW [12, 32]. Incarceration estimates focused on BMSM had values ranging from 25.5 to 84%, while those on YBMSM ranged from 2.4 to 53.5% [32, 33, 37,38,39, 44]. The three estimates among BTW of all ages had relatively small sample sizes, but estimates were uniformly high in contrast to estimates among BMSM of all ages, ranging from 46.9 to 80% [12].

Seven studies (15% of total) assessed more recent estimates of incarceration prevalence ranging from the past 2 months to 2 years [22, 35, 45,46,47,48,49]. Among these, two articles contained estimates specific to BTW [45, 46] and one among YBMSM in particular [49]. Six articles described prevalence estimates for history of arrest [25, 39, 50,51,52,53], varying from lifetime history to the past 6 or 12 months, though one had an unspecified window [50], Two analyses based on the same sample of YBMSM broadly assessed lifetime criminal justice involvement (inclusive of both arrest and incarceration) [13, 54].

Incarceration frequency and duration were relatively uncommon variables of interest with only three articles providing estimates of incarceration duration [29, 55, 56], four on incarceration frequency [11, 25, 32, 55], one on arrest frequency [51], and one on length of time since release [57]. Two estimates among these were specific to BTW [32, 55]. Only one analysis explored incarceration incidence, reporting an annualized incidence rate of 35% [31, 38] among a sample of 1278 BMSM [11]. Twelve articles (25.5%) also contained comparative incarceration estimates by race and/or ethnicity [22, 27, 30, 31, 33, 39, 41, 46, 47, 52, 53, 58]. Seven (14.9%) studies were conducted among detainees in a jail setting [56, 59,60,61,62,63,64], and an additional four reported on a post-incarcerated population in which all participants had been recently released from jail or prison [55, 57, 65, 66].

Key Findings Related to the Intersection Between HIV and Criminal Justice Involvement Among BMSM and BTW

Seventeen (17) studies examined the intersection between HIV and criminal justice involvement among BMSM and BTW. These articles are described in Table 2. Among these, 14 were focused primarily on BMSM [11, 13, 22, 28, 34, 37, 40, 44, 54, 56, 57, 62, 64, 67], two primarily on BTW [45, 65], and one on both BMSM and BTW [46]. We describe the specific studies in detail below by four categories of interest.

HIV Prevalence Among Current and Recently Incarcerated BMSM and BTW

Two studies reported HIV prevalence estimates among a large sample of Black detainees receiving HIV testing in the same protected jail unit for MSM and transgender women (consisting of primarily MSM), reporting prevalence estimates of 11.9% (91/767) and 16.3% (258/1,584) respectively [62, 64]. These estimates were somewhat low because detainees with diagnosed HIV disease generally were not screened for HIV. Three studies reported HIV prevalence estimates among criminal justice–involved BMSM. Harawa et al. [57] and Li et al. [66] both provided a self-reported HIV positivity estimate of 30.7% based on the same RCT sample of 212 BMSM with a recent 12-month incarceration. Schneider et al. [13] reported a higher HIV prevalence in a sample of YBMSM reporting a history of incarceration, among which 46.0% were HIV seropositive (n = 285).

Incarceration History Among BMSM and BTW Living with HIV

Four studies reported a high prevalence of incarceration among BMSM living with HIV [28, 34, 44, 46]. Three of these analyses reported lifetime incarceration history estimates ranging from 28 to 84% [28, 34, 44]. The fourth article examined the prevalence of recent incarceration among BMSM by HIV serostatus, reporting a prior 12-month incarceration prevalence of 21% among BMSM living with HIV compared to 14% among BMSM who were not living with HIV (n = 760) [46]. Only one article, by Bukowski et al. [45], provided estimates on incarceration prevalence specific to BTW living with HIV, finding that 42.3% (n = 111) of BTW with previously undiagnosed HIV were incarcerated in the past 2 years, compared to 31.5% of BTW who were not living with HIV (n = 232), aOR = 1.92 [1.15, 3.22], and 39.7% of BTW with diagnosed HIV (n =70). In summary, a high percentage of BMSM and BTW living with HIV have a high frequency of contact with the criminal justice system.

Independent Associations Between Criminal Justice Involvement and HIV

Four studies explicitly examined the independent associations between HIV and criminal justice involvement among BMSM and BTW [11, 22, 40, 46]. Three of those analyses utilized cross-sectional data [22, 40, 46] and one utilized longitudinal data [11]. Anderson-Carpenter et al. [46] found that recent incarceration history was associated with HIV-positive status (aOR = 1.69, CI = [1.31, 2.16], p ≤ .001). Oster et al. [22] found that 12% of newly diagnosed BMSM living with HIV were incarcerated in the past 12 months but did not find an independent association between incarceration and HIV infection. Similarly, Magnus et al. [40] did not find an independent association between HIV infection and incarceration history among 178 BMSM. The study by Brewer et al. [11] was the only analysis to longitudinally explore the relationship between incident incarceration and HIV acquisition. Brewer et al. [11] did not find an association between incarceration (last 6 months) and HIV acquisition (adjusted hazard ratio 1.69, CI = [0.64–4.44]) among a sample of 1278 BMSM followed for 12 months after controlling for demographics and unprotected receptive anal intercourse (URAI). In summary, findings related to the independent association between HIV and criminal justice involvement are mixed, but the current evidence favors no association between incarceration and HIV.

Criminal Justice Involvement and the HIV Care Continuum Components Among BMSM and BTW

Five studies examined the relationship between criminal justice involvement and the HIV care continuum, with four focused on BMSM and one on BTW [37, 45, 54, 56, 67]. Koblin et al. [37] found that intention to test for HIV was negatively associated with lifetime incarceration history (aOR = 0.37, CI = [0.16, 0.89], p = .008) in a sample of 169 consisting of 91.7% YBMSM and 8.3% YBTW. Yet Mimiaga et al. [67] found that in a sample of 197 BMSM, those who had not received an HIV test in the past 2 years were less likely to report a history of incarceration. This is likely due to an increased likelihood of receiving an HIV test during incarceration, as those who had ever obtained HIV testing in jail had a lower odds of not testing for HIV in the 2 years prior to study enrollment relative to those who were never tested in jail (OR = 0.53, CI = [0.13–0.99], p = .049) [67]. The sole article focused exclusively on BTW showed that incarceration was significantly and positively associated with undiagnosed HIV and negatively associated with viral suppression in multivariable models such that BTW who were virally suppressed had significantly lower odds of incarceration compared to those who were virally unsuppressed (aOR = 0.15, CI = [0.03, 0.71], p < .05).45 Thus, criminal justice settings may serve as an important venue for the first step (i.e., HIV testing/diagnosis) of the HIV care continuum and not subsequent stages (e.g., viral suppression) among BTW.

The findings from Schneider et al.’s [54] examination of the associations between criminal justice involvement and the HIV care continuum broadly aligned with those of Mimiaga et al. [67]. Such that, any history of criminal justice involvement was associated with engagement in the overall care continuum (aOR = 2.35, CI = [1.13–4.88]), as well as linkage and retention to care, adherence to anti-retroviral treatment, and viral suppression. Although longer incarceration duration events were associated with improved engagement, repeated incarceration events were associated with poorer HIV care engagement [54]. In contrast, Vagenas et al. [56] highlighted low rates of access to HIV treatment, retention in HIV care, and viral suppression among jail detainees including YBMSM. Thus, the relationship between criminal justice involvement and the HIV care continuum may be highly dependent upon the specific CJI measure assessed among BMSM. However, criminal justice settings may serve as a critical venue for HIV testing services for both BMSM and BTW.

Key Findings Related to Criminal Justice Involvement and STIs Among BMSM and BTW

Only 7 studies (15%) examined the relationship between criminal justice involvement and STIs among the populations of interest. Three studies were among BMSM [25, 29, 67], one specifically among YBMSM [13], and three on the same jail population that consisted primarily of MSM [62,63,64]. None of the studies reported specific relationships between criminal justice involvement and STIs among BTW. However, three of 7 studies specifically examined the independent association between one or more STIs and criminal justice involvement among BMSM [13, 25, 29]. One article found a significant relationship between incarceration history and syphilis seropositivity, and another between incarceration and self-reported lifetime history of STIs (including syphilis, chlamydia, gonorrhea, or herpes) [13, 29]. Nelson et al. [25] reported a significant relationship between a recent conviction and any STI in the past 6 months (aOR = 3.97, CI = [1.58, 9.94], p < 0.05). Contrary to the mixed findings between HIV and criminal justice involvement, these studies show that incarceration history is independently associated with a greater odds of STIs among BMSM with limited information particularly among BTW.

Key Findings Related to Criminal Justice Involvement and Sexual Risk Behaviors Among BMSM and BTW

Eleven of 47 articles (23%) assessed sexual risk behaviors [11,12,13, 25, 29, 36, 49, 57, 65, 66, 68]. A range of sexual risk behaviors were assessed to include the following: exchange sex, three articles [57, 65, 68]; jail sex, two articles [57, 68]; history of condomless or unprotected sexual acts (e.g., any condomless sex and condomless insertive and/or receptive anal intercourse), six articles [12, 25, 29, 49, 57, 65]; frequency of condomless or unprotected sex, four articles [36, 57, 66, 68]; age of first intercourse, one article [11]; number of sexual partners, eight articles [12, 13, 25, 29, 36, 57, 65, 68]; and gender identity of sexual partners (i.e., male, female, transgender), seven articles [12, 25, 29, 36, 57, 66, 68] with three trans-inclusive articles [36, 57, 65]. Overall, the analysis by Harawa et al. [57] was the most comprehensive assessment of sexual risk behaviors prior to entering a jail setting, but also surveyed the prevalence of sex within jail by partner gender. The analysis by Wohl et al. [68] was the most comprehensive assessment of sexual risk behaviors within a jail setting.

Only one study assessed sexual behaviors among transgender women and reported high rates (≥ 65%) of condomless sex, exchange sex, and multiple partners in the past 90 days among small CJI samples of primarily BTW [65]. None of the included studies specifically examined the direct relationship between incarceration and sexual risk behaviors among BTW. However, findings on the relationship between incarceration and sexual risk behaviors among BMSM were mixed. Two articles did not find any significant relationship between lifetime incarceration and having multiple partners or unprotected and/or condomless sex in the past 6 months among BMSM [29, 36]. Nelson et al. [25] actually reported a protective effect of lifetime incarceration for condomless receptive anal sex (aOR = 0.73, CI = [0.58, 0.91], p < 0.05) among BMSM. However, two articles found a relationship between more recent incarceration and unprotected sex among BMSM [29, 49], and the same article by Nelson et al. [25] reported a significant relationship between recent conviction and multiple male partners in the past 6 months (aOR = 0.56, CI = [0.33, 0.93], p < 0.01). Schneider et al. [13] also found a significant relationship between a history of criminal justice involvement and having a greater number of sexual partners among YBMSM (50.8% among YBMSM with a history of criminal justice involvement vs. 66.1% with no history of criminal justice involvement; p = 0.0001).

Interestingly, Schneider et al. [13] found that CJI YBMSM were more likely to have CJI homophily in their confidant and sexual networks such that 59% of CJI YBMSM reported at least one sexual partner who had been in jail during the past 6 months compared to 26.5% of YBMSM without a history of criminal justice involvement (p < 0.0001). Furthermore, network CJI history (i.e., confidants and sexual partners) was associated with greater encouragement of HIV/STI transmissible sexual behaviors [13].

Key Findings Related to Criminal Justice Involvement and Substance Use/Misuse Among BMSM and BTW

Nine studies (19.1%) assessed substance use/misuse behaviors among study participants. These ranged from three studies (6.4%) on general alcohol use [12, 56, 65], one (2.1%) on binge drinking (n=1) [65], and nine (19.1%) assessing substances such as marijuana, methamphetamine, crack/cocaine, opioids, stimulants, and multiple substances [11,12,13, 29, 46, 56, 57, 61, 65]. Two studies (4.3%) also assessed whether participants received an alcohol/drug treatment program during custody [29, 57]. One article utilized the Addiction Severity Index (n=1); the others did not address severity of substance use disorders [56]. Only one article had estimates focused primarily on BTW [65], finding high usage rates across a range of substances including crack/cocaine and binge alcohol, but in limited sample sizes of 20 or less. The remaining studies reported findings based on samples of all or nearly all BMSM and did not provide separate estimates for BTW. Overall, high levels of substance use were identified among BMSM and BTW.

Several studies found an independent association between previous criminal justice involvement and substance use among BMSM with no studies examining this relationship among BTW [11,12,13, 29, 41]. Bland et al. [29] found that incarceration history was significantly associated with injection drug use, lifetime history of substance use, and prior substance abuse treatment in a sample of 197 BMSM, with a higher odds among those incarcerated for ≥ 90 days. Interestingly, BMSM with longer incarceration trajectories (i.e., ≥ 90 days) were more likely to use crack cocaine during sex than those with shorter incarceration histories [24]. In Brewer et al.’s [12] two papers reporting on BMSM enrolled in the HIV Prevention Trials Network (HPTN) 061 study, any alcohol or drug in the past 6 months was significantly associated with a higher odds of lifetime incarceration history, while stimulant and marijuana use was linked to a higher odds of incident incarceration [11]. Vagenas et al. [56] examined age differences between younger and older BMSM, finding no significant differences between these groups in terms of drug type and addiction severity, except for cocaine usage, which was more frequent among older BMSM. Among YBMSM, network CJI history (i.e., confidants and sexual partners) was associated with a higher likelihood of drug use [13].

Key Findings Related to Criminal Justice Involvement and Mental Health Outcomes Among BMSM and BTW

Five articles (11%) assessed mental health outcomes with no reported estimates specific to BTW even though they were enrolled in the larger cohort studies [11,12,13, 29, 56]. The most common condition measured were depressive symptoms using the Center for Epidemiologic Studies Depression Scale (CES-D). Two of the five papers were based on subsets of participants from the larger cohort of BMSM enrolled in the HPTN 061 study, with final analytic samples of 1278 and 1521 participants [11, 12]. In these papers, the prevalence of depressive symptoms among BMSM reporting any incarceration during study follow-up or in their lifetime was 27% and 46%, respectively [11, 12]. One article was focused on a statewide sample of 197 BMSM and found a similarly high prevalence of depression (i.e., 46%) among those incarcerated for ≥ 90 days [29]. In short, relatively high levels of depressive symptoms among BMSM were reported.

Three of the five articles examined the relationship between a history of criminal justice involvement and depression among BMSM, with only one of the studies showing a statistically significant relationship between incarceration history (≥ 90 days) and depressive symptoms (AOR = 2.76, CI = [1.22, 6.24], p = .01) [29]. In terms of differences by age group, no significant differences between younger and older CJI BMSM were observed among included studies (50% for younger vs. 58.2% for older, p=0.54) [56]. However, interesting findings were observed for the relationship between   network history of criminal justice involvement and depression, such that network stability was associated with a reduction in respondent depression [13]. In summary, the findings on the relationship between criminal justice involvement and depression among BMSM remain mixed.

For other mental health outcomes, among YBMSM specifically, a higher prevalence of criminal justice involvement history among network members (i.e., confidants and sexual partners) was associated with higher levels of anxiety and distress [13]. Schneider et al. [13] concluded that network criminal justice involvement history may limit access to the types of relationships that support and maintain mental health [69] and/or create situations of instability that further exacerbate mental health problems among non-institutionalized YBMSM [70].

Key Findings Related to Criminal Justice Involvement and Other Socio-Structural Characteristics Among BMSM and BTW

Only six studies (12.8%) assessed one or more of the four other socio-structural characteristics of interest within the context of criminal justice involvement (i.e., stigma/discrimination, housing, employment, and education) [11, 13, 56, 57, 65, 66]. In terms of stigma/discrimination, among BMSM enrolled in the HPTN 061 study, incarceration during study follow-up was associated with high levels of perceived racism (aOR 1.82; 95% CI = 1.02–3.27) [11].

All six studies (12.8%) documented high instances of housing instability among CJI BMSM ranging from 28 to 62% of samples [11, 13, 56, 57, 65, 66]. Of these studies, two included estimates focused on YBMSM [13, 56]. Schneider et al. [13] found that CJI YBMSM were more likely to report unstable housing in the past 12 months compared with non-CJI YBMSM (28.9% vs 20.7%, p=0.063), though the difference was not significant. Among YBMSM, a history of criminal justice involvement was associated with housing instability. In addition, network criminal justice involvement history (i.e., confidants and sexual partners) was associated with a higher likelihood of housing instability [13]. Vagenas et al. [56] compared housing instability at baseline by age group among detained BMSM and did not find any significant differences (38.9% among younger vs. 32.7% among older, p = 0.63). Harawa et al. [57] and Li et al. [66] reported a high rate of housing instability of 62.7% in the past 12 months among post-incarcerated BMSM. Among BTW specifically, only one study reported a similarly high rate of housing instability (44%) among a small sample of 16 predominantly HIV-positive BTW with a history of criminal justice involvement [65].

Five articles (10.6%) provided estimates of educational attainment among CJI participants, including one focused on BTW [65], two on BMSM [11, 57], one on YBMSM [13], and one comparative analysis including estimates for older BMSM and YBMSM [56]. All five articles across these various groups reported over 10% (range = 10.4–38.9%) of their samples not having completed a high school education [11, 13, 56, 57, 65]. The lowest estimate of 10.4% was in a sample of 285 CJI YBMSM, but all other estimates across groups (YBMSM, BMSM, and BTW), were ≥ 25%. Of these articles, Harawa et al. [57] specified that having a GED, high school diploma, or some college were categorized as having completed a high school–level education, while the other four articles did not delineate a classification method.

Two articles (4.3%) examined the relationship between criminal justice involvement and education level among BMSM. The study by Brewer et al. [11] showed that BMSM with less education were more likely to be incarcerated during study follow-up (aOR=1.83, CI = [1.27, 2.62], p < 0.01). Schneider et al. [13], however, found no significant difference in education between CJI and non-CJI YBMSM at baseline. No studies examined the relationship between criminal justice involvement and education level among BTW.

Three studies (6.4%) provided employment estimates for CJI BMSM but none were identified specifically for BTW [11, 13, 57]. All were quantitative studies with city/county (i.e., Los Angeles County and Chicago) and multi-state (i.e., six HPTN sites) samples of BMSM ranging from 212 to 1278 participants. The Chicago-specific study by Schneider et al. [13] did not find any significant differences in employment between CJI and non-CJI YBMSM at baseline which may be point to high level of need regardless of incarceration status. The multi-city HPTN study showed a greater odds of incarceration during study follow-up among men who were unemployed compared with those who were employed at baseline (OR=2.22, 1.63–3.03) [35]. Thus, unemployment is a risk factor for future incarceration among BMSM. No studies examined the relationship between criminal justice involvement and employment among BTW.

Discussion

As of the date of this review, there was a paucity of research focused on the intersection of criminal justice involvement, HIV, and associated factors among BMSM and BTW. The objective of this systematic review was to provide a comprehensive overview of the intersections between criminal justice involvement, HIV, and HIV-related characteristics among BMSM and BTW. Included studies examined a wide range of criminal justice experiences, but the most common measure of criminal justice involvement was lifetime history of incarceration with a robust non-representative national sample composed primarily of BMSM enrolled in the HPTN 061 study reporting a high (i.e., 60%) lifetime history of incarceration among BMSM with a greater likelihood of incarceration history among BTW [12].

The dearth of studies up to 2018 specifically focused on CJI BTW across all of the categories of interest is quite evident. Even though several studies enrolled BTW, specific analyses were not conducted and/or reported. A limited number of trans-inclusive published studies at the date of this review showed that criminal justice involvement may be more pronounced for BTW compared with BMSM. The geographic diversity of studies was additionally limited with most studies occurring in Los Angeles, CA and Chicago, IL. Geographic limitations may partly be explained by investigator interest and presence in those locales as well as the challenges associated with building and maintaining academic/criminal justice setting research collaborations, which may be even more pronounced during the COVID-19 pandemic [71].

Consistent with national data documenting the high prevalence of HIV among CJI persons, BMSM and BTW living with HIV had a high frequency of contact with the criminal justice system [72]. Even though we observed mixed findings as it relates to the independent association between criminal justice involvement and HIV infection, the current evidence favors no association between criminal justice involvement and HIV among BMSM with the largest cohort of BMSM to date not finding an association between recent incarceration and HIV acquisition. Findings related to the independent association between criminal justice involvement and sexual risk behaviors among BMSM were mixed. Criminal justice involvement was positively associated with substance use among BMSM with a strong network influence particularly among YBMSM such that network CJI history (i.e., confidants and sexual partners) was associated with a higher likelihood of drug use. Thus, network interventions may be appropriate for YBMSM.

In terms of STIs, we observed an independent relationship between criminal justice involvement and STI prevalence among BMSM, but inability to establish a temporal relationship precludes assessment of evidence for causality. Regardless, it is plausible that criminal justice settings provide an opportunity for increased STI testing services which are coupled with opt-out HIV testing within these settings [17]. Incarceration may also disrupt sexual relationships leading to an increase in casual sexual partnerships and associated increases in the likelihood of being exposed to STIs [25, 73]. As of the date of this review, there remains a dearth of information about the relationship between criminal justice involvement and STIs among BTW.

A limited number of studies explored the relationship between criminal justice involvement and the HIV care continuum. Criminal justice–involved settings served as an important venue for HIV testing/diagnosis for both BMSM and BTW. The relationship between criminal justice involvement and the HIV care continuum particularly among BMSM seems highly dependent upon the specific CJI measure assessed. For example, any CJI history was associated with improved HIV care outcomes across the HIV care continuum. However incident jail stays and/or more frequent incarceration events may negatively impact HIV care. Interestingly, in one study of BTW living with HIV, those who were virally suppressed had significantly lower odds of lifetime incarceration compared to those who were virally unsuppressed. This may point to the syndemic nature of health disparities production among BTW [45] and existing barriers to viral suppression within criminal justice settings for this population [74,75,76,77]. For example, a qualitative study consisting primarily of transgender women of color described a non-affirming correctional culture in which their feminine identity was not recognized and a combination of provider bias, limited provider knowledge or inexperience caring for transgender patients, and restrictive correctional policies (e.g., sex-segregated environments based on genitalia and strict rules surrounding access to transition-related medical care) limited access to needed medical care [74].

In terms of the socio-structural characteristics, only one study examined the relationship between stigma/discrimination and criminal justice involvement showing that incarceration during study follow-up was associated with perceived racism among BMSM. This is not a surprising finding given the pervasive nature of racial inequality and systemic racism in the US [78]. African Americans are more likely to be stopped by police and feel somewhat or very unsafe during their interactions with law enforcement, more likely to experience threats or use of physical force by law enforcement, and twice as likely to die from a fatal shooting by an on-duty police officer than their White counterparts [79,80,81]. Furthermore, BMSM’s positioning at the intersection of multiple identities expose them to other forms of discrimination including homophobia [5, 82, 83].

Few studies comprehensively assessed mental health outcomes among the CJI priority populations. Depressive symptoms were the most commonly assessed outcome, with high estimates of depressive symptoms among BMSM, but inconsistent associations with criminal justice involvement among BMSM. Network stability was identified as a protective factor in terms of mental health particularly for YBMSM. This finding also highlights the importance of network factors among YBMSM.

High rates of housing instability were found among CJI BMSM and BTW, with no statistically observed differences between CJI vs. non-CJI participants. However, criminal justice history was strongly connected to housing instability particularly among YBMSM. Findings indicated that more than 10% of BMSM and BTW with CJI have not completed a high diploma or GED. However, we note that, nationally, 15% of Black men ages 25 and higher have not completed high school [84]. While we did not observe differences in education level and employment for CJI YBMSM vs. non-CJI YBMSM [13], the largest multi-site cohort of BMSM did find an independent association between criminal justice involvement and education as well as criminal justice involvement and employment, particularly as it relates to incarceration incidence [11, 12]. We note that research indicates that higher levels of both education and income offer relatively less protection against criminal justice involvement for African Americans than they do for Whites [85].

Conclusion

This systematic review provides some of the first comprehensive examinations of the burden of incarceration among Black sexual and gender minority populations as well as the intersection of criminal justice involvement, HIV, and associated factors among BMSM and BTW. Most studies were cross-sectional in nature, limiting the ability to make causal inferences between criminal justice involvement, HIV, and associated factors. Our review is limited to published scientific studies up to 2018 in the three scientific databases used. Additional studies have been published after the cutoff date used for this systematic review. For this review, we did not include studies focused on HIV pre-exposure prophylaxis (PrEP). Thus, a future research area may include examining the relationship between criminal justice involvement and PrEP outcomes among BMSM and BTW. In spite of these limitations, this analysis provides a comprehensive overview of the state of the science as it relates to criminal justice involvement, HIV, and HIV-related characteristics among Black sexual and gender minority populations in the US. It highlights the need for more focused studies on BTW to include the relationship between CJI and STIs among this population as well as the importance of CJI network dynamics particularly among YBMSM. It also underscores the importance of the intersection between criminal justice involvement and HIV in any response to Ending the HIV Epidemic in the US [86, 87].